Handbook of Neurosurgery 7th Ed

21. Tumor

21.1. General information

CLASSIFICATION OF NERVOUS SYSTEM TUMORS

WHO classification of tumors of the nervous system^1-5

The 2007 WHO classification⁵ identifies 7 categories of tumors of the nervous system (see Table 21-1) and a modified outline appears below (along with the unofficial category “intracranial and/or intraspinal embryonal remnants”, and pituitary adenomas (not part of the CNS)). Also to be considered: cysts (neurocysticercosis…), tumor-like masses (e.g. giant aneurysms), and local extension of regional tumors. Cytogenetic and molecular genetic information are playing an ever increasing role in the definitive classification of some tumors.

Table 21-1 Overview of WHO classification of nervous system tumors²

1. tumors of neuroepithelial tissue 2. tumors of cranial and paraspinal nerves 3. tumors of the meninges 4. lymphomas and hematopoietic neoplasms 5. germ cell tumors 6. tumors of the sellar region 7. metastatic tumors

	ICD-O*	Page
A. TUMORS OF NEUROEPITHELIAL TISSUE†
1. astrocytes ß astrocytic tumors‡		518
A. astrocytomas that are typically infiltrating (lower grade tumors in this category tend to progress in malignancy)		518
1. diffuse astrocytoma (WHO II§). Variants:	9400/3	518
a. fibrillary	9420/3
b. protoplasmic	9410/3
c. gemistocytic	9411/3
2. anaplastic (malignant) astrocytoma (WHO III)	9401/3	595
3. glioblastoma (WHO IV) (formerly glioblastoma multiforme (GBM)). Variants:	9440/3	596
a. giant cell glioblastoma	9441/3
b. gliosarcoma	9442/3
4. gliomatosis cerebri	9381/3	598
B. more circumscribed lesions (these do not tend to progress to anaplastic astrocytoma and GBM)
1. pilocytic astrocytoma	9421/1	603
• pilomyxoid astrocytoma (WHO II)	9425/3	606
2. pleomorphic xanthoastrocytoma (PXA)	9424/3	592
3. subependymal giant cell astrocytoma: associated with tuberous sclerosis	9384/1	725
2. oligodendrocytes → oligodendroglial tumors
A. oligodendroglioma (WHO II)	9450/3	609
B. anaplastic oligodendroglioma (WHO III)	9451/3	610
3. oligoastrocytic tumors (nee: mixed gliomas)
A. oligoastrocytoma (WHO II)	9382/3	612
B. anaplastic (malignant) oligoastrocytoma (WHO III)	9382/3	612
4. ependymocytes → ependymal tumors
A. ependymoma (WHO II). Variants:	9391/3	682
1. cellular	9391/3	683
2. papillary	9393/3	683
3. clear cell	9391/3	683
4. tanycytic	9391/3	683
B. anaplastic (malignant) ependymoma (WHO III)	9392/3	683
C. myxopapillary ependymoma: filum terminale only (WHO I)	9394/1	683
D. subependymoma (WHO I)	9383/1	683
5. choroid plexus tumors
A. choroid plexus papilloma	9390/0	695
B. atypical choroid plexus papilloma	9390/1	695
C. choroid plexus carcinoma	9390/3	695
6. other neuroepithelial tumors		696
A. astroblastoma	9430/3
B. chordoid glioma of the 3rd ventricle	9444/1	696
C. angiocentric glioma	9431/1
7. neuronal and mixed neuronal-glial tumors		612
A. gangliocytoma	9492/0
B. ganglioglioma	9505/1	677
C. dysembryoplastic neuroepithelial tumor (DNT)	9413/0	591
D. desmoplastic infantile astrocytoma/ganglioglioma (DIG)	9412/1	612
E. dysplastic gangliocytoma of cerebellum (Lhermitte-Duclos)	9493/0	593
F. anaplastic (malignant) ganglioglioma	9505/3
G. central neurocytoma	9506/1	612
H. extraventricular neurocytoma	9506/1
I. cerebellar liponeurocytoma	9506/1
J. papillary glioneuronal tumor	9509/1
K. rosette-forming glioneuronal tumor of the 4th ventricle	9509/1
L. paraganglioma (of the filum terminale	8680/1
8. pinealocytes → pineal parenchymal tumors
A. pineocytoma (pinealoma)	9361/1	692
B. pineoblastoma	9362/3	692
C. pineal parenchymal tumor of intermediate differentiation	9362/3
D. papillary tumor of the pineal region	9395/3
9. embryonal tumors		685
A. medulloblastoma. Variants:	9470/3	686
1. desmoplastic/nodular medulloblastoma	9471/3	687
2. anaplastic medulloblastoma	9474/3
3. large-cell medulloblastoma	9474/3	687
4. medulloblastoma with extensive nodularity	9471/3
B. CNS primitive neuroectodermal tumors (PNET)	9473/3	686
1. CNS neuroblastoma	9500/3
2. CNS ganglioneuroblastoma	9490/3
3. medulloepithelioma	9501/3
4. ependymoblastoma	9392/3	688
C. atypical teratoid/rhabdoid tumor (AT/RT)	9508/3	688
B. TUMORS OF CRANIAL, SPINAL AND PERIPHERAL NERVES
1. schwannoma (neurilemmoma, neurinoma) (vestibular schwannoma AKA acoustic neuroma, 620)	9560/0
A. cellular	9560/0
B. plexiform	9560/0
C. melanotic	9560/0
2. neurofibroma	9540/0	722
A. plexiform	9550/0	723
3. perineurioma	9571/0	696
A. perineurioma, NOS	9571/0	696
B. malignant perineurioma	9571/3	697
4. malignant peripheral nerve sheath tumor (MPNST) (neurogenic sarcoma, anaplastic neurofibroma, “malignant schwannoma”). Variants:	9540/3
A. epithelioid MPNST	9540/3
B. MPNST with mesenchymal differentiation	9540/3
C. melanotic MPNST	9540/3
D. MPNST with glandular differentiation	9540/3
C. TUMORS OF THE MENINGES
1. tumors of meningothelial cells
A. meningioma. Variants:	9530/0	613
1. meningothelial (WHO I)	9531/0
2. fibrous (fibroblastic) (WHO I)	9532/0
3. transitional (mixed) (WHO I)	9537/0
4. psammomatous (WHO I)	9533/0
5. angiomatous (WHO I)	9534/0
6. microcystic (WHO I)	9530/0
7. secretory (WHO I)	9530/0
8. lymphoplasmacyte-rich (WHO I)	9530/0
9. metaplastic (WHO I)	9530/0
• the following meningiomas exhibit more malignant behavior
10. clear cell (intracranial) (WHO II)	9538/1
11. chordoid (WHO II)	9538/1
12. atypical meningioma (WHO II)	9539/1	616
13. papillary meningioma (WHO III)	9538/3
14. rhabdoid meningioma (WHO III)	9538/3	615
15. anaplastic (malignant) meningioma (WHO III)	9530/3	616
2. mesenchymal, non-meningothelial tumors		620
A. lipoma (e.g. of corpus callosum, page 246)	8850/0
B. angiolipoma	8861/0
C. hiberneuroma	8880/0
D. liposarcoma (intracranial)	8850/3
E. solitary fibrous tumor	8815/0
F. fibrosarcoma	8810/3
G. malignant fibrous histiocytoma	8830/3
H. leiomyoma	8890/0
I. leiomyosarcoma	8890/3
J. rhabdomyoma	8900/0
K. rhabdomyosarcoma	8900/3
L. chondroma	9220/0
M. chondrosarcoma	9220/3
N. osteoma	9180/0
O. osteosarcoma	9180/3
P. osteochondroma	9210/0
Q. hemangioma	9120/0
R. epithelioid hemangioendothelioma	9133/1
S. hemangiopericytoma	9150/1	620
T. anaplastic hemangiopericytoma	9150/3
U. angiosarcoma	9120/3
V. Kaposi sarcoma	9140/3
W. Ewing sarcoma - PNET	9364/3
3. primary melanocytic lesions
A. diffuse melanocytosis	8728/0
B. melanocytoma	8728/1
C. malignant melanoma (primary CNS)	8720/3	697
D. meningeal melanomatosis	8728/3
4. other neoplasms related to the meninges
A. hemangioblastoma	9161/1	667
D. LYMPHOMAS AND HEMATOPOIETIC NEOPLASMS
1. malignant lymphoma (primary CNS lymphoma)	9590/3	672
2. plasmacytoma	9731/3
3. granulocytic sarcoma	9930/3
E. GERM CELL TUMORS
1. germinoma	9064/3	692
2. embryonal carcinoma	9070/3
3. endodermal sinus tumor (EST) (yolk sac tumor)	9071/3
4. choriocarcinoma	9100/3
5. teratoma (from all 3 germ-cell layers)	9080/1	692
A. mature	9080/0
B. immature	9080/3
C. teratoma with malignant transformation	9084/3
6. mixed germ cell tumors	9085/3
F. TUMORS OF THE SELLAR REGION
1. craniopharyngioma. Variants:	9350/1	663
A. adamantinomatous	9351/1
B. papillary	9352/1
2. adenohypophyseal cells → pituitary adenomaΔ	8272/0	633
A. prolactinomaΔ	8271/0	637
B. ACTH secreting adenoma		638
C. growth-hormone secreting adenoma		639
D. thyrotropin (TSH) secreting adenoma		640
E. gonadotropin (LH and/or FSH) secreting adenoma
3. neurohypophysis and infundibulum
A. granular cell tumor	9582/0	641
B. neurohypophyseal cells	9432/1	641
4. pituitary carcinoma		634
5. spindle cell oncocytoma of the adenohypophysis	8291/0
G. METASTATIC TUMORS: those commonly involving brain include:		702
1. lung cancer: especially small-cell (see page 703)		703
2. breast
3. melanoma		704
4. renal cell		705
5. lymphoma
6. GI
H. LOCAL EXTENSIONS FROM REGIONAL TUMORSΔ
1. paraganglioma (chemodectoma)
A. glomus jugulare tumor		680
2. notochord → chordoma		675
3. carcinoma
I. CYSTS AND TUMOR-LIKE LESIONSΔ
1. Rathke’s cleft cyst		665
2. ectodermal rests
A. epidermoid cyst		689
B. cholesteatoma		689
3. dermoid cyst		688
4. colloid cyst of the third ventricle		665
5. neurenteric/enterogenous cyst		227
6. neuroglial cyst
7. hypothalamic neuronal hamartoma		226
8. nasal glial heterotopia
9. plasma cell granuloma
J. UNCLASSIFIED TUMORSΔ

* ICD-O = morphology code of the International Classification of Diseases for Oncology (http://www.iarc.fr/WHO-BlueBooks/). The extension after the slash is the “behavior code”: /0 = benign, /1 = low or uncertain malignant potential or borderline malignancy, /2 = in situ lesions, /3 = malignant tumors

^† represents a significant portion of what is usually considered to be primary brain tumors

^‡ the term “glioma” is occasionally used to refer to all glial tumors (e.g. “low-grade glioma” is often used when discussing low-grade tumors of any glial lineage, see page 590), although in its usual sense glioma (especially “high grade gliomas”) refers only to astrocytic tumors

^§ “WHO II” means World Health Organization (WHO) grade II, “WHO III” means WHO grade III, etc.

Δ these tumors are not part of the 2007 WHO classification⁵

21.1.1. Brain tumors - general clinical aspects

PRESENTATION

The most common presentation of brain tumors is progressive neurologic deficit (68%), usually motor weakness (45%). Headache was a presenting symptom in 54% (see below), and seizures in 26%. For details of presentation, see sections below for supratentorial and infratentorial tumors.

SUPRATENTORIAL TUMORS⁶

Signs and symptoms include:

1. those due to increased ICP (see Infratentorial tumors below):

A. from mass effect of tumor and/or edema

B. from blockage of CSF drainage (hydrocephalus): less common in supratentorial tumors (may occur e.g. with colloid cyst, entrapped lateral ventricle)

2. progressive focal deficits: includes weakness, dysphasia (which occurs in 37-58% of patients with left-sided brain tumors⁷): see below

A. due to destruction of brain parenchyma by tumor invasion

B. due to compression of brain parenchyma by mass and/or peritumoral edema and/or hemorrhage

C. due to compression of cranial nerve(s)

3. headache: see below

4. seizures: not infrequently the first symptom of a brain tumor. Tumor should be aggressively sought in an idiopathic first time seizure in a patient > 20 years (if negative, the patient should be followed with repeat studies at later dates). Rare with posterior fossa tumors or pituitary tumors

5. mental status changes: depression, lethargy, apathy, confusion

6. symptoms suggestive of a TIA (dubbed “tumor TIA”) or stroke, may be due to:

A. occlusion of a vessel by tumor cells

B. hemorrhage into the tumor: any tumor may hemorrhage, see Hemorrhagic brain tumors, page 1123

C. focal seizure

7. in the special case of pituitary tumors (see Pituitary tumors, page 633):

A. symptoms due to endocrine disturbances

B. pituitary apoplexy: see page 635

C. CSF leak

Booking the case - craniotomy for supratentorial tumor

Also see defaults & disclaimers (page v). If awake craniotomy is required, then see page 151.

1. position: (depends on location of tumor)

2. pre-op embolization (by neuroendovascular interventionalist) for some vascular tumors including some meningiomas

3. equipment:

A. microscope

B. ultrasonic aspirator

C. image guidance system

4. blood availability: type and cross 2 U PRBC

5. post op: ICU

6. consent (in lay terms for the patient - not all-inclusive):

A. procedure: surgery through the skull remove as much of the tumor as is safely possible

B. alternatives: nonsurgical management, radiation therapy for some tumors

C. complications: (usual craniotomy complications - see page v) plus inability to remove all of the tumor

INFRATENTORIAL TUMORS

Seizures are rare (unlike supratentorial tumors, (seizures arise from irritation of cerebral cortex).

1. most posterior fossa tumors present with signs and symptoms of increased intracranial pressure (ICP) due to hydrocephalus (HCP). These include:

A. headache: (see below)

B. nausea/vomiting: due either to increased ICP from HCP, or from direct pressure on the vagal nucleus or the area postrema (“vomiting center”)

C. papilledema: estimated incidence is ≈ 50-90% (more common when the tumor impairs CSF circulation)

D. gait disturbance/ataxia

E. vertigo

F. diplopia: may be due to VI nerve (abducens) palsy which may occur with increased ICP in the absence of direct compression of the nerve (see page 836)

2. S/S indicative of mass effect in various locations within the p-fossa

A. lesions in cerebellar hemisphere may cause: ataxia of the extremities, dysmetria, intention tremor

B. lesions of cerebellar vermis may cause: broad based gait, truncal ataxia, titubition

C. brainstem involvement usually results in multiple cranial nerve and long tract abnormalities, and should be suspected when nystagmus is present (especially rotatory or vertical)

Booking the case - craniotomy for infratentorial tumor

Also see defaults & disclaimers (page v). For retromastoid surgery for vestibular schwannomas, see page 629.

1. position: (typically either prone or park bench, depending on tumor type/location and surgeon preference)

2. pre-op embolization (by neuroendovascular interventionalist) for some vascular tumors such as hemangioblastoma

3. equipment:

A. microscope

B. ultrasonic aspirator

C. image guidance system (optional)

4. blood availability: type and cross 2 U PRBC

5. post op: ICU

6. consent (in lay terms for the patient - not all-inclusive):

A. procedure: surgery through the skull remove as much of the tumor as is safely possible

B. alternatives: nonsurgical management, radiation therapy for some tumors

C. complications: (usual craniotomy complications - see page v) plus inability to remove all of the tumor, hydrocephalus, CSF leak

FOCAL NEUROLOGIC DEFICITS ASSOCIATED WITH BRAIN TUMORS

In addition to nonfocal signs and symptoms (e.g. seizures, increased ICP…), as with any destructive brain lesion tumors may produce progressive deficits related to the function of the involved brain. Some characteristic “syndromes”:

1. frontal lobe: abulia, dementia, personality changes. Often nonlateralizing, but apraxia, hemiparesis or dysphasia (with dominant hemisphere involvement) may occur

2. temporal lobe: auditory or olfactory hallucinations, déja vu, memory impairment. Contralateral superior quadrantanopsia may be detected on visual field testing

3. parietal lobe: contralateral motor or sensory impairment, homonymous hemianopsia. Agnosias (with dominant hemisphere involvement) and apraxias may occur (see Clinical syndromes of parietal lobe disease, page 113)

4. occipital lobe: contralateral visual field deficits, alexia (especially with corpus callosum involvement with infiltrating tumors)

5. posterior fossa: (see above) cranial nerve deficits, ataxia (truncal or appendicular)

HEADACHES WITH BRAIN TUMORS

Headache (H/A) may occur with or without elevated ICP. Present equally in patients with primary or metastatic tumor (≈ 50% of patients⁸). Classically described as being worse in the morning (possibly due to hypoventilation during sleep), this may actually be uncommon⁸. Often exacerbated by coughing, straining, or (in 30%) bending forward (placing head in dependent position). Associated with nausea and vomiting in 40%, may be temporarily relieved by vomiting (possibly due to hyperventilation during vomiting). These features along with the presence of a focal neurologic deficit or seizure were thought to differentiate tumor H/A from others. However, H/A in 77% of brain tumor patients were similar to tension H/A, and in 9% were migraine-like⁸. Only 8% showed the “classic” brain tumor H/A, two thirds of these patients had increased ICP.

Etiologies of tumor headache: The brain itself is not pain sensitive. H/A in the presence of brain tumor may be due to any combination of the following:

1. increased intracranial pressure (ICP): which may be due to

A. tumor mass effect

B. hydrocephalus (obstructive or communicating)

C. mass effect from associated edema

D. mass effect from associated hemorrhage

2. invasion or compression of pain sensitive structures:

A. dura

B. blood vessels

C. periosteum

3. secondary to difficulty with vision

A. diplopia due to dysfunction of nerves controlling extra-ocular muscles

1. direct compression of III, IV, or VI

2. abducens palsy from increased ICP (see diplopia on page 586)

3. internuclear ophthalmoplegia due to brainstem invasion/compression

B. difficulty focusing: due to optic nerve dysfunction from invasion/compression

4. extreme hypertension resulting from increased ICP (part of Cushing’s triad)

5. psychogenic: due to stress from loss of functional capacity (e.g. deteriorating job performance)

FAMILIAL SYNDROMES

Several familial syndromes are associated with CNS tumors as shown in Table 21-2 (with page number locations).

Turcot syndrome¹⁰: a rare inherited disorder characterized by multiple colorectal neoplasms (carcinomas or benign adenomatous polyps) together with neuroepithelial tumors of the CNS (GBM, AA, MB, pineoblastoma, ganglioglioma & ependymoma). Type 1: GBM without familial polyposis (but often with nonpolyposis colorectal cancer). Mean survival of Turcot patients with GBM is 27 months (longer than sporadic cases). Type 2: MB & familial adenomatous polyposis.

Li-Fraumeni syndrome: rare (< 400 families identified) inherited autosomal dominant mutation of the TP53 tumor suppressor gene. Patients have increased incidence of multiple types of tumors, including: sarcoma & osteosarcoma, breast cancer, astrocytoma and PNET, adrenocortical carcinoma, leukemia.

Table 21-2 Familial syndromes associated with CNS tumors

Syndrome	Page	CNS tumor
von Hippel-Lindau	667	hemangioblastoma
tuberous sclerosis	725	subependymal giant cell astrocytoma
neurofibromatosis type I	723	optic glioma, astrocytoma, neurofibroma
neurofibromatosis type II	724	vestibular schwannoma, meningioma, ependymoma, astrocytoma
Turcot syndrome (BTP syndrome)9	588	GBM, AA, & medulloblastoma, pineoblastoma
Li-Fraumeni	588	astrocytoma, PNET
Cowden	593	meningiomas
Lhermitte-Duclos	593

STEROID USE IN BRAIN TUMORS

The beneficial effect of steroids in metastatic tumors is often much more dramatic than with primary infiltrating gliomas.

Dexamethasone (Decadron®) dose for brain tumors (see page 33 for cautions):

• for patients not previously on steroids:

adult: 10 mg IVP loading, then 6 mg PO/IVP q 6 hrs^{11, 12}. In cases with severe vasogenic edema, doses up to 10 mg q 4 hrs may be used

peds: 0.5-1 mg/kg IVP loading, then 0.25-0.5 mg/kg/d PO/IVP divided q 6 hrs. NB: avoid prolonged treatment because of growth suppressant effect

• for patients already on steroids:

for acute deterioration, a dose of approximately double the usual dose should be tried

for “stress doses”, see page 32

PROPHYLACTIC ANTICONVULSANTS WITH BRAIN TUMORS

PRACTICE GUIDELINE 21-1 PROPHYLACTIC ANTICONVULSANTS WITH BRAIN TUMORS

Level I¹³: prophylactic AEDs should not be used routinely in patients with newly diagnosed brain tumors

Level II¹³: in patients with brain tumors undergoing craniotomy, prophylactic AEDs may be used, and if there has been no seizure, it is appropriate to taper off AEDs starting 1 week post-op

20-40% of patients with a brain tumor will have had a seizure by the time their tumor is diagnosed¹³. Antiepileptic drugs (AEDs) are indicated in these patients. 20-45% more will ultimately develop a seizure¹³. ProphylacticAEDs do not provide substantial benefit (reduction of risk > 25% for seizure-free survival) and there are significant risks involved.

CHEMOTHERAPY FOR BRAIN TUMORS

Some agents used for CNS tumors are shown in Table 21-3^{14, 15}.

Nitrosoureas: Excellent BBB penetration (see below). Significant hematopoietic, pulmonary and renal toxicity.

Blood-brain barrier (BBB):

Traditionally, the BBB has been considered to be a major hindrance to the use of chemotherapy for brain tumors. In theory, the BBB effectively excludes many chemotherapeutic agents from the CNS, thereby creating a “safe haven” for some tumors, e.g. metastases. This concept has been challenged¹⁶. Regardless of the etiology, the response of most brain tumors to systemic chemotherapy is usually very modest, with a no-table exception being a favorable response of oligodendrogliomas (see page 611). Considerations regarding chemotherapeutic agents in relation to the BBB include:

1. some CNS tumors may partially disrupt the BBB, especially malignant gliomas¹⁷

2. lipophilic agents (e.g. nitrosoureas) may cross the BBB more readily

3. selective intraarterial (e.g. intracarotid) injection¹⁸: produces higher local concentration of agents which increases penetration of the BBB, with lower associated systemic toxicities than would otherwise occur

4. the BBB may be iatrogenically disrupted (e.g. with mannitol) prior to administration of the agent

5. the BBB may be bypassed by intrathecal administration of agents via LP or ventricular access device (e.g. methotrexate for CNS lymphoma, see page 675)

6. biodegradable polymer wafers containing the agent may be directly implanted (see page 601)

Table 21-3 Chemotherapeutic agents used for CNS tumors

Agent	Mechanism
1. nitrosoureas: BCNU (carmustine), CCNU (AKA lomustine), ACNU (nimustine)	DNA crosslinks, carbamoylation of amino groups
2. alkylating (methylating) agents (procarbazine, temozolomide - see page 602)	DNA alkylation, interferes with protein synthesis
3. carboplatin, cisplatin	chelation via intrastrand crosslinks
4. nitrogen mustards: cyclophosphamide, isofamide, cytoxan	DNA alkylation, carbonium ion formation
5. vinca alkaloids: vincristine, vinblastine, paclitaxel	microtubule function inhibitors
6. epidophyllotoxins (ETOP-oside, VP16, teniposide, VM26)	topoisomerase II inhibitors
7. topotecan, irinotecan (CPT-11)	topoisomerase I inhibitors
8. tamoxifen	protein kinase C inhibitor
9. bevacizumab (Avastin®)	anti-VEGF antibody may be useful in vestibular neuromas (see page 624)
10. hydroxyurea 11. bleomycin 12. taxol (paxlitaxol) 13. methotrexate 14. cytosine, arabinoside 15. corticosteroids: dexamethasone, prednisone 16. fluorouracil (FU)

CAT SCAN FOLLOWING SURGICAL REMOVAL OF TUMOR

To assess degree of tumor removal, a post-op brain CT without and with contrast should either be obtained within 2-3 days¹⁹, or should be delayed at least ≈ 30 days. The non-contrast scan is important in the early post-op period to determine which areas of increased intensity are due to post-op blood and not enhancement. The contrast CT demonstrates areas of enhancement, which may represent residual tumor. After ≈ 48 hours, contrast enhancement due to post-operative inflammatory vascular changes ensues, which may be impossible to differentiate from tumor. This usually subsides by ≈ 30 days²⁰, but may persist for 6-8 weeks²¹. This recommendation regarding the timing of post-op CT does not apply to pituitary tumors (see Pituitary tumors, page 633). The effect of steroids on contrast enhancement is controversial^{22, 23}, and may depend on many factors (including tumor type).

POSTERIOR FOSSA (INFRATENTORIAL) TUMORS

See Posterior fossa lesions on page 1209 for differential diagnosis (includes non-neo-plastic lesions as well).

EVALUATION

In pediatric patients with a posterior fossa tumor, an MRI of the lumbar spine should be done pre-op to rule-out drop mets (post-op there may be artifact from blood).

In adults, most intraparenchymal p-fossa tumors will be metastatic, and work-up for a primary should be undertaken (see page 706).

TREATMENT OF ASSOCIATED HYDROCEPHALUS

In cases with hydrocephalus at the time of presentation, some authors advocate initial placement of VP shunt or EVD prior to definitive surgery (waiting ≈ 2 wks before surgery) because of possibly lower operative mortality²⁴. Theoretical risks of using this approach include the following:

1. placing a shunt is generally a lifelong commitment, whereas not all patients with hydrocephalus from a p-fossa tumor will require a shunt

2. possible seeding of the peritoneum with malignant tumor cells e.g. with medulloblastoma. Consider placement of tumor filter (may not be justified given the high rate of filter occlusion and the low rate of “shunt metastases”²⁵)

3. some shunts may become infected prior to the definitive surgery

4. definitive treatment is delayed, and the total number of hospital days may be increased

5. upward transtentorial herniation (see page 285) may occur if there is excessively rapid CSF drainage

Either approach (shunting followed by elective p-fossa surgery, or semi-emergent definitive p-fossa surgery) is accepted. At Children’s Hospital of Philadelphia, dexamethasone is started and the surgery is performed on the next elective operating day, unless neurologic deterioration occurs necessitating emergency surgery²⁶.

Many surgeons place a ventriculostomy at the time of surgery (see page 156). CSF is drained only after the dura is opened to help equilibrate the pressures between the infra- and supratentorial compartments. Post-op, the external ventricular drain is usually set at a low height (≈ 10 cm above the EAM) for 24 hours, and is progressively raised over the next 48 hrs and should be D/C’d by ≈ 72 hrs post-op.

21.2. Primary brain tumors

21.2.1. Low grade gliomas

This special section is included here because the following tumors are sometimes grouped together despite the fact that they have different cell lineages.

Cell lineages considered under the heading of low-grade gliomas (LGG) include:

1. WHO grade II infiltrating astrocytoma (fibrillary or protoplasmic) (see page 595) } comprise most low grade gliomas in adults

2. oligodendroglia (see page 609) } comprise most low grade gliomas in adults

3. mixed astrocytes & oligodendroglia (oligoastrocytoma) } comprise most low grade gliomas in adults

4. gangliogliomas (see page 677) } less frequent histologies

5. gangliocytomas } less frequent histologies

6. juvenile pilocytic astrocytoma (see page 603)} less frequent histologies

7. pleomorphic xanthoastrocytomas (see below)} less frequent histologies

8. dysembryoplastic neuroepithelial tumors (DNT) (see below)} less frequent histologies

Spatial definition^{27, 28}

Can be used to classify LGG into 3 types (independent of histologic group).

• Type 1: solid tumor only without infiltration of brain parenchyma. Most amenable to surgical resection. Most favorable prognosis. Includes gangliogliomas, pilocytic astrocytomas, pleomorphic xanthoastrocytomas, and some protoplasmic astrocytomas (no oligodendrogliomas are in this group)

• Type 2: solid tumor associated with surrounding tumor-infiltrated brain parenchyma. Surgical resection may be possible, depending on tumor location. Often low-grade astrocytomas

• Type 3: infiltrative tumor cells without solid tumor tissue. Risk of neurologic deficit may preclude surgical resection. Usually oligodendrogliomas

Clinical

Although there are differences among the specific histological types, these tumors generally occur in young adults or children, and are often diagnosed after a history of seizures.

Neuroradiology

MRI: Most LGG are hypointense on T1WI. T2WI shows high signal changes that extend beyond the tumor volume demonstrated on other imaging sequences²⁸. Only ≈ 30% enhance.

PET scans: usually shows reduced uptake of fluorodeoxyglucose compared to the rest of the brain, indicative of hypometabolism.

Diagnosis

Although imaging (and clinical) characteristics may suggest one specific tumor type, biopsy is usually required to definitively determine the diagnosis.

Treatment

Complete surgical excision is often sufficient for some of these tumors when it can be accomplished (e.g. with cystic cerebellar pilocytic astrocytomas (PCAs)). When this is not possible (e.g. with most hypothalamic PCAs and PCAs involving the optic nerves and chiasm), then further therapy is required, usually in the form of chemotherapy for younger children²⁹ (to defer the need for XRT until the patient is as old as possible).

DYSEMBRYOPLASTIC NEUROEPITHELIAL TUMORS (DNT) OR (DNET)^{30, 31}

Epidemiology

Incidence: not accurately known because the diagnosis may be missed. Estimated range: 0.8-5% of all primary brain tumors. Typically occurs in children and young adults.

Most common locations: temporal or frontal. Parietal and especially occipital lobe involvement is rare. DNTs have been reported in the cerebellum, pons & basal ganglia.

Clinical

Typically associated with longstanding medically intractable seizures, usually complex partial. Symptoms usually begin before age 20.

Imaging

Cortical lesions with no surrounding edema and no midline mass effect.

CT: hypodense with distinct margins. Deformity of overlying calvaria is common.

MRI: T1WI: hypointense. T2WI: hyperintense, septations may be seen. If there is enhancement, it is usually nodular.

PET scan: hypometabolic with [18F]-fluorodeoxyglucose. Negative [11C]-methionine uptake (unlike all other gliomas).

Pathology

A WHO Grade I glioma. Thought to arise embryologically from the secondary germinal layer (which includes subependymal layer, cerebellar external granular layer, hippocampal dentate fascia & subpial granular layer).

Multinodularity at low-power is a key feature, and the primary constituent cells are oligodendrocytes and to a lesser extent, astrocytes that are often pilocytic. Occasionally difficult to differentiate from oligodendroglioma.

Two distinct forms³² (do not appear to have different prognoses):

1. simple form: glioneural elements consisting of axon bundles perpendicular to the cortical surface, lined with oligodendroglial-like cells that are S-100 positive and GFAP negative. Normal appearing neurons floating in a pale eosinophilic matrix are scattered between these columns (no resemblance to ganglion cells, unlike gangliogliomas)

2. complex form: glioneural elements as described above in the simple form, with glial nodules scattered throughout. The glial component may mimic a low-grade fibrillary astrocytoma. Foci of cortical dysplasia occur

Outcome

Seizure control: usually improves after surgery. Degree of control seems to correlate with completeness of removal. Improvement in seizures correlates inversely with the duration of intractable seizures.

Recurrence/continued growth: recurrence after complete removal, or tumor growth after partial resection is rare. Adjuvant treatment (XRT, chemotherapy…) is of no benefit. Mitoses or endothelial proliferation, seen on occasion, do not affect outcome. Malignant transformation is very rare.

PLEOMORPHIC XANTHOASTROCYTOMA (PXA)

Key concepts:

• low-grade glioma, possibly from subpial astrocytes → superficial location, > 90% supratentorial, most common in children or young adults

• mural nodule with cystic component in 25%, meninges involved in > 67%

• pathology: pleomorphic cells (xanthomatous (lipid laden) cells, fibrillary and giant multinucleated astrocytes). Usually circumscribed, occasionally invasive

• WHO grade II unless high mitotic index or necrosis, which is WHO grade III

• treatment: maximal safe resection. XRT or chemo ≈ only for grade III

A low-grade glioma thought to arise from subpial astrocytes which may explain their superficial location and abundance of reticulin fibers. Over 90% are supratentorial. Predilection for temporal lobes (50%), followed by parietal, occipital & frontal lobes. Most have a cystic component (may be multiloculated, but > 90% have a large, single cyst).

Epidemiology:≈ 1% of astrocytomas. Usually occurs in children or young adults (most are < 18 years age). No gender difference.

Clinical: Usual presentation: seizures. May also produce focal deficit or increased ICP.

Imaging: The cyst, when present, may partially enhance on CT or MRI. A mural nodule is present in 25%. May have “dural tail” (67% show leptomeningeal involvement, 13% show involvement of all 3 meningeal layers). Peritumoral edema may be mild to moderate, calcifications are rare³³.

CT: solid portion of tumor is ill-defined and may be isodense to grey matter.

MRI: T1WI: hypointense cystic component with ill defined isointense solid component that strongly enhances with gadolinium. T2WI: hyperintense cystic component with ill-defined isointense solid component.

Pathology: WHO grade II (MIB is usually < 1%) unless there is a high mitotic index or necrosis which qualifies as WHO grade III “PXA with anaplastic features”³⁴. Compact, superficial tumor with marked cellular pleomorphism (fibrillary and giant multinucleated astrocytes, large xanthomatous (lipid laden) GFAP staining cells (bespeaking glial origin)), abundant reticulin and frequent perivascular chronic inflammatory cells. The reticulin fibers surround two cell types:

1. spindle cells: fusiform cell shape with elongate nuclei

2. pleomorphic cells: round cells with heterochromic, pleomorphic nuclei that may be mononucleated or multinucleated. Variable intracellular lipid content

Usually circumscribed, occasionally infiltrates cortex. Marked cellular pleomorphism may cause these tumors to be mistaken for anaplastic astrocytoma. Vascular proliferation and necrosis are absent³⁵, most but not all lack mitotic figures. Some PXAs undergo anaplastic change³⁶. There have also been several reported cases of malignant transformation to anaplastic astrocytoma or glioblastoma³⁷.

Differential diagnosis:

1. imaging: meningioma is also superficial with dural tail, may also resemble low grade fibrillary astrocytoma

2. pathology: may be confused with anaplastic astrocytoma

Treatment:

1. surgery: primary treatment

A. gross total resection if it can be accomplished without unacceptable neurologic deficit, otherwise subtotal resection

B. extent of resection: most strongly associated with recurrence free survival³⁸

C. incomplete resections should be followed since these tumors may grow very slowly over many years before retreatment is necessary, and repeat excision should be considered

2. radiation therapy: controversial

A. literature suggests either no difference in overall survival or possibly a trend toward prolonged survival³⁵

B. considered with: residual disease, high mitotic index, or necrosis

3. chemotherapy: role not defined

Prognosis: Overall survival with gross total resection or subtotal resection, with or without radiation and chemotherapy: 5 years = 80%, 10 years = 71%³⁴.

Extent of resection, mitotic index, and necrosis appear to be the best predictors of outcome^{33, 38}.

DYSPLASTIC GANGLIOCYTOMA OF CEREBELLUM (LHERMITTE-DUCLOS DISEASE)

AKA: ganglioneuroma of the cerebellum, purkinjoma, granular cell hypertrophy of the cerebellum, gangliocytoma dysplasticum, hamartoma of the cerebellum.

Rare (200 case reports³⁹) cerebellar lesion with features of both a malformation and a low grade (WHO I) neoplasm that has the propensity to progress (enlarge) and recur after surgery. May be focal or diffuse. Diffuse enlargement of cerebellar folia.

Strongly associated with Cowden syndrome: AKA multiple hamartoma syndrome. Autosomal dominant. Incidence: 1 in 250,000 live births⁴⁰. Associated with thyroid, breast & uterine Ca, mucosal neuromas & meningiomas).

Histology: Derangement of normal laminar cellular architecture of the cerebellum with:

1. thickening of the outer molecular cell layer

2. loss of middle Purkinje cell layer

3. infiltration of inner granular cell layer with dysplastic ganglion cells

Clinical: Typically a middle aged adult with signs and symptoms of a cerebellar mass. May also present with hydrocephalus or may be an incidental finding.

Imaging:

CT: hypo- to isodense, nonenhancing lesion with mass effect.

MRI: T1WI: hypo- to isointense. T2WI: hyperintense, heterogeneous. Nonenhancing. Characteristic striated appearance⁴¹ (tiger stripes) due to widened cerebellar folia. May contain calcifications. DWI: hyperintense. ADC map: hypointense.

NB: in a child with MRI findings of Lhermitte-Duclos disease (LDD) (even if classic), a medulloblastoma is statistically more likely^{42, 43} (especially medulloblastoma with extensive nodularity⁴⁴ (MBEN)).

Treatment: Controversial. A few cases with a benign course have been described⁴⁵. Shunting for hydrocephalus. Biopsy is recommended⁴³ particularly for pediatric cases to rule-out medulloblastoma. Surgical excision may be considered when there is significant mass effect⁴⁶. Efficacy of XRT is unknown.

21.2.2. Astrocytoma

The most common primary intra-axial brain tumor, ≈ 12,000 new cases/year in the U. S.

CLASSIFICATION BY CELL TYPE

The dominant cell types of astrocytomas allows their classification into one of the subdivisions shown in Table 21-4. The rationale for separating “ordinary” from “special” astrocytomas is based on a much different and more favorable behavior of the latter group which does not depend on grade within that group (these also tend to occur in younger patients). The notion that pilocytic and microcystic cerebellar astrocytomas are the same tumor as fibrillary astrocytomas but in a different location has been abandoned (see Pilocytic astrocytomas, page 603).

Table 21-4 Classification of astrocytomas by cell type

“Ordinary” astrocytomas	“Special” astrocytomas
fibrillary gemistocytic protoplasmic	pilocytic microcystic cerebellar subependymal giant cell

Gemistocytic astrocytomas: Gemistocytes are plump cells filled with eosinophilic, hyaline cytoplasm, seen almost exclusively in gemistocytic astrocytomas and GBM. Small numbers, however, may be seen in fibrillary astrocytomas (gemistocytes should account for > 20% of tumor cells for an astrocytoma to be considered a gemistocytic astrocytoma). Gemistocytic astrocytomas are comprised primarily of these cells, but rarely occur in pure form. Often meet grade III (malignant astrocytoma) criteria.

21.2.2.1. “Ordinary” astrocytomas

“Ordinary” here is meant to encompass Grade II-IV infiltrating astrocytomas, and to exclude special more-circumscribed astrocytomas such as pilocytic astrocytoma.

GRADING AND NEUROPATHOLOGY

Grading of astrocytomas remains controversial. Some special concerns:

1. sampling error: may have different degrees of malignancy in different areas

2. dedifferentiation: tumors tend to progress in malignancy over months or years (see Dedifferentiation, page 595)

3. histological criteria that affect prognosis include: cellularity, presence of giant cells, anaplasia, mitosis, vascular proliferation with or without endothelial proliferation, necrosis, and pseudopalisading⁴⁷

4. in addition to histology, issues that affect clinical behavior include:

A. patient age

B. extent of tumor

C. topography: tumor location, especially in relation to critical structures

NEUROPATHOLOGICAL GRADING

Kernohan system

The obsolete Kernohan system⁴⁹ divided these tumors into 4 grades (grade IV AKA glioblastoma multiforme) based on the degree of presence of a number of features such as anaplasia, nuclear pleomorphism, number of mitoses… Prognostically, this system distinguished only 2 clinically different groups (grades I/II, and grades III/IV) and is not used today. It is presented for completeness when reviewing older literature.

Table 21-5 Approximate equivalence of Kernohan grade (I-IV) to WHO system

Kernohan	WHO designation48
	I more circumscribed tumors: e.g. pilocytic astrocytomas
I II	} II: diffuse astrocytoma (low-grade)
III	III: anaplastic astrocytoma	} malignant astrocytoma
IV	IV: glioblastoma

Current grading systems

The 2 main systems is use today are shown below, and differ primarily in the definition of Grade I.

WHO system: The World Health Organization (WHO) system is shown in Table 21-5⁴⁸. In the WHO system, grade I is reserved for special types of astrocytomas that are more circumscribed, including pilocytic astrocytomas (see Low grade gliomas, page 590), while the more typical astrocytic neoplasms are graded II through IV. The approximate equivalence to the Kernohan grade is also shown.

St. Anne/Mayo grading system: The classification system known as the St. Anne/Mayo (SA/M) system⁵⁰ addresses histological considerations, and is reproducible and prognostically significant⁵¹. It is restricted to “ordinary” astrocytomas, as grade has not been shown to correlate with clinical behavior in pilocytic astrocytomas. It is similar to the WHO system except that SA/M grade I astrocytomas are a very rare diffuse astrocytoma without atypia⁴⁸.

Table 21-6 WHO classification of (“ordinary”) astrocytic tumors

Designation	Criteria
II: diffuse astrocytoma	cytological atypia alone
III: anaplastic astrocytoma	anaplasia and mitotic activity
IV: glioblastoma (GBM)	also show microvascular proliferation and/or necrosis

The SA/M system assesses the presence or absence of 4 criteria (see Table 21-7) and then assigns a grade based on the number of criteria present (see Table 21-8). When the presence of any criteria is uncertain, it is considered to be absent.

The criteria tended to occur in a predictable sequence: nuclear atypia occurred in all grade 2 tumors, mitotic activity was seen in 92% of grade 3 tumors (and in none of the grade 2 tumors), necrosis and endothelial proliferation were restricted almost only to grade 4 tumors (they were seen in only 8% of grade 3 tumors).

The frequencies among 287 astrocytomas were: grade 1 = 0.7%, grade 2 = 16%, grade 3 = 17.8%, and grade 4 = 65.5%.

Median survival was as follows⁵⁰: (there were only two grade 1 patients, one survived 11 years and the other was still alive after 15 years), grade 2 = 4 years, grade 3 = 1.6 years, and grade 4 = 0.7 years (8.5 months).

Table 21-7 St. Anne/Mayo criteria

• nuclear atypia: hyperchromatasia and/or obvious variation in size and shape

• mitoses: normal or abnormal configuration

• endothelial proliferation: vascular lumina are surrounded by “piled-up” endothelial cells. Does not include hypervascularity

• necrosis: only when obviously present. Does not include pseudopalisading when seen alone

Relative frequency of astrocytoma grades

Ratio of (glioblastoma):(anaplastic astrocytoma):(low-grade astrocytoma) is ≈ 5:3:2. Peak age incidence rises with increasing grade: 34 years for low-grade astrocytoma, 41 years for anaplastic astrocytoma, and 53 years for GBM⁵².

Low-grade astrocytoma (WHO II)

AKA low-grade diffuse astrocytoma. Three cell types:

1. fibrillary astrocytoma: the most common histological subtype of Grade II

2. gemistocytic astrocytoma: particularly prone to progress to Grade II & IV

3. protoplasmic astrocytoma

Table 21-8 St. Anne/Mayo grade

Grade	No. of criteria
1	0
2	1
3	2
4	3 or 4

These tumors tend to occur in children and young adults. Most present with seizures. There is a predilection for temporal, posterior frontal and anterior parietal lobes⁵³. They demonstrate low degrees of cellularity and preservation of normal brain elements within the tumor. Calcifications are rare. Anaplasia and mitoses are absent (a single mitosis is allowed). Blood vessels may be slightly increased in number. The ultimate behavior of these tumors is usually not benign. The most important favorable prognosticator is young age. Poor prognosis is associated with findings of increased ICP, altered consciousness, personality change, significant neurologic deficits⁵⁴, short duration of symptoms before diagnosis, and enhancement on imaging studies. Also see page 590.

Dedifferentiation: The major cause of morbidity with low-grade astrocytomas is dedifferentiation to a more malignant grade. Low grade fibrillary astrocytomas tend to undergo malignant transformation more quickly (with six-fold increased rapidity) when diagnosed after age 45 years than when diagnosed earlier⁵² (see Table 21-9). Gemistocytic astrocytomas tend to dedifferentiate more rapidly than fibrillary astrocytomas. > 60% of fibrillary astrocytomas have a mutation of the TP53 gene located on chromosome 17p; these tumors are more likely to dedifferentiate. Once dedifferentiation occurs, median survival is 2-3 years beyond that event. Genetic markers that correlate with a higher degree of malignant degeneration include:

1. loss of heterozygosity on chromosomes 10 & 17

2. alteration in tumor suppressor genes at 9p, 13q, 19q & 22q

3. changes in epidermal growth factor receptor (EGRF) and platelet-derived growth factor (PDGF)

4. transformation of the p⁵³ suppressor gene

Table 21-9 Dedifferentiation rate for low grade astrocytomas

	Patients diagnosed @ age < 45 yrs	Patients diagnosed @ age≥45 yrs
mean time to dedifferentiation	44.2 ± 17 mos	7.5 ± 5.7 mos
time to death	58 mos	14 mos

Malignant astrocytomas (WHO III & IV)

This category encompasses anaplastic astrocytoma (AA) and glioblastoma (GBM). Although both are “malignant”, AA and GBM have distinct differences. Among 1265 patients with malignant astrocytomas, the mean age was 46 yrs for AA, and 56 yrs for GBM. Mean duration of symptoms pre-op: 5.4 mos for GBM, and 15.7 mos for AA. Malignant astrocytomas may develop from low grade astrocytomas via dedifferentiation (see above), however they may also arise de novo.

Glioblastoma (multiforme) (WHO IV): The most common primary brain tumor, it is also the most malignant astrocytoma. Current nomenclature omits “multiforme”⁵.

Primary vs. secondary glioblastoma²: most GBM arise de novo (primary), others progress from less malignant astrocytomas (secondary). Although they evolve from different genetic precursors, there is no reliable distinguishing histopathologic marker, and the difference in prognosis and response to different therapies is unknown.

• primary glioblastoma: the majority of GBMs. Arise without evidence (clinical or histological) of a less malignant precursor. More common in older patients (mean age = 55 years) after a short (< 3 month) clinical history. Characterized by EGFR amplification (≈ 40% of cases) and/or overexpression (60%), PTEN mutations (30%), p¹⁶INK4a deletion (30–40%), MDM2 amplification (< 10%), and/or overexpression (50%), and in 50–80% of cases, loss of heterozygosity (LOH) on the entire chromosome 10

• secondary glioblastoma: develop by malignant degeneration of WHO grade II or III astrocytoma. Patients are younger (mean age = 40 years) and have a slower clinical course. 60% have TP53 mutations (> 90% of these are evident in the less malignant precursors). Malignant degeneration is characterized by allelic loss of chromosomes 19q and 10q

Histological findings associated with GBM (not all may be present, and this list does not follow any of the standard grading systems above):

• gemistocytic astrocytes

• neovascularization with endothelial proliferation

• areas of necrosis

• pseudopalisading around areas of necrosis

Infratentorial glioblastoma (GBM) is rare, and often represents subarachnoid dissemination of a supratentorial GBM (used as an argument for irradiation in all patients with p-fossa GBM)⁵⁵.

MISCELLANEOUS PATHOLOGICAL FEATURES

Glial fibrillary acidic protein (GFAP): Most astrocytomas stain positive for GFAP (however, may not stain positive in some poorly differentiated gliomas, and in purely gemistocytic astrocytomas since fibrillary astrocytes are required to be positive).

Cysts: Gliomas may have cystic central necrosis, but may also have an associated cyst even without necrosis. When fluid from these cysts is aspirated it can be differentiated from CSF by the fact that it is usually xanthochromic and often clots once removed from the body (unlike e.g. fluid from a chronic subdural). Although they may occur with malignant gliomas, cysts are more commonly associated with pilocytic astrocytomas (see page 604).

MIB-1 index: (see page 720). It has been suggested that a MIB-1 index ≥ 7-9% is indicative of an anaplastic tumor, while MIB-1 < 5% favors a low-grade tumor. However, variability between observers and institutions precludes using the MIB-1 index as a sole discriminant between grade II & III astrocytomas².

NEURORADIOLOGICAL GRADING AND FINDINGS

Astrocytomas typically arise in white matter (e.g. centrum semiovale) and traverse through white matter tracts (see below). For MR-spectroscopy findings, see page 133.

Table 21-10 Grading gliomas by CT or MRI

Kernohan grade	Typical radiographic findings
I	CT: low density MRI: abnormal signal on T2WI	no mass effect, no enhancement
II		mass effect, no enhancement
III	complex enhancement*
IV	ring enhancement (central necrosis or cyst)

* however, some may not enhance

CT scan & MRI grading

Grading gliomas by CT or MRI is imprecise⁵⁶, but may be used as a preliminary assessment (see Table 21-10). Neuroradiologic grading is not applicable to pediatric patients or special astrocytomas (e.g. pilocytic astrocytomas).

Low grade gliomas: usually hypodense on CT. Most are hypointense on T1WI MRI, and show high intensity changes on T2WI that extend beyond the tumor volume. Most do not enhance on CT or MRI (although up to 40% do⁵⁷, and these may have a worse prognosis). The UCSF preoperative grading system for low-grade infiltrating gliomas⁵⁸ assigns 1 point for the presence of each of 4 parameters shown in Table 21-11. The points are summed and the prognosis is shown in Table 21-12 (this scale needs to be validated at other institutions). Another study found poor prognosis associated with: age ≥ 40 years, tumor ≥ 6 cm dia, tumor crossing midline and the presence of neurologic deficit⁵⁹.

Table 21-11 Preoperative grading of low-grade gliomas⁵⁸

Item	Yes/No
age > 50 years	Yes = 1, No = 0
KPS* ≤ 80	Yes = 1, No = 0
located in eloquent brain†	Yes = 1, No = 0
maximal diameter > 4 cm	Yes = 1, No = 0

* KPS = Karnofsky performance score (page 1182)

^† for this study, eloquent brain is defined as any of: primary sensory or motor cortex, Wernicke’s or Broca’s area, basal ganglia/internal capsule, thalamus or primary visual cortex

Table 21-12 Sum of points from Table 21-11

Sum	5-year survival	5-year progression-free survival
0-1	97%	76%
2	81%	49%
3-4	56%	18%

Malignant gliomas: anaplastic astrocytomas (AA) may not enhance⁶⁰ (31% of highly anaplastic and 59% of moderately anaplastic astrocytomas do not enhance on CT⁶¹ (MRI not studied)). Calcifications and cysts occur in 10-20% of AA⁶⁰. Most glioblastomas enhance, but some rare ones do not^{56, 61}.

Ring-enhancement with glioblastoma (GBM): The nonenhancing center may represent necrosis or associated cyst (see above). The enhancing ring is cellular tumor, however, tumor cells also extend ≥ 15 mm beyond the ring⁶².

Positron emission tomography (PET) scan

Low grade fibrillary astrocytomas appear as hypometabolic “cold” spots with fluorodeoxyglucose PET scans. Hypermetabolic “hot” spots suggest high-grade astrocytomas.

Angiographic appearance

AA’s usually appear as an avascular mass. Tumor blush and AV-shunting with early draining veins are more characteristic of GBM.

SPREAD

Gliomas may spread by the following mechanisms⁶³ (note: < 10% of recurrent gliomas recur away from the original site⁶⁴):

1. tracking through white matter

A. corpus callosum (CC)

1. through genu or body of CC → bilateral frontal lobe involvement (“butterfly glioma”)

2. through splenium of CC → bilateral parietal or occipital lobes

B. cerebral peduncles → midbrain involvement

C. internal capsule → encroachment of basal ganglion tumors into centrum semiovale

D. uncinate fasciculus → simultaneous frontal and temporal lobe tumors

E. interthalamic adhesion → bilateral thalamic gliomas

2. CSF pathways (subarachnoid seeding): 10-25% frequency of meningeal and ventricular seeding by high grade gliomas⁶⁵

3. rarely, gliomas may spread systemically

MULTIPLE GLIOMAS

Discussion of multiple gliomatous masses has to acknowledge the concept that astrocytoma is a multifocal disease, not a focal one. Some terms are probably artificial, e.g. since gliomatosis cerebri probably represents a diffuse infiltrating glial tumor with areas that may dedifferentiate into higher grade and then is called multicentric glioma.

Settings in which multiple gliomatous masses are encountered:

1. conventional glioma that has spread by one of the mechanisms previously described (see above)

2. gliomatosis cerebri: a diffuse, infiltrating astrocytoma that invades almost all of the cerebral hemispheres and brainstem. Usually low-grade⁵⁷, areas of anaplasia and glioblastoma may also occur⁶⁶ and may present as focal mass⁶⁷. Occurs most frequently in 1st 2 decades

3. multiple primary gliomas: some of the following terms are inconsistently used interchangeably: “multicentric”, “multifocal”, and “multiple”. Reported range of occurrence is 2-20% of gliomas^{68, 69} (lower end of range ≈ 2-4% is probably more accurate, the higher end of the range is probably accounted for by infiltrative extension^{70 (p 3117)})

A. commonly associated with neurofibromatosis and tuberous sclerosis

B. rarely associated with multiple sclerosis and progressive multifocal leukoencephalopathy

4. meningeal gliomatosis: dissemination of glioma throughout the CSF, similar to carcinomatous meningitis (see page 711). Occurs in up to 20% of autopsies on patients with high-grade gliomas. May present with cranial neuropathies, radiculopathies, myelopathy, dementia, and/or communicating hydrocephalus

In a series of 25 patients with multicentric glioma⁷¹, glioblastoma was the most common pathology (48%), followed by anaplastic astrocytoma (20%), and glioblastoma with simultaneous AA (20%).

TREATMENT CONSIDERATIONS FOR MULTIPLE GLIOMAS

There is little data available. In a nonrandomized study of 25 patients with multi-focal glioma⁷¹, the 16 patients who underwent debulking did better than the 9 who did not. However, there was significant selection bias in choosing patients suitable for craniotomy.

Biopsy is generally required/recommended to confirm the diagnosis.

Σ

Once the diagnosis of multiple gliomatous masses has been ascertained, local therapies (e.g. surgery, interstitial radiation…) are impractical. Whole brain radiation and possibly chemotherapy are indicated. An exception would be to consider debulking tumor to prevent herniation in a patient deteriorating from mass effect.

TREATMENT

LOW-GRADE ASTROCYTOMAS (WHO GRADE II)

Treatment options:

1. no treatment: follow serial neurologic exams and imaging studies

2. radiation

3. chemotherapy

4. surgery

5. combinations of radiation and chemotherapy, with or without surgery

Analysis

No well-designed study has shown that any approach for supratentorial WHO grade II infiltrating astrocytomas in adults is clearly superior. Some treatments may simply expose the patient to the risk of treatment side effects. These tumors are slow growing, and until progression on imaging or malignant degeneration is documented, it may be no worse to not treat the patient⁷². Although this view has been challenged⁷³, a definitive study has yet to be performed. The following are associated with more aggressive tumors and should prompt consideration for some form of treatment:

1. extremely young patients, or patients > 50 yrs age (increasing age at diagnosis is associated with more rapid dedifferentiation, see Dedifferentiation, page 595)

2. large tumors that enhance (size is one of the most important prognosticators⁷⁴)

3. symptomatic patients, especially those with short clinical history

4. evidence of progression on imaging studies

Surgery for low grade gliomas

The role of surgery in low-grade gliomas is controversial, due in part to the fact that surgery is not curative for most infiltrating hemispheric gliomas, and many of these tumors are not completely resectable. There is a trend suggesting that “complete” surgical removal, when possible, is associated with a better prognosis^{28, 74}. However, this remains unproven.

Surgery is the principal treatment in the following situations of low-grade astrocytomas:

1. surgical biopsy or partial resection is recommended in almost all cases to establish the diagnosis since clinical and radiographic data are not definitive⁵³

2. pilocytic astrocytomas

A. cerebellar tumors occurring in children & young adults (see page 604)

B. supratentorial pilocytic astrocytomas

3. when herniation threatens from large tumors or tumor cysts

4. tumors causing obstruction of CSF flow

5. may help in seizure control with refractory seizures

6. in an attempt to delay adjuvant therapy and its side-effects in children (especially XRT in those < 5 yrs old)⁵³

7. smaller tumors are less aggressive than large ones⁷⁵ and may be better candi-dates for early surgery (also, see below)

The role of surgery is limited in the following situations of low-grade astrocytomas:

1. disseminated (poorly circumscribed) tumors

2. multifocal tumors

3. location in eloquent brain

Technical considerations at surgery:

Since the margins of low-grade gliomas may not be readily visible at the time of surgery, adjuncts such as stereotactic and image guided techniques may be advantageous for deep tumors or in areas bordering on eloquent brain⁷⁶. Awake surgery is an option for tumors bordering on eloquent brain.

Unresolved issues: whether the extent of tumor removal influences 1) time to tumor progression, 2) incidence of malignant degeneration, and 3) period of survival. One series⁷⁷ suggested that 5-year survival improved from 50% with incomplete resection to 80% with complete resection. Early radical surgery may reduce the rate of malignant degeneration, especially when tumor volume is < 30 ml⁵³.

Radiation therapy (XRT) for low grade gliomas

Background: Early XRT increases progression-free survival (PFS) and seizures, but has no effect on overall survival (OS)⁷⁸. Following incomplete resection, retrospective evidence suggests that PFS and OS are prolonged by XRT⁷⁹. Two prospective trials found no difference in OS or PFS between different XRT doses (EORTC trial⁷⁴: 45 Gy in 5 weeks vs. 59.4 Gy in 6.6 weeks; Intergroup study⁸⁰ 50.4 vs. 64.8 Gy). Side effects from WBXRT include: leukoencephalopathy & cognitive impairment (see Radiation injury and necrosis, page 771). The frequency of side effects may⁸⁰ or may not⁸¹ increase at higher XRT doses.

Recommendations for XRT in low-grade gliomas (modified⁸²):

1. dogmatic statements regarding XRT are unwarranted

2. when considered for use as a primary treatment, XRT may be best reserved for patients who are more likely to progress (older patients, involvement of corpus callosum). Dosage: 45-54 Gy (NCCN guidelines). Expectant management may be a better course for younger patients with asymptomatic lesions

3. in cases of gross total surgical removal, or incomplete removal in cases of pilocytic astrocytoma or cystic cerebellar astrocytoma, XRT may be withheld until tumor recurrence or progression that cannot be treated surgically is documented

4. in cases of incomplete removal of ordinary low-grade astrocytomas, post-op XRT may be considered, consisting of fractionated treatments to a maximum of 45 Gy to the tumor bed plus surrounding margin (2 cm for enhancing, and 1 cm around hypodense zone for nonenhancing tumors) instead of whole brain XRT

5. malignant degeneration of tumor should be treated with XRT, following reoperation when appropriate

Chemotherapy for low grade gliomas

Usually reserved for tumor progression. PCV (procarbazine, CCNU, and vincristine) frequently stabilizes tumor growth. Temozolomide (Temodar®) may be effective in progressive WHO grade II astrocytomas (off label use)⁸³.

MALIGNANT ASTROCYTOMAS (WHO GRADES III & IV)

Stereotactic biopsy

Due to sampling error, stereotactic biopsy may underestimate the occurrence of GBM by as much as 25%.

Indications for stereotactic biopsy (instead of initial resection) in suspected malignant astrocytomas⁸⁴:

1. tumors located in eloquent or inaccessible areas of brain

2. small tumors with minimal deficit

3. patients in poor medical condition precluding general anesthesia

4. to ascertain a diagnosis when one is not definitely established (including when considering a more definitive operation). Some CNS lymphomas mimic GBM radiographically (and without immunostaining, some have also been mistaken pathologically) biopsy should be given serious consideration (to avoid operating on a lymphoma)

Technique: Yield of biopsy is highest when targets within the low density (necrotic) center and enhancing rim are chosen⁶².

Outcome: In a study of 91 cases of malignant gliomas with “critical location” (i.e. deep, midline, or near eloquent brain), it was found that cytoreductive surgery may not improve survival (a limited number of patients underwent cytoreductive surgery with no obvious improvement in survival, but too few to tell if statistically significant), and that biopsy + XRT may be appropriate therapy for these non-lobarmalignant tumors (see Table 21-13). There was no significant difference in survival between AA and GBM when the tumors were not lobar. A Karnofsky rating ≥ 70 at presentation also portends a better prognosis (not statistically significant in this study).

Patients with left-sided tumors and dysphasia are at significant risk of worsening of language function following stereotactic biopsy (the risk of deterioration is low if there is no dysphasia before biopsy)⁸⁵.

Surgery for high grade gliomas

Cytoreductive surgery followed by external beam radiation (40 Gy whole-brain + 15-20 Gy to the tumor bed delivering a total of ≈ 60 Gy to the tumor) has become the standard against which other treatments are compared⁸⁶. Elderly patients (> 65 yrs age): the benefit conferred by surgery is modest (median survival of 17 weeks after biopsy + XRT, versus 30 weeks for surgery + XRT)⁸⁷.

Extent of resection: The extent of tumor removal and (in an inverse relationship) the volume of residual tumor on post-op imaging studies have a significant effect on time to tumor progression and median survival⁸⁸. However, while it has been shown that postop residual enhancing tumor is a marker for a worse prognosis (11.8 months median survival if there was enhancing GBM on post-op MRI, 16.7 months if none⁸⁹), it does not follow and has not been proven that being more aggressive to try and remove that last bit of enhancing tissue improves survival⁹⁰. A small randomized study showed improved survival with resection vs. biopsy in elderly patient with GBM⁹¹. Piepmeir⁹²stated it succinctly, “Ultimately, a significant improvement in survival for patients with malignant gliomas will not result from more extensive surgery…”.

Alternative views suggest that surgery may be justified to reduce significant mass effect but not for reducing tumor burden^{93, 94}. Since these tumors cannot be cured with surgery, the goal should be to prolong quality survival; this can usually be accomplished with tumor excision for lobar gliomas in patients in good neurologic condition.

Partial resection of a GBM carries significant risk of post-operative hemorrhage and/or edema (wounded glioma syndrome) with risk of herniation. Furthermore, the benefit of subtotal resection is dubious. Therefore, surgical excision should only be considered when the goal of gross total removal is feasible.

As a result of the above, the following are usually not candidates for surgical debulking

1. extensive dominant lobe GBM

2. lesions with significant bilateral involvement (e.g. large butterfly gliomas)

3. elderly patients

4. Karnofsky score < 70 (in general, with infiltrating tumors, the neurologic condition on steroids is as good as it is going to get, and surgery rarely improves this)

5. multicentric gliomas

Radiation therapy for high grade gliomas

The usual dose of XRT for malignant gliomas is 50-60 Gy. Whole brain XRT has not been shown to increase survival compared to focal XRT, and the risk of side effects is greater⁹⁵.

Brachytherapy has shown no significant benefit as an adjunct to EBRT in the initial treatment of malignant astrocytomas⁹⁶.

Stereotactic radiosurgery provided no additional benefit when added to conventional XRT and BCNU chemotherapy⁹⁷.

Chemotherapy for high grade gliomas

All agents in use have no more than a 30-40% response rate, and most have 10-20%⁹⁸. Although not positively proven, it appears that the more complete the surgical resection, the more value the chemotherapy has⁹⁸. When given before XRT, chemotherapy may also be useful. There was no survival advantage with combination PCV therapy when added to XRT vs. XRT alone⁹⁹.

Alkylating agents produce significant benefit in ≈ 10% of patients¹⁰⁰ (similar efficacy among all available agents: BCNU, CCNU, procarbazine…). Carmustine (BCNU) (BiCNU®)¹⁰¹ and cisplatinum (AKA cisplatin, Platinol®) have been the primary chemo-therapeutic agents used against malignant gliomas. The response may be enhanced by inhibition (via methylation) of the gene responsible for production of the DNA-repair enzyme O6-methylguanine-DNA methyltransferase (MGMT)¹⁰².

Σ

Following surgery + XRT, median survival is ≈ 9 mos, and 2-year survival is only 5-10%103. Meta-analysis showed an absolute increase in 1-year survival of 6% and a 2-month increase in median survival with chemotherapy. Nitrosoureas are fairly well tolerated and easy to administer. However, the quality of life during this modest increase is uncertain, making chemotherapy an option104.

carmustine (BCNU) (BiCNU®)DRUG INFO

In an attempt to reduce systemic effects, intraarterial injection of carmustine has been tried^{105, 106}, but side effects are significant, including progressive leukoencephalopathy and visual deterioration due to retinal toxicity (attempts to offset this by selectively injecting distal to the ophthalmic artery have been disappointing). BCNU containing wafers may also be surgically implanted following tumor resection (see below).

The only protocol to have been fully validated by Phase 3 study⁹⁸ is maximal surgical resection when possible, followed by XRT of 60 Gy, and then BCNU at 6-week intervals of 110 mg/M².

Implantable chemotherapy:

Gliadel® wafers: carmustine (BCNU) 7.7 mg in a 200 mg prolifeprosan 20 hydrophobic polymer carrier (wafer). Following tumor removal, up to 8 of the 1.4 cm X 1 mm wafers are applied to the resection bed at the time of surgery. The drug is released over ≈ 2-3 wks. This exposes the tumor to 113 times the concentration of BCNU that could be achieved with systemic administration¹⁰⁷. In animals, only trace amounts of the drug reach the systemic circulation. FDA approved in the U.S. for implantation in newly diagnosed and recurrent glioblastoma.

Implanting at initial surgery: median survival increased from 11.6 months to 13.8 months in a series of 240 patients with malignant glioma (207 with GBM)¹⁰⁸ when Gliadel® was added to surgical debulking and XRT, and 2-year survival was 16% vs. 8% in the placebo group.

Recurrent malignant glioma: median survival was 28 weeks with BCNU implants compared to 20 weeks with placebo, and 6 month survival was 64% compared to 44% with placebo¹⁰⁹.

No effect on blood counts occurred. The implants do increase cerebral edema, wound healing problems, and the incidence of seizures within 5 days of surgery. 8 wafers cost ≈ $12,500¹¹⁰.

temozolomide (Temodar® in the U.S., Temodal® globally)DRUG INFO

An oral alkylating agent that is given as a prodrug which undergoes rapid non-enzymatic conversion at physiologic pH to the active metabolite monomethyl triazenoimidazole carboxamide (MTIC). The cytotoxic effect of MTIC is associated with alkylation (methylation) of DNA at various sites including the O6 and N7 positions on guanine.

FDA approved for use in adults for:

• the initial relapse of anaplastic astrocytoma and progression of disease while on a regimen containing a nitrosourea (see Table 21-3, page 589) and procarbazine

• for newly diagnosed GBM: given as a low dose with concurrent XRT, followed by higher maintenance dose (see below)

Has also been used off-label for:

• newly diagnosed AA¹¹¹

• for patients with minimal post-surgery treatment as well as for progressive low grade astrocytomas⁸³

• Phase II trials have been published with oligoastrocytoma¹¹².

Rx for anaplastic astrocytoma: 150 mg/m²/d PO q d x 5 d. Dose for subsequent 5 day cycles every 28 days is adjusted according to nadir neutrophil and platelet counts (which occurs at day 21) during the previous cycle and the start of the next cycle (therefore check CBC on days 21 & 29).

Rx when given concurrently with XRT for newly diagnosed GBM: 75 mg/m²/d orally on an empty stomach x 42 days, 1 hr before XRT (on days without XRT, it is given in the morning)¹¹³. After completion of XRT, it is given orally as 150 mg/m²/d on days 1-5 every 28 days for at least 6 cycles.

SIDE EFFECTS: Most common side-effect is N/V which may be ameliorated by pretreatment with ondansetron (Zofran®) 30 minutes before temozolomide dose. Constipation and fatigue are also common. H/A and seizures have been reported. SUPPLIED: 5, 20, 100, 140, 180 & 250 mg capsules. 100-mg powder for injection. Cost: $1,300-1,500 per cycle.

Reoperation for recurrence

Less than 10% of recurrent gliomas recur away from the original tumor site⁶⁴. Re-operation extends survival by an additional 36 weeks in patients with GBM, and 88 weeks in AA^{114, 115} (duration of high quality survival was 10 weeks and 83 weeks respectively, and was lower with pre-op Karnofsky score < 70). In addition to Karnofsky score, significant prognosticators for response to repeat surgery include: age and time from the first operation to re-operation (shorter times → worse prognosis)¹⁰⁹. Morbidity is higher with reoperation (5-18%); the infection rate is ≈ 3 x that for first operation, wound dehiscence is more likely.

OUTCOME

Survival with various grades of astrocytoma

In general, with “optimal treatment” the survival of the various grades of astrocytoma are approximately given in Table 21-14 (more details may be found in other sections - also see below for recursive partitioning analysis (RPA) for GBM).

Low-grade astrocytomas (WHO grade II)

For low grade infiltrating gliomas, see page 597 for prognosis based on pre-op grading.

Table 21-14 Median survival for astrocytomas

Grade	Median survival
I	8-10 yrs
II	7-8 yrs
III	≈ 2-3 yrs
IV	< 1 yr

Malignant astrocytomas (WHO grades III & IV)

The following 3 statistically independent factors affect longevity:

1. patient age: consistently found to be the most significant prognosticator, with younger patients faring better. With GBM, 18 month survival is 50% for patients < 40 yrs, 20% for ages 40-60, and 10% for age > 60116

2. histological features: median survival is 36 mos for AA, and 10 mos for GBM¹¹⁶ (also, see below)

3. performance status (e.g. Karnofsky score (KPS) see page 1182) at presentation:

A. with GBM, 18 month survival is 34% for KPS > 70, vs. 13% for KPS < 60116

B. 5-year survival: 7.6% with KPS ≥ 70 pre-op, vs. 3.2% for KPS < 70⁶ With AA, smaller size and frontal location influence survival favorably.

Survival differences between AA and GBM:

Two large studies treated malignant gliomas by surgical resection, 60 Gy whole brain irradiation, and then various chemo-therapy regimens (BCNU, procarbazine, methylprednisolone…) resulted in the survival statistics shown in Table 21-15.

Recursive partitioning analysis (RPA) with glioblastoma

Analysis of 832 patients with GBM⁸¹⁸ identified 4 risk groups shown in Table 21-16.

For example, group 1 (low risk) consists of patient age ≤ 40 years AND who only have tumor located in frontal lobe.

Subgroup analysis found that inclusion of adjuvent chemo-therapy provided minimal increase in survival for patients older than 65 years, for patients > 40 years with KPS < 80, and for those treated with brachytherapy.

21.2.2.2. Pilocytic astrocytomas

Key concepts:

• a subgroup of astrocytomas with better prognosis (10-year survival: 94%) than infiltrating fibrillary or diffuse astrocytomas

• age ≤ 20 yrs in 75%, which is lower than for typical astrocytomas

• common locations: cerebellar hemisphere, optic nerve/chiasm, hypothalamus

• radiographic appearance: discrete appearing, contrast enhancing lesion, often cystic with mural nodule

• pathology: compacted and loose textured astrocytes with Rosenthal fibers and/or eosinophilic granular bodies

• danger of overgrading and overtreating if not recognized. Histology alone may be inadequate for diagnosis; knowledge of radiographic appearance is critical

BACKGROUND AND TERMINOLOGY

Pilocytic astrocytoma (PCA) is the currently recommended classification of these tumors that have been referred to for many years variously as cystic cerebellar astrocytomas, juvenile pilocytic astrocytomas, optic gliomas, and hypothalamic gliomas^{118 (p 77-96)}. PCAs differ markedly from infiltrating fibrillary or diffuse astrocytomas in terms of their ability to invade tissue and for malignant degeneration.

LOCATION

PCAs arise throughout the neuraxis and are more common in children and young adults:

1. optic gliomas & hypothalamic gliomas:

A. PCAs arising in the optic nerve are called optic gliomas (see page 606)

B. when they occur in the region of the chiasm they cannot always be distinguished clinically or radiographically from so-called hypothalamic gliomas (see page 606) or gliomas of the third ventricular region

2. cerebral hemispheres: tends to occur in older patients (i.e. young adults) than optic nerve/hypothalamic lesions. These PCAs are potentially confused with fibrillary astrocytomas possessing more malignant potential. PCAs are often distinguished by a cystic component with an enhancing mural nodule (would be atypical for a fibrillary astrocytoma), & some PCAs have dense calcifications¹¹⁸

3. brainstem gliomas: usually are fibrillary infiltrating type and only a small pro-portion are pilocytic. Those that are PCAs may comprise the majority of the prognostically favorable group described as “dorsally exophytic”¹¹⁹(see page 607)

4. cerebellum: formerly referred to as cystic cerebellar astrocytoma (see below)

5. spinal cord: PCAs may also occur here, but little information is available on these. Again, patients tend to be younger than with spinal cord fibrillary astrocytomas

PATHOLOGY

PCAs are composed of loosely knit tissue comprised of stellate astrocytes in microcystic regions containing eosinophilic granular bodies intermixed with regions of compact tissue consisting of elongated and fibrillated cells often associated with Rosenthal fiber^A formation¹¹⁸. These latter two distinctive features facilitate the diagnosis. Another characteristic finding is that the tumors easily break through the pia to fill the over-lying subarachnoid space. PCAs may also infiltrate into the perivascular spaces. Vascular proliferation is common. Multinucleated giant cells with peripherally located nuclei are common, especially in PCAs of the cerebellum or cerebrum. Mitotic figures may be seen, but are not as ominous as with fibrillary astrocytomas. Areas of necrosis may also be seen. In spite of well-demarcated margins grossly and on MRI, at least 64% of PCAs infiltrate the surrounding parenchyma, especially the white matter¹²⁰ (the clinical significance of this is uncertain, one study found no statistically significant decrease in survival¹²¹).

A. Rosenthal fibers: sausage or corkscrew shaped cytoplasmic eosinophilic inclusion bodies consisting of glial filament aggregates resembling hyaline. Stain bright red on Masson trichrome smears

Differentiating from a diffuse or infiltrating fibrillary astrocytoma: Unless some of the distinctive findings described above are seen, pathology alone may not be able to differentiate. This may be especially problematic with small specimens obtained e.g. with stereotactic biopsy. Factors that suggest the diagnosis include young age, and knowledge of the radiographic appearance is often critical (see below).

Malignant degeneration: Malignant degeneration has been reported, often after many years. This may occur without radiation therapy (XRT)¹²², although in most cases XRT had been administered¹²³.

RADIOGRAPHIC APPEARANCE

On CT or MRI, PCAs are usually well circumscribed, 94% enhance with contrast¹²⁰ (unlike most low-grade fibrillary astrocytomas), frequently have a cystic component with a mural nodule, and have little or no surrounding edema. Although they may occur anywhere in the CNS, 82% are periventricular¹²⁰. Calcifications are only occasionally present¹²⁰. 4 main imaging patterns of cerebellar or cerebral PCAs are shown in Table 21-17.

Table 21-17 Common imaging characteristics of cerebellar or cerebral PCAs

%	Description
21%	nonenhancing cyst with enhancing mural nodule	} over 66% are cystic with enhancing mural nodule
46%	enhancing cyst with enhancing mural nodule
16%	mass with nonenhancing central area (necrosis)
17%	solid mass with minimal or no cyst

EPIDEMIOLOGY

Usually presents during second decade of life (ages 10-20). 75% occur in age < 20 years¹²⁴. No evidence of gender predilection.

PILOCYTIC ASTROCYTOMA OF THE CEREBELLUM

Key concepts:

• often cystic, half of these have mural nodule

• usually presents during the second decade of life (ages 10-20 yrs)

• also, see Key concepts:, page 603 for pilocytic astrocytomas in general

Formerly referred to by the nonspecific and confusing term cystic cerebellar astrocytoma. One of the more common pediatric brain tumors (≈ 10%¹²⁵), comprising 27-40% of pediatric p-fossa tumors^{126 (p 367-74), 127 (p 3032)}. They may also occur in adults, where the mean age is lower and the post-operative survival is longer than for fibrillary astrocytomas¹²⁸.

PRESENTATION

Signs and symptoms of pilocytic astrocytoma (PCA) of the cerebellum are usually those of any p-fossa mass, i.e. those of hydrocephalus or cerebellar dysfunction (see Posterior fossa (infratentorial) tumors, page 590).

PATHOLOGY

The classic “juvenile pilocytic astrocytoma” of the cerebellum is a distinctive entity with its macroscopic cystic architecture and microscopic spongy appearance¹¹⁸. For other microscopic findings, see above.

These tumors may be solid, but are more often cystic (hence the older term “cystic cerebellar astrocytoma”), and tend to be large at the time of diagnosis (cystic tumors: 4-5.6 cm dia; solid tumors: 2-4.8 cm dia). Cysts contain highly proteinaceous fluid (averaging ≈ 4 Hounsfield units higher density than CSF on CT¹²⁵).

50% of cystic tumors have a mural nodule and a cyst lining of reactive, non-neoplastic cerebellar tissue or ependymal lining (non-enhancing on CT), whereas the remaining 50% lack a nodule and have a cyst wall of poorly cellular tumor¹²⁹ (enhances on CT).

Histological classification of Winston

The Winston classification system¹³⁰ is shown in Table 21-18. 72% of cerebellar PCAs tended to cluster with either Type A or B characteristics, 18% in his series had both, and 10% had neither.

Table 21-18 Classification of cerebellar astrocytoma

• Type A: microcysts, leptomeningeal deposits, Rosenthal fibers, foci of oligodendroglioma

• Type B: perivascular pseudorosettes, high cell density, mitosis, calcification

• common features of types A & B: hypervascularity, endothelial proliferation, parenchymal desmoplasia, pleomorphism

TREATMENT GUIDELINES

The natural history of these tumors is slow growth. Treatment of choice is surgical excision of the maximal amount of the tumor that can be removed without producing deficit. In some, invasion of brainstem or involvement of cranial nerves or blood vessels may limit resection. In tumors composed of a nodule with a true cyst, excision of the nodule is sufficient; the cyst wall is non-neoplastic and need not be removed. In tumors with a so-called “false cyst” where the cyst wall is thick and enhances (on CT or MRI), this portion must be removed also. Because of the high 5 and 10 year survival rates together with the high complication rate of radiation therapy over this time interval (see Radiation injury and necrosis, page 771) and the fact that many incompletely resected tumors enlarge minimally if at all over periods of 5, 10 or even 20 years, it is recommended to not radiate these patients post-op. Rather, they should be followed with serial CT or MRI and be reoperated if there is recurrence¹³¹. Radiation therapy is indicated for nonresectable recurrence (i.e. reoperation is preferred if possible) or for recurrence with malignant histology. Chemotherapy is preferable to XRT in younger patients²⁹.

Also, see Posterior fossa (infratentorial) tumors, page 590 for guidelines regarding hydrocephalus, etc.

PROGNOSIS

Children with Winston Type A cerebellar PCAs had 94% 10-yr survival, whereas those with Type B had only 29% 10-yr survival.

Tumor recurrence is relatively common, and although it has been said that they generally occur within ≈ 3 yrs of surgery¹³², this is controversial and very late recurrences (violating Collins’ law, which says that a tumor may be considered cured if it does not recur within a time period equal to the patient’s age at diagnosis + 9 months) are well known¹³¹. Also, some tumors excised partially fail to show further growth, representing a form of cure.

About 20% of cases develop hydrocephalus requiring treatment following surgery¹³³. So-called “drop metastases” are rare with PCAs.

OPTIC GLIOMA

Accounts for ≈ 2% of gliomas in adults, and 7% in children. The incidence is higher (≈ 25%) in neurofibromatosis (NFT) (see page 722).

May arise in any of the following patterns:

1. one optic nerve (without chiasmal involvement)

2. optic chiasm: less commonly involved in patients with NFT than in sporadic cases

3. multicentric in both optic nerves sparing the chiasm: almost only seen in NFT

4. may occur in conjunction with or be part of a hypothalamic glioma (see below)

Pathology

Most are composed of low-grade (pilocytic) astrocytes. Rarely malignant.

Presentation

Painless proptosis is an early sign in lesions involving one optic nerve. Chiasmal lesions produce variable and nonspecific visual defects (usually monocular) without proptosis. Large chiasmal tumors may cause hypothalamic and pituitary dysfunction, and may produce hydrocephalus by obstruction at the foramen of Monro. Gliosis of the optic nerve head may be seen on fundoscopy.

Evaluation

Plain x-rays: not usually helpful, although in some cases dilatation of the optic canal can be seen in optic canal views.

CT/MRI: CT scan is excellent for imaging structures within the orbit. MRI is helpful for demonstrating chiasmal or hypothalamic involvement. On CT or MRI, involvement of the optic nerve produces contrast enhancing fusiform enlargement of the nerve usually extending > 1 cm in length.

Treatment

Tumor involving a single optic nerve, sparing the chiasm, producing proptosis and visual loss should be treated with a transcranial approach with excision of the nerve from the globe all the way back to the chiasm (a transorbital (Kronlein) approach is not appropriate since tumor may be left in the nerve stump). In addition to the anticipated blindness in the involved eye, this may produce a junctional scotoma (see page 1071).

Chiasmal tumors are generally not treated surgically except for biopsy (especially when it is difficult to distinguish an optic nerve glioma from a hypothalamic glioma), CSF shunting, or to remove the rare exophytic component to try and improve vision.

Further treatment: Chemotherapy²⁹ (especially in younger patients) or XRT is used for chiasmal tumors, for multicentric tumors, post-op if tumor is found in the chiasmal stump end of the resected nerve, and for the rare malignant tumor. Typical XRT treatment planning is for 45 Gy given in 25 fractions of 1.8 Gy.

HYPOTHALAMIC GLIOMA

Pilocytic astrocytomas of the hypothalamus and third ventricular region occur primarily in children. Radiographically, the lesion may have an intraventricular appearance. Many of these tumors have some chiasmal involvement and the distinction from optic nerve glioma cannot be made (see above).

May present with so-called “diencephalic syndrome”, a rare syndrome seen in peds, usually caused by an infiltrating glioma of the anterior hypothalamus. Classically: cachexia (loss of subcutaneous fat) associated with hyperactivity, over-alertness and an almost euphoric affect. May also see: hypoglycemia, failure to thrive, macrocephaly.

When complete resection is not possible, further treatment may be needed as outlined under optic gliomas (see Astrocytoma above).

PILOMYXOID ASTROCYTOMA (PMA)

WHO grade II. Related to pilocytic astrocytomas (PCA) but more aggressive with greater tendency to recur and spread in CSF¹³⁴. May be an infantile form of PCA with a case report of “maturation” to a typical PCA¹³⁵. Typical onset in infancy (10 months).

Histologically: dominant mucoid matrix, monomorphic bipolar cells, and angiocentric cell arrangement. By definition, does not contain Rosenthal fibers.

May also occur in spinal cord, with a case report of extraneural peritoneal mets spread through a VP shunt¹³⁶.

21.2.2.3. Brainstem glioma

Key concepts:

• not a homogeneous group. MRI can differentiate malignant from benign lesions

• trend: lower grade tumors tend to occur in the upper brainstem, and higher grade tumors in the lower brainstem/medulla

• usually presents with multiple cranial nerve palsies and long tract findings

• most are malignant, have poor prognosis, and are not surgical candidates

• role of surgery primarily limited to dorsally exophytic lesions and shunting

Brainstem gliomas (BSG) tend to occur during childhood and adolescence (77% are < 20 yrs old, they comprise 1% of adult tumors¹³⁷). BSG are one of the 3 most common brain tumors in pediatrics (see Pediatric brain tumors, page 697), comprising ≈ 10-20% of pediatric CNS tumors¹¹⁹.

PRESENTATION¹³⁸

Upper brainstem tumors tend to present with cerebellar findings and hydrocephalus, whereas lower brainstem tumors tend to present with multiple lower cranial nerve deficits and long tract findings. Due to their invasive nature, signs and symptoms usually do not occur until the tumor is fairly extensive in size.

Signs and symptoms:

1. gait disturbance

2. headache (see page 587)

3. nausea/vomiting

4. cranial nerve deficits: diplopia, facial asymmetry

5. distal motor weakness in 30%

6. papilledema in 50%

7. hydrocephalus in 60%, usually due to aqueductal obstruction (often late, except with periaqueductal tumors, e.g. see Tectal gliomas below)

8. failure to thrive (especially in age ≤ 2 yrs)

PATHOLOGY

BSG is a heterogeneous group. There may be a tendency towards lower grade tumors in the upper brainstem (76% were low-grade) versus the lower brainstem (100% of the glioblastomas were in the medulla)¹³⁹. A cystic component is seen rarely. Calcifications are also rare. 4 growth patterns that can be identified by MRI¹⁴⁰ that may correlate with prognosis¹⁴¹:

1. diffuse: all are malignant (most are anaplastic astrocytomas, the rest are glioblastomas). On MRI these tumors extend into the adjacent region in vertical axis (e.g. medullary tumors extend into pons and/or cervical cord) with very little growth towards obex, remaining intraaxial

2. cervicomedullary: most (72%) are low-grade astrocytomas. The rostral extent of these tumors is limited to the spinomedullary junction. Most bulge into the obex of the 4th ventricle (some may have an actual exophytic component)

3. focal: extent limited to medulla (does not extend up into pons nor down into spinal cord). Most (66%) are low-grade astrocytomas

4. dorsally exophytic: may be an extension of “focal” tumors (see above). Many of these may actually be low grade gliomas including:

A. pilocytic astrocytomas: see page 603

B. gangliogliomas (see page 677): very rare, only 13 cases reported as of 1984. Compared to other BSGs, these patients tend to be slightly older and the medulla is involved more frequently¹⁴²

EVALUATION

MRI

The diagnostic test of choice. MRI evaluates status of ventricles, gives optimal assessment of tumor (CT is poor in the posterior fossa) and detects exophytic component. T1WI: almost all are hypointense, homogeneous (excluding cysts). T2WI: increased signal, homogeneous (excluding cysts). Gadolinium enhancement is highly variable¹⁴⁰.

CT

Most do not enhance on CT, except possibly an exophytic component. If there is marked enhancement, consider other diagnoses (e.g. high grade vermian astrocytoma).

TREATMENT

SURGERY

Biopsy: should not be performed when the MRI shows a diffuse infiltrating brainstem lesion¹⁴³ (does not change treatment or outcome).

Treatment is usually non-surgical. Exceptions where surgery may be indicated:

1. tumors with a dorsally exophytic component¹¹⁹: see below these may protrude into 4th ventricle or CP angle, tend to enhance with IV contrast, tend to be lower grade

2. some success has been achieved with non-exophytic tumors that are not malignant astrocytomas (surgery in malignant astrocytomas is without benefit)¹⁴¹ (detailed follow-up is lacking)

3. shunting for hydrocephalus

Dorsally exophytic tumors

These tumors are generally histologically benign (e.g. gangliogliomas) and are amenable to radical subtotal resection. Prolonged survival is possible, with a low incidence of disease progression at short-term follow-up¹¹⁹.

Surgical goals in exophytic tumors include:

1. enhanced survival by subtotal removal of exophytic component¹⁴⁴: broad attachment to the floor of 4th ventricle is typical and usually precludes complete excision (although some “safe entry” zones have been described¹⁴⁵). An ultrasonic aspirator facilitates debulking

2. establishing diagnosis: radiographic differentiation of exophytic brainstem gliomas tumors from other lesions (e.g. medulloblastoma, ependymoma and dermoids) may be difficult

3. tumors that demonstrate recurrent growth after resection remained histologically benign and were amenable to re-resection¹¹⁹

Complications of surgery generally consisted of exacerbation of pre-operative symptoms (ataxia, cranial nerve palsies…) which usually resolved with time.

MEDICAL

No proven chemotherapeutic regimen. Steroids are usually administered. In pediatrics, there is some indication of response to Temodar® (temozolomide, see page 602).

RADIATION

Traditionally given as 45-55 Gy over a six week period, five days per week. When combined with steroids, symptomatic improvement occurs in 80% of patients.

Possible improved survival with so called “hyperfractionation” where multiple smaller doses per day are used.

PROGNOSIS

Most children with malignant BSG will die within 6-12 months of diagnosis. XRT may not prolong survival in patients with grade III or IV tumors. A subgroup of children have a more slowly growing tumor and may have up to 50% five-year survival. Dorsally exophytic tumors comprised of pilocytic astrocytomas may have a better prognosis.

TECTAL GLIOMAS

A topically defined diagnosis generally consisting of low-grade astrocytomas. Considered a benign subgroup of brainstem glioma. Because of location, tends to present with hydrocephalus. Focal neurologic findings are rare (diplopia, visual field deficits, nystagmus, Parinaud’s syndrome (see page 114), ataxia, seizures…) and are often reversible after the hydrocephalus is corrected.

Epidemiology

Comprises ≈ 6% of surgically treated pediatric brain tumors¹⁴⁶. Presents primarily in childhood. Median age of patients becoming symptomatic = 6-14 years¹⁴⁶.

Pathology

Since many of these are not biopsied, meaningful statistical analysis is not possible. Pathologies identified include: WHO II diffuse astrocytoma, pilocytic astrocytomas, WHO II ependymoma, anaplastic astrocytoma, oligodendroglioma & oligoastrocytoma.

Radiographic evaluation

CT scan detects the hydrocephalus, but may miss the tumor in ≈ 50%¹⁴⁷. Calcification on CT has been described in 9-25%^{147, 148}.

MRI is the study of choice for diagnosis and follow-up. Typically appears as a mass projecting dorsally from the quadrigeminal plate. Isointense on T1WI, iso- or hyperintense on T2WI^{146, 149}. Enhancement with gadolinium occurs in 18% and is of uncertain prognostic significance.

Treatment

Due to the indolent course, open surgery is not recommended. Options include:

1. VP shunt: the standard treatment for years. Long-term results are good with a functioning shunt

2. endoscopic third ventriculostomy: may avoid the need for a shunt. Endoscopic biopsy¹⁵⁰ may be done at the same time through the same burr hole if it is technically feasible (requires a dilated foramen of Monro, which is often present). Long-term results unknown

3. endoscopic aqueductoplasty (with or without stenting): an option for some. Long-term results unknown

Stereotactic radiosurgery: May be offered for tumor progression (criteria are not defined: radiographic progression may not be associated with clinical deterioration¹⁴⁹). Dosing should be limited to ≤ 14 Gray at the 50-70% isodose line to avoid radiation-induced side effects¹⁵¹.

Prognosis

Tumor progression: described in 15-25%.

Follow-up: no accepted guidelines. Serial neurologic exams and MRIs every 6-12 months has been suggested¹⁴⁶.

21.2.3. Oligodendroglioma

Key concepts:

• frequently presents with seizures

• predilection for the frontal lobes

• histology: classic features of “fried egg” cytoplasm (on permanent pathology) & “chicken wire” vasculature are unreliable. Calcifications are common

• grading: controversial. Recommendation: low grade and high grade

• recommended treatment: surgery for mass effect or low grade lesions (high-grade lesions are controversial). Chemotherapy for all (with or without surgery), XRT only for anaplastic transformation

Table 21-19 Location of oligodendrogliomas

Location	%
supratentorial	> 90%
frontal lobes	45%
hemisphere (outside frontal lobes)	40%
within third or lateral ventricle	15%
infratentorial + spinal cord	< 10%

EPIDEMIOLOGY

Oligodendroglioma (ODG) have long been thought to comprise only ≈ 2-4% of primary brain tumors^{152, 153} or 4-8% of cerebral gliomas¹⁵³; but recent evidence indicates these tumors have been underdiagnosed (many are misinterpreted as fibrillary astrocytomas, especially the infiltrative portion of these tumors) and ODGs may represent up to 25-33% of glial tumors^{154, 155}. Ratio of male:female = 3:2. Primarily a tumor of adults: average age ≈ 40 years (peak between 26-46 years), but with a smaller earlier peak in childhood between 6-12 years¹⁵⁶. CSF metastases reportedly occur in up to 10%, but 1% may be a more realistic estimate¹⁵². Spinal ODGs comprise only ≈ 2.6% of intramedullary tumors of the cord and filum.

CLINICAL

Classic presentation of ODG: a patient with seizures for many years prior to the diagnosis being made when they would present with an apoplectic event due to peri-tumoral intracerebral hemorrhage. This scenario is less common in the post CT era.

Seizures are the presenting symptoms in ≈ 50-80% of cases^{152, 156}. The remainder of presenting symptoms are non-specific for ODG, and are more often related to local mass effect and less commonly to ↑ ICP. Presenting symptoms are shown in Table 21-20.

Table 21-20 Presenting symptoms in 208 oligodendrogliomas¹⁵²

Symptom	%
seizures	57%
headache	22%
mental status changes	10%
vertigo/nausea	9%

EVALUATION

Calcifications are seen in 28-60% of ODGs on plain radiographs¹⁵², and on 90% of CTs.

PATHOLOGY

73% of tumors have microscopic calcifications¹⁵⁷. Isolated tumor cells consistently penetrate largely intact parenchyma, an associated solid tumor component may or may not be present¹⁵⁵. The solid portion, when present, classically demonstrates lucent peri-nuclear halos giving a “fried egg” appearance (actually an artifact of formalin fixation, which is not present on frozen section and may make diagnosis difficult on frozen). A “chicken-wire” vascular pattern has also been described¹⁵⁸. These features are felt to be unreliable, and cells with monotonous round nuclei (often in cellular sheets) with an eccentric rim of eosinophilic cytoplasm lacking obvious cell processes are more consistent features¹⁵⁹.

16% of hemispheric ODGs are cystic¹⁵⁷ (cysts form from coalescence of microcysts from micro-hemorrhages, unlike astrocytomas which actively secrete fluid).

33-41% have a component of ependymal or neoplastic astrocytic cells (so called oligoastrocytomas or mixed gliomas¹⁶⁰ or collision tumors) (see page 612).

GFAP staining: Since most ODGs contain microtubules instead of glial filaments¹⁶¹, ODGs usually do not stain for GFAP (see page 720) although some do¹⁶². In mixed gliomas, the astrocytic component may stain for GFAP.

GRADING

A work-in-progress. Historically, a number of attempts at grading ODGs have been proposed and then abandoned because of lack of prognostic significance (for a review, see reference¹⁵⁹). Necrosis does not appear to reliably predict a poor prognosis¹⁵⁹.

For prognostic purposes, it is suggested that ODGs be stratified into two groups:

• oligodendroglioma (WHO grade II) or low grade

• anaplastic oligodendroglioma (WHO grade III) or high grade^{153, 159}

Although there is not uniform agreement on the means for differentiating the two, the factors shown in Table 21-21 should be taken into account as they have been demonstrated to have prognostic significance. Using the spatial grading system for low grade gliomas (see page 591), no ODGs are of the Type 1 tumor (solid tumor without infiltrative component).

Table 21-21 Features associated with low-grade and high-grade oligodendrogliomas

Feature	WHO II (low grade)	WHO III (high grade)
contrast enhancement on CT or MRI	absent	present
endothelial proliferation on histology	absent	present
pleomorphism (large variability in nuclear and cytoplasmic size and shape)	absent	present
tumor proliferation (evidenced by mitotic figures or high MIB-1 index*)	absent	present
astrocytic component	absent	present

* see page 720 for information on the MIB-1 index

TREATMENT

Σ	Recommendation: (see text for details). Following an appropriate surgical procedure (if indicated), chemotherapy is the primary treatment modality. XRT is reserved for anaplastic transformation, if it should occur159

CHEMOTHERAPY

Most ODGs respond to chemotherapy, usually in < 3 mos, often with a reduction in size. The response is variable in degree and duration¹⁶³. No pathological or clinical feature of high-grade ODGs has been identified that reliably predicts response to chemo-therapy. However, allelic loss of chromosome 1p, and combined loss of chromosome arms 1p and 19q, are associated with response to chemo; and losses of both 1p and 19q were associated with longer tumor-free survival after chemo¹⁶⁴.

The most experience is with PCV (procarbazine 60 mg/m² IV, CCNU AKA lomustine (CeeNU®) 110 mg/m² PO, and vincristine 1.4 mg/m² IV, all given on a 29 day cycle repeated every 6 weeks)^{165, 166}. Also studied: temozolomide for recurrent anaplastic oligoastrocytoma showed some efficacy¹¹².

SURGERY

Indications for surgery:

1. ODGs with significant mass effect regardless of grade: surgery decreases the need for corticosteroids, reduces symptoms and prolongs survival¹⁵⁹

2. tumors without significant mass effect:

A. low-grade ODGs and oligoastrocytoma: surgery is recommended for resectable lesions. Gross total removal should be attempted when possible (survival is improved even more than with astrocytomas¹⁶⁷), but not at the expense of neurologic function

B. high-grade ODGs: data for improved survival is less convincing, and some studies show no advantage of gross total removal over partially resected or biopsied-only high-grade lesions¹⁵⁹

Grossly, the tumor appears as a pink to red, friable mass. There may be a false plane of demarcation between tumor and what appears to be normal brain.

POSTOPERATIVE RADIATION

Benefits of postoperative irradiation is controversial¹⁵⁶. In a retrospective analysis with no set selection criteria, survival was better in patients receiving > 45 Gy (1 Gy = 100 cGy)¹⁶⁸. In another series, no difference in 5 year survival following surgery was seen with or without XRT (amount of radiation not specified)¹⁶⁹. Radiation side effects of memory loss, dementia and personality changes are more common with the longer survival seen in many of these cases¹⁷⁰.

PROGNOSIS

Pure ODGs have a better prognosis than mixed oligoastrocytomas which are better than pure astrocytomas (an oligodendroglial component, no matter how small, confers a better prognosis).

10 year survival of 10-30% has been quoted for tumors that are completely or predominantly ODGs¹⁶⁸. As a group, median survival for surgically treated lesions is given as 35 months post-op (mean 52 months)¹⁵².

The presence of calcifications is debated as a prognosticator; in one series, calcified ODG on plain films had a longer median survival of 108 months (vs. 58 months for noncalcified)¹⁵².

Frontal lobe ODGs survived longer than those in temporal lobes (37 months vs. 28 months postoperative survival)¹⁵², possibly due to increased ease of radical resection with the former.

Chromosomal 1p loss (or combined 1p and 19q loss) is also associated with longer survival^{164, 171}.

21.2.4. Mixed gliomas

21.2.4.1. Oligoastrocytoma

Molecular biology: May show changes typical for diffuse astrocytoma (TP⁵³ mutation & LOH on 17p) or for ODG (LOH on 1p and 19q). No molecular genetic markers have been identified to distinguish oligoastrocytoma from either astrocytoma or ODG. Unlike ODG, the prognostic/therapeutic value of LOH on 1p is less clear¹⁷¹.

Oligoastrocytoma (WHO grade II)

Two distinct neoplastic cell types, 1 type resembles oligodendroglioma cells, and the other resembles cells in diffuse astrocytomas. Some cells may have features of both. The 2 cell types may be segregated or diffusely admixed.

Anaplastic oligoastrocytoma (WHO grade III)

Increased cellularity, nuclear atypia, pleomorphism, and high mitotic activity. Necrosis and microvascular proliferation may be present. Differentiating from GBM may be difficult since GMS may have areas resembling anaplastic ODG (the term “glioblastoma with oligodendroglioma component” is a disputed term suggested for these - unproven suggestion that survival may be better than for ordinary GBM¹⁷²).

21.2.5. Neuronal and mixed neuronal-glial tumors

21.2.5.1. Desmoplastic infantile astrocytoma/ganglioglioma

The former entities “desmoplastic cerebral astrocytoma of infancy” and “desmoplastic infantile ganglioglioma” have been combined to “desmoplastic infantile astrocytoma and ganglioglioma” (DIG)⁵. A lesion with either astrocytic or dual glio-neuronal differentiation. Prognosis is usually favorable.

21.2.5.2. Central neurocytoma

Rare. Usually considered benign, but malignant variation/behavior has been described¹⁷³. Slow-growing well circumscribed tumor usually located in the lateral ventricles or at the septum pellucidum^{174, 175}. Tends to affect young adults, usually males. Histologically, resembles oligodendrogliomas. Ultrastructure shows neuronal differentiation. Molecular oncogenesis is not known.

Usually curable by total resection¹⁷⁵. Subtotal resection and histologic atypia are associated with an increased risk of recurrence, but early recurrence may occur even without malignant histologic features¹⁷³.

Variants:

1. “extraventricular neurocytomas”: neurocytic neoplasms located within brain parenchyma. Not as well characterized as intraventricular type¹⁷⁵

2. central liponeurocytoma: extremely rare. Classified as a glioneuronal tumor. Usually occurs in the posterior fossa of older adults¹⁷⁶. Once considered a variant of medulloblastoma (called medullocytoma), has more indolent behavior¹⁷⁷ and characteristic morphologic features of well-differentiated neurons with the cytology of neurocytes in addition to a population of lipidized cells resembling mature adipose tissue¹⁷⁷

Treatment

1. ideal: total resection if possible

2. stereotactic radiosurgery may be effective for recurrence¹⁷⁸ or for incompletely removed or biopsied tumors¹⁷⁹

3. chemotherapy with etoposide, cisplatin and cyclophosphamide, has been reported for recurrent progressive tumor¹⁸⁰

21.2.5.3. Cerebellar liponeurocytoma

Née lipomatous medulloblastoma. Occurs exclusively in cerebellum of adults (mean age: 50 years). No gender preference.

Histology: clusters of neoplastic neurocytes with lipidization (resembling adipocytes) with background of small neoplastic cells with morphological features more suggestive of neurocytes. Synaptophysin (see page 721) and MAP-2 immunostaining is consistent and diffuse, focal GFAP staining is common. Usually no mitotic figures. MIB-1 index 1-3%.

21.2.6. Meningiomas

Key concepts:

• slow growing, extra-axial tumor, usually benign, arise from arachnoid (not dura)

• imaging (MRI or CT): classically broad based attachment on dura often with dural tail, typically enhance densely, may cause hyperostosis of adjacent bone

• MRI: isointense on T1WI, hypodense on T2WI

• 32% of incidentally discovered meningiomas do not grow over 3 years follow-up

• surgical indications: documented growth on serial imaging and/or symptoms referable to the lesion that are not satisfactorily controlled medically

• most (but not all) are cured if completely removed, which is not always possible

• most commonly located along falx, convexity, or sphenoid bone

• frequently calcified. Classic histological finding: psammoma bodies

Usually slow growing, circumscribed (non-infiltrating), benign lesions. Histologically malignant (incidence: ≈ 1.7% of meningiomas¹⁸¹) and/or rapidly growing varieties are also described (a rapidly growing lesion that looks like a meningioma may be a hemangiopericytoma, see page 620). Actually arise from arachnoid cap cells (not dura). May be multiple in up to 8% of cases¹⁸², this finding is more common in neurofibromatosis. Occasionally forms a diffuse sheet of tumor (meningioma en plaque). This section considers intracranial meningiomas.

May occur anywhere that arachnoid cells are found (between brain and skull, within ventricles, and along spinal cord). Ectopic meningiomas may arise within the bone of the skull (primary intraosseous meningiomas)¹⁸³ and others occur in the subcutaneous tissue with no attachment to the skull. Most are asymptomatic (see below).

EPIDEMIOLOGY

As many as 3% of autopsies on patients > 60 yrs age reveals a meningioma¹⁸⁴. Meningiomas account for 14.3-19% of primary intracranial neoplasms¹⁸⁵. Incidence peaks at 45 years age. Female:male ratio is 1.8:1.

1.5% occur in childhood and adolescence, usually between 10-20 years age^{127 (p 3263)}. 19-24% of adolescent meningiomas occur in patients with neurofibromatosis type I.

LOCATION

Table 21-22 lists common locations. Other locations include: CP-angle, clivus, planum sphenoidale and foramen magnum. ≈ 60-70% occur along the falx (including parasagittal), along sphenoid bone (including tuberculum sellae), or over the convexity. Childhood meningiomas are rare, 28% are intraventricular, and the posterior fossa is also a common site.

Table 21-22 Location of adult meningiomas (series of 336 cases¹⁸⁶)

Location	%
parasagittal	20.8
convexity	15.2
tuberculum sellae	12.8
sphenoidal ridge	11.9
olfactory groove	9.8
falx	8
lateral ventricle	4.2
tentorial	3.6
middle fossa	3
orbital	1.2
spinal	1.2
intrasylvian	0.3
extracalvarial	0.3
multiple	0.9

Sphenoid wing (or ridge) meningiomas

Three basic categories¹⁸⁷:

1. lateral sphenoid wing (or pterional): behavior and treatment are usually similar to convexity meningioma

2. middle third (or alar)

3. medial (clinoidal): tend to encase the ICA and the MCA as well as cranial nerves in the region of the superior orbital fissure and the optic nerve. May compress brainstem. Total removal is often not possible

Parasagittal and falx meningiomas

Up to 50% invade the superior sagittal sinus (SSS). Grouped based on location along AP direction of SSS as:

1. anterior (ethmoidal plate to coronal suture): 33%. Most often present with H/A and mental status changes

2. middle (between coronal and lambdoidal sutures): 50%. Most often present as Jacksonian seizure and progressive monoplegia

3. posterior (lambdoidal suture to torcular Herophili): 20%. Most often present with H/A, visual symptoms, focal seizures, or mental status changes

Classification systems for the extent of SSS invasion include one by Bonnal and Brotchi¹⁸⁸, and a more recent one by Sindou et al.¹⁸⁹ shown in Figure 21-1.

Parasagittal meningiomas may originate at the level of the motor strip, and a common initial manifestation of these is a contralateral foot drop¹⁹⁰.

Olfactory groove meningiomas

Presentation (usually asymptomatic until they are large) may include:

1. Foster Kennedy syndrome: anosmia (patient is usually unaware of this), ipsilateral optic atrophy, contralateral papilledema (see page 112)

2. mental status changes: often with frontal lobe findings (apathy, abulia…)

3. urinary incontinence

4. posteriorly located lesions may compress the optic apparatus causing visual impairment

5. large lesions may compress the fornix and cause short-term memory loss

6. seizure

Figure 21-1 Grading system for meningioma invasion of the superior sagittal sinus Modified from Sindou MP et al., J Neurosurg, 105: pp 514-25, 2006 Shown: schematic coronal section through superior sagittal sinus (SSS). Type I = attachment to lateral wall of sinus; Type II = invasion of lateral recess; Type III = invasion of lateral wall; Type IV = invasion of lateral wall and roof; Type V = total sinus occlusion, contralateral wall spared; Type VI = total sinus occlusion, invasion of all walls

The morbidity, mortality and difficulty in achieving total removal increase significantly for tumors > 3 cm in size¹⁹¹.

Pre-op MRA, CTA or angiogram may be helpful to assess location of anterior cerebral arteries relative to the tumor. 70-80% of these get the majority of their blood supply from the anterior ethmoidal artery, which is usually not embolized due to risk to ophthalmic artery (and blindness). If there are substantial middle meningeal feeders, these may be embolized, but the benefit tends to be small.

Planum spehnoidale meningiomas

Arise from the flat part of the sphenoid bone anterior to the chiasmatic sulcus in the posterior part of the anterior cranial fossa.

Tuberculum sellae meningiomas (TSM)

The site of origin of these tumors is only about 2 cm posterior to that of olfactory groove meningiomas¹⁹¹. The tuberculum sellae is the bony elevation between the chiasmatic sulcus and the sella turcica. By definition, the anterior margin of the chiasmatic sulcus (the limbus sphenoidale) is the demarcation between the anterior and middle cranial fossa. Therefore these tumors originate in the middle fossa (unlike planum sphenoi-dale meningiomas which are in the anterior fossa).

TSMs are notorious for producing visual loss (chiasmal syndrome = primary optic atrophy + bitemporal hemianopsia). When a TSM grows posteriorly into the sella turcica it may be mistaken for a pituitary macroadenoma (see page 1216 for MRI and differentiating features).

Foramen magnum meningiomas

As with any foramen magnum (FM) lesion, the neurologic symptoms and signs can be very confusing and often do not initially suggest a tumor in this location (see page 711).

In the French Cooperative Study, there were 106 FM meningiomas¹⁹², 31% arose from the anterior lip, 56% were lateral, and 13% arose from the posterior lip of the FM. Most are intradural, but they can be extradural or a combination (the latter 2 have a lateral origin and are often invasive, which makes total removal more difficult)¹⁹³. They may be above, below, or on both sides of the vertebral artery¹⁹³.

ASYMPTOMATIC MENINGIOMAS

Meningiomas are the most common primary intracranial tumors, and most remain asymptomatic throughout the patient’s life¹⁹⁴. The routine use of CT & MRI for numerous indications inevitably results in the discovery of incidental (asymptomatic) meningiomas. In a population^A based study¹⁹⁴, incidental meningiomas were seen in 0.9% of MRIs. In another series, 32% of primary brain tumors seen on imaging studies were meningiomas, and 39% of these were asymptomatic¹⁹⁵. Of 63 cases followed for > 1 year with non-surgical management, 68% showed no increase in size over an average follow-up of 36.6 mos, whereas 32% increased in size over 28 mos average follow-up¹⁹⁵. Asymptomatic meningiomas with calcification seen on CT and/or hypointensity on T2WI MRI appeared to have a slower growth rate¹⁹⁵.

A. the study population was middle class caucasians and result may not be generalized to other groups

Data is lacking to make evidence-based management guidelines. A suggestion is to obtain a follow-up imaging study 3-4 months after the initial study to rule-out rapid progression, and then repeat annually for 2-3 years. The development of symptoms would prompt performing a study at that time.

Treatment is indicated for lesions that produce symptoms that cannot be satisfactorily controlled medically, or for those that demonstrate significant continued growth on serial imaging studies. When surgery was performed, the perioperative morbidity rate was statistically significantly higher in patients > 70 years old (23%) than in those < 70 (3.5%)¹⁹⁵.

PATHOLOGY

Four critical histopathological variables:

1. grade } see Table 21-23

2. histological subtype } see Table 21-23

3. proliferation indices: see page 616

4. brain invasion: see page 616

Table 21-23 WHO classification of meningiomas

WHO Grade I	meningothelial fibrous (fibroblastic) transitional (mixed) psammomatous angiomatous microcystic secretory lymphoplasmacyte-rich metaplastic
WHO II	chordoid clear cell (intracranial) atypical
WHO III	papillary rhabdoid (see text) anaplastic

There are a number of pathologic classification systems^{3, 196, 197 (p 465)}, and transitional forms between the major types exist. More than one histological pattern may be seen in a given tumor. The WHO 2000 classification is shown in Table 21-23.

1. meningiomas with low recurrence risk and/or aggressive growth (WHO grade I)

A. meningothelial or meningotheliomatous, AKA syncytial: the most common. Sheets of polygonal cells. Some use the term angiomatous for meningotheliomatous variety with closely packed blood vessels

B. fibrous or fibroblastic: cells separated by connective tissue stroma. Consistency is more rubbery than meningotheliomatous or transitional

C. transitional: intermediate between meningotheliomatous and fibrous. Cells tend to be spindle shaped, but areas of typical meningotheliomatous cells occur. Whorls, some of which are calcified (psammoma bodies)

D. psammomatous: calcified meningothelial whorls

E. angiomatous

F. microcystic: AKA “humid” or vacuolated meningioma. The characteristic dilated extracellular spaces are usually empty, but occasionally contain substance that stains positive for PAS (? glycoprotein) or contain fat¹⁹⁸. The cysts may coalesce and form grossly or radiologically visible cysts and may resemble astrocytomas

G. secretory

H. lymphoplasmacyte-rich

2. meningiomas with greater recurrence risk and/or aggressive growth include

A. atypical meningioma: increased mitotic activity (1-2 mitotic figure/high-powered field), increased cellularity, focal areas of necrosis, giant cells. Cellular pleomorphism is not unusual but is not significant in and of itself. Increasing atypia appears to correlate with increasing aggressiveness

B. rhabdoid meningiomas: usually have malignant features and behave aggressively. Behavior in the absence of malignant features is undetermined²

C. malignant meningiomas: AKA anaplastic, papillary or sarcomatous. Characterized by frequent mitotic figures, cortical invasion, rapid recurrence even after apparent total removal¹⁹⁹, and, rarely, metastases (see below). Frequent mitotic figures (≥ 4 mitoses per high-power field) or the presence of papillary features are strong predictors of malignancy. May be more common in younger patients

Obsolete terms (in the current WHO classification) presented for context in older literature: metaplastic, myxomatous, xanthomatous (abundant cytoplasmic lipids; appear vacuolated), lipomatous, granular, chondroblastic, osteoblastic, melanotic. Angioblastic or (meningeal) hemangiopericytomas (true hemangiopericytomas are sarcomas, see page 620). (others use the term “angioblastic” for tumors histologically similar to hemangioblastoma. Angioblastic meningiomas were felt to have more malignant clinical characteristics than other forms^{197 (p 479-83)}.

Proliferation indices

Due to variation between institutions and observers, it is advised that proliferation indices (e.g. Ki-67 or MIB-1) not be used as the sole discriminant for grading. However, these indices do correlate with prognosis (see Table 21-24). Adding the phrase “with high proliferative activity” is suggested for tumors with a very high index².

Brain invasion

The presence of brain invasion increases the likelihood of recurrence to levels similar to atypical meningiomas (not anaplastic)²⁰¹, but is not an indicator of malignant grade. Brain invasion in atypical meningiomas does not dictate malignant behavior. Adding the phrase “with brain invasion” is suggested to denote higher risk of recurrence².

Table 21-24 Ki-67 proliferation index in meningiomas²⁰⁰*

Description & WHO grade	Mean Ki-67 index*	Recurrence rate
Common meningioma (WHO grade I)	0.7%	9%
Atypical meningioma (WHO grade II)	2.1%	29%
Anaplastic meningioma (WHO grade III)	11%	50%

* not recommended for grading (see text)

Metastases

Very rarely a meningioma may metastasize outside the CNS. Most of these are angioblastic or malignant. Lung, liver, lymph nodes and heart are the most common sites.

DIFFERENTIAL DIAGNOSIS/DIAGNOSTIC CONSIDERATIONS OF MENINGIOMA

1. multiple meningiomas: suggests neurofibromatosis 2 (NF2)

2. pleomorphic xanthastrocytoma (PXA): may mimic meningiomas since they tend to be peripherally located and may have a dural tail (see page 592)

3. Rosai-Dorfman disease: especially if extracranial lesions are also identified. A connective tissue disorder with sinus histiocytosis and massive painless lymphadenopathy (most have cervical lymphadenopathy). Usually in young adults. Isolated intracranial involvement is rare. MRI: dural-based enhancing mass with signal characteristics similar to meningioma, may have dural tail. Most common intracranial locations: cerebral convexities, parasagittal, suprasellar, cavernous sinus. Pathology: dense fibrocollagenous connective tissue with spindle cells and lymphocytic infiltration, stains for CD68 & S-100. Histiocytic proliferation without malignancy. Foamy histiocytes are characteristic. Surgery and immunosuppressive therapy not effective. Low-dose XRT may be the best option

EVALUATION

MRI

Occasionally may be isointense with brain on T1WI and T2WI, but most enhance with gadolinium. Brain edema may or may not be present. Calcifications appear as signal voids on MRI. Gives information regarding patency of dural venous sinuses (accuracy in predicting sinus involvement is ≈ 90%²⁰²). “Dural tail” is a common finding²⁰³.

CT

Homogeneous, densely enhancing mass with broad base of attachment along dural border. Non-contrast Hounsfield numbers of 60-70 in a meningioma usually correlates with presence of psammomatous calcifications. There may be little cerebral edema, or it may be marked and may extend throughout the white matter of the entire hemisphere.

Intraventricular meningiomas: 50% produce extraventricular edema. On angio, these may falsely appear malignant.

Prostate cancer may mimic meningioma (prostate mets to brain are rare, but prostate frequently goes to bone, and may go to skull and can cause hyperostosis).

Angiography

Classic pattern: “comes early, stays late” (appears early in arterial phase, blush persists beyond venous phase). Meningiomas characteristically have external carotid artery feeders. Exceptions: low frontal median (e.g. olfactory groove) meningiomas which feed from the ICA (ethmoidal branches of the ophthalmic artery). Suprasellar meningiomas may also be fed by large branches of the ophthalmic arteries. Parasellar meningiomas tend to feed from the ICA. Secondary vascular supply may be derived from pial branches of the anterior, middle, and posterior cerebral arteries.

Artery of Bernasconi & Cassinari AKA artery of tentorium (a branch of the meningohypophyseal trunk) AKA the “Italian” artery: enlarged in lesions involving tentorium (e.g. tentorial meningiomas).

Angiography also gives information about occlusion of dural venous sinuses, especially for parasagittal/falx meningiomas. Oblique views are often best for evaluating patency of the superior sagittal sinus (SSS). Angiography can also help confirm diagnosis by the distinctive prolonged homogeneous tumor blush. Angiography also provides an opportunity for pre-op embolization (see below).

Pre-op embolization: Reduces the vascularity of these often bloody tumors, facilitating surgical removal. Timing of subsequent surgery is controversial. Some advocate waiting 7-10 days to permit tumor necrosis which simplifies resection^{204, 205}. Complications include: hemorrhage (intratumoral and SAH), cranial nerve deficits (usually transient), CVA from embolization through ICA or VA anastomoses, scalp necrosis, retinal embolus, and potentially dangerous tumor swelling. Some meningiomas (e.g. olfactory groove) are less amenable to embolization.

Plain x-rays

May show: calcifications within the tumor (in ≈ 10%), hyperostosis or blistering of the skull (including floor of frontal fossa with olfactory groove meningiomas), enlargement of vascular grooves (especially middle meningeal artery).

TREATMENT

Surgery is the treatment of choice for symptomatic meningiomas. Incidental meningiomas with no brain edema or those presenting only with seizures that are easily controlled medically may be managed expectantly with serial imaging as meningiomas tend to grow slowly, and some may “burn out” and cease growing (see page 615).

SURGICAL TECHNIQUE

Often very bloody. Preoperative embolization and autologous blood donation may be helpful. General principles of meningioma surgery²⁰⁶:

1. early interruption of the blood supply to the tumor

2. internal decompression (using ultrasonic aspirator, cautery loops…)

3. dissection of the tumor capsule from the brain by cutting and coagulating vascular and arachnoid attachments while infolding the tumor into the area of decompression with minimal retraction on adjacent brain

4. removal of attached bone and dura when possible

Position

As usual, the head should be elevated ≈ 30° above the right atrium. For meningiomas involving the superior sagittal sinus (SSS):

• for tumors involving the anterior third of the SSS: supine semi-sitting position

• for tumors of the middle third of the SSS: lateral position with the side of the tumor down, the neck tilted 45° toward the upward shoulder

• for tumors of the posterior third of the SSS: prone position

Sinus involvement

IMHO

Attempting to occlude or bypass the middle third of the superior sagittal sinus involved with meningioma is treacherous. Even in expert hands, there is significant risk of venous infarction/sinus occlusion with 8% morbidity and 3% mortality189, and complete removal is still not assured207. Venous collaterals may be found in the dura adjacent to the sinus, and even the tumor itself may participate. It is almost always preferable to leave residual tumor (and use radiosurgery when appropriate) than to cause a venous infarction.

Alternatives for treatment of dural sinus involvement include:

1. superior sagittal sinus (SSS)

A. if the tumor occludes the SSS, it has been suggested that the sinus can be resected carefully preserving veins draining into the patent portions of the sinus. ✖ However, this should be undertaken with great trepidation since patients still not infrequently develop venous infarcts, probably as a result of loss of minimal sinus flow and venous channels in the dura. Before ligating the sinus, the lumen should be inspected for a tail of tumor within

B. partial occlusion of superior sagittal sinus:

1. anterior to the coronal suture, the sinus may usually be divided safely

2. posterior to the coronal suture, it must not be divided or else severe venous infarction will occur

a. with superficial involvement (Type I, Figure 21-1, page 614), tumor may be dissected off the sinus with care to preserve patency

b. with extensive involvement:

i. sinus reconstruction: hazardous. Thrombosis rate using venous graft approaches 50%, and is close to 100% with artificial grafts (e.g. Gore-Tex) which should not be used

ii. it may be best to leave residual tumor, and follow with CT or MRI. If the residual tumor grows, or if the Ki-67 score is high (see page 616), SRS may be used (SRS may also be used as initial treatment for tumors that are < 2.3-3 cm, see page 774)

2. transverse sinus (TS): a patent dominant TS must not be suddenly occluded

Sphenoid wing, parasagittal or falx meningiomas (general principles)

Once tumor is exposed a partial internal debulking is performed. Then the point of attachment (to the falx or sphenoid bone) is peeled away using bipolar cautery to divide feeding vessels. Then the main portion of the tumor may be separated from brain, with the tumor being avascular once the vascular pedicle has been transected.

Parasagittal and falx meningiomas

The inferior portion of the tumor may adhere to branches of the anterior cerebral artery. Middle or posterior third tumors are exposed using a horseshoe incision based in the direction of the major scalp feeding vessels. The patient may be placed in a lateral position, or the sitting position may be used with doppler monitoring for air embolism (see page 153). Anterior third tumors are approached using a bicoronal skin incision with the patient supine. For tumors that cross the midline, burr holes are placed to straddle the SSS. For managing superior sagittal sinus involvement, see above.

Sphenoid wing meningiomas

A pterional craniotomy is utilized (see page 160). The neck is extended to allow gravity to retract the brain off of the floor of the skull.

Lateral sphenoid wing meningiomas: These tumors are often similar to convexity meningiomas. The head is turned 60° to the side (see page 159). The height of the skin incision and bone opening should be high enough to encompass the tumor.

Medial sphenoid wing meningiomas: A lumbar drain is used. The head is turned 30° off the vertical. Aggressive extradural removal of sphenoid wing is performed. An FTOZ approach may provide additional exposure. The sylvian fissure is split widely. The ICA and MCA are often encased by tumor (look for the appearance of “grooves” on the surface of the tumor on MRI, which indicates vessels, e.g. MCA). To locate the ICA, identify MCA branches and follow them proximally into the tumor. The optic nerve is best identified at the optic canal. Avoid excessive retraction of the optic apparatus. The deep portion of the tumor often has numerous small parasitic vessels from the ICA (which makes this part very bloody), and may also invade the lateral wall of the cavernous sinus (which creates risk of cranial nerve deficits with attempted removal). Therefore, the recommendation is to leave some tumor behind and use radiosurgery to deal with it.

Olfactory groove meningiomas

Approached via a bifrontal craniotomy (preserving the periosteum to cover the frontal air sinus and floor of frontal fossa at the end of the case). Small tumors may be approached via unilateral craniotomy on the side with the most tumor)^{127 (p 3284)}. For large tumors, a lumbar CSF drain will help with brain relaxation¹⁹¹. The head is rotated 20° to one side to facilitate dissection of the anterior cerebral arteries and optic nerve while pre-serving visualization of both sides of the tumor involvement²⁰⁸. The neck is slightly extended. The dura is opened low, and the superior sagittal sinus is ligated and divided at this location. Amputation of the frontal pole should be done if necessary to avoid excessive retraction. Vascular feeding arteries come through the floor of the frontal fossa in the midline. Initially, the anterior tumor capsule is opened and the tumor debulked from within heading towards the floor of the frontal fossa to interrupt the blood supply. The posterior capsule of the tumor is dissected carefully as this portion of the tumor may encase branches of the anterior cerebral artery, and/or optic nerves and chiasm. A large tumor with suprasellar extension usually displaces the optic nerve and chiasm inferiorly¹⁹¹. If necessary, the frontopolar branch and other small branches may be sacrificed without problem²⁰⁹. Post-op risks include CSF leak through the ethmoid sinuses.

Tuberculum sellae meningiomas

These tumors typically displace both optic nerves posteriorly and laterally¹⁹¹. Occasionally, the nerves are completely engulfed by tumor.

Cerebellopontine angle meningiomas

Usually arise from the meninges covering the petrous bone. May be divided into those that occur anterior to, and those that occur posterior to the IAC.

Foramen magnum meningiomas

Tumors arising from the posterior or posterolateral lip of the foramen magnum (FM) are removed relatively easily. Anterior and lateral FM tumors may be operated by the posterolateral approach, and for anterior tumors¹⁹³, a transcondylar approach may alternatively be used²¹⁰.

With meningiomas below the vertebral artery (VA), the lower cranial nerves are displaced superiorly with the VA. However, when the tumor is above the VA, the position of the lower cranial nerves cannot be predicted¹⁹³.

Large tumors may adhere to or encase neurovascular structures, and these should be internally debulked and then dissected free.

Posterior suboccipital approach: Used for meningiomas arising from the posterior lip of the FM or slightly posterolateral.

The patient is positioned prone or three-quarter prone. Neck flexion should be kept to a minimum to avoid brainstem compression by the tumor²¹¹. The surgeon must remain vigilant for the PICA and vertebral arteries, which may be encased.

Radiation therapy (XRT)

Generally regarded as ineffective as primary modality of treatment. Many prefer not to use XRT for “benign” lesions. Efficacy of XRT in preventing recurrence is controversial (see below under Recurrence); some surgeons reserve XRT for malignant (invasive), vascular, rapidly recurring (“aggressive”), or non-resectable meningiomas.

For recurrent atypical or anaplastic meningioma with residual disease: post-op, XRT with 55-60 Gy is recommended.

Table 21-25 Simpson grading system for removal of meningiomas²¹²

Grade	Degree of removal
I	macroscopically complete removal with excision of dural attachment and abnormal bone (including sinus resection when involved)
II	macroscopically complete with endothermy coagulation (Bovie, or laser) of dural attachment
III	macroscopically complete without resection or coagulation of dural attachment or of its extradural extensions (e.g. hyperostotic bone)
IV	partial removal leaving tumor in situ
V	simple decompression (± biopsy)

OUTCOME

5 year survival for patients with meningioma⁶: 91.3%.

RECURRENCE

The extent of surgical tumor removal is the most important factor in the prevention of recurrence. The Simpson grading system for the extent of meningioma removal is shown in Table 21-25. Recurrence after gross total tumor removal occurred in 11-15% of cases, but was 29% when removal is incomplete (length of follow-up not specified)¹⁸⁶; 5-year recurrence rates of 37%²¹³-85%²¹⁴ after partial resection are also quoted. The overall recurrence rate at 20 years was 19% in one series²¹⁵, and 50% in another²¹⁴. Malignant meningiomas have a higher recurrence rate than benign ones.

Value of XRT

A retrospective series of 135 non-malignant meningiomas followed 5-15 years postop at UCSF revealed a recurrence rate of 4% with total resection, 60% for partial resection without XRT, and 32% for partial resection with XRT²¹⁶. Mean time to recurrence was longer in the XRT group (125 mos) than in the non-XRT group (66 mos). These results suggest that XRT may be beneficial in partially resected meningiomas. Alternatively, one can follow these patients with CT or MRI and use XRT for progression.

In addition to the usual side effects of XRT (see Radiation injury and necrosis, page 771), there is also a case report of a malignant astrocytoma developing after XRT was used to treat a meningioma²¹⁷.

21.2.7. Mesenchymal, non-meningothelial tumors

Hemangiopericytoma

A sarcoma arising from pericytes (surrounding blood vessels). May metastasize (usually to bone, lung or liver). Occur ≈ anywhere (soft tissues, muscles, thoracic aorta, kidney, omentum…). May mimic meningioma on CT or MRI (MRS may help distinguish²¹⁸). Recurrence is common, sometimes late. Neurosurgically relevant sites:

1. intracranial: includes intraventricular

2. spinal

Treatment: Surgery is primary treatment. XRT may reduce recurrence rate. Chemo-therapy is used for metastases or for tumors failing local control measures.

Primary cerebral sarcoma

Rare. May result from sarcomatous change in preexisting tumor such as meningioma, glioblastoma, or oligodendroglioma.

21.2.8. Vestibular schwannoma

Key concepts:

• histologically benign tumor. Usually arises from superior vestibular nerve in CPA

• 3 most common early symptoms (clinical triad): hearing loss (insidious and progressive), tinnitus (high pitched) and dysequilibrium (true vertigo is uncommon)

• W/U: All patients: MRI (without & with contrast), audiometrics (pure tone audiogram and speech discrimination). In addition for small VSs (≤ 15 mm dia): ENG, VEMP, ABR

• histology: comprised of Antoni A (narrow elongated bipolar cells) and Antoni B fibers (loose reticulated)

• management options (observation, surgery, XRT or chemotherapy (Avastin®)) depend heavily on tumor size, growth, hearing status, and presence of NF2

Vestibular schwannoma is currently preferred^{220, 221} over the older term acoustic neuroma since most of these tumors arise not from the acoustic nerve but from the schwann-cell sheath of the superior division of the vestibular nerve (not the cochlear portion). Histologically benign. VSs arise as a result of the loss of a tumor-suppressor gene on the long arm of chromosome 22 (in sporadic cases this is a somatic mutation; in neurofibromatosis Type 2 (NF2)this is either inherited or represents a new mutation that may then be transmitted to offspring).

Epidemiology: One of the most common intracranial tumors, comprising 8-10% of tumors in most series²²². Annual incidence is probably about 1.5 cases per 100,000 population - over the past couple decades this estimate has increased and the typical size at diagnosis has decreased as a result of the proliferation of MRI scans²²³. VSs typically become symptomatic after age 30. At least 95% are unilateral.

Neurofibromatosis Type 2

The incidence of vestibular schwannomas (VS) is increased in neurofibromatosis (NFT), with bilateral VS being pathognomonic of neurofibromatosis Type 2 (NFT2) (central NFT, see page 724). Any patient < 40 yrs old with unilateral VS should also be evaluated for NFT2. Cytologically, the VSs of NFT2 are identical to sporadic cases, however in NFT2 the tumors form grape-like clusters that may infiltrate the nerve fibers (unlike most sporadic VSs which displace the eighth nerve).

CLINICAL

SYMPTOMS

Symptoms are shown in Table 21-26. The type of symptoms are closely correlated with tumor size. Most initially cause the triad of ipsilateral sensorineural hearing loss, tinnitus and balance difficulties. Larger tumors can cause facial numbness, weakness or twitching, and possibly brainstem symptoms. Rarely, large tumor may produce hydrocephalus. With current imaging modalities (CT and especially MRI), increasing numbers of smaller lesions are being detected.

Table 21-26 Symptoms in vestibular schwannoma (131 patients²²²)

Symptom	%
hearing loss	98%
tinnitus	70%
dysequilibrium*	67%
H/A	32%
facial numbness	29%
facial weakness	10%
diplopia	10%
N/V	9%
otalgia	9%
change of taste	6%

* or vertigo

Symptoms from 8th nerve compression

Unilateral sensorineural hearing loss, tinnitus and dysequilibrium are related to pressure on the eighth nerve complex in the IAC. These are the earliest symptoms, and by the time of diagnosis, virtually all tumors have caused otologic symptoms.

Hearing loss is insidious and progressive in most (c.f. the hearing loss in Meniere’s disease which fluctuates), however 10% report sudden hearing loss (see Sudden hearing loss below). 70% have a high frequency loss pattern, and word discrimination is usually affected (especially noticeable in telephone conversation).

The tinnitus is usually high pitched.

Unsteadiness manifests primarily as balance difficulty; true vertigo occurs in < 20%.

Sudden hearing loss: The differential diagnosis for sudden hearing loss (SHL) is extensive²²⁴. Idiopathic SHL (i.e. no identified etiology: must rule out neoplasm, infection, autoimmune, vascular and toxic causes) occurs in an estimated 10 per 100,000 population²²⁵. 1% of patients with SHL will be found to have a VS, and SHL may be the presenting symptom in 1-14% of patients with VS²²⁶. SHL with VS is presumably due to an infarction of the acoustic nerve, or acute occlusion of the cochlear artery. Treatment options for SHL include:

1. steroids: e.g. prednisone 60 mg PO q d x 10 d then tapered²²⁶

2. ✖ heparin has been shown not to be of help

3. conservative treatment: rest, restriction of salt, alcohol and tobacco²²⁷

4. experimental: thrombolytic therapy (e.g. rt-PA) (see page 1016)

Symptoms from 5th and 7th nerve compression

Otalgia, facial numbness and weakness, and taste changes occur as the tumor enlarges and compresses the fifth and seventh nerves. These symptoms usually do not occur until the tumor is > 2 cm. This highlights an interesting paradox: facial weakness is a rare or late occurrence, even though the 7th nerve is almost always distorted early; whereas facial numbness occurs sooner once trigeminal compression occurs (often in the presence of normal facial movement), despite the fact that the 5th nerve is farther away²²⁸. This may be due to the resiliency of motor nerves relative to sensory nerves.

Symptoms from compression of brainstem and other cranial nerves

Larger tumors cause brainstem compression (with ataxia, H/A, N/V, diplopia, cerebellar signs, and if unchecked, coma, respiratory depression and death) and lower cranial nerve (IX, X, XII) palsies (hoarseness, dysphagia…). Obstruction of CSF circulation by larger tumors (usually > 4 cm) may produce hydrocephalus with increased ICP.

Rarely, 6th nerve involvement may cause diplopia.

SIGNS

Hearing loss due to VIII involvement is the earliest cranial nerve finding. 66% of patients have no abnormal physical finding except for hearing loss (for other findings, see Table 21-28).

Since hearing loss is sensorineural, Weber test will lateralize to the uninvolved side, and if there is enough preserved hearing, Rinne test will be positive (i.e. normal; air conduction > bone conduction) on both sides (see page 848 for these tests).

Table 21-27 Clinical grading of facial nerve function (House and Brackmann²²⁹)

Grade	Function	Description
1	normal	normal facial function in all areas
2	mild dysfunction	1. gross: slight weakness noticeable on close inspection; may have very slight synkinesis 2. at rest: normal symmetry and tone 3. motion: A. forehead: slight to moderate movement B. eye: complete closure with effort C. mouth: slight asymmetry
3	moderate dysfunction	1. gross: obvious but not disfiguring asymmetry: noticeable but not severe synkinesis 2. motion: A. forehead: slight to moderate movement B. eye: complete closure with effort C. mouth: slightly weak with maximal effort
4	moderate to severe dysfunction	1. gross: obvious weakness and/or asymmetry 2. motion: A. forehead: none B. eye: incomplete closure C. mouth: asymmetry with maximum effort
5	severe dysfunction	1. gross: only barely perceptible motion 2. at rest: asymmetry 3. motion: A. forehead: none B. eye: incomplete closure
6	total paralysis	no movement

Table 21-28 Signs in 131 vestibular schwannomas (excluding hearing loss)²²²

Sign	%
abnormal corneal reflex	33
nystagmus	26
facial hypoesthesia	26
facial weakness (palsy)	12
abnormal eye movement	11
papilledema	10
Babinski sign	5

Facial nerve (VII) dysfunction is uncommon before treatment. When present, it is usually graded clinically on the House and Brackmann scale (see Table 21-27).

Vestibular involvement causes nystagmus (may be central or peripheral) and abnormal electronystagmography (ENG) with caloric stimulation.

DIFFERENTIAL DIAGNOSIS

See Cerebellopontine angle (CPA) lesions on page 1210. The major differentials are: meningioma, or neuroma of an adjacent cranial nerve (e.g. trigeminal).

PATHOLOGY

Tumors are composed of Antoni A fibers (narrow elongated bipolar cells) and Antoni B fibers (loose reticulated). Verocay bodies are also seen, and consist of acellular eosinophilic areas surrounded by parallel arrangement of spindle shaped schwann cells.

EVALUATION

1. brain MRI without and with contrast. FIESTA MRI if available. If MRI is contraindicated, then a CT scan without and with contrast

2. audiometric evaluation:

A. pure tone audiogram (see below)

B. speech discrimination evaluation (see below)

C. patients with small VSs (≤ 15 mm dia) also get:

1. ENG: assesses superior vestibular nerve (see page 624)

2. VEMP: assesses inferior vestibular nerve (see page 624)

3. ABR: prognosticates chance of hearing preservation (see page 624)

AUDIOMETRIC AND AUDIOLOGIC STUDIES

Baseline studies are helpful for management treatment decisions and for later comparison and to assess the contralateral ear.

Pure tone audiogram (PTA)

May be useful as first-step screening test. Air conduction assesses the entire system, bone conduction assesses from the cochlea and proximally. PTA assesses the functionality of hearing (to help in treatment decision making) and acts as a baseline for future comparison. The single numerical score is an average of the thresholds for frequencies across the audio spectrum. On a standard audiogram, X’s denote the left ear (AS) and O’s denote the right ear (AD).

Progressive unilateral or asymmetric sensorineural hearing loss of high tones occurs in > 95% of VSs²³⁰. High-frequency hearing loss also happens to be the most common type of hearing loss with age or with noise induced sensorineural hearing loss, but is usually symmetrical. Only ≈ 1 in 1000 patients with asymmetric hearing have a VS²²⁰. Other causes of asymmetrical sensorineural hearing loss²³¹: other CPA lesions (e.g. meningioma), inner ear lesions, intraaxial lesions (including 9 infarctions), multiple sclerosis. On hearing screening tests, an unexplained PTA difference from one ear to the other > 10-15 dB is suspicious and should be investigated further.

Speech discrimination evaluation

Speech discrimination is maintained in conductive hearing loss, moderately impaired in cochlear hearing loss, and worst with retrocochlear lesions. No longer used for diagnostic purposes (a score of 4% suggests a retrocochlear lesion, as does a score that is worse than would be predicted based on PTA testing (the speech recognition threshold should be similar to PTA thresholds below 4 kHz)). Has found usefulness in determining serviceability of hearing and prognosticating for hearing preservation surgery. Open-set word recognition score (WRS) (see Table 21-29) is a more sensitive measure of communication ability than PTA.

Table 21-29 Open-set word recognition score

Class	WRS%
I	70-100%
II	50-69%
III	1-49%
IV	0

Definition of serviceable hearing

There are many definitions of what constitutes serviceable hearing. Also, even non-serviceable hearing can offer some benefit. If WRS is good (≥ 70%) but PTA is poor, a hearing aid may provide significant benefit.

Some definitions of serviceable hearing (see text that follows for details):

1. AAO-HNS class A or B

2. “50/50 rule”: Gardner-Robertson class I or II (pure tone audiogram threshold ≤ 50 dB and speech discrimination score ≥ 50%)

3. some prefer a 70/30 rule (70% WRS, 30 dB PTA)

4. in a patient with good hearing in the contralateral ear, a speech discrimination score (SDS) of < 70% in the affected ear is not considered good hearing; whereas if the contralateral ear is totally deaf, a SDS of ≥ 50% can be useful²³²

Modified Gardener-Robertson system for grading hearing: shown in Table 21-30. Class I patients may use a phone on that side, class II patients can localize sounds.

The American Academy of Otolaryngology - Head and Neck Surgery Foundation (AAO-HNS) hearing classification system²³³: shown in Table 21-31.

Additional tests that are helpful with small VSs (≤ 15 mm diameter)

The ENG and VEMP evaluate the superior and inferior division of the vestibular nerve (VN) respectively. The inferior VN is closer to the cochlear nerve than the superior VN (see Figure 5-7, page 90), and small tumors (≤ 4 mm) of the inferior VN tend to be deeper and closer to the cochlear nerve than similar sized tumors of the superior division which tend to be more superficial and more easily removed.

Electronystagmography (ENG): Only tests the horizontal semicircular canal assesses the superior vestibular nerve which innervates it. Normally, each ear contributes an equal portion of the response. The ENG is considered abnormal if there is > 20% difference between the two sides. Response may be normal with a small tumor arising from the inferior division of vestibular nerve. NB: the vestibular nerve may continue to function until almost all of the nerve fibers are affected.

Vestibular evoked myogenic potential (VEMP): Assesses inferior vestibular nerve by testing the saccule²³⁶. Independent of hearing (can be done even with severe sensorineural hearing loss).

Auditory brainstem responses (ABR): AKA BAER (see page 267). The most common findings are prolonged I-III and I-V interpeak latencies. No longer used for diagnostic purposes (sensitivity is only ≈ 88-90% (i.e. will miss 10-12% of VSs) and specificity is only 85%. ABR is useful for prognostication - poor wave morphology correlates with lower chance of preserving hearing (even with good hearing).

RADIOGRAPHIC EVALUATION

MRI: Thin slice axial plane gadolinium enhanced MRI is the diagnostic procedure of choice with sensitivity close to 98% and almost 0% false positive rate. Characteristic findings: round or oval enhancing tumor centered on IAC. Large VSs (> 3 cm dia) may show cystic appearing areas on CT or MRI; in actuality these areas are usually solid. Adjacent trapped CSF cisterns may also give cystic appearance.

FIESTA MRI (fast imaging employing steady state acquisition): uses CSF as the contrast agent ( does not use gadolinium).

CT scan: CT with IV contrast is second choice for imaging modality. If normal, and clinical suspicion of VS is strong, small lesions may be visualized by introducing 3-4 ml of subarachnoid air via lumbar puncture, and scanning the patient with the affected side up (to trap air in region of IAC), non-filling of the IAC is indicative of an intracanalicular mass. Even with air contrast, CT was normal in 6% in Mayo series²²². Although many VSs enlarge the ostium of the IAC (called trumpeting) (normal diameter of the IAC is = 5-8 mm), 3-5% of VSs do not enlarge the IAC on CT; this percent may increase as patients are scanned earlier with smaller tumors. Advantage over MRI: shows bony anatomy (including mastoid air cells) which is often helpful for planning translabyrinthine approach.

MANAGEMENT

Options for management include:

1. expectant management: follow symptoms, hearing (audiometrics) and tumor growth on serial imaging (MRI or CT). Intervention is performed for progression. Growth patterns observed:

A. little or no growth: applies to most (83%) VSs confined within the IAC and 30% extending into CPA (see natural history of growth below)

B. slow growth ≈ 2 mm/yr

C. rapid growth: ≥ 10 mm/yr

D. a few actually shrink²²³

2. radiation therapy (alone, or in conjunction with surgery)

A. external beam radiation therapy (EBRT)

B. stereotactic radiation

1. stereotactic radiosurgery (SRS): single dose (see page 775)

2. stereotactic radiotherapy (SRT): fractionated (see page 776)

3. surgery: approaches include the following (see below for details)

A. retrosigmoid (AKA suboccipital): may be able to spare hearing

B. translabyrinthine (and its several variations): sacrifices hearing, may be slightly better for sparing VII

C. middle fossa approach (extradural subtemporal): only for small lateral VSs

4. chemotherapy: some early promise for NF2-related vestibular schwannomas with bevacizumab (Avastin®), an anti-VEGF (vascular endothelial growth factor) monoclonal antibody. In 6 such patients, 4 had radiographically significant tumor shrinkage and 4 had improvement in auditory word recognition score²³⁷

Patient/tumor factors influencing management decisions

In addition to the usual factors involved in the decision process with brain tumors, e.g. the patient’s general medical condition, age, natural history, etc., elements unique to VSs include: chances of preserving VII & V nerve function and hearing (in those with serviceable hearing) (all of which are related to tumor size), and the presence of NF2.

Specifics:

1. natural history of growth

A. usual quoted range: ≈ 1-10 mm/yr. However this can be quite variable

B. strictly intracanalicular tumors: only 17% grew outside the meatus^A

C. extrameatal tumors (with extension into CP angle): 30% grew > 2 mmA

D. VSs that did not grow in the 5 years after diagnosis did not grow after that

E. 6% actually decrease in size²³⁸

2. natural history of hearing function in untreated intracanalicular VSs in AAOHNS Group A (see Table 21-31) patients²³⁹

A. 50% deteriorated to a lower class over 4.6 years (loss of ≥ 10 dB PTA or ≥ 10% SDS)

B. after 4.6 years of observation, the proportion of patients eligible for hearing preservation treatment (as determined by a word recognition score class I (70-100% SDS)) was reduced to 28% (a 44% reduction) and by AAO-HNS class A to 9% (a 53% reduction)

C. the risk of losing hearing was not related to: age, gender, acoustic tumor size (all tumors were intracanalicular) or tumor sublocalization (fundus, central, porus)

D. hearing loss was positively correlated to the absolute volumetric tumor growth rate (tumors that eventually expand out of the IAC have a faster rate and degree of hearing loss compared to tumors remaining in the IAC

E. the risk of losing hearing was significantly lower for patients with 100% word recognition score. Over 4.6 years observation, 89% remained in WRS class I (see Table 21-29) compared to only 43% for patients with only a small (1-10%) loss of WRS at diagnosis

3. size: as tumors exceed 15 mm diameter, treatment complications increase

A. significantly lower chance for hearing preservation

B. increased incidence of VII injury

4. presence of cysts: cystic tumors may display sudden and dramatic growth²²³

5. serviceable hearing: see Definition of serviceable hearing, page 623

6. hearing in contralateral ear

A. in 522 VSs over 3.6 years mean follow-up 223

Management algorithm

1. intracanalicular or CPA tumors ≤ 20 mm diameter that are noncystic & non-NF2: observation with serial imaging and hearing tests (“wait and scan” scheme). Follow-up schedule

A. imaging: F/U CT or MRI (treatment for > 2 mm growth between studies)

1. q 6 mos x 2 yrs (i.e. 1 & 2 years after diagnosis)

2. if stable, then annually until year 5 after diagnosis

3. if stable, then at years 7, 9 & 14 after diagnosis²²³

B. annual audiology evaluations: in patients with small tumors and normal SDS, comparing the results of hearing preservation with surgery or SRS to the natural history, the conclusion is that established tumor growth should be the main determinant for treatment²³²

2. tumors > 15-20 mm should be treated^{223, 232}. However, this also must take into account the patient’s age, hearing…

3. NF2 patients present a challenge and should be evaluated individually. In general the success rate in the management of their tumors is lower (higher cranial nerve deficit and higher recurrence rate)^{240, 241}. Early management is considered more favorable for good outcome²⁴². Recent attempts with chemotherapy using bevacizumab (Avastin®) appears to be a promising albeit still investigational option²³⁷ (see above)

Selection of treatment option

Comparison of microsurgery vs. radiosurgery (SRS)

1. hearing preservation

A. for patients with testable pre-operative hearing

1. summary: radiosurgery or stereotactic radiation appears to be better at preserving hearing than microsurgery. The difference is minor for tumors < 10 mm and very good pre-operative hearing (70% SDS, and 30 dB PTA). The advantage of radiation is more pronounced for larger tumors and greater pre-operative hearing loss. Details:

2. SRS: overall, at 3, 5, and 10 years, 81%, 77% and 66% of the patients maintained their GR hearing class (see Table 21-30, page 623). For patients receiving a tumor margin dose of 13 Gy or less, those same percentages were 93%, 87% and 87% 243. Hearing preservation appears to be related to the radiation dose to the cochlea rather than to the tumor itself²⁴⁴

3. microsurgery: hearing preservation is significantly related to the tumor size and to the experience of the surgical team. Hearing preservation in Samii’s series of 1000 VS²⁴¹ improved from 24% in the first 200 cases to 49% in later cases. Hearing preservation in microsurgery has improved with the use of direct cochlear nerve monitoring²⁴⁵ compared to auditory brainstem responses monitoring only. Hearing preservation in patients with class A, small tumors and direct cochlear nerve monitoring (compound nerve action potential) was 91%²⁴⁶

2. facial nerve preservation

A. preservation has been excellent with both microsurgery and radiosurgery

B. microsurgery: 98.5% overall²⁴⁷ and 100% in tumors not touching the brainstem. Staged resection has been advocated by some to improve facial nerve preservation in giant VS (> 4-4.5 cm)²⁴⁸

C. radiosurgery: 98% of patients²⁴³. The incidence of facial neuropathy has significantly decreased since the SRS dose was decreased to 12-13 Gy. Facial neuropathy in the recent series occurred in patients having received 18-20 Gy

3. trigeminal neuropathy (TGN)

A. a complication classically feared in large tumors especially following SRS

B. SRS: 7% incidence of TGN (mainly in patients receiving higher doses, i.e. 18 Gy). No patients who received a dose < 13 Gy developed TGN²⁴³

C. microsurgery: post-op TGN is not reported in most series

4. tumor control (local control rates (LCR)):

A. tumor control has been a concern with radiosurgery and with the more recent decrease in dose from 18-20 Gy to 12-14 Gy, long term data are lacking

B. microsurgery: tumor recurrence has been poorly studied. Quoted rates in the literature vary between 0.5% at 6 years²⁴⁷ to 9.2%²⁴⁹

C. SRS: tumor recurrence requiring retreatment at 5 years was 4%²⁴³ but 18% of patients presented with transient increase in the size of the tumor (“pseudogrowth”) at a mean of 8 months, with later regression in half and stabilization to the new size in the other half

Vertigo: For patients with episodic vertigo or balance difficulties as the predominant symptom (also, see points under Selection of treatment option on page 625):

1. remember: patients with VS are susceptible to other causes of vertigo as well, and patients should undergo ENG and functional balance assessment

2. vertigo that is due to the VS is often self-limited, and improves in 6-8 weeks to a reasonably tolerable level with no treatment (patients may do better with so-called “vestibular rehab”)

3. residual dizziness and balance disturbances are common whether stereotactic radiosurgery (SRS) or microsurgery (MS) is used, but are typically less after MS

4. after SRS: beneficial effects require a minimum of 5-6 mos, and sometimes may require up to eighteen months

5. following MS: symptoms are usually immediately worsened, but then gradually improve in most cases (except perhaps when the balance difficulties are due to brainstem compression). Symptoms are improved more rapidly than with SRS

6. conclusion: observation may be the best choice for ≈ 20% of patients. When treatment is desired, surgery is the best choice for most VSs producing vertigo. SRS may be the right choice for some, especially: elderly patients (> 70 yrs) with other health problems, for recurrence of VS, and for individual preference

Hydrocephalus: When hydrocephalus is present, it may require separate treatment with a CSF shunt (see Surgical considerations, page 627), and may possibly be done at the same time as surgery for the VS (if indicated).

SURGICAL TREATMENT

Approaches

Three basic surgical approaches:

1. those with possibility of hearing preservation

A. middle fossa (MF): poor access to posterior fossa (see below)

B. retrosigmoid (RS) (see page 628) AKA retrosigmoid-transmeatal approach

2. translabyrinthine (TL): non hearing preserving (see below)

Excellent results have been reported with each of these approaches. These guide-lines assume that the surgical team is comfortable with all three approaches.

Decision algorithm:

The choice of approach is dictated by hearing salvageability and tumor size as follows:

1. salvageable hearing (see Table 21-32 for definition and guidelines)

A. if tumor is intracanalicular (no extension beyond a few mm into the posterior fossa (CPA)^A): middle fossa approach

B. if tumor extends > few mm into the posterior fossa: retrosigmoid approach

2. non salvageable hearing (see Table 21-32 for definition and guidelines)

A. translabyrinthine approach for most

B. if the neurotologist feels that the tumor is too large to remove via a translabyrinthine approach^B: retrosigmoid approach

A. differences of opinion exist regarding how much tumor in the CPA can be removed via MF

B. may be due to a small presigmoid space and/or a large tumor

Table 21-32 Hearing salvageability

Definition of serviceable hearing

A generous definition of serviceable hearing: PTA < 50 dB and SDS > 50%*

Unsalvageable hearing

Serviceable hearing is unlikely to be preserved post-op when 1. pre-op SDS < 75% 2. or pre-op PTA loss > 25 dB 3. or pre-op BAER has abnormal wave morphology 4. or tumor > 2-2.5 cm diameter

* for other definitions of serviceable hearing see page 623

SURGICAL CONSIDERATIONS

The superior vestibular division of VIII is the usual origin of the tumor. The facial nerve is pushed forward by the tumor in ≈ 75% of cases (range: 50-80%), but may occasionally be pushed rostrally, less often inferiorly, and rarely posteriorly. It may even continue to function while it is flattened to a mere ribbon on the tumor capsule surface.

Anesthesia with minimal muscle relaxants allows intra-op seventh nerve monitoring. In only ≈ 10% of large tumors is the cochlear nerve a separate band on the tumor capsule, in the remainder it is incorporated into the tumor.

Total excision of tumor is usually the goal of surgery. The only indications for planned subtotal resection is a large tumor on the side of the only ear with good hearing or those patients requiring debulking with little chance of recurrence because of limited life expectancy, especially if the facial nerve is densely adherent to the tumor^{250, 251}.

If hydrocephalus is present, it used to be standard practice to place a CSF shunt and wait ≈ 2 weeks before the definitive operation²⁵². While still acceptable, this is less commonly done at present.

• large tumors may be approached by a combined translab-retrosigmoid approach to debulk tumor and preserve facial nerve; a two stage approach (with 1-2 weeks in between) may improve results with very large tumors²⁵³

MIDDLE FOSSA APPROACH

• indications:

A. hearing preservation

B. laterally placed tumors

C. small tumors (usually < 2.5 cm)

• pros:

A. allows drilling and exposure of the IAC all the way to the geniculate ganglion (good for laterally placed tumors)

B. basically an extradural subtemporal operation

• cons:

A. potential damage to temporal lobe with risk of seizures

B. facial nerve is the most superficial nerve in this exposure and therefore the surgeon works “around” the facial nerve (possibility of injury)

• technique summary

A. lumbar drain

B. usually straight incision, starting in front of the tragus, extending cephalad for 6 cm, held open with a self retaining retractor

C. the temporalis muscle is incised vertically (along the muscle fibers) along the most posterior aspect of the exposure, as well and reflected anteriorly

D. craniotomy: 4 cm x 3 cm

E. elevate the middle fossa dura, section the middle meningeal artery. Identify and preserve the greater superficial petrosal nerve (GSPN), arcuate eminence, V3, and true edge of the petrous bone (the false edge is the groove occupied by the superior petrosal sinus

F. drill and expose the internal auditory canal all the way to Bill’s bar (for tumors extending laterally)

G. localize the facial nerve with the nerve stimulator

H. open the IAC dura along the main axis of the IAC, avoiding VII

I. identify the vestibular, cochlear and facial nerves

J. dissect the tumor off the nerves

TRANSLABYRINTHINE APPROACH

Especially useful for tumors with primarily intracanalicular component with little CPA extension. Often preferred by neurootologists.

• pros & cons: see Table 21-33

• technique summary

A. skin incision should be tailored to the location of the sigmoid sinus (observe location of the sigmoid sinus and pinna of the ear on the pre-op MRI). Usually smaller opening than retrosigmoid approach

B. does not require a craniotomy. For large tumors requiring an “extended translab”, 1-2 cm of retrosigmoid dura should be exposed during the mastoidectomy to allow for retraction of the sigmoid sinus

C. dural opening along the IAC after identification of VII with stimulator

D. for large tumor: section the superior petrosal sinus and section the tentorium to gain better intradural exposure

E. closure requires fat graft

Table 21-33 Pros & cons of translab approach

Disadvantages	Advantages
• sacrifices hearing (acceptable when hearing is already nonfunctional or unlikely to be spared by other approach) • limited exposure (limits maximal tumor size that can be approached) • may take longer than retrosigmoid approach • possibly higher rate of post-op CSF leak	• early identification of VII may result in higher preservation rate • less risk to cerebellum and lower cranial nerves • patients do not get as “ill” from blood in cisterna magna, etc. (essentially an extracranial approach)

RETROSIGMOID APPROACH

AKA posterior fossa, AKA suboccipital approach^{254, 255}.

• pros:

A. familiar to most neurosurgeons often preferred by neurosurgeons

B. quick access to the tumor

C. hearing preservation possible

D. NOTE: this approach is very versatile. Samii²⁴¹ resected all his acoustic tumors via a retrosigmoid approach^A

• cons:

A. cerebellar retraction: not a problem for tumors < 4 cm, provided the craniotomy is sufficiently lateral and the cisterna magna and the CP angle cistern has been opened

B. headaches: it has been suggested that headaches are more common in retrosigmoid craniotomy than in the translabyrinthine craniotomy

A. he achieved a significant amount of brain relaxation and improved exposure by using in the sitting position, which is generally not used in the USA because of associated complications (see page 153)

Booking the case - retrosigmoid craniotomy for vestibular schwannoma

Also see defaults & disclaimers (page v).

1. position: supine with shoulder roll

2. equipment:

A. microscope

B. ultrasonic aspirator

C. image guided navigation system (if used) (may be helpful for placing skin incision and craniotomy more than for tumor localization except with large tumors)

3. some surgeons use ENT to assist with the IAC and for follow-up

4. neuromonitoring: facial EMG (does not require EEG tech), direct cochlear nerve monitoring and SSEPs (if used, requires EEG tech)

5. post op: ICU

6. consent (in lay terms for the patient - not all-inclusive):

A. procedure: surgery through an incision behind the ear to remove a tumor in the skull on the nerve to the ear; possible need for post-op lumbar drain

B. alternatives: nonsurgical management with follow-up MRIs, other surgical approaches, radiation (stereotactic radiosurgery)

C. complications: CSF leak with possible meningitis, loss of hearing in ipsilateral ear (if not already lost), paralysis of facial muscles on the side of surgery with possible need for surgical procedures to help correct (correction is often far from perfect), post-op balance difficulties/vertigo, brainstem injury with stroke

Technique summary

1. position: 30° elevation of the head is paramount (see Posterior fossa (suboccipital) craniectomy, Lateral oblique position on page 154)

2. percutaneous lumbar drain (optional)

3. incision is shaped like the pinna of the ear, 3 finger breaths behind the external auditory canal

4. the craniotomy has to be lateral enough to expose part of the sigmoid and part of the transverse sinuses.

5. to prevent CSF leak, seal all bone edges with bone wax

6. dural opening along the lines of the craniotomy

7. exposure is enhanced by opening the cerebello-pontine angle cistern and the cisterna magna under the microscope and draining CSF (20-40 ml of CSF can also be drained via a lumbar subarachnoid catheter)

8. the petrosal vein is often sacrificed at the beginning of the procedure to allow the cerebellum to relax and fall back and to avoid tearing off the transverse sinus. Be careful not to coagulate the SCA that often runs with the petrosal vein

9. using the facial nerve stimulator, the posterior aspect of the tumor is inspected to make sure the facial nerve has not been pushed posteriorly

10. the thin layer of arachnoid that covers most tumors is identified. Vessels within the arachnoid may contribute to cochlear function and may be preserved by keeping them with the arachnoid

11. the plane between tumor and cerebellum may be followed to the brainstem, and occasionally to the VII nerve (this plane is harder to follow once bleeding from tumor debulking occurs)

12. to help locate the origin of the VII nerve at the brainstem see Table 21-34 and CPA anatomy in Figure 5-8, page 90

13. the posterolateral tumor capsule is opened, and internal decompression is performed. The tumor is collapsed inward and the capsule is kept intact and is rolled laterally off of VII and is eventually removed. The most difficult area to separate VII from tumor is just proximal to the entrance to the porus acusticus

Table 21-34 Aids in localizing VII nerve origin²⁵⁶

• VII nerve originates in the pontomedullary sulcus near the lateral end of the sulcus, 1-2 mm anterior to the VIII nerve • the pontomedullary sulcus ends just medial to the foramen of Luschka (extending from the lateral recess of the IV ventricle) see Figure 5-8 • a tuft of choroid plexus usually extends out of the foramen of Luschka on the posterior surface of IX and X nerve, just inferior to the origin of VII • the flocculus of the cerebellum projects from the lateral recess into the CPA just posterior to the origin of VII and VIII • VII origin is 4 mm cephalad and 2 mm anterior to that of the IX nerve

NB: large tumors: in some large tumors, the capsule may be adherent to the brainstem and so portions of tumor must be left; recurrence rate among these is ≈ 10-20%¹²⁷. Large tumors may also involve V superiorly (sometimes VII is pushed up against V), and inferiorly may involve IX, X, and XI. The lower cranial nerves can usually be spared by dissecting them off of the tumor capsule, and protecting them with cottonoids.

14. after the extracanalicular portion of tumor is removed, the dura over the IAC is incised, and the IAC is drilled open and tumor is removed from this portion. To preserve hearing, the bony labyrinth must not be violated. The posterior semicircular canal (SCC) is the most vulnerable structure (see Figure 21-2). The vestibule of the SCCs is also at risk but is less likely to be entered. The maximal amount of temporal bone drilling that can be accomplished without entering the posterior SCC can be determined from the pre-op CT. There is no exact anatomical landmark, some say that the IAC must not be opened lateral to the transverse crest which is ≈ 1 cm deep within the IAC, others recommend measuring the distance to the SCC on a pre-op CT and not opening the lateral 1-2 mm of the IAC²⁵⁷. However, opening the labyrinth cannot always be avoided; and any opening should be plugged with bone-wax or muscle²⁵⁷. If the facial nerve is not intact and is not going to be grafted, then the IAC should be plugged, e.g. by bone wax covered with a small piece of hammered muscle (hammering makes the muscle sticky by activating extrinsic clotting factors) and Gelfoam®.

Figure 21-2 Structures of the left temporal bone. CT scan (left petrous bone, axial slice) provided courtesy of Chris Danner, M.D.

POST-OP CARE & CARE FOR COMPLICATIONS

Cranial nerve and brainstem dysfunction

Facial nerve (VII): If eye closure is impaired due to VII dysfunction: Rx natural tears 2 gtts to affected eye q 2 hrs and PRN. Apply Lacrilube® to affected eye and tape it shut q hs. If there is complete VII palsy with little chance of early recovery, or if facial sensation (Vth nerve) is also impaired, tarsorrhaphy is performed within a few days.

Facial re-animation (e.g. hypoglossal-facial anastomosis) is performed after 1-2 months if VII was divided, or if no function returns after 1 year with an anatomically intact nerve.

Vestibular nerve (VIII): Vestibular dysfunction is common post-op, nausea and vomiting due to this (and also intracranial air) is common. Balance difficulties due to this clear rapidly, however, ataxia from brainstem dysfunction may have a permanent component.

Lower cranial nerves: The combination of IX, X and XII dysfunction creates swallowing difficulties and creates a risk of aspiration.

Brainstem dysfunction: Brainstem dysfunction may occur from dissection of tumor off of the brainstem. This may produce ataxia, contralateral paresthesias in the body… Although there may be improvement, once present, there is often some permanent residual.

CSF fistula

Also, see CSF fistula (cranial), page 300 for general information. CSF fistula may develop through the skin incision, the ear (CSF otorrhea) through a ruptured tympanic membrane, or via the eustachian tube and then through the nose (rhinorrhea) or down the back of the throat.

Rhinorrhea may occur through any of the following routes (circled numbers in Figure 21-3):

via the apical cells to the tympanic cavity (TC) or eustachian tube (the most common path)

• entry into the bony labyrinth - in order to reach the middle ear would require rupture e.g. of the oval window by overpacking bone wax into the labyrinth)

through the vestibule of the horizontal semicircular canal (SCC)

through the posterior SCC (the posterior SCC is the most common area that is entered by drilling)

follows the perilabyrinthine cells and tracts to the mastoid antrum

through the mastoid air cells at the craniotomy site

Most leaks are diagnosed within 1 week of surgery, although 1 presented 4 years post-op²⁵⁸. They appear to be more common with more lateral unroofing of the IAC²⁵⁸. Meningitis complicates a CSF leak in 5-25% of cases, and usually develops within days of the onset of leak²⁵⁸. Hydrocephalus may promote the development of a CSF fistula.

Figure 21-3 Possible routes for CSF rhinorrhea following vestibular schwannoma surgery (see text) (right petrous bone, axial slice). Adapted from Surgical Neurology, Vol. 43, Nutik S L, Korol H W, Cerebrospinal Fluid Leak After Acoustic Neuroma Surgery, 553-7, 1995, with permission from Elsevier Science

Treatment:^25-35% of leaks stop spontaneously (one series reported 80%)²⁵⁸. Treatment options include:

1. non-surgical:

A. elevate HOB

B. if leak persists: a percutaneous lumbar subarachnoid drain may be tried^{259, 260}, although some debate its efficacy²⁵⁵, and there is a theoretical risk of drawing bacteria into the CNS

2. surgical treatment for persistent leaks:

A. re-explore and coat mastoid air cells with bone wax, free-muscle grafts, or fibrin glue. Rewaxing the exposed air cells fails to stop the leak in ≈ 38% of cases, but is still the procedure of choice if hearing is preserved²⁵⁸(current production methods of bone wax may result in higher leakage rates if the dural closure is not watertight than with the older formulation²⁶¹)

B. cover bone surface with dural flap, pericranium or fascia lata

C. when the IAC has been drilled and if the VII & VIII nerves are completely lost: pack IAC with muscle

D. if no hearing on that side (the following is usually done in collaboration with an ENT or skull-base surgeon)

1. reuse the same incision, but stay extradural and utilize a middle-ear approach to perform a mastoidectomy and pack the area with fat²⁶². Fails in 4-23% of cases²⁵⁸

2. more aggressive treatment: fat obliteration of the eustachian tube, middle ear, and mastoid²⁶³ (occludes drainage from apical air cells or oval window)²⁵⁸

E. if the leak occurred because of the development of hydrocephalus, adjunctive CSF shunting is usually also necessary

OUTCOME & FOLLOW-UP

Complete surgical removal was reported in 97-99% of cases²⁶⁴.

SURGICAL MORBIDITY AND MORTALITY

Also see Post-op considerations for p-fossa crani’s, page 157. Estimated frequency of some complications²⁶⁵: CSF leakage in 4-27%²⁵⁸ (see above), meningitis in 5.7%, CVA in 0.7%, subsequent requirement for CSF shunt (for hydrocephalus or to treat leak) in 6.5%.

The mortality rate is ≈ 1% at specialized centers^{241, 264, 266}.

CRANIAL NERVE DYSFUNCTION

Table 21-35 shows statistics of VII and VIII cranial nerve preservation following suboccipital removal of VSs. For more details, see below.

Post-radiation cranial neuropathies generally appear 6-18 months following stereotactic radiosurgery (SRS)²⁶⁸, and since more than half resolve within 3-6 months after the onset the recommendation is treat these with a course of corticosteroids.

Table 21-35 Cranial nerve preservation in retrosigmoid removal of VSs*

Size of tumor	Preserved function
	VII nerve	VIII nerve
< 1 cm	95-100%	57%
1-2 cm	80-92%	33%
> 2 cm	50-76%	6%

* series of 135 VSs^{267 (p 729)} and other sources^{127 (p 3337), 264}

Facial nerve (VII)

See Table 21-27, page 622 for the House and Brackmann grading scale. Grades 1-3 are associated with acceptable function. In one surgical series, the facial nerve was preserved with all tumors ≤ 2 cm; it was preserved only in 29% of tumors > 4 cm²²². Continuous recording of spontaneous EMG activity and responses to electrical stimulation during surgery may improve preservation of VII nerve^{269, 270}. If VII is anatomically preserved, partial post-op facial weakness will usually resolve, but may take up to one year. In ≈ 13% of cases, anatomic preservation of VII is not possible.

SRS for tumors ≤ 3 cm diameter²⁷¹: transient VII weakness occurred in 15%, and V dysfunction (usually temporary) developed in 18%. In another series²⁷², 92% of cases had grade 1-2 function post op (compared to 90% for microsurgery²⁷³). Stereotactic radiation therapy (SRT) had 2% incidence of new facial palsy²⁷⁴.

Vestibulo-acoustic nerve (VIII)

Patients with unilateral VS and Class I or II hearing (see Table 21-30, page 623) comprised ≈ 12% of cases in a large series²⁷⁵. Preservation of hearing is critically dependent on tumor size, with little chance of preservation with tumors > 1-1.5 cm diameter. Chances of preserving hearing may possibly be improved by intraoperative brainstem auditory evoked potential monitoring²⁷⁶. In centers treating large numbers of VSs, hearing preservation rates of 35-71% can be achieved with tumors < 1.5 cm^{275, 277} (although a range of 14-48% may be more realistic²⁷⁸). Hearing may rarely be improved post-op²⁷⁹.

SRS: for tumors ≤ 3 cm diameter²⁷¹, hearing was preserved in 26% of 65 cases with pre-op pure tone threshold < 90 dB. Hearing loss has been correlated with increase in tumor size²⁸⁰. • NB: there is a highrate of hearing loss at 1 year. SRT: useful hearing was preserved in 93%²⁷⁴.

Vestibular nerve function is rarely normal post-op. Attempts at “vestibular” sparing surgery have shown no better results than surgery not specifically addressing this issue. Most patients with unilateral loss of vestibular nerve function will learn to compensate to a significant degree with input from the contralateral side, if normal. Patients with ataxia as a result of brainstem injury from the tumor or the surgery will have more difficulties post-op. Some patients will seem to do well initially post-op with respect to vestibular nerve function, only to undergo a delayed deterioration several months post-op. These cases likely represent aberrant regeneration of the vestibular nerve fibers and may be extremely difficult to manage. Some experts advocate cutting the vestibular nerve (as for Meniere’s disease, see page 840).

Trigeminal nerve (V)

Postoperative trigeminal nerve symptoms occur transiently in 22% and permanently in 11% following microsurgery, similar to the results of SRS²⁷². New facial numbness occurred in 2% with SRT²⁷⁴.

Lower cranial nerves

Injuries to IX, X and XI occur infrequently following surgery on large tumors that distort the nerves and displace them inferiorly against the occipital bone.

RECURRENCE

Following microsurgery (MS)

Recurrence is highly dependent on extent of removal. However, recurrence can develop in tumors that were apparently totally removed, or when subtotal resection was performed. This can occur many years after treatment. Tumor progression rate following subtotal resection is ≈ 20%²⁷⁸. All patients should be followed with imaging (CT or MRI). In older series with up to 15 yrs follow-up, local control rate (LCR) after “total resection” is ≈ 94%. More recent series with MRI follow-up indicate recurrence rates of 7-11% (3-16 yrs follow-up)²⁷⁸.

Use of EBRT

EBRT may improve LCR in incompletely resected tumors as shown in Table 21-36 (note: with the long survival expected with benign tumors, post XRT complications may occur).

Table 21-36 Local control rates of surgery vs. surgery + EBRT for VSs²⁸¹

Extent of surgical removal	Local control rate (LCR)
	Surgery	Surgery + EBRT*
gross total	60/62 (97%)	no data
near total (90-99%)	14/15 (93%)	2/2 (100%)
subtotal (< 90%)	7/13 (54%)	17/20 (85%)*
biopsy only	no data	3/3 (100%)

* with doses < 45 Gy, LCR was 33%; with > 45 Gy LCR was 94%

Microsurgery vs. SRS

The long-term results for SRS using the current recommended dose of 14 Gy are still not known²⁸². In a non-randomized retrospective study²⁷² of VSs < 3 cm dia, the short-term LCR (median 24 mos follow-up) was 97% for microsurgery vs. 94% for stereotactic radiosurgery (SRS). However, for benign tumors, long-term followup is critical, and this study suggests that the long-term LCR will be better for MS than SRS. SRS studies with long-term follow-up²⁸³ are not directly comparable because in the cases with longest follow-up, higher radiation doses were used with a resultant higher incidence of radiation complications, and an anticipated better LCR.

Initially there may be temporary enlargement of the tumor accompanied by loss of central contrast enhancement following SRS in ≈ 5% of patients²⁸⁴ (with up to 2% of patients showing actual initial tumor growth), and so the need for further treatment after SRS should be postponed until there is evidence of sustained growth²⁸⁵. Surgery should be avoided during the interval from 6 to 18 months after SRS because this is time of maximum damage from the radiation²⁸⁵.

Although the numbers are small, there have been indications that the rate of VII nerve injury may be higher in patients undergoing microsurgery following SRS failure^{286, 287}, however, this has been disputed²⁸⁵. Lastly, there is a potential for malignant trans-formation following SRS including triton tumors^{288, 289} (malignant neoplasms with rhabdoid features) or the induction of skull base tumors (reported with external beam radiation²⁹⁰), as well as the risk of late arterial occlusion (the AICA lies near the surface of VSs), any of which may occur many years later.

Treatment for recurrence following microsurgery

Repeat surgery for recurrent VS is an option. One series of 23 patients²⁹¹ showed that 6 of 10 patients with moderate or normal VII function maintained at least moderate function after reoperation, 3 patients had increased ataxia, and 1 patient had a cerebellar hematoma. The use of SRS has been endorsed by some for recurrence of VS following one or more MS procedures²⁷⁸. Using SRS for recurrent VSs resulted in worsening of facial nerve function in 23% of patients with Grade I-III function before SRS (median follow-up = 43 mos), and 14% developed new trigeminal symptoms²⁷⁸. 6% of patients developed tumor progression after SRS.

HYDROCEPHALUS

May occur following treatment (MS or SRS) for VS, and may even occur years later. The increased CSF pressure may also predispose to development of a CSF fistula.

21.2.9. Pituitary tumors

Key concepts:

• most are benign adenomas arising from the anterior pituitary (adenohypophysis)

• presentation (see below): most commonly present due to hormonal effects (includes: hyperprolactinemia, Cushing’s syndrome, acromegaly…), mass effect (most commonly: bitemporal hemianopsia from compression of optic chiasm), as an incidental finding, or infrequently with pituitary apoplexy (see page 635)

• work-up for a newly diagnosed intrasellar lesion: see Table 21-40, page 642

• treatment (see page 649): some prolactinomas may be treated medically (DA agonists). Other options include transsphenoidal or transcranial surgery, or XRT

• post-op concerns include: diabetes insipidus, adrenal insufficiency, CSF leak

For a review of pituitary embryology & neuroendocrinology, see page 109.

Pituitary adenomas

Most primary pituitary tumors are benign adenomas which arise from the anterior pituitary gland (adenohypophysis). Neurohypophyseal tumors are rare (see Pituicytoma, page 641). Adenomas may be classified by a number of schemes, including: by endocrine function (aided by immunostaining), by light microscopy with routine histological staining (see page 641), and by electron microscopic appearance.

Microadenoma: A pituitary tumor < 1 cm diameter. Currently, 50% of pituitary tumors are < 5 mm at time of diagnosis. These may be difficult to find at the time of surgery.

Macroadenomas: Tumors > 1 cm diameter.

Pituitary carcinoma²⁹²

Rare (< 140 reports). Usually invasive and secretory (most common hormones: ACTH, PRL). Can metastasize, at which point prognosis is poor (66% 1-year mortality). Little improvement with further surgery, XRT, or chemotherapy.

EPIDEMIOLOGY

Pituitary tumors represent ≈ 10% of intracranial tumors (incidence is higher in autopsy series). They are most common in the 3rd and 4th decades of life, and affect both sexes equally. The incidence is increased in multiple endocrine adenomatosis or neoplasia (MEA or MEN)

DIFFERENTIAL DIAGNOSIS OF PITUITARY TUMORS

See page 1215 which also includes non-neoplastic etiologies.

CLINICAL PRESENTATION OF PITUITARY TUMORS

Classically, pituitary tumors are divided into functional (or secreting), and nonfunctional (AKA endocrine-inactive, which are either nonsecretory, or else secrete products such as gonadotropin that do not cause endocrinologic symptoms). Nonsecreting tumors usually do not present until of sufficient size to cause neurologic deficits by mass effect, whereas the former frequently present earlier with symptoms caused by physiologic effects of excess hormones that they secrete²⁹³.

Presentation

Main modes: endocrine syndromes, mass effect, incidental finding, pituitary apoplexy.

1. endocrinologic disturbance:

A. hormone oversecretion (secretory tumor): ≈ 65% of adenomas secrete an active hormone (48% prolactin, 10% GH, 6% ACTH, 1% TSH)²⁹⁴:

1. prolactin (PRL): can cause amenorrhea-galactorrhea syndrome in females), impotence in males (see page 637). Etiologies:

a. prolactinoma: neoplasia of pituitary lactotrophs (see page 637)

b. stalk effect: pressure on the pituitary stalk may reduce the inhibitory control over PRL secretion (see page 644)

2. growth hormone (GH): elevated GH is due to a pituitary adenoma > 95% of the time

a. in adults: causes acromegaly (see page 639)

b. in prepubertal children (before epiphyseal plate closure): produces pituitary gigantism (very rare)

3. corticotropin AKA adrenocorticotropic hormone (ACTH):

a. Cushing’s disease (endogenous hypercortisolism): see below

b. Nelson syndrome: can develop only in patients who have had an adrenalectomy (see page 639)

4. thyrotropin (TSH): secondary (central) hyperthyroidism (see page 640)

5. gonadotropins (leuteinizing hormone (LH) and/or follicle stimulating hormone (FSH)): usually does not produce a clinical syndrome

B. underproduction of pituitary hormones

1. from compression of the normal pituitary by large tumors. More common with non-secretory tumors than with secretory tumors. In order of sensitivity to compression: GH, gonadotropins (LH & FSH), TSH, ACTH (mnemonic: Go Look For The Adenoma). Chronic deficiency of all pituitary hormones (panhypopituitarism) may produce pituitary cachexia (AKA Simmonds’ cachexia)

a. growth hormone deficiency^A:

i. in children: produces growth delay

ii. in adults: produces vague symptoms with metabolic syndrome (decreased lean body mass, centripetal obesity, reduced exercise tolerance, impaired sense of well-being)

b. hypogonadism: amenorrhea (women), loss of libido, infertility

c. hypothyroidism: cold intolerance, myxedema, entrapment neuropathies (e.g. carpal tunnel syndrome), weight gain, memory disturbance, integumentary changes (dry skin, coarse hair, brittle nails), constipation, increased sleep demand

d. hypoadrenalism: orthostatic hypotension, easy fatigability

2. ✖ NB: selective reduction of a single pituitary hormone is very atypical of pituitary adenomas. May occur with autoimmune hypophysitis (see page 1217) which most commonly involves ACTH or ADH (causing DI²⁹⁵- see below))

3. diabetes insipidus: almost never seen pre-operatively with pituitary tumors (except possibly with pituitary apoplexy, see below). If DI is present, other etiologies should be sought (e.g. (see page 1217)) including

a. autoimmune hypophysitis: see page 1217

b. hypothalamic glioma

c. suprasellar germ cell tumor

4. gonadotropin deficiency (hypogonadotrophic hypogonadism) with anosmia is part of Kallmann syndrome²⁹⁶

2. mass effect (other than compression of the pituitary). Because they tend to get to a larger size before detection, this is more common with nonfunctioning tumors. Of functional tumors, prolactinoma is the most likely to become large enough to cause mass effect (ACTH tumor is least likely). Nonspecific symptoms include headaches. Structures commonly compressed and manifestations include:

A. optic chiasm: classically produces bitemporal hemianopsia (non-congruous). May also cause decreasing visual acuity

B. involvement of third ventricle may produce obstructive hydrocephalus

C. cavernous sinus

1. pressure on cranial nerves contained within (III, IV, V₁, V₂, VI): ptosis, facial pain, diplopia (see Invasive pituitary adenomas below)

2. occlusion of the cavernous sinus: proptosis, chemosis

3. encasement of the carotid artery by tumor: may cause slight narrowing, but complete occlusion is rare

3. incidental finding on imaging study done for other reasons

4. pituitary apoplexy (see below)

5. invasive adenomas may rarely present with CSF rhinorrhea²⁹⁷ (see page 637)

6. macroadenomas may produce H/A possibly via increased intrasellar pressure

7. seizures are rarely attributable to pituitary adenomas

A. a growth hormone stimulation test (see page 648) is more sensitive and specific for GH deficiency than measuring basal GH levels

PITUITARY APOPLEXY

Key concepts:

• due to expansion of a pituitary adenoma from hemorrhage or necrosis

• typical presentation: paroxysmal H/A with endocrinologic and/or neurologic deficit (usually ophthalmoplegia or visual loss)

• management: immediate administration of glucocorticoids, and transsphenoidal decompression within 7 days in most cases

Definition: Neurologic and/or endocrinologic deterioration due to sudden expansion of a mass within the sella turcica.

Etiology: Sudden intrasellar expansion may occur as a result of hemorrhage, necrosis^{298, 299} and/or infarction within a pituitary tumor and adjacent pituitary gland. Occasionally, hemorrhage occurs into a normal pituitary gland or Rathke’s cleft cyst³⁰⁰.

Clinical features of pituitary apoplexy: Patients often present with abrupt onset of H/A, visual disturbance, and loss of consciousness. Neurologic involvement includes:

1. visual disturbances: one of the most common findings. Includes:

A. ophthalmoplegia (unilateral or bilateral): opposite the situation with a pituitary tumor, ophthalmoplegia occurs more often (78%) than visual pathway deficits (52-64%)³⁰¹

B. one of the typical field cuts seen in pituitary tumors (see page 642)

2. reduced mental status: due to ↑ ICP or hypothalamic involvement

3. cavernous sinus compression can cause venous stasis and/or pressure on any of the structures within the cavernous sinus

A. trigeminal nerve symptoms

B. proptosis

C. ophthalmoplegia (Cr. N. III palsy is more common than VI)

D. ptosis may be an early symptom³⁰²,³⁰³

E. pressure on carotid artery

F. compression of sympathetics within the cavernous sinus may produce a form of Horner’s syndrome with unilateral ptosis, miosis, & anhidrosis limited to the forehead

G. carotid artery compression may cause CVA or vasospasm

4. when hemorrhage breaks through the tumor capsule and the arachnoid membrane into the chiasmatic cistern, signs and symptoms of SAH may be seen

A. N/V

B. meningismus

C. photophobia

5. increased ICP may produce lethargy, stupor or coma

6. hypothalamic involvement may produce

A. hypotension

B. thermal dysautoregulation

C. cardiac dysrhythmias

D. respiratory pattern disturbances

E. diabetes insipidus

F. altered mental status: lethargy, stupor or coma

7. suprasellar expansion can produce acute hydrocephalus

Epidemiology

In Wilson’s series, 3% of his patients with macroadenomas had an episode of pituitary apoplexy. In another series of 560 pituitary tumors, a high incidence of 17% was found (major attack in 7%, minor in 2%, asymptomatic in 8%)³⁰⁴. It is common for apoplexy to be the initial presentation of a pituitary tumor³⁰⁵.

Evaluation

CT or MRI shows hemorrhagic mass in sella turcica and/or suprasellar region, often distorting the anterior third ventricle.

Cerebral angiography should be considered in cases where differentiating pituitary apoplexy from aneurysmal SAH is difficult.

Management of pituitary apoplexy

Pituitary function is consistently compromised, necessitating rapid administration of corticosteroids and endocrine evaluation.

In the absence of visual deficits, prolactinomas may be treated with bromocriptine.

Rapid decompression is required for: sudden constriction of visual fields, severe and/or rapid deterioration of acuity, or neurologic deterioration due to hydrocephalus. Surgery in ≤ 7 days of pituitary apoplexy resulted in better improvement in ophthalmoplegia (100%), visual acuity (88%) and field cuts (95%) than surgery after 7 days³⁰⁶. Decompression is usually via a transsphenoidal route (transcranial approach may be advantageous in some cases). Goals of surgery:

1. to decompress the following structures if under pressure: optic apparatus, pituitary gland, cavernous sinus, third ventricle (relieving hydrocephalus)

2. obtain tissue for pathology

3. complete removal of tumor is usually not necessary

4. for hydrocephalus: ventricular drainage is generally required

INVASIVE PITUITARY ADENOMAS

About 5% of pituitary adenomas become locally invasive. The genetic make-up of these tumors may differ from more benign adenomas³⁰⁸, even though the histology is similar. Numerous classifications systems have been devised, Wilson’s system³⁰⁷ (modified from Hardy^{309, 310}) is shown in Table 21-37.

The clinical course is variable, with some tumors being more aggressive than others. Occasionally, these tumors grow to gigantic sizes (> 4 cm dia), and these are often very aggressive and follow a malignant course³¹¹.

At times, an adenoma may push the medial wall of the cavernous sinus ahead of it without actually perforating this dural structure³¹². This is difficult to reliably identify on MRI, and the most definitive sign of cavernous sinus invasion is carotid artery encasement³¹³.

Table 21-37 Anatomic classification of pituitary adenoma (modified Hardy system)³⁰⁷

Extension

• Suprasellar extension 0: none A: expanding into suprasellar cistern B: anterior recesses of 3rd ventricle obliterated C: floor of 3rd ventricle grossly displaced • Parasellar extension D*: intracranial (intradural) E: into or beneath cavernous sinus (extradural)

Invasion/Spread

• Floor of sella intact I: sella normal or focally expanded; tumor < 10 mm II: sella enlarged; tumor ≥ 10 mm • Sphenoid extension III: localized perforation of sellar floor IV: diffuse destruction of sellar floor • Distant spread V: spread via CSF or blood-borne

* specify: 1) anterior, 2) middle, or 3) posterior fossa

Presentation:

1. visual system

A. most present due to compression of the optic apparatus, usually producing gradual visual deficit (however, sudden blindness is not unheard of)

B. extraocular muscle deficits may occur with cavernous sinus invasion, and usually develop after visual loss

C. exophthalmos may occur with orbital invasion due to compromise of orbital venous drainage

2. hydrocephalus: suprasellar extension may obstruct one or both foramen of Monro

3. invasion of the skull base may lead to nasal obstruction or CSF rhinorrhea, which occasionally may be precipitated by tumor shrinkage in response to bromocriptine³¹⁴

4. tumors that secrete prolactin often present with findings of hyperprolactinemia (see page 634) and with these, the prolactin levels are usually > 1000 ng/ml (caution: giant invasive adenomas with very high PRL production may have a falsely low PRL level due to “hook effect”, see page 644)

HORMONALLY ACTIVE PITUITARY TUMORS

PROLACTINOMAS

The most common secretory adenoma. Arise from neoplastic transformation of anterior pituitary lactotrophs. See Table 21-42, page 644 for DDx of hyperprolactinemia.

Manifestations of prolonged hyperprolactinemia:

1. females: amenorrhea-galactorrhea syndrome (AKA Forbes-Albright syndrome, AKA Ahumada-del Castillo syndrome). Variants: oligomenorrhea, irregular menstrual cycles. 5% of women with primary amenorrhea will be found to have a PRL-secreting pituitary tumor³¹⁵. Remember: pregnancy is the most common cause of secondary amenorrhea in females of reproductive potential. The galactorrhea may be spontaneous or expressive (only on squeezing the nipples)

2. males: impotence, decreased libido. Galactorrhea is rare (estrogen is also usually required). Gynecomastia is rare. Prepubertal prolactinomas may result in small testicles and feminine body habitus

3. either sex:

A. infertility is common

B. bone loss (osteoporosis in women, and both cortical and trabecular osteopenia in men) due to a relative estrogen deficiency, not due to the elevated prolactin itself

At the time of diagnosis, 90% of prolactinomas in women are microadenomas, vs. 60% for males (probably due to gender specific differences in symptoms resulting in earlier presentation in females). Some tumors secrete both PRL and GH.

CUSHING’S DISEASE

Cushing’s syndrome (CS) is a constellation of findings caused by hypercortisolism. Cushing’s disease (endogenous hypercortisolism due to hypersecretion of ACTH by an ACTH secreting pituitary adenoma) is one cause of CS. The most common cause of CS is iatrogenic (administration of exogenous steroids). Possible etiologies of endogenous hypercortisolism are shown in Table 21-38. To determine the etiology of CS, see Dexamethasone suppression test on page 646.

Prevalence of Cushing’s disease: 40 cases/million population. ACTH-producing adenomas comprise 10-12% of pituitary adenomas³¹⁶. Cushing’s disease is 9 times more common in women, whereas ectopicACTH production is 10 times more common in males. Non-iatrogenic CS is 25% as common as acromegaly.

At the time of presentation, over 50% of patients with Cushing’s disease have pituitary tumors < 5 mm in diameter, which are very difficult to image with CT or MRI. Most are basophilic, some (especially the larger ones) may be chromophobic. Only ≈ 10% are large enough to produce some mass effect, which may cause enlargement of the sella turcica, visual field deficit, cranial nerve involvement and/or hypopituitarism.

Conversion factors³¹⁷ for ACTH and cortisol between U.S. units and SI units are shown in Eq 21-1 and Eq 21-2.

Clinical findings in Cushing’s disease (and also Cushing’s syndrome) include:

1. weight gain

A. generalized in 50% of cases

B. centripetal fat deposition in 50%: trunk, upper thoracic spine (“buffalo hump”), supraclavicular fat pad, neck, “dewlap tumor” (episternal fat), with round plethoric face (“moon facies”) and slender extremities

2. hypertension

3. ecchymoses and purple striae, especially on flanks, breasts and lower abdomen

4. amenorrhea in women, impotence in men, reduced libido in both

5. hyperpigmentation of skin and mucous membranes: due to MSH cross-reactivity of ACTH. Occurs only with elevated ACTH, i.e. Cushing’s disease (not Cushing’s syndrome) or ectopic ACTH production (also see Nelson’s syndrome (or Nelson syndrome) (NS) below)

6. atrophic, tissue-paper thin skin with easy bruising and poor wound healing

7. psychiatric: depression, emotional lability, dementia

8. osteoporosis

9. generalized muscle wasting with complaints of easy fatigability

10. elevation of other adrenal hormones: androgens may produce hirsutism and acne

11. sepsis: associated with advanced Cushing’s syndrome

Laboratory findings in Cushing’s disease:

1. hyperglycemia: diabetes or glucose intolerance

2. hypokalemic alkalosis

3. loss of diurnal variation in cortisol levels

4. normal or elevated ACTH levels

5. failure to suppress cortisol with low-dose (1 mg) dexamethasone test: see page 646

6. elevated 24-hour urine free-cortisol

7. CRH levels will be low (not commonly measured)

Ectopic ACTH secretion

Usually secreted by tumors, most commonly small-cell carcinoma of the lung, thymoma, carcinoid tumors, pheochromocytomas, and medullary thyroid carcinoma. In addition to findings of Cushing’s syndrome, patients are typically cachectic due to the malignancy which is usually rapidly fatal.

NELSON’S SYNDROME (OR NELSON SYNDROME) (NS)

Key concepts:

• a rare condition that follows 10-30% of total bilateral adrenalectomies performed for Cushing’s disease

• classic triad: hyperpigmentation (skin & mucus membranes), abnormal ↑ ACTH, and progression of pituitary tumor (the last criteria is now controversial)

• treatment options: surgery (transsphenoidal or transcranial), XRT, medication

A rare condition that occurs in 10-30% of patients following total bilateral adrenalectomy (TBA) for treatment of Cushing’s syndrome. NS is due to continued growth of corticotroph (ACTH-secreting) adenoma cells. Usually occurs 1-4 years after TBA (range: 2 mos-24 years)³¹⁶. Theoretical explanation (unproven)³¹⁸: following TBA, hypercortisolism resolves, and CRH levels increase back to normal from the (reduced) suppressed state; corticotroph adenomas in patients with NS have an increased & prolonged response to CRH resulting in increased growth. Also, corticotrophs in NS and CD show reduced inhibition by glucocorticoids. It is controversial if some cases may be related to insufficient glucocorticoid replacement after TBA³¹⁶.

Manifestations³¹⁸

1. hyperpigmentation (due to melanin stimulating hormone (MSH) cross reactivity of ACTH and actual increased levels of MSH due to increased propiomelanocortin production). Often the earliest sign that Nelson’s syndrome is developing. Look for linea nigra (midline pigmentation from pubis to umbilicus) and hyperpigmentation of scars, gingivae, and areolae. DDx of hyperpigmentation includes: primary adrenal insufficiency (high levels of ACTH), ectopic ACTH secretion, hemochromatosis (more bronze color), jaundice (yellowish)

2. tumor growth → increased mass effect or invasion: the most serious consequence. These corticotroph tumors are among the most aggressive of pituitary tumors^{319 (p 545)}. May produce any of the problems associated with macroadenomas (optic nerve compression, cavernous sinus invasion, pituitary insufficiency, H/A, bony invasion…) as well as necrosis with precipitous intracranial hypertension³²⁰ (pituitary apoplexy, see page 635)

3. malignant transformation of the corticotroph tumor (very rare)

4. hypertrophy of adrenal tissue rests: may be located in the testes → painful testicular enlargement and oligospermia. Rarely the rests can secrete enough cortisol to normalize cortisol levels or even cause a recurrence of Cushing’s disease despite the adrenalectomy

Laboratories & tests

1. ACTH > 200 ng/L (usually thousands of ng/L) (normal: usually < 54 ng/L)

2. exaggerated ACTH response to CRH (not required for diagnosis)

3. other pituitary hormones may be affected as with any macroadenoma causing mass effect (see page 635) and endocrine screening should be done (see page 642)

4. formal visual field testing should be done in patients with suprasellar extension or in those being considered for surgery (as a baseline) (see page 642).

Treatment

For treatment, see page 650.

ACROMEGALY

Key concepts:

• abnormally high levels of growth hormone in an adult. > 95% of cases are due to a benign pituitary somatotroph adenoma, > 75% are > 10 mm at time of diagnosis

• effects include soft tissue and skeletal changes, cardiomyopathy, colon Ca

• work-up (page 647): endocrine tests (page 642), cardiology consult, colonoscopy

• treatment (see page 652): surgery for most, and then if necessary, medical therapy (page 652) and/or XRT (see page 656)

• suggested criteria for biochemical cure (page 662): normal IGF-1, growth hormone level < 5 ng/ml, AND GH nadir of < 1 ng/ml after OGST (see page 662)

Incidence: 3 cases/1-million persons/year. > 95% of cases of excess GH result from a pituitary somatotroph adenoma. Growth hormone carcinoma is extremely rare. Ectopic GH secretion may occur uncommonly with: carcinoid tumor, lymphoma, pancreatic islet-cell tumor. By the time of diagnosis, > 75% of pituitary GH tumors are macroadenomas (> 10 mm dia) with cavernous sinus invasion and/or suprasellar extension.

25% of acromegalics have thyromegaly with normal thyroid studies. 25% of GH adenomas also secrete prolactin. Acromegaly occurs rarely as part of a genetic syndrome, including: multiple endocrine neoplasia type 1 (MEN 1), McCune-Albright syndrome, familial acromegaly, and Carney complex³²¹.

Clinical

Elevated levels of GH in children before closure of the epiphyseal plates in the long bones produces gigantism. Usually presents in the teen years.

In adults, elevated GH levels produces acromegaly (age: usually > 50 yrs) with findings that may include^{322, 323} (also see Table 21-39):

1. skeletal overgrowth deformities

A. increasing hand and foot size

B. thickened heel pad

C. frontal bossing

D. prognathism

2. cardiovascular

A. cardiac findings (structural and functional): arrhythmias, valvular disease, concentric myocardial hypertrophy

B. hypertension (30%)

3. soft tissue swelling (includes macroglossia)

4. glucose intolerance

5. peripheral nerve entrapment syndromes (including carpal tunnel syndrome)

6. debilitating headache

7. excessive perspiration (especially palmar hyperhidrosis)

8. oily skin

9. joint pain

10. sleep apnea

11. fatigue

12. colon cancer: risk is ≈ 2 x risk of general population³²⁴

Patients with elevated levels of GH (including partially treated cases) have 2-3 times the expected mortality rate³²⁵, primarily due to hypertension, diabetes, pulmonary infections, cancer, and cardiovascular disease (see Table 21-39). Soft-tissue swelling and nerve entrapment may be reversible with normalization of GH levels, however many disfiguring changes and health risks are permanent (see Table 21-39).

Table 21-39 Risks of long-term exposure to excess growth hormone (GH)³²⁵

Arthropathy

• unrelated to age of onset or GH levels • usually with longstanding acromegaly • reversibility*: • rapid symptomatic improvement • bone & cartilage lesions irreversible

Peripheral neuropathy

• intermittent anesthesias, paresthesias • sensorimotor polyneuropathy • impaired sensation • reversibility*: • symptoms may improve • onion bulbs (whorls) do not regress

Cardiovascular disease

• cardiomyopathy • reduced LV diastolic function • increased LV mass and arrhythmias • fibrous hyperplasia of connective tissue • HTN: exacerbates cardiomyopathic changes • reversibility*: may progress even with normal GH

Respiratory disease

• upper airway obstruction: caused by soft tissue overgrowth and decreased pharyngeal muscle tone with sleep apnea in ≈ 50% • reversibility*: generally improves

Neoplasia

• increased risk of malignancies (especially colonCa) & soft-tissue polyps • reversibility*: unknown

Glucose intolerance

• occurs in 25% of acromegalics (more common with family history of DM) • reversibility*: improves

* reversibility with normalization of GH levels

THYROTROPIN (TSH)-SECRETING ADENOMAS

Rare: comprise ≈ 0.5-1% of pituitary tumors^{294, 326}. Produces central (secondary) hyperthyroidism^A: elevated circulating T₃ and T₄ levels, with elevated or inappropriately normal TSH³²⁷ (TSH should be undetectable in primary hyperthyroidism). Up to 33% of tumors positive for TSH immunostaining are nonsecretory³²⁷. Many of these tumors are plurihormonal, but the secondary hormone is usually clinically silent. Most of these tumors are aggressive and invasive and are large enough at presentation to also produce mass effect (especially if prior thyroid ablative procedures have been done, which occurs in up to 60% of cases due to lack of recognition of pituitary abnormality^{327, 328}).

A. NB: central hyperthyroidism may also occur with pituitary resistance to thyroid hormones³²⁷

Symptoms of hyperthyroidism: anxiety, palpitations (due to a-fib), heat intolerance, hyperhidrosis, and weight gain despite normal or increased intake. Signs: hyperactivity, lid lag, tachycardia, irregular rhythm when a-fib is present, hyperreflexia, tremor. Exophthalmous and infiltrative dermopathy (e.g. pretibial myxedema) are present only in Grave’s disease.

PATHOLOGICAL CLASSIFICATION OF PITUITARY TUMORS

LIGHT MICROSCOPIC APPEARANCE OF ADENOMAS

In order of decreasing frequency:

• chromophobe: most common (ratio of chromophobe to acidophil is 4-20:1). Originally considered “non-secretory”, in actuality may produce prolactin, GH, or TSH

• acidophil (eosinophilic): produce prolactin, TSH, or usually GH

• basophil → gonadotropins, ß-lipotropin, or usually ACTH → Cushing’s disease

CLASSIFICATION OF ADENOMAS BASED ON SECRETORY PRODUCTS

1. endocrine-active tumors: ≈ 70% of pituitary tumors produce 1 or 2 hormones that are measurable in the serum and cause defined clinical syndromes, these are classified based on their secretory product(s)

2. endocrine-inactive (nonfunctional) tumors³²⁹

A. null-cell adenoma } constitute the bulk of endocrine-inactive adenomas

B. oncocytoma } constitute the bulk of endocrine-inactive adenomas

C. gonadotropin-secreting adenoma

D. silent corticotropin-secreting adenoma

E. glycoprotein-secreting adenoma

TUMORS OF THE NEUROHYPOPHYSIS AND INFUNDIBULUM

The most common tumors encountered in the posterior pituitary are metastases (owing to the rich blood supply).

Granular cell tumors

AKA (infundibular) granular cell tumor (GCT). WHO grade I. Obsolete terms: choristoma³³⁰, granular cell myoblastoma, pituicytoma (this term is now reserved for a circumscribed glial neoplasm - see below). Tumors with nests of large cells having granular, eosinophilic cytoplasm.

While rare, GCTs are the most common primary tumor of the neurohypophysis and pituitary stalk/infundibulum³³¹ with a predilection for the stalk (these result in suprasellar extension). GCTs have been identified in the gastrointestinal tract, genitourinary tract, orbital region as well as in other locations of the central nervous system with no connection to the pituitary gland or hypothalamus (e.g. spinal meninges³³²). Female:male ratio ≥ 2:1. Asymptomatic microscopic clusters of granular cells (tumorettes) are more common, with an incidence up to 17%³³³.

The most common presentation is with visual field deficits due to optic chiasm compression³³⁰. However, any symptom typical of a hormonally inactive sellar mass may occur.

Imaging: may appear radiographically identical to adenomas. Rarely considered in the differential diagnosis pre-op. Isodense on CT and isointense on T1WI MRI, dense homogeneous enhancement on CT & MRI.

Treatment: if GCT is suspected pre-op, a transcranial approach is preferred over transsphenoidal because of the vascularity which has prevented total resection in 60-70% of reported cases³³⁴. XRT may be considered for subtotal resection³³¹.

Pituicytoma

Less favored alternate terms include posterior pituitary astrocytoma. Rare (mostly case reports). Circumscribed tumor with spindle cells, arising from the neurohypophysis or infundibulum³³⁵. WHO grade I. Reported only in adults.

Treatment: surgical excision. Subtotal removal may be followed by recurrence over several years.

EVALUATION

HISTORY AND PHYSICAL

Directed to look for signs and symptoms of:

1. endocrine hyperfunction (see Functional pituitary tumors above), including:

A. prolactin: amenorrhea (women), nipple discharge (primarily in women since estrogen is also required), impotence (males),

B. thyroid: heat intolerance

C. growth hormone: change in ring size or shoe size or coarsening of facial features, gigantism (children)

D. cortisol: hyperpigmentation, Cushingoid features

2. endocrine deficits (due to mass effect on pituitary) (see page 635)

3. visual field deficit: bedside confrontational testing to rule-out visual field deficit (classically bitemporal hemianopsia, see below)

4. deficits of cranial nerves within cavernous sinus (III, IV, V₁, V₂, VI)

DIAGNOSTIC TESTS

Initial tests to work-up a patient presenting with a known or suspected pituitary mass are shown in Table 21-40. Further testing is indicated for abnormal results or for strong suspicion of specific syndromes (see indicated page for details).

VISUAL FIELDS

Formal visual field testing: by perimetry with a tangent screen (using the small red stimulus since desaturation of color is an early sign of chiasmal compression) or by Goldman or automated Humphrey perimeter (the latter requires good cooperation from the patient to be valid).

Visual field deficit patterns

Depends in part on location of chiasm with respect to sella turcica: the chiasm is located above the sella in 79%, posterior to the sella turcica (postfixed chiasm) in 4%; in front of the sella (pre-fixed) in 5%³³⁷(p 2135)

1. compression of the optic chiasm:

A. bitemporal hemianopsia that obeys the vertical meridian: classic visual field deficit associated with a pituitary tumor. Due to impingement on crossing nasal fibers in the chiasm (see page 829)

B. other reported patterns that occur rarely: monocular temporal hemianopsia

2. optic nerve compression: more likely in patients with a postfixed chiasm

A. loss of vision in the ipsilateral eye. If carefully sought, there is usually a superior outer (temporal) quadrantanopsia in the contralateral eye^{337 (p 2135)} (so-called junctional scotoma AKA “pie in the sky” defect) from compression of the anterior knee of Wilbrand (see page 1071) (may also be an early finding even without a post-fixed chiasm)

B. may produce central scotoma or monocular reduction in visual acuity

3. compression of the optic tract: may occur with a pre-fixed chiasm. Produces homonymous hemianopsia

ENDOCRINOLOGIC EVALUATION

BASELINE ENDOCRINE EVALUATION (modified³³⁸)

Also, see Table 21-40. May give indication of tumor type, determines whether any hormones need to be replaced, and serves as a baseline for comparison following treatment. Includes clinical assessment for signs and symptoms, as well as laboratory tests. Screening tests should be checked in all patients with pituitary tumors. Note: selective loss of a single pituitary hormone together with thickening of the pituitary stalk is strongly suggestive of autoimmune hypophysitis (see page 1217).

1. adrenal axis screening (for tests to assess cortisol reserve, see page 647):

A. 8 AM cortisol level: better for hypocortisolism³³⁶. Normal: 6-18 μg/100 ml. Note: AM cortisol may normally be slightly elevated

B. in questionable cases, including to distinguish pseudo-Cushing states from Cushing’s syndrome, see page 646

C. 24-hour urine free cortisol: more accurate for hypercortisolism³³⁶ (almost 100% sensitive and specific, false negative rare except in stress or chronic alcoholism). If not elevated several times above normal, at least 2 additional determinations should be made³³⁹

2. prolactin levels (PRL): see page 111 for prolactin neurophysiology

A. interpretation is shown in Table 21-41. See Table 21-42 for differential diagnosis of hyperprolactinemia. Prolactin level correlates with size of prolactinomas³⁴³: if PRL is < 200 ng/ml, ≈ 80% of tumors are microadenomas, and 76% of these will have normal PRL after surgery; if PRL > 200, only ≈ 20% are microadenomas

B. blood samples should be obtained midmorning (i.e. not soon after awakening) and not after stress, breast stimulation, or physical examination, which may increase PRL levels

C. be aware of the following when interpreting PRL levels:

• because of variations in secretion (daily fluctuations can be as high as 30%) and intrinsic inaccuracies of radioimmunoassay, PRL levels should be rechecked if there is a reason to question a specific result

• heterophilic antibodies (seen in individuals routinely exposed to animal serum products) can cause anomalous results

• stalk effect: PRL is the only pituitary hormone primarily under inhibitory regulation (see page 111). Injury to the hypothalamus or pituitary stalk can cause modest elevation of PRL due to decrease in prolactin inhibitory factor (PRIF). Rule of thumb: the percent chance of an elevated PRL being due to a prolactinoma is equal to one half the PRL level. Persistent post-op PRL elevation may occur even with total tumor removal as a result of injury to stalk (usually ≤ 90 ng/ml; stalk effect doubtful if PRL > 150). For stalk effect, follow these patients, do not use bromocriptine

• “prolactin level > 200 ng/ml”: if the lab re-ports the prolactin level as “> 200” (or some other high value) instead of an actual number, it usually indicates a very high prolactin level that exceeds the upper limits of the assay. Call the lab and ask them to determine the actual value. This usually requires serial dilutions until the PRL is in a range that their assay can quantify. The reasons this is important:

1. treatment decisions: PRL > 500 usually indicates that surgery alone will not be able to normalize the PRL (see page 651)

2. to assess response: it is essential to know what value you are starting with to determine response to therapy (medication, surgery, XRT…)

• hook effect: extremely high PRL levels may overwhelm the assay and produce falsely low results. Therefore, for large adenomas with a normal PRL level, have the lab perform several dilutions of the serum sample and re-run the PRL, especially in patients with clinical hyperprolactinemia

• macroprolactinemia: a situation where prolactin molecules polymerize and bind to immunoglobulins. Prolactin in this form has reduced biologically activity but produces a laboratory finding of hyperprolactinemia. Clinical significance is controversial³⁴⁷, asymptomatic patients usually do not require treatment

3. thyroid axis: the basis for thyroid screening is shown in Table 21-43

A. screening: T₄ level (total or free), thyroid-stimulating hormone (TSH) (AKA thyrotropin). Normal values: free T₄ index is 0.8-1.5, TSH 0.4-5.5 μU/ml, total T₄ 4-12 μg/100ml (NB: be sure to check both T₄ AND TSH)

B. further testing: thyrotropin-releasing hormone (TRH) stimulation test (indicated if T₄ is low or borderline): check baseline TSH, give 500 μg TRH IV, check TSH at 30 & 60 mins. Normal response: peak TSH twice baseline value at 30 mins. Impaired response with a low T₄ indicates pituitary deficiency. Exaggerated response suggests primary hypothyroidism

4. growth hormone:

A. IGF-1 (somatomedin-C) level (see page 647) is the recommended initial test (testing for elevated IGF-1 is extremely sensitive for acromegaly)

B. checking a single random GH level may not be a reliable indicator and is therefore not recommended (see page 648)

5. gonadal axis

A. screening:

1. serum gonadotropins: FSH & LH

2. sex steroids

a. estradiol in women

b. testosterone in men: measure total testosterone

B. further testing: none dependable in differentiating pituitary from hypothalamic disorders

6. neurohypophysis (posterior pituitary): deficits are rare with pituitary tumors

A. screening: check adequacy of ADH by demonstrating concentration of urine with water deprivation (see page 17)

B. further testing: measurement of serum ADH in response to infusion of hypertonic saline

Table 21-41 Significance of prolactin levels*

PRL (ng/ml)	Interpretation	Situations observed in
3-30†	normal	non-pregnant female
10-400		pregnancy (see Table 21-42)
2-20		postmenopausal female
25†-150	moderate elevation	• prolactinoma • “stalk effect” (see text) • other causes§
> 150‡	significant elevation	prolactinoma§

* Note: ectopic sites of prolactin secretion have rarely been reported (e.g. in a teratoma³⁴⁴)

^† normal values vary, use your lab’s reference range

^‡ some authors recommend 200 ng/ml as the cutoff for probable prolactinomas³⁴⁵

^§ for DDx of hyperprolactinemia see Table 21-42

Table 21-42 Differential diagnosis of elevated prolactin (PRL) level (hyperprolactinemia)*

1. pregnancy-related A. during pregnancy†: 10-400 ng/ml B. postpartum: PRL decreases ≈ 50% (to ≈ 100 ng/ml) in the first week postpartum, and is usually back to normal in 3 weeks C. in the lactating female: suckling increases PRL, which is critical for lactogenesis (once initiated, nonpregnant PRL levels can maintain lactation). First 2-3 months postpartum: basal PRL = 40-50 ng/ml, suckling → increases x 10-20. 3-6 months postpartum: basal PRL levels become normal or slightly elevated, and double with suckling. PRL should normalize by 6 months after weaning 2. pituitary adenoma A. prolactinoma: larger prolactin microadenomas and macroadenomas usually produce PRL > 100 ng/ml B. stalk effect: rule of thumb, the percent chance of an elevated PRL being due to a prolactinoma is equal to one half the PRL level (see page 644) C. some tumors secrete both PRL and GH 3. drugs: dopamine receptor antagonists (e.g. phenothiazines, metoclopramide), oral contraceptives (estrogens), tricyclic antidepressants, verapamil, H2 antagonists (e.g. ranitidine), some SSRIs in particular paroxetine (Paxil®)346… 4. primary hypothyroidism: TRH (a prolactin releasing factor (PRF)) will be elevated (see page 111) 5. empty sella syndrome: see page 719 6. post-ictal: PRL usually normalizes within 1-2 hrs after a seizure (see page 401) 7. breast or chest-wall trauma/surgery: usually ≤ 50 ng/ml 8. excessive exercise: usually ≤ 50 ng/ml 9. stress: in some cases the stress of having the blood test is enough to elevate PRL, anorexia nervosa 10. ectopic secretion: reported in renal cell or hepatocellular tumors, uterine fibroids, lymphomas 11. infiltrating hypothalamic tumors 12. renal failure 13. cirrhosis 14. macroprolactinemia: see text

* hyperprolactinemia from causes other than prolactinomas rarely exceeds 200 ng/ml

^† always R/O pregnancy as a cause of amenorrhea & hyerprolactinemia in a female with reproductive potential

Table 21-43 Basis for thyroid screening

Primary hypothyroidism* (problem with thyroid gland itself)	T4	TSH
• chronic primary hypothyroidism may produce secondary pituitary hyperplasia (pituitary pseudotumor) indistinguishable from adenoma on CT or MRI. Must be considered in any patient with a pituitary mass340, 341 • pathophysiology: loss of negative feedback from thyroid hormones causes increased TRH release from the hypothalamus producing secondary hyperplasia of thyrotrophic cells in the adenohypophysis (thyrotroph hyperplasia). The patient may present due to pituitary enlargement (visual symptoms, elevated PRL from stalk effect, enlarged sella turcica on x-rays…) • chronic stimulation from elevated TRH may rarely produce thyrotroph adenomas • labs: T4 low or normal, TSH elevated (> 90-100 in patients presenting with thyrotroph hyperplasia), prolonged and elevated TSH response to TRH stimulation test (see text)	↓	↑
Secondary hypothyroidism* (insufficient TSH stimulation of thyroid)	↓	↓ or nl
• pituitary hypothyroidism accounts for only ≈ 2-4% of all hypothyroid cases342 • ≈ 23% of patients with chromophobe adenomas develop secondary hypothyroidism if untreated (pituitary compression causes reduced TSH) • labs: T4 low, TSH low or normal, reduced response to TRH stimulation test (see text)
Primary hyperthyroidism (problem with thyroid gland itself)	↑	↓
• etiologies: localized hyperactive thyroid nodule, circulating antibody that stimulates the thyroid, or diffuse thyroid hyperplasia (Graves’ disease, AKA ophthalmic hyperthyroidism) • labs: T4 elevated, TSH subnormal (usually undetectable)
Secondary hyperthyroidism (central hyperthyroidism)	↑	↑ or nl
• etiologies • TSH-secreting pituitary adenoma (rare) • pituitary resistance to thyroid hormones (disrupts negative feedback loop) • labs: T4 elevated, TSH elevated or inappropriately normal

* ✖ Caution: replacing thyroid hormone with inadequate cortisol reserves (as may occur in panhypopituitarism) can precipitate adrenal crisis) (see page 650 for management)

SPECIALIZED ENDOCRINOLOGIC TESTS

Cushing’s syndrome

A. tests to determine if hypercortisolism (Cushing’s syndrome, (CS)) is present or not, regardless of etiology if the screening 24-hr urine free cortisol (see page 643) is equivocal (the basis of these tests is shown in Table 21-44)

1. overnight low-dose dexamethasone (DMZ) suppression tests³⁴⁸:

A. overnight low dose test: give DMZ 1 mg PO @ 11 P.M. and draw serum cortisol the next day at 8 A.M. Results:

1. cortisol < 1.8 μg/dl^A: Cushing’s syndrome is ruled out (except for a few patients with CS who suppress at low DMZ doses, possibly due to low DMZ clearance³⁴⁹)

2. cortisol 1.8 -10 μg/dl: indeterminate, retesting is necessary

3. cortisol > 10 μg/dl: CS is probably present. False positives can occur in the so-called pseudo-Cushing’s state where ectopic CRH secretion produces hyperplasia of pituitary corticotrophs that is clinically indistinguishable from pituitary ACTH producing tumors (requires further testing³⁴⁹). Seen in: 15% of obese patients, in 25% of hospitalized and chronically ill patients, in high estrogen states, in uremia, and in depression. The combined DMZ-CRH test can be used to identify this (see reference³⁴⁹). False positives also may occur in alcoholics or patients on phenobarbital or phenytoin due increased metabolism of DMZ caused by induced hepatic microsomal degradation

B. 2 day low dose test (used when overnight test is equivocal): give DMZ 0.5 mg PO q 6 hrs for 2 days starting at 6 A.M.; 24 hr urine collections are obtained prior to test and on the 2nd day of DMZ administration. Normal patients suppress urinary 17-hydroxycorticosteroids (OHCS) to less than 4 mg/24 hrs, whereas ≈ 95% of patients with CS have abnormal response (higher amounts in urine)³⁴⁹

2. 11 PM salivary cortisol: this is the time of the usual cortisol nadir. Test must be run at NIH approved lab. Accuracy is as good as low-dose DMZ suppression test

B. tests to distinguish primary Cushing’s disease (CD) (pituitary ACTH hypersecretion) from ectopic ACTH production and adrenal tumors (40% of CD patients have normal MRI³³⁶)

1. random serum ACTH: if < 5 ng/L indicates ACTH independent CS (e.g. adrenal tumor). Not sensitive or specific due to variability of ACTH levels

2. abdominal CT: usually shows unilateral adrenal mass with adrenal tumors, or normal or bilateral adrenal enlargement in ACTH-dependent cases

3. high-dose dexamethasone (DMZ) suppression test: (NB: up to 20% of patient’s with CD do not suppress with high-dose DMZ. Phenytoin may also interfere with high-dose DMZ suppression³⁵⁰)

A. overnight high-dose test: obtain a baseline 8 A.M. plasma cortisol level

B. then give DMZ 8 mg PO @ 11 P.M. and measure plasma cortisol level the next morning at 8 A.M.

C. in 95% of CD cases plasma cortisol levels are reduced to < 50% of baseline, whereas in ectopic ACTH or adrenal tumors it will usually be unchanged

4. metyrapone (Metopirone®) test: performed on an inpatient basis. Give 750 mg metyrapone (suppresses cortisol synthesis) PO q 4 hrs for 6 doses. Most patients with CD will have a rise in 17-OHCS in urine of 70% above baseline, or an increase in serum 11-deoxycortisol 400-fold above baseline

5. corticotropin-releasing hormone (CRH) stimulation test: CD responds to exogenous CRH 0.1 μg/kg IV bolus with even further increased plasma ACTH and cortisol levels; ectopic ACTH and adrenal tumors do not³⁵¹

6. inferior petrosal sinus (IPS) sampling (or, cavernous sinus sampling is preferred by some): done by interventional neuroradiologist. Uses a microcatheter to measure ACTH levels on each side at baseline, and then at 2, 5 & 10 minutes after stimulation with IV CRH (with simultaneous peripheral ACTH levels at each interval). General information:

A. IPS sampling is not needed when the following criteria of CD are met³¹⁷:

1. ACTH-dependent Cushing’s disease

2. suppression with high-dose dexamethasone test (see above)

3. visible pituitary adenoma on MRI

B. may also determine likely side of a microadenoma within the pituitary (thus may be able to avoid bilateral adrenalectomy which requires lifelong gluco- and mineralo-corticoid replacement and risks Nelson’s syndrome in 10-30% - see page 639). 15-30%³³⁶ of the time this test falsely lateralizes the tumor due to the communication through the circular sinus

C. a baseline IPS ACTH to peripheral ACTH ratio > 1.4:1 is consistent with primary Cushing’s disease

D. a post CRH ratio > 3 is also consistent with primary Cushing’s disease

E. complication rate: 1-2%, includes puncture of the sinus wall

A. this is the currently accepted normal value; previously it was 5 μg/dl

Table 21-44 Basis for biochemical tests in Cushing’s syndrome (CS)

• normally, low DMZ doses suppress ACTH release through negative feedback on hypothalamic-pituitary axis, reducing urine and serum corticosteroids • in ≥ 98% of cases of Cushing’s syndrome, suppression occurs, but at a much higher threshold • adrenal tumors and most (85-90%) cases of ectopic ACTH production (especially bronchial Ca) will not suppress even with high dose DMZ • ACTH response to CRH is exaggerated in CS • DMZ does not interfere with measurement of urinary and plasma cortisol and 17-hydroxycorticosteroids

To check cortisol reserve:

1. cosyntropin stimulation test 352:

A. draw a baseline cortisol level (fasting is not required; test can be performed at any time of day)

B. give cosyntropin (Cortrosyn®) (a potent ACTH analogue) 1 ampoule (250 mcg) IM or IV

C. then check cortisol levels at 30 mins (optional) and at 60 mins

D. normal response: peak cortisol level > 18 μg/dl AND an increment > 7 μg/dl, or a peak > 20 μg/dl regardless of the increment

E. subnormal response: indicates adrenal insufficiency. In primary adrenal insufficiency, pituitary ACTH secretion will be elevated. In secondary adrenal insufficiency, chronically reduced ACTH causes adrenal atrophy and unresponsiveness to acute stimulation with this exogenous ACTH analogue

F. normal response: rules out primary and overt secondary adrenal insufficiency, but may be normal in mild cases of reduced pituitary ACTH or early after pituitary surgery where adrenal atrophy has not occurred. In these cases further testing may be positive: see metyrapone test (page 661) or ITT (see below)

2. insulin tolerance test (ITT): “gold standard” for assessing integrity of the hypothalamic-pituitary-adrenal axis. Cumbersome to do. Abnormal in 80% of CS. Assesses ACTH, cortisol & GH reserve

A. rationale: an appropriate cortisol increment in response to insulin-induced hypoglycemia suggests patient will also be able to respond to other stresses (acute illness, surgery…)

B. contraindications: seizure disorder, ischemic cardiac disease, untreated hypothyroidism

C. pre-test preparation: D/C estrogen replacement for 6 weeks prior to test. Have 50 ml of D50 and 100 mg IV hydrocortisone available during test

D. protocol: give regular insulin 0.1 U/kg IV push, and draw blood for glucose, cortisol and GH at 0, 10, 20, 30, 45, 60, 90 and 120 mins (monitor blood sugar by fingerstick during test, and give IV glucose if patient becomes symptomatic). If fingerstick blood sugar is not < 50 mg/dl by 30 minutes and patient is asymptomatic, give additional regular insulin 5 U IVP. There must be 2 specimens after adequate hypoglycemia

E. results:

1. if adequate hypoglycemia (< 40 mg/dl) was not accomplished: cortisol or GH deficiency cannot be diagnosed

2. normal: cortisol increment > 6 μg/dl to a peak > 20

3. peak cortisol = 16-20: steroids needed only for stress

4. peak cortisol < 16: glucocorticoid replacement needed

5. Cushing syndrome: increment < 6

Acromegaly

For suspected acromegaly, the most useful test is an IGF-1 level.

1. insulin-like growth factor-1 (IGF-1) (formerly somatomedin-C) level: an excellent integrative marker of average GH secretion. Normal levels depend on age (peaking during puberty), gender, pubertal stage and lab. Typical fasting levels by age are shown in Table 21-45. Estrogen may suppress IGF-1 levels

2. growth hormone (GH): normal basal fasting level is < 5 ng/ml. In patients with acromegaly, GH is usually > 10 ng/ml but can be normal. Normal basal levels do not reliably distinguish normal patient from GH deficiency³⁵³. Furthermore, due to pulsatile secretion of GH, normal patients may have sporadic peaks up to 50 ng/ml³²². Occasionally acromegaly may be present even with GH levels as low as 37 pg/ml³⁵⁴. random GH levels are not generally useful for diagnosing acromegaly (see above for IGF-1)

3. other tests used uncommonly

A. oral glucose suppression test (OGST): less precise and more expensive than measuring IGF-1, however may be more useful than IGF-1 for monitoring initial response to therapy. GH levels are measured at 0, 30, 60, 90 & 120 minutes after a 75 gm oral glucose load. If the GH nadir is not < 1 ng/ml, the patient is acromegalic^{321, 323}. GH suppression may also be absent with liver disease, uncontrolled DM & renal failure. ✖ Relatively contraindicated in patients with DM and high glucose levels

B. growth-hormone releasing hormone (GHRH) levels: may help diagnose ectopic GH secretion in a patient with proven acromegaly with no evidence of pituitary tumor on imaging. If an extrapituitary source is suspected, chest and abdominal CT and/or MRI should also be obtained³⁵⁵

C. GHRH stimulation test: results may be discordant in up to 50% of patients with acromegaly³²¹ and is thus rarely used (as of this writing, pharmaceutical production of GHRF has been discontinued)

4. octreotide scan: SPECT imaging 4 and 24 hours after injection with 6.5 mCi of indium-111 Octreo-Scan, a somatostatin receptor imaging agent

Table 21-45 Normal IGF-1 by age

Age (yrs)	Level (ng/ml)
1-5	49-327
6-8	52-345
9-11	74-551
12-15	143-996
16-20	141-903
21-39	109-358
40-54	87-267
>54	55-225

RADIOGRAPHIC EVALUATION

Requires either CT or MRI. MRI has an advantage in large tumors and when evaluating for recurrence. A lateral skull x-ray may help define anatomy of sphenoid sinus in cases where transsphenoidal surgery is contemplated. ≈ 50% of pituitary tumors causing Cushing’s syndrome are too small to be imaged on CT or MRI (therefore endocrinologic testing is required to prove the pituitary origin). See page 1215 for differential diagnosis of intrasellar lesions (some are indistinguishable radiographically).

Normal AP diameter of pituitary gland: female of childbearing age (≈ 13-35 yrs)^A: ≤ 11 mm, for all others normal is ≤ 9 mm.

A. pituitary glands in adolescent girls may be physiologically enlarged (mean height: 8.2 ± 1.4 mm) as a result of hormonal stimulation of puberty³⁵⁶

MRI

Imaging test of choice for pituitary tumors. Gives information about invasion of cavernous sinus, and about location and/or involvement of para-sellar carotids. MRI may fail to demonstrate tumor in 25-45% of cases of Cushing’s disease³⁵⁷. 3T vs. 1.5T MRI: based on 5 cases of Cushing’s disease, a 3T MRI showed the adenoma more clearly in 2 cases, in 1 case it showed the tumor on the correct side opposite to where the 1.5T MRI showed it, and in 2 cases neither 1.5T nor 3T MRI could show the microadenoma)³⁵⁸.

Microadenoma: 75% are low signal on T1WI, and high signal on T2WI (but 25% can behave in any way, including completely opposite to above). Enhancement is time-dependent. Imaging must be done with 5 minutes of contrast administration to see a discrete microadenoma. Initially, gadolinium enhances the normal pituitary (no blood brain barrier) but not the pituitary tumor. After ≈ 30 minutes, the tumor enhances about the same. Dynamic MRI scans have been used to increase the sensitivity (contrast is injected while the MRI scanner is running).

Neurohypophysis: normally is high signal on T1WI³⁵⁹ (possibly due to phospholipids). Absence of this “bright spot” often correlates with diabetes insipidus as may occur with autoimmune hypophysitis (see page 1217).

Deviation of the pituitary stalk may also indicate the presence of a microadenoma. Normal thickness of the pituitary stalk is approximately equal to basilar artery diameter. Thickening of stalk is usually NOT adenoma, differential diagnosis here: lymphoma, autoimmune hypophysitis (see page 1217), granulomatous disease, hypothalamic glioma.

CT

Generally superseded by MRI. May be appropriate when MRI is contraindicated (e.g. pacemaker). When done, should include direct coronal imaging, or coronal reconstructions from thin-cut axial CT. If MRI cannot be done, consider also cerebral angiography to demonstrate parasellar carotid arteries and to R/O aneurysm as a possibility.

Calcium in pituitary usually signifies hemorrhage or infarction within tumor.

Enhancement (with IV contrast):

1. normal pituitary enhances densely (no BBB)

2. macroadenomas enhance more than normal pituitary

3. microadenomas enhance less (may just be slower). Diagnostic criteria:

A. must have attenuation change on CT PLUS

B. 2 or more of the following:

1. focal bone erosion of sella turcica

2. focal superior bulge of gland

3. displacement of stalk (this is unreliable, and may actually deviate to opposite side)

ANGIOGRAPHY

Sometimes used in cases considered for transsphenoidal surgery (e.g. as a complement to CT) to localize the parasellar carotids (note: MRI provides this information, and evaluates involvement of cavernous sinuses, usually obviating the need for angiography).

MANAGEMENT/TREATMENT

See page 635 for treatment of pituitary apoplexy. For large invasive adenomas, see below. Prolactinoma is the only pituitary tumor for which medical therapy (dopamine agonist) is the primary treatment modality (in certain cases).

HORMONALLY INACTIVE MACROADENOMAS - MANAGEMENT

Due to poor response rates to medication, when treatment is indicated, surgery and/or XRT are usually the initial treatment of choice (see below for XRT).

MEDICAL MANAGEMENT

Non-secreting adenomas

Bromocriptine has been tried with mild reductions in tumor size in only ≈ 20% of patients. The poor results are probably due to the paucity of dopaminergic receptors on cell membranes in these tumors. Octreotide reduces tumor volume in ≈ 10% of cases. These agents have been used pre-op in some cases to decrease tumor size for surgery.

Follow-up: For asymptomatic microadenomas (< 1 cm dia), recommend: F/U pituitary MRI at years 1, 2, 5 and ± 10 (can stop F/U after 10 and possibly 5 years if no growth).

For tumors > 1 cm, recommend: check visual fields, pituitary bloodwork (to R/O pituitary insufficiency) and pituitary MRI at years 0.5, 1, 2 & 5, and any time symptoms develop.

Gonadotropin-secreting tumors

Rarely, a non-functional tumor may secrete gonadotropins (FSH, LH). This does not produce a clinical syndrome. Normal and neoplastic pituitary gonadotrophs have gonadotropin-releasing hormone (GnRH)receptors, and may respond to long-acting GnRH agonists (by down-regulating receptors) or GnRH antagonists, but significant reductions in tumor size does not occur.

SURGICAL MANAGEMENT

Surgical indications for hormonally inactive pituitary macroadenomas:

1. tumors causing symptoms by mass effect: visual field deficit (classically: bitemporal hemianopsia, panhypopituitarism

2. some surgeons recommend surgery for macroadenomas that elevate the chiasm even in the absence of endocrine abnormalities or visual field deficit because of the possibility of injury to the optic apparatus (see below for invasive pituitary macroadenomas)

3. acute and rapid visual or other neurologic deterioration. May represent ischemia of the chiasm, or tumor hemorrhage/infarction causing expansion (pituitary apoplexy). The major danger is blindness (hypopituitarism can be treated with replacement therapy). Visual loss usually requires emergent decompression. Some surgeons feel that a transcranial approach is necessary, but transsphenoidal decompression is usually satisfactory^{311, 329}

4. to obtain tissue for pathological diagnosis in questionable cases

5. Nelson’s syndrome (see page 639):

A. surgery (transsphenoidal or transcranial): the primary treatment. The aggressiveness of the tumor sometimes requires total hypophysectomy

B. XRT (possibly SRS) is used following subtotal excision

C. medical therapy is usually ineffective. Agents that could be considered include³¹⁶: dopamine agonists, valproic acid, somatostatin analogues, rosiglitazone, and serotonin agonists

MANAGEMENT OF LARGE, INVASIVE ADENOMAS³¹¹

1. prolactinomas

A. dopamine agonists (DA) (see page 651) unless there is unstable deficit

B. for unstable deficit, or if the tumor does not respond to DAs: debulk the tumor transsphenoidally and then rechallenge with DA therapy

2. tumors secreting growth hormone or ACTH: an aggressive surgical approach is indicated with these tumors since the secretion product is harmful and effective medical adjuvants are lacking

A. pre-treat invasive GH-secreting tumors with somatostatin analogue therapy before surgery to reduce surgical risks (general and cardiac)

B. elderly patients or tumors > 4 cm diameter: debulk tumor transsphenoidally and/or adjuvant therapy (XRT and/or medications)

C. young age and size < 4 cm: radical surgery (may be curative)

3. nonfunctional adenomas:

A. elderly patient: expectant management is an option, with intervention for signs of progression (radiographic or neurologic)

B. central tumor or elderly patient with progression: transsphenoidal debulking and/or XRT (residual tumor in the region of the cavernous sinus may show little or no change over several years, and with these nonfunctional tumors, there is less harm in following them than if there is a harmful secretion product)

C. parasellar tumor and/or young age: radical surgery (often not curative)

HORMONE REPLACEMENT THERAPY (HRT)

Critical issues:

1. corticosteroids

A. indications: inadequate cortisol reserve as demonstrate by failing a cosyntropin stimulation test (failure to achieve a peak cortisol level > 18 μg/dl in response to cosyntropin (see page 647))

B. may start cortisol immediately after bloodwork for cosyntropin test is drawn (do not need to wait for test results) - then, when test results available, continue therapy based on test results

C. Rx physiologic replacement dose: cortisol 20 mg po q AM and 10 mg po q 4 PM. Stress doses may be needed in some situations (see page 32)

2. thyroid hormone replacement

A. ✖ can precipitate adrenal crisis if started before cortisol in a patient with adrenal insufficiency (as may occur in panhypopituitarism) (see page 34)

1. do a cosyntropin stimulation test (see page 647) and start cortisol

2. thyroid replacement may be initiated after 1 full day of cortisol. Rx: start with synthroid 125 μg/d

B. although there are warnings not to do surgery on a hypothyroid patient, the reality is that it takes 3-4 weeks for adequate replacement and hypothyroid patients frequently undergo surgery before then with no untoward effect

3. testosterone replacement: may increase intratumoral levels of estradiol which may promote tumor growth. wait for stabilization of tumor before starting

PROLACTINOMAS - MANAGEMENT

1. prolactin level (PRL) < 500 ng/ml in tumors that are not extensively invasive (see below for invasive tumors): PRL may be normalized with surgery

2. PRL > 500 ng/ml: the chances of normalizing PRL surgically are very low[Barrow, 1988 #2669. Algorithm:

A. if no acute progression (worsening vision…), an initial attempt at purely medical control should be made as the chances of normalizing PRL surgically with pre-op levels > 500 ng/ml are very low³⁶⁰ (these tumors may shrink dramatically with bromocriptine)

B. response should be evident by 4-6 weeks

C. if tumor not controlled medically (≈ 18% will not respond to bromocriptine): surgery followed by reinstitution of medical therapy may normalize PRL

MEDICAL MANAGEMENT

Dopamine agonists

SIDE EFFECTS:³⁶¹ (may vary with different preparations) nausea, H/A, fatigue, orthostatic hypotension with dizziness, cold induced peripheral vasodilatation, depression, nightmares and nasal congestion. Side effects are more troublesome during the first few weeks of treatment. Tolerance may be improved by bedtime dosing with food, slow dose escalation, sympathomimetics for nasal congestion, and acetaminophen 1-2 hrs before dosing to reduce H/A. Psychosis and vasospasm are rare side effects that usually necessitates discontinuation of the drug.

bromocriptine (Parlodel®)DRUG INFO

A semi-synthetic ergot alkaloid that binds to dopamine receptors (dopamine agonist) on normal and tumor lactotrophs, inhibiting synthesis and secretion of PRL and other cell processes resulting in decreased cell division and growth. Bromocriptine lowers prolactin level regardless of the whether the source is an adenoma or normal pituitary (e.g. as a result of stalk effect) to < 10% of pretreatment values in most patients. It also frequently reduces the tumor size in 6-8 weeks in 75% of patients with macroadenomas, but only as long as therapy is maintained and only for tumors that actually produce prolactin. Only ≈ 1% of prolactinomas continue to grow while the patient is on bromocriptine. Prolactinomas may enlarge rapidly upon discontinuation of the drug. However, permanent normoprolactinemia can occur (see below).

Pregnancy issues: bromocriptine can restore fertility. Continued therapy during pregnancy is associated with a 3.3% incidence of congenital anomalies and 11% spontaneous abortion rate which is the same as for the general population. Estrogen elevation during pregnancy stimulates hyperplasia of lactotrophs and some prolactinomas, but the risk of symptomatic enlargement of microadenomas and totally intrasellar macroadenomas is < 3%, vs. 30% risk for macroadenomas³⁶².

Prolonged treatment with bromocriptine may reduce the chances of surgical cure if this should be chosen at a later date. With a microadenoma, one year of bromocriptine may reduce the surgical cure rate by as much as 50%, possibly due to induced fibrosis³⁶³. Thus, it is suggested that if surgery is to be done that it be done in the first 6 months of bromocriptine therapy. Shrinkage of large tumors due to bromocriptine may cause CSF rhinorrhea³¹⁴. SIDE EFFECTS: see above.

Rx: start with 1.25 mg (half of a 2.5 mg tablet) PO q hs (nighttime dosing reduces some side effects) (vaginal administration is an alternative). Add additional 2.5 mg per day as necessary (based on PRL levels), making a dosage change every 2-4 weeks for microadenomas, or every 3-4 days for macroadenomas causing mass effect. Initial recheck of prolactin level after about 4 weeks at a reasonable dose to verify response. To shrink large tumors or for extremely high PRL levels, higher doses are usually needed initially (e.g. 7.5 mg TID for ≈ 6 mos), and then lower doses may be able to maintain normal levels (typical maintenance dosage: 5-7.5 mg daily (range: 2.5-15 mg) which may be given as a single dose or divided TID). SUPPLIED: 2.5 mg scored tabs; 5 mg capsules.

cabergoline (Dostinex®)DRUG INFO

An ergot alkaline derivative that is a selective D₂ dopamine agonist (bromocriptine (see above) affects both D₂ and D₁ receptors)³⁶⁴. The elimination half-life is 60-100 hrs which usually permits dosing 1-2 times weekly. Control of PRL and resumption of ovulatory cycles may be better than with bromocriptine³⁶⁵. SIDE EFFECTS: (see above) H/A and GI symptoms are reportedly less problematic than with bromocriptine. ✖ Cardiac valve disease³⁶⁶affecting the mitral, aortic, and tricuspid valves possibly leading to regurgitation which has not been observed at doses used for prolactinomas (is associated with doses used for Parkinson’s disease which are > 10 x pituitary doses): recommendation: do not discontinue cabergoline for this reason if dose is < 2 mg/wk. ✖ Contraindications: eclampsia or pre-eclampsia, uncontrolled HTN. Dosage should be reduced with severe hepatic dysfunction.

Rx: Start with 0.25 mg PO twice weekly, and increase each dose by 0.25 mg every 4 weeks as needed to control PRL (up to a maximum of 3 mg per week). Typical dose is 0.5-1 mg twice weekly. Some combine the total dose and give it once weekly. Initial recheck of prolactin level after about 4 weeks to verify response. SUPPLIED: 0.5 mg scored tablets.

pergolide (Permax®)DRUG INFO

A long-acting ergot alkaloid dopamine agonist that reduces PRL levels for > 24 hrs. Not FDA approved for hyperprolactinemia. Once daily dosing improves compliance. SIDE EFFECTS: see above. ✖ Risk of cardiac valve disease (see cabergoline (Dostinex®) above).

Rx Start with 0.05 mg PO q hs, and increase by 0.025-0.05 increments (up to a maximum of ≈ 0.25 mg/d) until desired PRL levels are achieved.

Response to medical treatment

Treatment response to DA is assessed with serial prolactin levels as shown in Table 21-46. It is uncommon for a prolactinoma to enlarge without an increase in prolactin level³⁴³.

Discontinuation of dopamine agonists: Long-term therapy with DA agonists has some cytocidal effect on pituitary tissue. In an early report, discontinuation of treatment after 24 months was associated with > 95% recurrence rate³⁶⁷. Recent literature suggests a 20-30% chance of normoprolactinemia off medication in select patients³⁶⁸.

Recommendations³⁶⁸: if response to DA agonist is satisfactory, treat for 1-4 years (microadenomas: check prolactin yearly, macroadenomas are more likely to grow and should be checked more often). Microadenomas or macroadenomas that are no longer visible on MRI are candidates for DA agonist withdrawal. For microadenomas: discontinue the drug; for macroadenomas: slowly taper the drug then discontinue. Recurrence rate is highest during 1st year, check prolactin levels and clinical symptoms every 3 months during the 1st year. Long-term follow up is required, especially for macroadenomas.

Table 21-46 Prolactin level with DA agonist treatment

PRL level (ng/ml)	Recommendation
< 20	maintain
20-50	reassess dose
> 50	consider surgery

ACROMEGALY - MANAGEMENT^{323, 325, 369}

Surgery is the primary treatment modality when treatment is indicated.

1. asymptomatic elderly patients do not require treatment since there is little evidence that intervention alters life expectancy in this group

2. if no contraindications, surgery (usually transsphenoidal) is currently the best initial therapy (worse prognosis with macroadenomas) providing more rapid reduction in GH levels and decompression of neural structures (e.g. optic chiasm) and improves the efficacy of subsequent somatostatin analogues³⁶⁹. Surgery is not recommended for elderly patients

3. medical therapy (see page 652): reserved for:

A. patients not cured^A by surgery (reoperation doesn’t work very often for acromegaly)

B. or for those who cannot tolerate surgery (e.g. due to cardiomyopathy, severe hypertension, airway obstruction…, these contraindications may improve with medical therapy and then surgery can be reconsidered)

C. or for recurrence after surgery or XRT

4. XRT (see page 656): for failure of medical therapy. Not recommended as initial treatment. NB: some practitioners use XRT for surgical failure, and employ medical therapy while waiting for XRT to have an effect (GH levels decline very slowly after XRT - see page 655 for details & for side effects)

Estimated (one-time) cost of transsphenoidal resection: $30,000 in the U.S.

A. the definition of “biochemical cure” with acromegaly is not standardized (see page 662), surgery is still helpful for those “not cured” and improves efficacy of other therapies; IGF-1 may take months to normalize after surgery

MEDICAL THERAPY

1. dopamine agonists (DAs): although not mentioned in the AACE guidelines³²¹, it may be worth trying a DA to see if the tumor responds (≈ 20% respond). If responsive, DAs are especially well suited for GH tumors that cosecrete PRL

A. bromocriptine: (see below) although it benefits only a minority, a first line drug since it is cheaper than pegvisomant or octreotide and is given PO

B. cabergoline (see above)

C. pergolide (see above)

D. others: lisuride, depo-bromocriptine (bromocriptine-LAR)

2. somatostatin analogues: indications: as initial medical therapy, or if no response to DAs, some also use this pre-op to improve surgical success rate

A. octreotide & octreotide-LAR (see below)

B. lanreotide, lanreotide SR & long-acting aqueous gel lanreotide (Autogel)

3. GH antagonists: pegvisomant (see below) considered for failures to above (not a primary therapy)

4. combination therapy: may be more effective than individual drugs. Pegvisomant or octreotide + dopamine agonist if no response to 1 drug alone

bromocriptine (Parlodel®)DRUG INFO

Neoplastic somatotrophs may respond fortuitously to dopamine agonists and reduce growth hormone (GH) secretion. Bromocriptine lowers GH levels to < 10 ng/ml in 54% of cases, to < 5 ng/ml in only ≈ 12%. Tumor shrinkage occurs in only < 20%. Higher doses are usually required than for prolactinomas. If effective, the drug may be continued but should be periodically withdrawn to assess the GH level. SIDE EFFECTS: see page 651. Estimated annual cost: $3,200 in the U.S.

Rx For growth hormone tumors that respond to bromocriptine, the usual dosage is 20-60 mg/d in divided doses (higher doses are unwarranted). The maximal daily dose is 100 mg. For dose escalation regimens, see page 651.

octreotide (Sandostatin®)DRUG INFO

A somatostatin analogue that is 45 times more potent than somatostatin in suppressing GH secretion but is only twice as potent in suppressing insulin secretion, has a longer half-life (≈ 2 hrs after SQ injection, compared to ≈ minutes for somatostatin), and does not result in rebound GH hypersecretion. GH levels are reduced in 71%, IGF-1 levels are reduced in 93%. 50-66% have normal GH levels, 66% achieve normal IGF-1 levels. Tumor volume reduces significantly in about 30% of patients. Many symptoms including H/A usually improve within the first few weeks of treatment. Annual cost to the patient: at least ≈ $7,800 in the U.S. Usually given in combination with bromocriptine.

After 50 μg SQ injection, GH secretion is suppressed within 1 hr, nadirs at 3 hrs, and remains reduced for 6-8 hrs (occasionally up to 12 hrs). SIDE EFFECTS: reduced GI motility and secretion, diarrhea, steatorrhea, flatulence, nausea, abdominal discomfort (all of these usually remit in 10 days), clinically insignificant bradycardia in 15%, cholesterol cholelithiasis (in 10-25%) or bile sludge. Asymptomatic stones require no treatment and routine ultrasonography is not required. Mild hypothyroidism or worsening of glucose intolerance may occur.

Rx: Start with 50-100 μg SQ q 8 hrs. Increase up to a maximum of 1500 μg/d (doses > 750 μg/d are rarely needed). Average dose required is 100-200 μg SQ q 8 hrs.

Sandostatin LAR Depot: long acting release (LAR) form given by IM injection. Rx: give a test dose of short acting octreotide SQ in the office, and if no reaction (e.g. N/V…) begin LAR injections with 20 mg IM q 4 weeks, increase to 30 mg if GH > 5 mU/L just before 4th dose. Control can achieved in some with dosing q 8-12 weeks³⁷⁰.

pegvisomant (Somavert®)DRUG INFO

A competitive GH-receptor antagonist. Treatment for ≥ 12 mos results in normal IGF-1 levels in 97% of patients³⁷¹. No change in pituitary tumor size has been observed³⁷². Indications: failure of somatostatin in patient with GH secreting adenoma (patient is switched to pegvisomant, it is not added to regimen). SIDE EFFECTS: significant but reversible liver function abnormalities occur in < 1%. Serum GH increases, probably as a result of loss of negative feedback on IGF-1 production.

Rx: 5-40 mg/d SQ (dose must be titrated to keep IGF-1 in the normal range, to avoid GH deficiency conditions).

CUSHING’S DISEASE - MANAGEMENT

Overall scheme:

1. if pituitary MRI shows a mass: transsphenoidal surgery

2. if pituitary MR is negative (up to 40% of patients with Cushing’s disease have negative MRI): perform IPS sampling (see page 646)

A. if IPS sampling is positive: surgery

B. if IPS sampling is negative: look for source of ACTH (abdominal CT)

3. if biochemical cure (see page 662 for criteria) is not obtained with surgery:

A. unlike acromegaly, a partial reduction is not helpful to the patient

B. consider re-exploration if pituitary source is still suspected

C. stereotactic radiosurgery or medical therapy (see below)

D. adrenalectomy in appropriate patients (see below)

Transsphenoidal surgery

Transsphenoidal surgery is the treatment of choice for most (medical therapy is inadequate as initial therapy since there is no effective pituitary suppressive medication). Cure rates are ≈ 85% for microadenomas (i.e. tumors ≤ 1 cm dia), but are lower for larger tumors. Even with microadenomas, hemihypophysectomy on the side of the tumor is usually required for cure (the tumor is difficult to completely extirpate) with attendant increased risk of CSF leak. If this fails, consideration should then be for total hypophysectomy. Failure of total hypophysectomy prompts consideration for bilateral adrenalectomy (total hypophysectomy virtually eliminates risk of Nelson’s syndrome following adrenalectomy - see below).

Stereotactic radiosurgery

Often normalizes serum cortisol levels. Useful for: recurrence after surgery, inaccessible tumors (e.g. cavernous sinus)³⁷³…

Adrenalectomy

Total bilateral adrenalectomy (TBA) corrects hypercortisolism in 96-100%³¹⁶ (unless there is an extra-adrenal remnant), but lifelong gluco- and mineralo-corticoid replacement are required and up to 30% develop Nelson’s syndrome (see page 639) (incidence reduced by total hypophysectomy or possibly by pituitary XRT). Indications: continued hypercortisolism with:

1. non-resectable pituitary adenoma

2. failure of medical therapy to control symptoms after transsphenoidal surgery

3. life-threatening Cushing’s disease (CD)

4. CD with no evidence of pituitary tumor (testing should include high-dose DMZ suppression test (see page 646) and/or inferior petrosal sinus sampling (see page 646))

Follow-up after TBA to rule-out Nelson’s syndrome: there is no standardized regimen. Suggestion: check serum ACTH levels q 3-6 months x 1 year, q 6 months x 2 years, q year thereafter. A pituitary MRI is done if an ACTH level is > 100 ng/L, otherwise, annual MRIs are sufficient³¹⁸ x 3 years and then if ACTH levels remain low, get an MRI every other year.

Medical therapy

For patients who fail surgical therapy or for whom surgery cannot be tolerated, medical therapy and/or radiation are utilized. Occasionally may be used for several weeks prior to planned surgery to control significant manifestations of hypercortisolism (e.g. diabetes, HTN, psychiatric disturbances…, see page 638).

Ketoconazole (Nizoral®)³⁶¹: an antifungal agent that blocks adrenal steroid synthesis. The initial drug of choice. Over 75% of patients have normalization of urinary free cortisol and 17-hydroxycorticosteroid levels. SIDE EFFECTS: reversible elevations of serum hepatic transaminase (in 15%), GI discomfort, edema, skin rash. Significant hepatotoxicity occurs in 1 of 15,000 patients. Watch for evidence of adrenal insufficiency (see page 32).

Rx Start with 200 mg PO BID. Adjust dosage based on 24-hr urine free cortisol and 17-hydroxycorticosteroid levels. Usual maintenance doses 400-1200 mg daily in divided doses (maximum of 1600 mg daily).

Aminoglutethimide (Cytadren®)³⁶¹: inhibits the initial enzyme in the synthesis of steroids from cholesterol. Normalizes urinary free cortisol in ≈ 50% of cases. SIDE EFFECTS: dose-dependent reversible effects include sedation, anorexia, nausea, rash and hypothyroidism (due to interference with thyroid hormone synthesis).

Rx Start with 125-250 mg PO BID. Effectiveness may diminish after several months and dose escalation may be needed. Generally do not exceed 1000 mg/d.

Metyrapone (Metopirone®): inhibits 11-ß-hydroxylase (involved in one of the final steps of cortisol synthesis) may be used alone or in combination with other drugs. Normalizes mean daily plasma cortisol in ≈ 75%. SIDE EFFECTS: lethargy, dizziness, ataxia, N/V, primary adrenal insufficiency, hirsutism and acne.

Rx Usual dose range is 750-6000 mg/d usually divided TID with meals. Initial effectiveness may diminish with time.

Mitotane (Lysodren®): related to the insecticide DDT. Inhibits several steps in glucocorticoid synthesis, and is cytotoxic to adrenocortical cells (adrenolytic agent). 75% of patients enter remission after 6-12 months of treatment, and the medication may sometimes be discontinued (however hypercortisolism may recur). SIDE EFFECTS: may be limiting, and include anorexia, lethargy, dizziness, impaired cognition, GI distress, hypercholesterolemia, adrenal insufficiency (which may necessitate supernormal doses of glucocorticoids for replacement due to induced glucocorticoid degradation).

Rx Start with 250-500 mg PO q hs, and escalate dose slowly. Usual dose range is 4-12 gm/d usually divided TID-QID. Initial effectiveness may diminish with time.

Cyproheptadine (Periactin®): a serotonin receptor antagonist that corrects the abnormalities of Cushing’s disease in a small minority of patients, suggesting that some cases of “pituitary” Cushing’s disease are really due to a hypothalamic disorder. Combined therapy with bromocriptine may be more effective in some patients. SIDE EFFECTS: sedation & hyperphagia with weight gain usually limit usefulness.

Rx Usual dosage range: 8-36 mg/d divided TID.

THYROTROPIN (TSH)-SECRETING ADENOMAS - MANAGEMENT

1. transsphenoidal surgery has been the traditional first-line treatment³²⁷. These tumors may be fibrous and difficult to remove³⁷⁴

2. for incomplete resection: post-op XRT is employed

3. if hyperthyroidism persists: medical therapy is added with agents including octreotide, bromocriptine (more effective for tumors that cosecrete PRL), and oral cholecystographic agents (which inhibit conversion of T₄ to T₃) e.g. iopanoic acid

MEDICAL THERAPY

Normal and neoplastic anterior hypophyseal thyrotroph cells possess somatostatin receptors and most respond to octreotide (see below). Occasionally, beta-blockers or low-dose antithyroid drugs (e.g. Tapazole® (methimazole) ≈ 5 mg PO TID for adults) may additionally be required.

Octreotide (Sandostatin®)

Doses required are usually < than with acromegaly. TSH levels decline by > 50% in 88% of patients, and become normal in ≈ 75%. T₄ and T₃ levels decrease in almost all, with 75% becoming normal. Tumor shrinkage occurs in ≈ 33%.

Rx Start with 50-100 μg SQ q 8 hrs. Titrate to TSH, T₄ and T₃ levels.

RADIATION THERAPY FOR PITUITARY ADENOMAS

Conventional EBXRT usually consists of 40-50 Gy administered over 4-6 weeks.

Side effects: Radiation injury to the remaining normal pituitary results in hypocortisolism, hypogonadism, or hypothyroidism in 40-50% of patients after 10 years. It may also injure the optic nerve and chiasm (possibly causing blindness), cause lethargy, memory disturbances, cranial nerve palsies, and tumor necrosis with hemorrhage and apoplexy. Cure rates but also complications are higher after proton beam therapy.

Recommendation: Radiation therapy should not be routinely used following surgical removal. Follow patient with yearly MRI. Treat recurrence with repeat operation. Consider radiation if recurrence cannot be removed and tumor continues to grow.

Nonfunctional tumors

In one series of 89 nonfunctioning pituitary tumors ranging 0.5-5 cm diameter (mean = 2 cm) not totally resected because of involvement of cavernous sinus (or other inaccessible sites), half were treated with radiation therapy (XRT). The recurrence rate was neither lower (and was actually higher) nor later in the XRT group²⁹³. However, another series of 108 pituitary macroadenomas found the recurrence rates shown in Table 21-47 which tend to favor radiation therapy.

When used, doses of 40 or 45 Gy in 20 or 25 fractions, respectively, is recommended³⁷⁶. The oncocytic variant of null cell pituitary tumors appears to be more radioresistant than the nononcocytic undifferentiated cell adenoma³⁷⁶.

Acromegaly

Not the preferred initial treatment. Works better with lower initial GH levels. In most patients, GH levels begin to fall during the first year after XRT, dropping by ≈ 50% after 2 years, and decrease gradually thereafter, reaching ≤ 10 ng/ml in 70% of patients after 10 years. It takes up to 20 years for 90% of patients to achieve GH levels < 5 ng/ml. During this latency period, patients are exposed to unacceptably high levels of GH (octreotide may be used while waiting). Patients are also still at risk for radiation side effects mentioned above. Options include: EBRT, stereotactic radiosurgery (about equally effective). Estimated cost: $20,000.

Cushing’s disease

XRT corrects hypercortisolism in 20-40%, and produces some improvement in another 40%. Improvement may not be seen for 1-2 yrs post treatment.

Table 21-47 Recurrence rate of pituitary tumors removed transsphenoidally*

Extent of removal	Post-op XRT?	Recurrence rate
subtotal	no	50%
gross total		21%
subtotal	yes	10%
gross total		0

* 108 macroadenomas, 6 mos to 14 years follow-up³⁷⁵

SURGICAL TREATMENT FOR PITUITARY ADENOMAS

MEDICAL PREPARATION FOR SURGERY

1. stress dose steroids: given to all patients during and immediately after surgery

2. hypothyroidism: ideally, hypothyroid patients should have > 4 weeks of replacement to reverse hypothyroidism, however:

✖ do not replace thyroid hormone until the adrenal axis is assessed; giving thyroid replacement to a patient with hypoadrenalism can precipitate adrenal crisis. If hypoadrenal, begin cortisol replacement first, may begin thyroid hormone replacement after 24 hours of cortisol

• surgery is done frequently on patients with hypothyroidism and appears to be tolerated well in the vast majority of cases

SURGICAL APPROACHES

1. transsphenoidal: an extra-arachnoid approach, requires no brain retraction, no external scar (aside from where a fat graft is procured, if used). Usually the procedure of choice. Indicated for microadenomas, macroadenomas without significant extension laterally beyond the confines of the sella turcica, patients with CSF rhinorrhea, and tumors with extension into sphenoid air sinus

A. sublabial

B. trans-nares: an alotomy may be used to enlarge the exposure through the nares if necessary

2. transethmoidal approach^{377 (p 343-50)}

3. transcranial approaches:

A. indications: most pituitary tumors are operated by the transsphenoidal technique (see above), even if there is significant suprasellar extension. However, a craniotomy may be indicated for the following³²⁹:

1. minimal enlargement of the sella with a large suprasellar mass, especially if the diaphragma sellae is tightly constricting the tumor (producing a “cottage loaf” tumor) and the suprasellar component is causing chiasmal compression^{378 (p 124)}

2. extrasellar extension into the middle fossa that is larger than the intrasellar component

3. unrelated pathology may complicate a transsphenoidal approach: rare, e.g. a parasellar aneurysm

4. unusually fibrous tumor that could not be completely removed on a previous transsphenoidal approach

5. recurrent tumor following a previous transsphenoidal resection

B. choices of approach

1. subfrontal: provides access to both optic nerves. May be more difficult in patients with prefixed chiasm

2. frontotemporal (pterional): places optic nerve and sometimes carotid artery in line of vision of tumor. There is also incomplete access to intrasellar contents. Good access for tumors with significant lateral extrasellar extension

3. ✖ subtemporal: usually not a viable choice. Poor visualization of optic nerve/chiasm and carotid. Does not allow total removal of intrasellar component

TRANSSPHENOIDAL SURGERY

Booking the case - transsphenoidal surgery

Also see defaults & disclaimers (page v) and pre-op orders (page 660).

1. position: supine, horseshoe head rest or (especially if image guided navigation is used) pin headholder

2. equipment:

A. microscope

B. C-arm (if used)

C. image guided navigation system (if used)

D. endoscopy cart for cases performed endoscopically (surgeon preference)

3. instrumentation: transsphenoidal instrument set (usually includes speculum, curettes, long instruments including bipolars)

4. some surgeons use ENT to perform the approach and closure and for follow-up

5. post op: ICU

6. consent (in lay terms for the patient - not all-inclusive):

A. procedure: removal of pituitary tumor through the nose, possible placement of fat graft from abdomen

B. alternatives: surgery through the skull (trans-cranial), radiation

C. complications: CSF leak with possible meningitis, problems with pituitary hormones which may sometimes be permanent (which would require life-time replacement therapy), injury to optic nerve with visual loss, injury to carotid artery with possible bleeding and/or stroke

Technique

For pre- and post-op orders, see below.

Details of the surgery are beyond the scope of this text, see references^378-381. A brief summary of the procedure:

1. lumbar drain: may be used with some macroadenomas to inject fluid in order to help bring the tumor down (see below), also may be used for post-op CSF drainage following transsphenoidal repair of CSF fistula

2. medications (in addition to pre-op meds, see below): intraoperatively 100 mg hydrocortisone IV q 8 hrs

3. positioning

A. elevate thorax 10-15°: reduces venous pressure

B. position option 1: surgeon standing to right of patient

1. shoulder-roll

2. top of head canted slightly to left

3. extend neck slightly with the head in either a Mayfield head-holder (mandatory if image-guided navigation is to be used) placed low or in AP orientation (to prevent obscuring the sella turcica on lateral fluoroscopy) or on a horse-shoe headrest

4. ET tube positioned down and to patient’s left (to get it out of the way)

5. microscope: observer’s eyepiece on the left

C. position option 2: surgeon standing above patient’s head

1. head pointing straight up towards ceiling, neck slightly extended

D. abdomen or right thigh is prepped for fat graft

4. C-arm fluoro: image-guided navigation can eliminate the need for fluoro. Orient the C-arm for a true lateral by aligning the mandibular rami and/or by superim-posing the floor of the left and right frontal fossae. If this proves difficult, lay a Penfield 4 on the inion oriented from lateral canthus to lateral canthus, then aim the fluoro to shoot “down the barrel” of the Penfield 4

5. after approach to floor of sella is complete (see below), outline the upper and lower boundaries of the sella using an instrument (e.g. suction tip) under fluoro (obtain hard-copy of images for documentation purposes)

6. opening the sellar floor:

A. starting the opening: open exactly in the midline using the nasal septum as a landmark (NB: the septum of the sphenoid sinus is unreliable as a midline indicator, and often curves inferiorly towards one of the carotid arteries).

1. macroadenomas may have thinned the bone to the point that it just flakes off

2. otherwise, use a bayoneted chisel or high-speed diamond burr to start the opening

B. use a Kerrison rongeur to expand the opening. ✖ CAUTION: stay away from the extreme lateral sella to avoid entering the cavernous sinus or injuring the carotid artery

7. coagulate the dura centrally in an “X” pattern (NOT “+” pattern) with bipolar cautery. Macroadenomas may cause yellowish discoloration of the dura directly over the tumor

8. consider aspirating through dura with a 20 gauge spinal needle to R/O large venous sinus (dura often has bluish discoloration), aneurysm, or empty sella

9. incise the dura in the “X” pattern the midline with a #11 scalpel on a bayonetted handle

10. tumor removal

A. macroadenoma:

1. gently bring tumor into the field with ring curettes, and remove with pituitary rongeurs or aspirate with suction. Some tumors are very fibrous and may be difficult to remove

2. do not pull on the lateral component of the tumor with pituitary rongeur due to risk of injuring carotid artery

3. if the suprasellar component will not come down, it may be brought down by having the anesthesiologist inject 5 ml aliquots of saline into a lumbar drain while monitoring blood pressure and pulse^{378 (p 135),382}

4. once the tumor is debulked internally, try to develop a plane between the tumor capsule and the pituitary. A good place to start looking is inferiorly where the dura can be separated from the tumor capsule and then followed on the surface. Sometimes the tumor capsule cannot be removed due to severe bleeding

5. complete tumor removal is often not possible, and the goal of the surgery then is “containment”

6. endoscopic techniques and image guided navigation may be employed to assist in removal of macroadenomas

B. microadenoma:

1. if the side of the tumor is known, begin exploration of the gland on that side by making incision with #11 blade and using a dissector to try and locate the tumor (like a “grain of rice in a blueberry”)

2. for Cushing’s disease, if no tumor is identified on pre-op MRI³¹⁷:

a. intraoperative ultrasound may help localize tumor in ≈ 70% of cases³⁵⁷ but a specialized U/S probe is required

b. if IPS sampling showed a lateralizing ACTH gradient

i. start with a paramedian incision on the side of the higher ACTH gradient

ii. if no adenoma is encountered, the contralateral paramedian and then midline incisions are used to explore the gland

c. if IPS sampling and MRI do not suggest tumor location: the gland is explored sequentially with 2 paramedian incisions and then a midline incision

d. if the adenoma cannot be found, a hemihypophysectomy is performed on the side of higher ACTH levels if IPS sampling shows a lateralizing gradient, or on the side with more suspicious tissue on frozen section. Total hypophysectomy is not routinely performed³¹⁷

C. most adenomas are purplishgrey and easily aspirated, however some may be more fibrous. The normal pituitary gland is firm and rubbery (the adenohypophysis is orange-pink, the neurohypophysis is a whitishgrey), and normally does not curette very easily

D. use image guidance or fluoro to determine approximate location of diaphragma sellae. Do not go cephalad to this to avoid a CSF leak, to avoid entering the circular venous sinus in the dura here and to avoid trauma to the optic chiasm

11. after removal of macroadenoma, check depth of tumor bed on fluoro or image guidance, and make sure it correlates with approximate tumor volume on MRI

12. the sella may be packed in a number of ways³⁸¹, one method:

A. place muscle or fat in defect within sella: some recommend against the use of muscle because it always putrifies^{378 (p 129)}. Do not overpack to avoid recreating mass effect with the graft

B. recreate the floor of the sella using nasal cartilage placed within the sella. Alternatively, a nonporous Medpor® polyethylene transsphenoidal sellar implant (Porex Surgical Products http://www.porexsurgical.com) may be used

C. pack sphenoid sinus with fat from abdomen (option: fat with fascia on surface)

D. fibrin glue may optionally be used to help hold any of these components in place

Approach to sphenoid sinus: often done by ENT. One method:

1. insert temporary speculum into nose. For this discussion, the right nares will be described

2. use endoscope to locate middle concha. Follow this posteriorly to identify os into sphenoid sinus

3. inject local anesthetic with epinephrine to blanch mucosa

4. insert sickle knife into os with sharp side facing the septum (medially) and incise the mucosa as the knife is drawn outward

5. use a Freer to dissect the thusly created mucosal flaps off the medial septum (pull one up, the other down)

6. break through the posterior part of the septum so that both sides of the floor of the sphenoid sinus are exposed. Cartilage or bone from this step is saved to use later in reconstructing the floor of the sella if desired

7. open the floor of the sphenoid and take it all the way to the right os (you will probably not see the left os)

8. place the Hardy speculum or equivalent

9. strip mucosa off the walls of the sphenoid sinus using a Blakely and slow pulling motion

Intraoperative disasters: usually related to loss of landmarks³⁷⁸.

1. entry into carotid artery: heralded by profuse arterial bleeding that can usually be packed off (using fat/fascia graft from thigh or abdomen may help). The operation is halted, and a post-op arteriogram must be done. If a pseudoaneurysm is identified angiographically, it must be eliminated before a potentially lethal rupture; accomplished either by endovascular techniques or by surgical trapping with clips above and below

2. opening through the clivus and erroneous biopsy of the pons

3. opening through the floor of the frontal fossa with entry into inferior frontal lobes

Peri-operative complications

1. hormonal imbalance:

A. acute post-op concerns:

1. alterations in ADH: transient abnormalities are common (see page 661 for typical post-op patterns) including DI. DI lasting > 3 mos is uncommon

2. cortisol deficiency → hypocortisolism → Addisonian crisis if severe

B. long-term: hypopituitarism in ≈ 5% (retrospective series³⁸³)

1. TSH deficiency → hypothyroidism → (rarely) myxedema coma if severe

2. adrenal insufficiency

3. deficiency of sex hormones → hypogonadotrophic hypogonadism

2. secondary empty sella syndrome (chiasm retracts into evacuated sella → visual impairment)

3. hydrocephalus with coma³⁸⁴: may follow removal of tumors with suprasellar extension (transsphenoidally or transcranially). Consider ventriculostomy placement if hydrocephalus is present (even if not symptomatic). Possible etiologies:

A. traction on the attached 3rd ventricle

B. cerebral edema due to vasopressin release from manipulation of the pituitary and/or stalk

C. tumor edema following resection

4. infection

A. pituitary abscess^{385, 386}

B. meningitis

5. CSF rhinorrhea (fistula): 3.5% incidence³⁷⁵

6. carotid artery rupture: rare. May occur intraoperatively (see above) or in delayed fashion after surgery, often ≈ day 10 post-op (due to breakdown of fibrin around carotid, or possibly due to rupture of a pseudoaneurysm created at surgery)

7. entry into cavernous sinus with possible injury of any structure within

8. nasal septal perforation

FRONTOTEMPORAL (PTERIONAL) APPROACH

A right sided approach is usually employed (less risk to dominant hemisphere). Exceptions: when the left eye is the side of worse vision; if there is predominant left sided tumor extension; if there is other pathology on the left (e.g. aneurysm).

Positioning is that same as for ACoA aneurysm with the head turned 60° to the side (see page 159). The frontal lobe is elevated, and the temporal tip is gently retracted posteriorly. Bridging veins to the temporal tip must be coagulated to avoid rupture, as for any pterional approach. The approach is similar to that for an ACoA aneurysm (i.e. more emphasis is placed on frontal lobe elevation than on temporal tip retraction), except that unlike ACoA aneurysm, exposure of the ICA is not needed because proximal control is not necessary.

The tumor capsule can usually be seen between the two optic nerves. The capsule is coagulated with bipolar cautery, and is incised. The tumor is then debulked from within. By staying within the capsule, risk of injury to the pituitary stalk and optic chiasm is minimized. Significant amounts of tumor can be removed by aspiration if it is soft and suckable.

✖ Caution: the blood supply to the optic chiasm is from the inferior aspect. Skeletonizing the chiasm or attempting to tease away tumor adherent to it may worsen vision.

PERI-OPERATIVE MANAGEMENT

Pre-op orders

1. Polysporin® ointment (PSO) applied in both nostrils the night before surgery

2. antibiotics, one of the following regimens may be used:

• chloramphenicol 500 mg IVPB at 11 PM & 6 AM

OR

• chloramphenicol 500 mg PO at MN & IV at 6 AM; ampicillin 1 gm PO at MN & IV at 6 AM

OR

• Unasyn® 1.5 gm (1 gm ampicillin + 0.5 gm sulbactam) IVPB at MN & 6 AM

3. steroids, either:

• hydrocortisone sodium succinate (Solu-Cortef®) 50 mg IM at 11 PM & 6 AM. On call to OR: hang 1 L D₅ LR + 20 mEq KCl/l + 50 mg Solu-Cortef at 75 ml/hr

OR

• hydrocortisone 100 mg PO at MN & IV at 6 AM

4. intra-op: continue 100 mg hydrocortisone IV q 8 hrs

Post-op orders

1. intake & output (I’s & O’s) q 1 hr; urine specific gravity (SG) q 4° and anytime urine output (UO) > 250 ml/hr

2. activity: BR with HOB @ 30°.

3. diet: ice chips PRN. Patient is not to drink through a straw^A

4. no incentive spirometry^A

5. IVF: base IV D5 1/2 NS + 20 mEq KCl/L at appropriate rate (75-100 ml/hr) PLUS: replace UO > base IV rate ml for ml with 1/2 NS.

NB: if patient receives significant fluids intraoperatively, then they may have an appropriate post-op diuresis, in which case consider replacing only ≈ 2/3 of UO > base IV rate with 1/2 NS

6. meds

A. antibiotics: continue chloramphenicol 500 mg IVPB q 6 hr (also continue ampicillin if used pre-op), change to PO when tolerated, D/C when nasal packing removed

B. steroids (post-op steroids are required until the adequacy of endogenous steroids is established, especially with Cushing’s disease, see below). Either:

• hydrocortisone 50 mg IM/IV q 6 hrs, on POD #2 change to prednisone 5 mg PO q 6 hrs x 1 day, then 5 mg PO BID, D/C on POD #6

OR

• hydrocortisone 50 mg IM/IV/PO BID, taper 10 mg/dose/day to physiologic dose of 20 mg q AM and 10 mg q PM until adrenal axis assessed

C. diabetes insipidus (DI): see below for typical patterns. Criteria: U.O. > 250 ml/hr x 1-2 hrs, and SG < 1.005 (usually < 1.003). If DI develops, attempt to keep up with fluid loss with IVF (see above); if rate is too high for IV or PO replacement (> 300 cc/hr x 4 hrs or > 500 cc/hr x 2 hrs), check urine S.G. and if < 1.005 then give a vasopressin preparation (see below, or see Table 2-7, page 17). Caution: danger of overtreating in case of triphasic response (see below), therefore use EITHER:

• 5 U aqueous vasopressin (Pitressin®) IVP/IM/SQ q 6 hrs PRN

OR

• desmopressin (DDAVP®) injection SQ/IV titrated to UO. Usual adult dose: 0.5-1 ml (2-4 μg) daily in 2 divided doses

AVOID

✖ avoid tannate oil suspension, because of erratic absorption and it is a long acting preparation

THEN: when nasal packs out, either

• intranasal DDAVP (100 μg/ml): range 0.1-0.4 ml (10-40 μg) intranasally BID (typically 0.2 ml BID) PRN

OR

• clofibrate (Atromid S®) 500 mg PO QID (does not always work)

7. labs: renal profile with osmolarity q 6 hrs, 8 A.M. serum cortisol

8. nasal packs: remove on post-op day 3-6

A. to avoid negative pressure on sphenoid sinus with risk of aggravating CSF fistula

Urinary output: patterns of postoperative diabetes insipidus

Manage diabetes insipidus (DI) as described above. Post-op DI generally follows one of three patterns³⁸⁷ (see Diabetes insipidus, page 15 for details):

1. transient DI: lasts until ≈ 12-36 hrs post-op

2. “prolonged” DI: lasts months, or may be permanent

3. “triphasic response” (least common). Summary: DI → normalization or SIADH-like picture → DI (again)

Assessment for postoperative ACTH (corticotropin) reserve

If patient was not hypocortisolemic pre-op:

1. taper and stop hydrocortisone 24-48 hrs post-op. Then, check 6 AM serum cortisol level 24 hrs after discontinuing hydrocortisone and interpret the results as shown in Table 21-48 352. If there is any question about reserve, the patient can be discharged on hydrocortisone 50 mg PO q AM and 25 mg PO q PM until adrenal reserve can be formally assessed

2. alternatively, if the patient goes home on hydrocortisone and was not on it pre-op, taper it over 2-3 weeks to 20 mg po q AM and 10 mg q 4 PM (a little above maintenance to provide for some stress coverage) for several days, then hold the PM dose and check an 8 AM serum cortisol then next day and as soon as the blood is drawn have the patient take their morning dose and resume regular dosing until the test results are available. If this 8 AM cortisol shows any significant function, then taper the patient off hydrocortisone

3. metyrapone (Metopirone®) test: useful if there is suspicion of reduced reserve of pituitary ACTH production. All patients should have a cosyntropin stimulation test first to rule-out primary adrenal insufficiency (see page 647). ✖ Do not do this test if there is primary adrenal insufficiency. ✖ Do not do this test as an outpatient. Metyrapone inhibits 11-ß-hydroxylation in the adrenal cortex, reducing production of cortisol and corticosterone with concomitant increase of serum 11-deoxycortisol precursors and its 17-OHCS metabolites which appear in the urine. In response, a normal pituitary increases ACTH production. Test: give 2-3 grams metyrapone PO at midnight; a normal response is a serum 11-deoxycortisol level > 7 μg/dl the next morning. CAUTION: in patients with very little reserve, the reduced cortisol may provoke adrenal insufficiency (this test is safer than the higher doses used for urinary 17-OHCS testing)

Table 21-48 Interpretation of 6 AM cortisol levels

6 AM cortisol	Interpretation	Management
≥ 9 μg/dl	normal	no further tests or treatment
3-9 μg/dl	possible ACTH deficiency	place patient on hydrocortisone* (see page 31)
≤ 3 μg/dl	ACTH deficient

* perform cosyntropin stimulation test (see page 647) 1 month post-op; D/C steroids if normal; if subnormal, then permanent replacement required

Postoperative CT/MRI scan

A study using CT in 12 patients with macroadenomas following transsphenoidal surgery without radiation therapy demonstrated that the maximal height of the pituitary “mass” did not return to normal immediately post-op (even with total tumor removal), rather a period of 3-4 months was required³⁸⁸.

Σ	The optimal timing of the initial post-op CT or MRI to function as a baseline to ruleout future recurrence after transsphenoidal surgery is ≈ 3-4 months post-op.

OUTCOME

FOLLOWING TRANSSPHENOIDAL SURGERY

In cases with compression of the optic apparatus, there can be significant improvements in vision following surgery^{375, 389}.

General information:

1. endocrinologic cure was attained in 25% of prolactin-secreting tumors, and in 20% of growth hormone-secreting tumors^A (see below)

2. gross total removal was unusual in tumors with > 2 cm suprasellar extension^A

3. recurrence incidence: ≈ 12%, with most recurring 4-8 years post-op^A

4. Cushing’s disease: surgical cure rates are ≈ 85% for microadenomas (i.e. tumors ≤ 1 cm dia), but are lower for larger tumors (see below)

BIOCHEMICAL OUTCOME

Acromegaly Criteria of biochemical cure: The criteria for biochemical cure of acromegaly is not standardized. There may be a discord between IGF-1 levels and mean GH levels³⁷⁰. Many use a GH cutoff level; range of levels described: < 2.5-5 ng/ml. Others feel that an elevated IGF-1 represents lack of cure even if GH < 5. However, normal IGF-1 levels may not be mandatory³⁹⁰. Still others require a normal IGF-1 AND a normal response to an oral glucose suppression test (OGST) (see page 648).

Low GH levels that do not also suppress to < 1 ng/ml after an OGST are considered controlled but not cured (even with normal IGF-1 levels)³²¹. If asymptomatic, expectant management with close follow-up is recommended³²¹.

Σ	Biochemical cure criteria for acromegaly is not standardized. Recommendation321: 1) IGF-1 levels within age-matched reference range 2) basal (morning) serum GH level < 5 ng/ml, AND GH nadir < 1 ng/ml in OGST

Outcome: Transsphenoidal surgery results in biochemical cure in 85% of cases with adenomas < 10 mm diameter, no evidence of local invasion, and random GH levels < 40 ng/ml pre-op. Overall, ≈ 50% of all acromegalics undergoing transsphenoidal surgery had a biochemical cure³⁹¹. Only 30% of macroadenomas and very few with marked suprasellar extension have surgical cure. Patients not cured with surgery require lifelong medical suppression. These tumors may also recur years later after apparent cure. Patients should be monitored every 6-12 months for recurrence³²¹.

Cushing’s disease

There are numerous methodologies for determining biochemical cure for Cushing’s disease. One difficulty is that exogenous steroids are often given post-op to avoid potential hypoadrenalism or Addisonian crisis or for nausea. Some options:

1. immediate post-op early morning cortisol levels³¹⁷:

A. all steroids are withheld post-op (including dexamethasone as an antiemetic) unless biochemical and/or clinical^B evidence of hypocortisolism. ✖ Requires close monitoring and administration of steroids if symptoms develop

B. serum ACTH and cortisol levels are drawn between 6-9:00 AM on post-op days 1 & 2

C. early remission defined as a lowest cortisol level ≤ 140 nmol/L (≤ 5 μg/dl)

1. 97% (31/32) patients with early remission had sustained remission with mean follow-up of 32 months

2. only 12.5% (1/8) without early remission showed evidence of sustained remission

3. this has been used to select patients for possible early re-exploration

4. early ACTH levels usually drop, but do not consistently become sub-normal and are not reliable in predicting sustained remission³¹⁷

2. provocative tests

A. overnight low-dose dexamethasone suppression test: an AM cortisol level on post-op day 3 that is ≤ 8 μg/dl after an overnight 1 mg dexamethasone suppression test is predictive of sustained remission in 97%³⁹²

B. CRH stimulation test³⁹³

3. measurements usually conducted 3 days to 2 weeks post-op following 24 hours of steroid cessation after initial post-op coverage with glucocorticoids

A. 24-hour urinary free cortisol

B. serum cortisol: the criteria of a cortisol level < 50 nmol/l (< 1.8 μg/dl)^394-396is probably too stringent^{317, 397, 398}

C. serum ACTH

A. based on a series of 108 macroadenomas³⁷⁵

B. clinical signs of hypocortisolism: nausea, anorexia, H/A, arthralgias

The overall remission rate since 1980 is 64-93%, with the highest rates (86-98%) in patients with noninvasive microadenomas identifiable on MRI³¹⁷.

Following effective treatment, all of the following usually improve but may not normalize:

1. HTN and hyperglycemia: within ≈ 1 year

2. osteoporosis related to CD: over ≈ 2 years

3. psychiatric symptoms

Thyrotropin (TSH)-secreting adenomas

Following debulking, small amounts of residual tumor may continue to produce sufficient TSH for hyperthyroidism to persist³⁷⁴. Following surgery + XRT, only ≈ 40% achieve a cure (defined as no residual tumor at surgery or on imaging, and normal free T₃ with TSH levels at or below normal).

MANAGEMENT OF RECURRENT PITUITARY ADENOMAS

For tumors demonstrating significant regrowth or symptoms following initial resection, consideration for re-resection may be given. Once the tumor is debulked, consideration should be given to XRT, either immediately following the second operation, or, if recurrence after a second operation then almost certainly after a third debulking.

21.2.10. Craniopharyngioma

Craniopharyngiomas (CP) are tumors that develop from residual cells of Rathke’s pouch (see page 109), and tend to arise from the anterior superior margin of the pituitary. They are lined with stratified squamous epithelium. Some CP may arise primarily within the third ventricle³⁹⁹. Almost all CP have solid and cystic components; fluid in the cysts varies, but usually contains cholesterol crystals. CP do not undergo malignant degeneration; but difficulty in cure makes them malignant in behavior^{267 (p 905-15)}. CP are distinct from Rathke’s cleft cyst, but share some similarities (see below).

Calcification: microscopically 50%. Plain x-ray: 85% in childhood, 40% in adults.

EPIDEMIOLOGY

Incidence: 2.5-4% of all brain tumors; about 50% occur in childhood (9% of Matson’s series). Peak incidence: age 5-10 yrs.

ANATOMY

Arterial supply: usually small feeders from ACA and A-comm, or from ICA and Pcomm (do not receive blood from PCA or BA-bifurcation unless blood supply of floor of third ventricle is parasitized).

SURGICAL TREATMENT

Pre-op endocrinologic evaluation

As for pituitary tumor (see page 643). Hypoadrenalism may be corrected rapidly, but hypothyroidism takes longer; either condition can increase surgical mortality.

Approach

Usually via large right frontotemporal flap as low as possible along base of frontal fossa (lateral sphenoid wing rongeured/drilled). Approach to tumor is extra-axial, whether subfrontal or frontotemporal. Alltumors should be aspirated (even if they appear solid radiographically). Then, with microscope, possible approaches include:

1. subchiasmatic: through space between optic nerves and anterior to chiasm. It was thought that a “prefixed chiasm” (i.e. congenitally short optic nerves with chiasm unusually close to the planum sphenoidale) was more common in patients with CP, making this approach more difficult. However, in reality the chiasm is probably bowed anteriorly by the tumor within the third ventricle giving the illusion of a prefixed chiasm in most cases

2. opticocarotid (between right ICA and right optic nerve/tract)

3. lamina terminalis (tumor often needs to be brought down and removed subchiasmatically)^{399, 400}

4. lateral to carotid artery

5. transfrontal-transsphenoidal: drill off tuberculum sellae

Alternative approaches to frontotemporal

1. pure transsphenoidal: if dark fluid is aspirated with no CSF evident, it is possible to leave a stent from the tumor cavity to the sphenoid air-sinus to permit continued drainage

2. transcallosal: strictly for tumors limited to the third ventricle

3. a combined subfrontal/pterional approach capitalizes on the advantages of each (head is positioned with slight lateral rotation)

Spare the following structures: small arterial feeders to undersurface of the chiasm (major supply) and tract; at least a remnant of pituitary stalk (recognized by unique pattern of longitudinal striations which are the long portal veins). If the tumor easily pulls down from above then this is permissible, however do not pull too hard or else hypothalamic injury may result.

Post-op

1. steroids: these patients are all considered hypoadrenal. Give hydrocortisone in physiologic doses (for mineralocorticoid activity) in addition to dexamethasone (glucocorticoid that treats edema) taper (see page 31). Taper steroids slowly to avoid aseptic (chemical) meningitis

2. diabetes insipidus (DI): often shows up early. May be part of a “triphasic response” (see Urinary output: patterns of postoperative diabetes insipidus, page 661). Best managed initially with fluid replacement. If necessary, use short acting vasopressin (prevents iatrogenic renal shutdown if a SIADH-like phase develops during vasopressin therapy)

Outcome

5-10% mortality in most series, most from hypothalamic injury (unilateral hypothalamic lesions are rarely clinically evident; bilateral injuries may produce hyperthermia and somnolence; damage to anterior osmoreceptors may → loss of thirst sensation). Five year survival is ≈ 55-85% (range from 30-93% has been reported).

RADIATION

Controversial. SIDE EFFECTS: include endocrine dysfunction, optic neuritis, dementia. Post-op XRT probably helps prevent regrowth when residual tumor is left behind⁴⁰¹, however, in pediatric cases it may be best to postpone XRT (to minimize deleterious effect on IQ, see page 771), recognizing that reoperation may be necessary for recurrence.

RECURRENCE

Most recurrences are in < 1 year, few > 3 yrs (very delayed recurrence usually follow what was thought to be “total” removal). Morbidity/mortality is higher with re-operation.

21.2.11. Rathke’s cleft cyst

Rathke’s cleft cyst (RCC) are nonneoplastic lesions that are thought to be remnants of Rathke’s pouch. They are primarily intrasellar, and are found incidentally in 13-23% of necropsies⁴⁰². The adenohypophysis arises from proliferation of the anterior wall of Rathke’s pouch, and so RCC have a similar lineage to pituitary adenomas and are rarely found together⁴⁰³. RCC are often discussed in contrast to craniopharyngiomas (CP) (see above). Some features are compared in Table 21-49.

RCC usually appear as low-density cystic lesions on CT. One half show capsular enhancement. MRI appearance is variable⁴⁰⁴.

Table 21-49 Comparison of craniopharyngioma to Rathke’s cleft cyst

Feature	Craniopharyngioma	Rathke’s cleft cyst
site of origin	anterior superior margin of pituitary	pars intermedia of pituitary
cell lining	stratified squamous epithelium	single layer cuboidal epithelium
cyst contents	cholesterol crystals	resembles motor oil
surgical treatment	total removal is the goal	partial excision and drainage404
cyst wall	thick	thin

21.2.12. Colloid cyst

Key concepts:

• slow-growing benign tumor comprising < 1% of intracranial tumors

• classically occurs in the anterior 3rd ventricle, blocking foramina of Monro → obstructive hydrocephalus involving only the lateral ventricles (≈ pathognomonic)

• enhances minimally or not at all on CT/MRI

• natural history: risk of sudden death has been described, but is controversial

• treatment is surgical. Main options: transcallosal, transcortical/transventricular (only if hydrocephalus), ventriculoscopic

AKA neuroepithelial cysts. Comprise 2% of gliomas, and about 0.5-1% of all intracranial tumors⁴⁰⁵. Usual age of diagnosis: 20-50 yrs.

PATHOGENESIS

Origin: unknown. Implicated structures include: paraphysis (evagination in roof of third ventricle, rudimentary in humans), diencephalic ependyma in the recess of the postvelar arch, ventricular neuroepithelium.

Comprised of a fibrous epithelial-lined wall filled with either mucoid or dense hyloid substance. A slow growing, benign tumor.

Most commonly found in the third ventricle in the region of the foramina of Monro, but may be seen elsewhere, e.g. in septum pellucidum⁴⁰⁶.

PRESENTATION

Symptoms are shown in Table 21-50. Signs are shown in Table 21-51, most commonly presents either with signs of intermittent acute intracranial hypertension (classically attributed to movement of the cyst on its pedicle causing episodic obstruction of the foramina of Monro, rarely born out at operation) or with chronic hydrocephalus (from chronic obstruction). Most clinically significant cysts are > 1.5 cm in diameter.

Table 21-50 Symptoms of colloid cyst at presentation*

Symptom	No.	%
headache	26	68%
gait disturbance	18	47%
disturbed mentation	14	37%
vomiting (± nausea)	14	37%
blurred vision	9	24%
incontinence	5	13%
dizziness	5	13%
tinnitus	5	13%
seizures	4	10%
acute deterioration	4	10%
diplopia	3	8%
“drop attacks”	1
diabetes insipidus	1
asymptomatic	1

* 38 patients, pre-CT era⁴⁰⁵

SUDDEN DEATH

A high rate of sudden death has been reported with colloid cysts (20% in pre-CT era⁴⁰⁷) but is probably overestimated. The obsolete theory was that these tumors are mobile and thus could shift position and acutely block CSF flow with resultant herniation. Progressive obstruction from tumor growth does often produce chronic hydrocephalus, and it is possible that at some point the brain may decompensate in some cases. Changes in CSF dynamics resulting from procedures (LP, ventriculography…) may have also contributed⁴⁰⁸. Another proposed mechanism is disturbance of hypothalamic-mediated cardiovascular reflex control⁴⁰⁸.

DIAGNOSIS

Imaging (MRI or CT) demonstrates the tumor usually located in the anterior 3rd ventricle. Here, it often blocks both foramina of Monro causing almost pathognomonic hydrocephalus involving only the lateral ventricles (sparing the 3rd and 4th).

MRI: usually the optimal imaging technique. However, there are cases where cysts are isointense on MRI and CT is superior⁴⁰⁹. When the lesion is identifiable, MRI clearly demonstrates the location of the cyst and relation to nearby structures, usually obviating an angiogram. MRI appearance: variable. Usually hyperintense on T1WI, hypointense on T2WI. Enhancement: minimal, sometimes involving only capsule.

CT scan: findings are variable. Most are hyperdense (however, iso- and hypodense colloid cysts occur), and about half enhance slightly. Density may correlate with viscosity of contents, hyperdense cysts were harder to drain percutaneously⁴¹⁰. CT is usually not quite as good as MRI, especially with isodense cysts. These tumors calcify only rarely.

✖ LP: contraindicated prior to placement of shunt due to risk of herniation.

Table 21-51 Signs at presentation*

Sign	No.	%
papilledema	18	47%
gait disturbance	12	32%
normal exam	10	26%
hyperreflexia	9	24%
Babinski reflex	8	21%
incoordination	5	13%
nystagmus	5	13%
tremor	4	10%
hyporeflexia	3	8%
6th nerve palsy	2	5%

* 38 patients with colloid cysts, pre-CT era⁴⁰⁵

TREATMENT

Optimal treatment remains controversial. Initially, shunting without treating the cyst was advocated⁴¹¹. The nature of the obstruction (both foramina of Monro) requires bilateral ventricular shunts (or, unilateral shunt with fenestration of the septum pellucidum). Presently, one form or another of direct surgical treatment is usually recommended for some or all of the following reasons:

• to prevent shunt dependency

• to reduce the possibility of tumor progression

• since the mechanism of sudden neurologic deterioration may be due to factors such as cardiovascular instability from hypothalamic compression and not due to hydrocephalus

Surgical management options (also see Approaches to the third ventricle, page 168):

1. transcallosal approaches: not dependent on dilated ventricles. Higher incidence of venous infarction or forniceal injury (see below)

2. transcortical approach: higher incidence of post-op seizures (≈ 5%). Not feasible with normal sized ventricles (e.g. in patient with VP shunt) see page 172

3. stereotactic drainage: see below

4. ventriculoscopic removal: see below

TRANSCALLOSAL APPROACH

Access to the 3rd ventricle via either the foramen of Monro or by interfornicial approach. Since colloid cysts tend to occur exactly at the foramen of Monro, it is rarely necessary to enlarge the foramen to locate the tumor. See Transcallosal approach to lateral or third ventricle on page 169.

STEREOTACTIC DRAINAGE OF COLLOID CYSTS

May be useful⁴¹², especially in patients with normal ventricles from shunting, but the contents may be too viscous⁴¹³, and the tough capsule may make blind penetration difficult. Total or even subtotal aspiration may not require further treatment in some patients; however, recurrence rate is higher than with surgical removal⁴¹⁴.

Early morbidity was relatively high from this procedure possibly from vascular injury or mechanical trauma; this has improved. May be more feasible with intraoperative ventriculography⁴¹⁵ or with a ventriculoscope⁴¹⁶ (some say this is the initial procedure of choice⁴¹⁷, with craniotomy reserved for treatment failures).

Two features that correlate with unsuccessful stereotactic aspiration⁴¹⁸:

1. high viscosity: correlates with hyperdensity on CT (low viscosity correlated with hypo- or isodense CT appearance; no MRI finding correlated with viscosity)

2. deflection of the cyst from tip of aspirating needle due to small size

Stereotactic technique⁴¹⁹:

1. insertion point of stereotactic needle is just anterior to right coronal suture

2. start with sharp-tipped 1.8 mm probe, and advance to 3-5 mm beyond target site (to accommodate for displacement of cyst wall)

3. use a 10 ml syringe and apply 6-8 ml of negative aspiration pressure

4. if this does not yield any material, repeat with a 2.1 mm probe

5. although complete cyst evacuation is desirable, if this cannot be accomplished an acceptable goal of aspiration is re-establishment of patency of the ventricular pathways (may be verified by injecting 1-2 cc of iohexol)

21.2.13. Hemangioblastoma

Key concepts:

• highly vascular well-circumscribed solid or cystic neoplasm of CNS or retina

• the most common primary intra-axial tumor in the adult posterior fossa

• may occur sporadically or as part of von Hippel-Lindau disease

• on imaging, may be solid, or cystic with enhancing mural nodule

• CBC: may be associated with erythrocytosis (polycythemia)

Hemangioblastomas^{267 (p 772-82)} (HGB) are histologically benign tumors. Intracranially, they occur almost exclusively in the p-fossa (the most common primary intra-axial pfossa tumor in adults). May occur in cerebellar hemisphere, vermis or brainstem. Less than 100 supratentorial cases have been reported. May also occur in spinal cord (1.5-2.5% of spinal cord tumors) - see page 732. Relationship and/or identity with angioblastic meningiomas is controversial. Also difficult to distinguish histologically from a renal cell carcinoma.

HGB may occur sporadically, but 20% occur as part of von Hippel-Lindau disease (see below). Retinal HGB and/or angiomas occur in 6% of patients with cerebellar HGBs.

21.2.13.1. von Hippel-Lindau disease (VHL)

Key concepts:

• disorder with hemangioblastomas (HGB) 1° of cerebellum, retina, brainstem & spinal cord, as well as renal cysts/tumors, pheochromocytomas (among others)

• autosomal dominant, due to inactivation of tumor suppressor gene on 3p²⁵

• expression and age of onset are variable, but ≈ always manifests by age 60

• mean age of developing HGBs is at least 10 years younger than sporadic HGBs

A multisystem neoplastic disorder characterized by a tendency to develop hemangioblastomas (HGB) of the retina, brain and spinal cord, renal clear cell carcinoma (RCC), pheochromocytomas, endolymphatic sac tumors, and others^{420, 421} (retinal location is 2nd most common after cerebellar) (see Table 21-52). The variability of von Hippel-Lindau disease (VHL) has lead some to suggest the use of the term hemangioblasomatosis.

Epidemiology

Incidence: 1 in 31,000 to 36,000 live births. ≈ 30% of patients with cerebellar HGB have VHL¹⁰.

Genetics

Autosomal dominant inheritance with ≈ 95% penetrance at age 60 yrs^{421, 424}. 4% of VHL are asymptomatic carriers. The VHL gene is a tumor suppressor gene on chromo-some 3p²⁵, and biallelic inactivation is required for tumor development¹⁰. Most patients inherit a VHL gene (allele) with the germline mutation from the affected parent and a normal somatic (wild-type) VHL gene from the unaffected parent.

Subtypes of VHL⁴²⁵

Type I may have any manifestation of VHL except pheochromocytoma

Type II pheochromocytoma is characteristic

Type IIA have low risk of renal cell Ca and neuroendocrine pancreatic tumor

Type IIB higher risk of renal cell Ca and neuroendocrine pancreatic tumor

Type IIC risk of pheochromocytoma only (without risk of HGB or RCC)

Diagnostic criteria

Suggested diagnostic criteria for VHL:

1. in 80% of patients with VHL there is a multigenerational family history, and only 1 manifestation (CNS HGB or visceral lesion) is necessary to make the diagnosis

2. if no family history (20% of VHL, many of these represent a de novo mutation): 2 manifestations including 1 CNS or retinal HGB are required⁴²⁶

3. genetic testing in uncertain cases (see below)

Table 21-52 Associations with von Hippel-Lindau disease*

Common lesions	Frequency in VHL
hemangioblastomas
cerebellum (solid or cystic)	80%
retina	41-59%
brainstem	10-25%
spinal cord	10-50%
pancreatic tumors or cysts	22-80%
renal clear cell Ca & cysts	14-60%
polycythemia	9-20% of intracranial HGBs
Rare lesions (pertinent to nervous system)	Frequency in VHL
supratentorial hemangioblastoma	3-6%
cystadenomas of the broad ligament	10% of ♀
papillary cystadenomas of epididymis	25-60% ♂
endolymphatic sac tumors	10-15%
adrenal medullary pheochromocytoma (tends to be bilateral)	7-24%

* see references^421-423 for more

Tumors associated with VHL

1. cerebellar hemangioblastomas (HGB):

A. prevalence: 44-72% of VHL patients

B. mean age of diagnosis in patients with cerebellar hemangioblastomas is at least 10 years younger than sporadic cerebellar hemangioblastomas

C. cysts are commonly associated with cerebellar, brainstem and spinal HGBs

D. cysts grow at a faster rate than the HGBs, symptoms related to mass effect are frequently secondary to the cysts

E. cerebellar HGBs were located in the superficial, posterior and superior half of the cerebellar hemispheres⁴²⁷

F. 93% of the cerebellar HGBs were located in the cerebellar hemispheres and 7% in the vermis

G. the HGBs are also more frequently found in the superficial posterior half of the brainstem and the spinal cord

H. the HGBs have multiple sequential growth and quiescent phases

2. spinal cord hemangioblastomas

A. occur in 13-44% of VHL patients

B. 90% are located rostrally within the cervical and thoracic cord. Almost all (96%) of the tumors are located in the posterior half of the spinal cord, 4% are located in the ventral half of the spinal cord. 1-3% are found in the lumbosacral nerve roots

C. by way of comparison, 80% of spinal cord HGB are associated with VHL, whereas only 5-31% of cerebellar HGB are associated with VHL

D. 95% of symptom-producing spinal HGBs are associated with syringomyelia

3. brainstem hemangioblastomas

A. usually located in the posterior medulla oblongata usually around the obex and the region within the area prostrema

4. pheochromocytomas (PCC): 20% of PCC are associated with VHL. PCC occur in 7-20% of families with VHL

5. endolymphatic sac tumors (ELST):

A. locally invasive benign tumors that occur in 10-15% of VHL patients (30% of these will develop bilateral ELSTs - VHL is the only disease with bilateral ELSTs). Rarely metastasize

B. presents with hearing loss in 95% (may be acute (86%) or insidious (14%)), tinnitus (90%), vertigo or imbalance (66%), aural fullness (30%), and facial paresthesias (8%)

C. mean age of onset of hearing loss: 22 years (range: 12-50)⁴²⁸

6. retinal hemangioblastomas⁴²⁹

A. occur in > 50% of VHL patients. Mean age a presentation: 25 years

B. frequently bilateral, multifocal and recurrent

C. often asymptomatic. Visual symptoms occur with progressive growth, edema, retinal detachments and hard exudates

D. typically located in the periphery and near or on the optic disc

E. microangiomas measuring a few hundred microns without dilated feeding vessels may be located in the periphery

F. retrobulbar HGB are rare(5.3% in NIH cohort)⁴³⁰

G. severity of optic disease correlates with CNS and renal involvement

H. early diagnosis and treatment with laser photocoagulation, and cryotherapy can prevent visual loss. Low dose external XRT may be an option for refractory cases

7. renal-cell carcinoma (RCC)^423,431-437

A. the most common malignant tumor in VHL. Usually a clear cell carcinoma

B. lifetime risk for RCC in VHL: ≈ 70%.

C. the growth rate of RCC is high variable

D. RCC is the cause of death in 15-50% of patients

E. metastases respond poorly to chemotherapy and radiation

F. bilateral and multiple lesions are common

G. partial nephrectomy or tumor enucleation is preferred to avoid/delay dialysis and transplantation

H. nephron- or renal-sparing surgery recommended for tumors less < 3 cm

I. promising techniques: cryo- and radiofrequency ablation of tumors < 3 cm

8. renal cysts^{423, 433, 436-438}

A. 50-70% of VHL patients have bilateral and multiple renal cysts

B. rarely cause profound renal impairment

C. chronic renal failure or renal hypertension not as common as with polycystic kidney disease

9. epididymal cystadenomas

A. benign lesions that arise from the epididymal duct

B. found in 10-60% of male VHL patients

C. typically appear in the teenage years

D. may cause infertility if bilateral

E. may be multiple

10. broad ligament cystadenomas

A. arise from the embryonic mesonephric duct

B. true incidence unknown

C. rarely reported and usually not recognized in women with VHL

11. pancreatic neuroendocrine tumors and cysts

A. 35 to 70% of patients with VHL develop an endocrine tumor or cyst

B. pancreatic cysts are generally asymptomatic and often multiple

C. pancreatic neuroendocrine tumors are usually non-functional and 8% of them are malignant

D. differential diagnosis: pancreatic islet cell tumors, MEN2

Table 21-53 Health-care provider’s surveillance guidelines for patients with or at risk for VHL*

Age	Surveillance
Any age	DNA testing for VHL marker is available to identify family members at risk
From birth	check for neurologic deficit, nystagmus, strabismus, white pupil… & refer to retinologist for abnormal findings. Newborn hearing screening
1 year	retina exam† (especially if positive for VHL mutation)
2-10 years	Annual: • PE‡ including orthostatic blood pressure measurement, neurologic exam, retina exam† • blood test or 24° urine for catecholamines & metanephrines (see page 679). If elevated: abdominal MRI or MIBG scan (see page 679) • abdominal U/S starting at age 8 Every 2-3 years: complete audiology exam. Annually if hearing loss, tinnitus or vertigo
11-19 years	Every 6 months: retina exam† Annual: • PE (including scrotal exam in males), neuro exam • 24° urine for catecholamines & metanephrines (see page 679). If elevated: abdominal MRI or MIBG scan (see page 679) • abdominal U/S (kidneys, pancreas & adrenals). If abnormal: abdominal MRI or CT (except in pregnancy) Every 1-2 years or if symptoms develop: • gadolinium MRI of brain & spine. Annually at onset of puberty or before and after pregnancy (only for emergencies during pregnancy) • complete audiology exam. If abnormal, or if tinnitus or vertigo at any time: MRI of IAC to look for ELST
≥ 20 years	Annual: • dilated retina exam† • PE (including scrotal exam in males), neuro exam • blood test or 24° urine for catecholamines & metanephrines (see page 679). If elevated: abdominal MRI or MIBG scan (see page 679) • check kidneys, pancreas & adrenals with abdominal U/S and at least every other year unenhanced/enhanced abdominal CT (not during pregnancy) Every 2 years: • (or before and after pregnancy, except for emergencies) gadolinium MRI of brain & spine • complete audiology exam. If abnormal, or if tinnitus or vertigo at any time: MRI of IAC to look for ELST
Prior to surgery or childbirth	• blood test or 24° urine for catecholamines & metanephrines (see page 679) to rule out pheochromocytoma

* adapted⁴⁴¹

^† indirect ophthalmoscope exam by retinologist familiar with VHL

^‡ abbreviations: PE = physical exam by physician familiar with VHL, ELST = endolymphatic sac tumor

Treatment

Resection of individual CNS tumors is usually reserved until symptomatic to decrease the number of operations over a lifetime since the tumors in VHL are usually multiple, tend to recur, and the growth pattern is saltatory. Surgery is the treatment of choice for accessible cystic HGBs. For details, see page 672 under Hemangioblastoma.

Stereotactic radiosurgery (SRS)⁴³⁹ ): May provide local control rates of > 50% over 5 years. SRS has been recommended for asymptomatic HBG > 5 mm diameter if they are cystic or progressing in size during surveillance⁴⁴⁰. Cranial treatment plan: using a median dose of 22 Gy (range: 12-40Gy) prescribed to the median 82% isodose line in 1-4 sessions. In cystic lesions, treatment is confined to the contrast enhancing mural nodule (the cyst wall is not treated). Spinal treatment plan: median dose of 21 Gy (range 20-25 Gy) prescribed to the median 77% isodose line in 1-3 sessions. Radiosurgery is usually contraindicated in hemangioblastomas with a cyst.

Surveillance

Because of the lifetime risk of developing tumors, regular surveillance is needed. Various protocols have been proposed^{442, 443}, including those by the NIH⁴²³ and the Danish clinical recommendations⁴⁴⁴. The algorithm recommended by the VHL Family Alliance for patients with VHL and atrisk relatives^A is shown in Table 21-53.

Individuals who do not carry the altered gene on DNA testing do not require surveillance.

Prognosis

The lifespan of patients with VHL is decreased. 30-50% die of renal cell Ca (RCC). Metastases from RCC and neurologic complications from cerebellar HGB are the primary causes of death.

Metastases respond poorly to chemotherapy and XRT.

Resources

Genetic screening for VHL can be done at a few centers. Information for patients and families can be found at www.vhl.org/.

21.2.13.2. Hemangioblastomas (in general)

EPIDEMIOLOGY

HGB represent 1-2.5% of intracranial tumors. Comprise 7-12% of primary p-fossa tumors⁴⁴⁵. 5-30% of cases of cerebellar HGB and 80% of spinal HGB are associated with VHL (see above).

A. screening atrisk relatives can be stopped at age 60 years if no abnormalities have been detected

Sporadic cases tend to present in the 4th decade, whereas VHL cases present earlier (peak in 3rd decade). In sporadic cases, the HGB are solitary and originate in the cerebellum (83-95%), spinal cord (3-13%), medulla oblongata (2%)⁴²¹ or cerebrum (1.5%)⁴⁴⁵. ≈ 30% of patients with cerebellar HGB have VHL¹⁰.

PRESENTATION

S/S of cerebellar HGB are usually those of any p-fossa mass (H/A, N/V, cerebellar findings… see Posterior fossa (infratentorial) tumors, page 590) and obstructive hydrocephalus may occur. HGB is rarely documented as a cause of apoplexy due to intracerebral hemorrhage (ICH) (lobar or cerebellar), however, some studies indicate that if cases of ICH are carefully examined, abnormal vessels consistent with HGB (and occasionally misidentified as AVM) may be found with surprising frequency (in spite of negative CT and/or angiography)⁴⁴⁶.

Retinal HGBs tend to be located peripherally, and may hemorrhage and cause retinal detachment. Erythrocytosis may be due to erythropoietin liberated by the tumor.

PATHOLOGY

No report of malignant change. May spread thru CSF after surgery, but remain benign. No true capsule, but usually well circumscribed (narrow zone of infiltration). May be solid, or cystic with a mural nodule (70% of cerebellar lesions are cystic; nodules are very vascular, appear red, are often located near pial surface, and may be as small as 2 mm; cyst fluid is clear yellow with high protein). In cystic lesions, the cyst wall is lined with non-neoplastic compressed cerebellum. The cyst develops because the vessel walls are so thin that they leak water, proteins don’t cross as readily.

Cardinal feature: numerous capillary channels, lined by a single layer of endothelium, surrounded by reticulin fibers. Macrophages stain PAS positive.

Three types of cells:

1. endothelial

2. pericytes: surrounded by basement membrane

3. stromal: polygonal. Foamy clear cytoplasm, often lipid laden. Origin controversial

Three types of HGB recognized⁴⁴⁷:

1. juvenile: thin walled capillaries & dilated vessels tightly packed

2. transitional: thin walled capillaries & dilated vessels intermingled with stromal cells, some of which are lipid laden (sudanophilic)

3. clear cell: neoplasm made up almost entirely of sheets of xanthoma cells with a rich vascular stroma

Cyst patterns⁴²⁷:

1. no associated cysts: 28%

2. peritumoral cyst alone: 51%

3. intratumoral cyst: 17%

4. peritumoral AND intratumoral cysts: 4%

EVALUATION

Patients with a p-fossa HGB (radiologically suspected or histologically proven) should undergo MRI of entire neuraxis because of possibility of spinal HGBs (may be distant from p-fossa lesion; may suggest possibility of VHL).

CT: solid lesions are usually isodense with intense contrast enhancement. Cystic HGBs remain low density with contrast, with the nodule enhancing.

MRI: preferable to CT due to the tumor’s predilection for the p-fossa. May show serpentine vascular signal voids, especially in the periphery of the lesion. Also, peripheral hemosiderin deposits may occur from previous hemorrhages⁴⁴⁵.

Vertebral angiography: usually demonstrates intense vascularity (most other tumors of the p-fossa are relatively avascular). May be required in HGBs where nodule is too small to be imaged on CT/MRI. 4 patterns: 1) vascular mural nodule on side of avascular cyst, 2) vascular lesion surrounding avascular cyst, 3) solid vascular mass, & 4) multiple, separate vascular nodules.

Labs: often discloses polycythemia (no hematopoietic foci within tumor). In cases with suggestive history, labwork to rule-out catecholamine production from pheochromocytoma may be indicated (see Endocrine/laboratory studies, page 681).

TREATMENT

Surgery

Surgical treatment may be curative in cases of sporadic HGB, not in VHL.

Pre-operative embolization may help reduce the vascularity.

Cystic HGBs require removal of mural nodule (otherwise, cyst will recur). The cyst wall is not removed unless there is evidence of tumor within the cyst wall on MRI (typically thick-walled cysts) or visually at the time of surgery⁴²⁷. 5-ALA fluorescence may aid in visual localization of small hemangioblastomas within the cyst wall⁴⁴⁸.

Solid HGBs tend to be more difficult to remove. They are treated like AVMs (avoid piecemeal removal), working along margin and devascularizing blood supply. A helpful technique is to shrink the tumor by laying a length of bipolar forceps along tumor surface and coagulating. HGBs with attachment to floor of 4th ventricle may be hazardous to remove (cardiorespiratory complications).

Multiple lesions: if ≥ 0.8-1 cm diameter: may treat as in solitary lesion. Smaller and deeper lesions may be difficult to locate at time of surgery.

Radiation treatment

Effectiveness is dubious. May be useful to reduce tumor size or to retard growth, e.g. in patients who are not surgical candidates, for multiple small deep lesions, or for inoperable brainstem HGB. Does not prevent regrowth following subtotal excision.

Chemotherapy

At the time of this writing there is an ongoing phase II trial with sunitnib, an inhibitor of vascular endothelial growth factor and platelet-derived growth factor.

21.2.14. CNS lymphoma

Key concepts:

• may be primary or secondary (pathologically identical)

• suspected with homogeneously enhancing lesion(s) in the central gray matter or corpus callosum (on MRI or CT) especially in AIDS patients

• may present with multiple cranial-nerve palsies

• diagnosis highly likely if tumor seen in conjunction with uveitis

• very responsive initially to steroids (may produce “ghost tumors”)

• treatment: usually XRT ± chemotherapy. Role of neurosurgery usually limited to biopsy and/or placement of ventricular access reservoir for chemotherapy

CNS involvement with lymphoma may occur secondarily from a “systemic” lymphoma, or may arise primarily in the CNS. It is controversial whether most intracranial malignant lymphomas are primary⁴⁴⁹ or secondary⁴⁵⁰.

SECONDARY CNS LYMPHOMA

Non CNS lymphoma is the fifth most common cause of cancer deaths in the U.S., 63% of new cases are non-Hodgkin’s. Secondary CNS involvement usually occurs late in the course. Metastatic spread of systemic lymphoma to the cerebral parenchyma occurs in 1-7% of cases at autopsy⁴⁵¹.

PRIMARY CNS LYMPHOMA

A rare, malignant primary CNS neoplasm comprising 0.85-2% of all primary brain tumors and 0.2-2% of malignant lymphomas⁴⁵³. Occasionally metastasizes outside the CNS. Older names include: reticulum cell sarcoma and microglioma⁴⁵².

EPIDEMIOLOGY

The incidence of primary CNS lymphoma (PCNSL) is rising relative to other brain lesions, and will likely exceed that of low-grade astrocytomas and approach meningiomas. This is in part due to the occurrence of PCNSL in AIDS and transplant patients, but the incidence has also increased in the general population over the past 20 years⁴⁵⁴.

Male:female ratio = 1.5:1 (based on literature review⁴⁵⁵).

Median age at diagnosis: 52 yrs⁴⁵⁵ (younger among immunocompromised patients: ≈ 34 yrs).

Most common supratentorial locations: frontal lobes, then deep nuclei; periventricular also common. Infratentorially: cerebellum is the most common location.

Conditions with increased risk of primary CNS lymphomas (PCNSL)

1. collagen vascular disease

A. systemic lupus erythematosus

B. Sjögren’s syndrome: an autoimmune connective tissue disorder

C. rheumatoid arthritis

2. immunosuppression

A. chronic immunosuppression in transplantation patients⁴⁵⁶

B. severe-congenital immunodeficiency syndrome (“SCIDS”)

C. AIDS^{457, 458}: CNS lymphoma occurs in ≈ 10% of AIDS patients, and is the first presentation in 0.6%

D. possibly increased incidence in the elderly due to reduced competency of immune system

3. Epstein-Barr virus⁴⁵⁹ is associated with a broad spectrum of lymphoproliferative disorders, and is detectable in ≈ 30-50% of systemic lymphomas, however, it has been associated with almost 100% of PCNSL⁴⁶⁰, especially AIDS-related cases^{461 (p 317)}

PRESENTATION

Presentation is similar with primary or secondary CNS lymphoma: the two most common manifestations are those due to epidural spinal cord compression and those of carcinomatous meningitis (multiple cranial nerve deficits, see Carcinomatous meningitis, page 711). Seizures occur in up to 30% of patients⁴⁴⁹.

Symptoms

1. presents with non-focal non-specific symptoms in over 50% of patients; at time of presentation most commonly includes:

A. mental status changes in one third

B. symptoms of increased ICP (H/A, N/V)

C. generalized seizures in 9%

2. focal symptoms in 30-42% of cases:

A. hemimotor or hemisensory symptoms

B. partial seizures

C. multiple cranial-nerve palsies (due to carcinomatous meningitis)

3. combination of focal and non-focal symptoms

Signs

1. non-focal in 16%:

A. papilledema

B. encephalopathy

C. dementia

2. focal findings in 45% of cases:

A. hemimotor or hemisensory deficits

B. aphasia

C. visual field deficits

3. combination of focal and non-focal signs

Uncommon but characteristic syndromes

1. uveocyclitis, coincident with (in 6% of cases) or preceding the diagnosis of (in 11% of cases) lymphoma

2. subacute encephalitis with subependymal infiltration

3. MS-like illness with steroid-induced remission

PATHOLOGY

Characteristic sites: corpus callosum, basal ganglia, periventricular.

The neoplastic cells are identical to those of systemic lymphomas. Most are bulky tumors that are contiguous with the ventricles or meninges.

Histologic distinguishing features: tumor cells form cuffs around blood vessels which demonstrate multiplication of basement membranes (best demonstrated with silver reticulum stain).

Frozen section distorts the cells and may lead to a misdiagnosis of malignant glioma^{461 (p 320)}.

Immunohistochemical stains differentiates B-cell lymphomas from T-cell lymphomas (B-cell types are more common, especially in PCNSL and in AIDS).

EM shows absence of junctional complexes (desmosomes) that are usually present in epithelial derived tumors.

Intravascular lymphomatosis⁴⁶²: Formerly: (malignant) angioendothelomatosis. A rare lymphoma with no solid mass in which malignant lymphoid cells are found in the lumen of small blood vessels in affected organs. CNS involvement is reported in most cases. Presentation is non specific: patients are often febrile, and may present with progressive multifocal cerebrovascular events (including stroke or hemorrhage), spinal cord or nerve root symptoms (including cauda equina syndrome, see page 446), encephalopathy or peripheral or cranial neuropathies⁴⁶³. Initial transient cerebral symptoms may mimic TIAs or seizures. The ESR is often elevated prior to initiation of steroids. Lymphoma cells may be seen in the CSF.

Painful skin nodules or plaques occur in ≈ 10% of cases, generally involving the abdomen or lower extremities, and these cases may be diagnosed with skin biopsy. Otherwise, diagnosis often requires brain biopsy (open or stereotactic), in which involved areas on imaging studies are targeted. Pathology: malignant lymphoid cells distend and occlude small arteries, veins and capillaries with little or no parenchymal extension^{461 (p 324)}. Treatment with combination chemotherapy can result in long-term remission in some patients, but early diagnosis before permanent damage occurs is critical (diagnosis is rarely made pre-mortem).

DIAGNOSIS

On imaging (CT or MRI) 50-60% occur in one or more cerebral lobes (in grey or white matter). 25% occur in deep midline structures (septum pellucidum, basal ganglion, corpus callosum). 25% are infratentorial. 10-30% of patients have multiple lesions at the time of presentation. In contrast, systemic lymphomas that spread to the CNS tend to present with leptomeningeal involvement instead of parenchymal tumors⁴⁶⁴.

CT: Non-AIDS-related cases tend to enhance homogeneously, whereas AIDS-related cases often have a necrotic center and appear as multifocal ring-enhancing lesions⁴⁶⁵ (the wall is thicker than with an abscess).

Non-AIDS related cases: CNS lymphomas should be suspected with homogeneously enhancing lesion(s) in the central gray or corpus callosum. 75% are in contact with ependymal or meningeal surfaces (this together with dense enhancement may produce a “pseudomeningioma pattern”, however lymphomas lack calcifications and tend to be multiple).

60% are hyperdense to brain, only 10% are hypodense. Characteristically, > 90% of these tumors enhance; this is densely homogeneous in over 70%. As a result, when rare non-enhancing cases occur it often leads to a delay in diagnosis⁴⁶⁶. The appearance of enhanced PCNSL on CT has been likened to “fluffy cotton balls”. There may be surrounding edema⁴⁶⁷ and there is usually mass effect.

There is an almost diagnostic tendency of rapid partial to complete resolution on CT (and even at the time of surgery) following the administration of steroids, earning the nickname of “ghost-cell tumor”^468,469 or disappearing tumor.

MRI: No pathognomonic feature. May be difficult to discern if tumor is located subependymally (signal characteristics similar to CSF); proton-weighted image may avoid this pitfall. Nonenhancing lymphoma (on MRI or CT) is rare⁴⁷⁰ (some of these may enhance after XRT) but may be underreported. Bright on DWI (restricted diffusion), isointense to hypointense on ADC map.

CSF: Should only be obtained if no mass effect. Usually abnormal, but non-specific. Most common abnormalities are elevated protein (in > 80%), and increased cell count (in 40%). Cytology is positive for lymphoma cells (pre-operatively) in only 10% (sensitivity may be higher with leptomeningeal involvement as in non-AIDS patients than with parenchymal involvement commonly seen in AIDS). Repeating up to 3 LPs may increase yield.

Angiography: Rarely helpful. 60% of cases show only an avascular mass. 30-40% show diffuse homogeneous staining or blush.

EVALUATION

All patients should be assessed (history, physical, and if appropriate, laboratory tests) for any of the conditions associated with lymphoma (see page 673). Since primary CNS lymphoma is very rare, any patient with CNS lymphoma should have work-up for occult systemic lymphoma including:

1. careful physical exam of all lymph nodes (LN)

2. evaluation of perihilar and pelvic LN (CXR, CT of chest & abdomen)

3. routine blood and urine testing

4. bone marrow biopsy

5. MRI of the entire spine

6. testicular ultrasound in males

7. ophthalmologic examination (including slit-lamp evaluation of both eyes) in all

A. for possible uveitis

B. ≈ 28% of patients with PCNSL will also have intraocular lymphoma. Often resistant to methotrexate, but responds to low dose ocular XRT (7-8 Gy)

TREATMENT

Surgery

Surgical decompression with partial or gross total removal does not alter patient’s prognosis. The main role for surgery is for tumor biopsy, and stereotactic techniques are often well-suited for these often deep tumors⁴⁷¹.

Radiation therapy

The standard treatment after tissue biopsy is whole-brain radiation therapy. Doses used tend to be lower than for other primary brain tumors. ≈ 40-50 Gy total are usually given in 1.8-3 Gy daily fractions.

Chemotherapy

In non-AIDS cases: survival with chemotherapy + XRT is greater than XRT alone⁴⁷².

Methotrexate (MTX): The addition of intraventricular MTX (rather than just intrathecal via LP) delivered through a ventricular access device (6 doses of 12 mg twice a week, with IV leucovorin rescue) may result in even better survival⁴⁷³. In the event of an intrathecal MTX overdose (OD), interventions recommended⁴⁷⁴: ODs of up to 85 mg can be well tolerated with little sequelae; immediate LP with drainage of CSF can remove a substantial portion of the drug (removing 15 ml of CSF can eliminate ≈ 20-30% of the MTX within 2 hrs of OD). This can be followed by ventriculolumbar perfusion over several hours using 240 ml of warmed isotonic preservative-free saline entering through the ventricular reservoir and exiting through a lumbar subarachnoid catheter. For major OD of > 500 mg, add intrathecal administration of 2,000 U of carboxypeptidase G₂ (an enzyme that inactivates MTX). In cases of MTX OD, systemic toxicity should be prevented by treating with IV dexamethasone and IV (not IT) leucovorin.

Rituximab: Available since 1997 for treatment of refractory systemic B-cell non-Hodgkins lymphoma. Intrathecally, may be more effective for CD33+ lymphomas.

PROGNOSIS

With no treatment, median survival is 1.8-3.3 months following diagnosis.

With radiation therapy⁴⁴⁹, median survival is 10 months, with 47% 1-year median survival, and 16% 2-year median survival. 3-year survival is 8%, and 5-year survival is 3-4%. With intraventricular MTX, median time to recurrence was 41 mos⁴⁷³. Occasionally, prolonged survival may be seen⁴⁷⁵.

About 78% of cases recur, usually ≈ 15 months after treatment (late recurrences also are seen). Of these recurrences, 93% are confined to the CNS (often at another site if the original site responded well), and 7% are elsewhere.

In AIDS-related cases, the prognosis appears worse. Although complete remission occurs in 20-50% following XRT, the median survival is only 3-5 months^{476, 477}, usually related to AIDS-related opportunistic infection. However, neurologic function and quality of life improve in ≈ 75%⁴⁷⁶.

Although there are individual studies that show trends, there are no prognostic features that consistently correlate with survival.

21.2.15. Chordoma

Key concepts:

• primary malignant tumor, usually of clivus or sacrum, with high recurrence rate

• histology: characteristic physaliphorous cells (containing intracellular mucin)

• generally slow-growing and radioresistant

• treatment of choice: wide en bloc resection when possible (piecemeal removal carries risk of inducing metastases), proton-beam radiation may help

Rare tumors (incidence of ≈ 0.51 cases/million) of the remnant of the primitive noto-chord (which normally differentiates into the nucleus pulposus of the intervertebral disks). Can arise anywhere along the neuraxis where there is remnant of notochord, however, cases tend to cluster at the two ends of the primitive notochord: 35% cranially⁴⁷⁸ in the spheno-occipital region (clivus), and 53%⁴⁷⁸ in the spine at the sacrococcygeal region⁴⁷⁹. Less commonly, they may occur in the spine above the sacrum⁴⁸⁰. The metastatic rate is low (5-20%)⁴⁸¹, but there is a high recurrence rate of 85% following surgery, and therefore aggressive RTX is usually employed post-op.

PATHOLOGY

Histologically, these tumors are considered low-grade malignancies. However, their behavior is more malignant because of the difficulty of total removal, a high recurrence rate, and the fact that they can metastasize (usually late). They are slow growing, locally aggressive and osteodestructive. Metastases occur in about 10% of sacral tumors, usually late and after multiple resections, and most often to lung, liver and bone. Malignant transformation into fibrosarcoma or malignant fibrous histiocytoma is rare. Physaliphorous cells are distinctive, vacuolated cells on histology that probably represent cytoplasmic mucus vacuoles seen ultrastructurally.

RADIOGRAPHIC APPEARANCE

Usually lytic with frequent calcifications⁴⁸². Enhances on CT with contrast⁴⁸². Rarely, may appear as a sclerotic vertebra⁴⁸³ (“ivory vertebra”).

CRANIAL CHORDOMAS

Peak incidence of cranial chordomas is 50-60 years of age. These tumors are rare in patients < 30 years of age⁴⁸⁴. Male:female distribution is ≈ equal.

Differential diagnosis: Primarily between other cartilaginous tumors of the skull base (for differential diagnosis of other foramen magnum region tumors, see page 1212):

1. chondrosarcomas

2. chondromas

Presentation: Usually produces cranial nerve palsies (usually oculomotor or abducens).

SPINAL CHORDOMAS

Occur primarily in the sacrococcygeal region. Unlike cranial chordomas, sacrococcygeal chordomas show a male predominance⁴⁷⁸, and these patients tend to be older. May also arise in C2. Chordomas constitute over 50% of primary bone tumors of the sacrum. May produce pain, sphincter disturbance or nerve root symptoms from local nerve root compression. It may occasionally extend cephalad into the lumbar spinal canal. It is usually confined anteriorly by the presacral fascia, and only rarely invades the wall of the rectum⁴⁸⁵. A firm fixed mass may be palpable between the rectum and the sacrum on rectal exam.

EVALUATION

Characteristic radiographic findings: centrally located destruction of several sacral segments, with an anterior soft-tissue mass that occasionally has small calcifications. CT and MRI show the bony destruction. This is usually difficult to see on plain x-rays. MRI also shows the soft-tissue mass.

Open or CT guided percutaneous posterior biopsy can confirm the diagnosis. Transrectal biopsy should be avoided because of the potential of rectal spread⁴⁸⁶.

Chest CT and bone scan: to R/O mets for staging purposes.

TREATMENT

Surgery

Wide en-bloc excision with postoperative radiation is usually the best option, although this may also be only temporarily effective. Decompression is best avoided since entering the mass serves to spread tumor (surgically induced metastases). Chordomas located in C2 are usually not amenable to en bloc resection⁴⁸⁷.

Sacral chordomas: The particulars of the surgical procedure are highly dependent on the extent of the lesion. These tumors may spread through the gluteal musculature, and if significant muscular excision is required, then a pedicle based rectus abdominis flap may be employed. A diverting colostomy may be required if it is necessary to resect the rectum or if a cephalic sacral resection is anticipated⁴⁸⁸.

For chordomas caudal to the third sacral segment, most agree that a posterior approach is satisfactory. For more rostral lesions, some advocate a combined anterior-posterior approach. However, a posterior approach has been also been used for these⁴⁸⁸.

Adverse effects of sacrectomy: if S2 nerve roots are the most caudal nerve roots spared, there is ≈ 50% chance of normal bladder and bowel control⁴⁸⁸. If S1 or more cephalic roots are the most caudad nerve roots spared, most will have impaired bladder control and bowel problems⁴⁸⁸.

Radiation therapy (XRT)

Best results were obtained with en bloc excision (even if marginal), sometimes combined with high-dose XRT^{480, 489} (conventional XRT did not prevent recurrence when incorporated with palliative or debulking surgery⁴⁸⁰), but it did lengthen the interval to recurrence⁴⁸⁹). Early radiation was associated with longer survival⁴⁹⁰. Higher XRT doses can be used in the sacrococcygeal region (4500-8000 rads) than in the cervical spine (4500-5500 rads) because of concerns of radiation injury to the spinal cord. IMRT and stereotactic radiosurgery have also been used⁴⁸⁷.

Proton beam therapy, alone⁴⁸¹ or combined with high-energy x-ray (photon) therapy^{491, 492} may be more effective than conventional XRT alone. However, proton beam therapy requires travel to one of a very limited number of facilities with a cyclotron (in the U.S.: Boston, or Loma Linda, California) which may be difficult to arrange for what is typically ≈ 7 weeks of fractionated treatments.

Chemotherapy

Imatinib (Gleevec®) (a tyrosine kinase inhibitor) has some antitumor effect in chordoma⁴⁹³.

Outcome

Median survival is 6.3 years⁴⁸⁷.

21.2.16. Ganglioglioma

Key concepts:

• composed of two cell types: ganglion cells (neurons) and glial cells

• extremely rare (< 2% of intracranial neoplasms)

• seen primarily in the first 3 decades of life

• characterized by slow growth and a tendency to calcify

A tumor composed of two types of cells: ganglion cells (neurons) which may arise from primitive neuroblasts, and glial cells, usually astrocytic in any phase of differentiation⁴⁹⁴.

EPIDEMIOLOGY

Location: May occur in various parts of the nervous system (cerebral hemispheres, spinal cord, brainstem, cerebellum, pineal region, thalamus, intrasellar, optic nerve, and peripheral nerve have been reported⁴⁹⁵). Most occur above the tentorium, primarily in or near the 3rd ventricle, in the hypothalamus or in the temporal or frontal lobes¹⁴². Brain-stem gangliogliomas occur rarely (see page 607).

Incidence: Typically quoted⁴⁹⁵ as 0.3-0.6%. One series⁴⁹⁶ found gangliogliomas in 1.3% of all brain tumors (including mets), or 3% of primary brain tumors. Considering only children and young adults, incidence ranges from 1.2-7.6% of brain tumors⁴⁹⁵.

Demographics: Occurs primarily in children and young adults (peak age of occurrence: 11 yrs).

PRESENTATION

Most common presenting symptom was seizure, or a change in a pre-existing seizure pattern. Often, the seizures are difficult to control medically.

RADIOLOGIC EVALUATION

Neuroradiologic findings are not specific for this tumor.

Plain skull x-ray: calcification was noted in 2 of 6 patients⁴⁹⁵.

CT: all of 10 patients had a low density lesion on non-contrast CT; 8 enhanced slightly with contrast; 5 of the 10 had calcification on CT⁴⁹⁶. 6 of the 10 were in temporal lobe (this predilection has been noted in many but not all series), and 4 were in frontal lobe. Frequently appears cystic on CT, but still may be found to be solid at operation. Mass effect rare (suggests slow growth).

MRI: high signal on T1WI, low signal on T2WI. Calcifications appear as low signal on both⁴⁹⁵.

Angiography: shows either an avascular or a minimally vascular mass.

PATHOLOGY

Mixture of 2 types of neoplastic cells: neuronal (ganglion) and astrocytic (glial). Very slow growing.

Two major classifications: ganglioneuromas (less common, more benign; predominance of neuronal component) and gangliogliomas (preponderance of glial cells).

Grossly: white matter mass; well-circumscribed, firm, with occasional cystic areas and calcified regions. Most dissect easily from brain, but the solid portion may show an infiltrative tendency⁴⁹⁵.

Microscopically: ganglion cells must demonstrate nerve cell differentiation, e.g. Nissl substance and axons or dendrites. Pitfall: differentiating neoplastic neurons from neurons entrapped by an invading astrocytoma may be difficult. Also, neoplastic astrocytes may resemble neurons on light microscopy. 2 of 10 patients had areas of oligodendroglioma. One series found necrotic areas in 7 of 14 patients, minimal calcification, and Rosenthal bodies⁴⁹⁷. Suggested criteria for diagnosis⁴⁹⁸:

1. clusters of large cells potentially representing neurons (required for diagnosis)

2. no perineural clustering of glial cells around the suspected neoplastic neurons

3. fibrosis (desmoplasia)

4. calcification

Aggressive malignant changes in the glial component may dictate a poor outcome, although an “aggressive” background is not unusual and may not indicate malignancy.

TREATMENT

Recommendation is wide radical excision when possible (may be more limited in spinal cord and brainstem tumors). Close follow-up is recommended, and re-resection should be considered for recurrence. The role of XRT is unknown, and due to the deleterious effects together with the good long-term prognosis, it is not recommended initially but may be considered for recurrence⁴⁹⁹.

PROGNOSIS

Russell and Rubinstein⁵⁰⁰ first proposed that the grade of the astrocytic component of the tumor determines the prognosis. This has been supported by some case reports, but clinical series have not been able to correlate histology with outcome⁴⁹⁹. Thus, anaplasia is not significantly associated with a worse prognosis⁴⁹⁹.

The majority of patients did well and were asymptomatic after resection. 1 patient in a series of 10 died 3 days post-op from cerebral edema.

In 58 patients, 5-year survival was 89% and 10-year survival was 84%⁴⁹⁹. In 9 brainstem gangliogliomas, 5-year survival was 78%.

The value of radiation therapy is not known. Consider radiation when growth is evident on follow-up CT, or when infiltration is felt to occur at time of surgery.

1 patient had degeneration to glioblastoma when a recurrence was discovered 5 years after removal (this patient received radiation therapy).

The prognosis with following subtotal resection of brainstem gangliogliomas is better than for brainstem gliomas as a group¹⁴².

21.2.17. Paraganglioma

AKA chemodectoma, AKA glomus tumors. Table 21-54 shows the designation of these tumors in various sites.

These tumors arise from paraganglion cells (not chemoreceptor cells as previously thought, therefore the term chemodectoma is losing favor). Slow growing tumors (< 2 cm in 5 years).

Histologically benign (< 10% associated with lymph node involvement or distant spread). Most contain secretory granules on EM (mostly epinephrine & nor-epinephrine, and these tumors may occasionally secrete these catecholamines with risk of life-threatening HTN and/or cardiac arrhythmias).

Table 21-54 Designation based on site of origin

Site	Designation
carotid bifurcation (most common)	carotid body tumors
auricular branch of vagus (middle ear)	glomus tympanicum
superior vagal ganglion (jugular foramen)	glomus jugulare
inferior vagal (nodose) ganglion (nasopharynx at skull base) (least common)	glomus intravagale (AKA glomus vagale)
adrenal medulla & sympathetic chain	pheochromocytoma

Glomus tumors may occur in 2 patterns:

1. familial: non multicentric. Up to 50%

2. nonfamilial. may be multicentric (metachronous) 5%

PHEOCHROMOCYTOMA

Located in the adrenal gland. May be sporadic, or as part of familial syndrome (von Hippel-Lindau disease - see page 667, MEN 2A & 2B, & neurofibromatosis). Consider genetic testing if age at diagnosis is < 50 years for mutations of VHL and other genetic abnormalities.

Laboratory studies

1. fractionated plasma metanephrines: 96% sensitivity, 85% specificity⁵⁰¹. More sensitive than serum catecholamines with sporadic elevations. Pheochromocytoma is ruled out if plasma normetanephrine (NMN) < 112 pg/ml and metanephrine (MN) < 61 pg/ml. Highly suspicious if NMN > 400 pg/ml or MN > 236 pg/ml

2. 24 hr urine collection for: total catecholamines (epinephrine and nor-epinephrine) and metanephrines (88% sensitivity, 99.7% specificity⁵⁰²)

3. where elevation is found, a clonidine suppression test can be done. Normal response consists of a fall in plasma catecholamines to ≤ 50% of baseline and below 500 pg/ml (there will be a reduction in essential hypertension, but no change with pheochromocytoma or other tumor production)

Imaging

Indicated when laboratory tests confirm pheochromocytoma.

MRI with contrast is preferred over CT.

CT may be used when MRI is contraindicated, but is less sensitive, especially for lesions < 1 cm diameter.

123I MIBG (iodine-123-meta-iodobenzylguanidine) scintigraphy detects extra-adrenal pheochromocytomas with 83-100% sensitivity, 95-100% specificity. If not available 131I MIBG may be used with 77-90% sensitivity, 95-100% specificity.

CAROTID BODY TUMORS

Possibly the most common paraganglioma (pheochromocytoma may be more common). Approximately 5% are bilateral; the incidence of bilaterality increases to 26% in familial cases (these are probably autosomal dominant).

CLINICAL

Usually present as painless, slow growing mass in upper neck. Large tumors may → cranial nerve involvement (especially vagus and hypoglossal). May also cause stenosis of ICA → TIAs or stroke.

EVALUATION

1. carotid angiogram: demonstrates predominant blood supply (usually external carotid, with possible contributions from vertebral and thyrocervical trunk). May also detect bilateral lesions. Characteristic finding: splaying of bifurcation

2. MRI (or CT): evaluates extent, and assesses for intracranial extension

TREATMENT

Resection reported to carry a high complication rate, including stroke (8-20%) and cranial nerve injury (33-44%). Mortality rate is 5-13%.

GLOMUS TUMORS

Glomus tumors may be subdivided into glomus jugulare and glomus tympanicum tumors. Glomus jugulare tumors arise from the jugular bulb (in the jugular foramen at the junction of the sigmoid sinus and jugular vein). Glomus tympanicum tumors are centered higher than glomus jugulare. Glomus tumors are rare (0.6% of all head and neck tumors), yet the glomus tympanicum is the most common neoplasm of the middle ear. Glomus jugulare tumors (GJT)arise from glomus bodies, usually in the area of the jugular bulb, and track along vessels. May have finger-like extension into the jugular vein (which may embolize during resection)⁵⁰³. Most are slow growing, although rapidly growing tumors do occur.

Vascular supply: very vascular. Main feeders of GJT are from the external carotid (especially inferior tympanic branch of ascending pharyngeal artery, and branches of posterior auricular, occipital, and internal maxillary), with additional feeders from petrous portion of the ICA. Glomus tympanicum tumors feed from the auricular artery.

CLINICAL

Epidemiology

Female:male ratio is 6:1. Bilateral occurrence is almost nonexistent.

Symptoms

Patients commonly present with hearing loss and pulsatile tinnitus. Dizziness is the third most common symptoms. Ear pain may also occur.

Signs

Hearing loss may be conductive (e.g. due to obstruction of the ear canal) or sensorineural due to invasion of the labyrinth often with accompanying vertigo (the eighth nerve is the most common cranial nerve involved). Various combinations of palsies of cranial nerves IX, X, XI & XII occur (see Jugular foramen syndromes, page 115) with occasional VII palsy (usually from involvement within the temporal bone). Ataxia and/or hydrocephalus can occur with massive lesions that cause brainstem compression. Occasionally patients may present with symptoms due to secretory products (see below).

Otoscopic exam → pulsatile reddish-blue mass behind eardrum (occasionally, lamentably biopsied by ENT physician with possible ensuing massive blood loss).

PATHOLOGY

Histologically indistinguishable from carotid body tumors. May invade locally, both through temporal bone destruction and especially along pre-existing pathways (along vessels, eustachian tube, jugular vein, carotid artery). Intradural extension is rare. Malignancy may occur, but is rare. These tumors rarely metastasize.

Secretory properties

These tumors usually possess secretory granules (even the functionally inactive tumors) and may actively secrete catecholamines (similar to pheochromocytomas, occurs in only 1-4% of GJT⁵⁰⁴). Norepinephrine will be elevated in functionally active tumors since glomus tumors lack the methyltransferase needed to convert this to epinephrine. Alternatively, serotonin and kallikrein may be released, and may produce a carcinoid-like syndrome (bronchoconstriction, abdominal pain and explosive diarrhea, violent H/A, cutaneous flushing, hypertension, hepatomegaly and hyperglycemia)⁵⁰⁵. During surgical manipulation, these tumors may also release histamine and bradykinin, causing hypotension and bronchoconstriction⁵⁰⁶.

DIFFERENTIAL DIAGNOSIS

See Cerebellopontine angle (CPA) lesions on page 1210. The major differential is neurilemmomas (vestibular schwannomas), both enhance on CT. A cystic component and extrinsic compression of the jugular bulb are characteristic of neurilemmomas. Angiography will differentiate difficult cases.

EVALUATION

Neurophysiologic testing

Audiometric and vestibular testing should be performed.

Imaging

1. CT or MRI used to delineate location and extent of tumor; CT is better for assessing bony involvement of the skull base

2. angiography: confirms diagnosis (helping to rule out vestibular schwannoma), and ascertains patency of contralateral jugular vein in event that jugular on side of tumor must be sacrificed; jugular bulb and/or vein are usually partially or completely occluded

Endocrine/laboratory studies

(see page 679)

CLASSIFICATION

A number of classification schemes have been proposed. The modified Jackson classification is shown in Table 21-55.

Table 21-55 Modified Jackson classification⁵⁰⁷

Type	Description	Intracranial extension
I	small; involves jugular bulb, middle ear & mastoid	none
II	extends under IAC	possible
III	extends into petrous apex	possible
IV	extends beyond petrous apex into clivus or infratemporal fossa	possible

TREATMENT

Surgical resection is usually simple and effective for small tumors confined to the middle ear. For larger tumors that invade and destroy bone, the relative role of surgery and/or radiation is not fully determined. With large tumors, surgery carries the risk of significant cranial nerve palsies.

MEDICAL

For tumors that actively secrete catecholamines, medical therapy is useful for palliation or as adjunctive treatment before embolization or surgery. Alpha and beta blockers given before embolization or surgery blocks possibly lethal blood pressure lability and arrhythmias. Adequate blockade takes ≈ 2-3 weeks of alpha blocker and at least 24 hours of beta blocker therapy; in emergency, 3 days of treatment may suffice.

Alpha blockers

Reduce BP by preventing peripheral vasoconstriction.

• phenoxybenzamine (Dibenzyline®): long acting; peak effect 1-2 hrs. Start with 10 mg PO BID and gradually increase to 40-100 mg per day divided BID

• phentolamine (Regitine®): short acting. Usually used IV for hypertensive crisis during surgery or embolization.

Rx: 5 mg IV/IM (peds: 1 mg) 1-2 hrs pre-op, repeat PRN before and during surgery

Beta blockers

Reduces catecholamine induced tachycardia and arrhythmias (may also prevent hypotension that might occur if only alpha blockade is used). These drugs are not always needed, but when used ✖ NB: these drugs must not be started before starting alpha-blockers (to prevent hypertensive crisis and myocardial ischemia).

• propranolol (Inderal®): Rx: oral dose is 5-10 mg q 6 hrs. IV dose for use during surgery is 0.5-2 mg slow IVP

• labetalol (Normodyne®): may have some efficacy in blocking α₁ selective and ß non-selective (potency < propranolol), see page 20

Serotonin, bradykinin, histamine release blockers

These agents may provoke bronchoconstriction that does not respond to steroids, but may respond to inhaled ß-agonists or inhaled anticholinergics. Somatostatin may be used to inhibit release of serotonin, bradykinin, or histamines. Since this drug has a short half-life, it is preferable to give octreotide 100 μg sub-Q q 8 hrs (see page 653).

RADIATION THERAPY

XRT may relieve symptoms and stop growth in spite of persistence of tumor mass. 40-45 Gy in fractions of 2 Gy has been recommended⁵⁰⁸. Lower doses of ≈ 35 Gy in 15 fractions of 2.35 Gy appear as effective and have fewer side effects⁵⁰⁹. Generally used as primary treatment only for large tumors or in patients too elderly or infirmed to undergo surgery. Some surgeons pretreat 4-6 mos pre-operatively with XRT to decrease vascularity⁵¹⁰ (controversial).

EMBOLIZATION

• generally reserved for large tumors with favorable blood supply (i.e. vessels that can be selectively embolized with no danger of particles passing thru to normal brain)

• post-embolization tumor swelling may compress brainstem or cerebellum

• may be used preoperatively to reduce vascularity. Performed 24-48 hours pre-op (not used prior to that, because of post-embolization edema)

• caution with actively secreting tumors which may release vasoactive substances (e.g. epinephrine) upon infarction from the embolization

• may also be used as primary treatment (± radiation) in patients who are not surgical candidates. In this case, is only palliative, as tumor will develop new blood supply

• absorbable (Gelfoam®) and non-absorbable (Ivalon®) materials have been used

SURGICAL TREATMENT

The tumor is primarily extradural, with extremely vascular surrounding dura.

Suboccipital approach may cause dangerous bleeding and usually results in incomplete resection. Team approach by a neurosurgeon in conjunction with a neuro-otologist and possibly head and neck surgeon has been advocated³⁷⁷. This approach utilizes an approach to the skull base through the neck.

ECA feeders are ligated early, followed rapidly by draining veins (to prevent systemic release of catecholamines).

Sacrifice of the jugular vein (JV) is tolerated if the contralateral JV is patent (often, the ipsilateral JV will already be occluded).

Complications and outcome

The most common complications are CSF fistula, facial nerve palsy, and varying degrees of dysphagia (from dysfunction of lower cranial nerves). Dysfunction of any of the cranial nerves VII thru XII can occur, and a tracheostomy should be performed if there is any doubt of lower nerve function, and a gastrostomy feeding tube may be needed temporarily or permanently. Lower cranial nerve dysfunction also predisposes to aspiration, the risk of which is also increased by impaired gastric emptying and ileus that may occur due to reduced cholecystokinin (CCK) levels post-op. Excessive blood loss can also occur.

Even after gross total tumor removal, recurrence rate may be as high as one third^{510, 511}.

21.2.18. Ependymoma

Ependymomas arise from ependymal cells lining the cerebral ventricles and the central canal of the spinal cord. They may occur anywhere along the neuraxis, in pediatrics they are most common in the posterior fossa (see below), in adults they tend to be intraspinal (see page 730).

Epidemiology:

• intracranial: comprises only ≈ 5-6% of intracranial gliomas, 69% occur in children⁵¹², comprise 9% of pediatric brain tumors⁵¹³. Incidence of pediatric intracranial ependymomas: ≈ 200 cases/yr in the U.S.

• spinal: ≈ 60% of spinal cord gliomas, 96% occur in adults⁵¹², especially those of filum terminale (see myxopapillary ependymoma below)

Ependymomas have the potential to spread via the CSF through the neuraxis, a process known as “seeding”, resulting in so-called “drop mets” in 11%. The incidence is higher with higher grade⁵¹³. Systemic spread occurs on rare occasion.

The mean age at diagnosis is shown in Table 21-56.

Table 21-56 Mean age at diagnosis of ependymoma⁵¹²

Location (in 101 patients)	All patients (yrs)	Children (yrs) (age < 15 yrs)
intracranial	17.5	5
infratentorial	14.5	4.5
supratentorial	22	6.5
intraspinal	40
intramedullary	47
cauda region	32

PATHOLOGY

Although they are usually circumscribed with a covering layer of ependyma, ependymomas may be invasive.

Classification is a work in progress. Ependymomas from different locations (pfossa, supratentorial, spinal cord) are genetically distinct⁵¹⁴. The World Health Organization (WHO) classification of ependymal tumors:

1. ependymoma (WHO II) - variants:

A. cellular

B. papillary: “classic lesion” occurring in brain or spinal cord. Can metastasize in up to 30% of cases. Dark, small nuclei. 2 cytoplasmic patterns:

1. differentiation along glial line: these form perivascular pseudorosettes (areas of radiating processes lacking nuclei surrounding blood vessels) which, when they occur, are diagnostic

2. cuboidal cells: these form true rosettes (ependymal tubules around a central blood vessel)

C. clear cell

D. tanycytic: rare. Tumor cells appear similar to “ependymoglia” or “tanycytes” (stretched cells present to a limited degree in the normal CNS). True rosettes are absent. No preference for age, sex or location within CNS2. Treatment of choice: gross total resection²

2. myxopapillary ependymoma: (WHO I) distinctive, occurs only in filum terminale. Papillary, with microcystic vacuoles and mucosubstance (see page 731)

3. subependymomas: (WHO I) typically occur in anterior lateral ventricles or posterior fourth ventricle, with prominent role of subependymal glial cells. Classically do not enhance (see page 1226). Not uncommon at autopsy, rarely surgical

4. anaplastic ependymomas: (WHO III) pleomorphism, multinucleation, giant cells, mitotic figures, vascular changes and areas of necrosis (the term ependymoblastoma has occasionally been used for more anaplastic lesions, but this term is best reserved for a distinct, rare childhood primitive neuroectodermal tumor, see page 688). It is unclear if the degree of anaplasia has any effect on outcome

INTRACRANIAL EPENDYMOMAS

Key concepts:

• usually benign tumors, often fibrillary with epithelial appearance. Perivascular pseudorosettes or true rosettes may be seen in classic (papillary) form

• most often occur in the floor of the 4th ventricle, presenting with hydrocephalus (increased ICP) and cranial nerve VI & VII palsies

• evaluation: includes imaging the entire neuraxis (usually with enhanced MRI: cervical, thoracic, lumbar & brain) because of potential for seeding through CSF

• worse prognosis the younger the patient (especially age < 24 months)

• treatment: the best outcomes are associated with gross total removal (no enhancing tumor on post-op MRI) followed by XRT. XRT may be withheld for age < 3

• do LP ≈ 2 weeks post op to send ≈ 10 cc of CSF for cytology for prognostication

Usually well circumscribed and benign (although anaplastic (malignant) ependymomas do occur), commonly arises in the floor of the fourth ventricle (60-70% are infratentorial, all of these occur near 4th ventricle⁵¹², they comprise 25% of tumors in region of 4th ventricle^{515 (p 2792)}). Children with p-fossa ependymomas often have anaplastic tumors with a higher risk of spread through the neuraxis. Supratentorial ependymomas are often cystic. Rarely occur outside the CNS in: mediastinum, lung or ovaries. Although not as malignant histologically as medulloblastomas, ependymomas have a worse prognosis due to their propensity to invade the obex which precludes complete removal.

CLINICAL

Symptoms

Mostly those of posterior fossa mass with increased ICP^{515 (p 2795)} (from hydrocephalus) and cranial nerve involvement.

Symptoms of increased ICP:

1. headache: 80%

2. N/V: 75%

3. ataxia or vertigo: 60%

4. seizures: only in ≈ 30% of supratentorial lesions; comprise only 1% of patients with intracranial tumors presenting with seizures

Cranial nerve involvement: invasion of the floor of the 4th ventricle may involve the facial colliculus producing facial nerve palsy (involvement of internal genu of VII, see page 844) and abducens palsy (from VI nucleus).

EVALUATION

MRI: imaging study of choice. Image the entire craniospinal axis with and without contrast because of possibility of drop mets. Usually appears as a mass in the floor of fourth ventricle, often with obstructive hydrocephalus. May be difficult to distinguish from medulloblastoma (MBS) radiographically, see page 1210 for differentiating features.

CT: not as detailed for evaluation of posterior fossa.

Myelogram: water-soluble contrast myelography is about as sensitive as gadolinium enhanced MRI in detecting “drop mets”. Myelography also provides CSF for cytology for staging.

MICROSCOPIC FEATURES

See Pathology on page 683.

TREATMENT

Surgical resection

Goal of surgery: maximal possible resection of intracranial portion without causing neurological deficits. Gross total resection may not possible when invasion of the floor is extensive, or when tumor extends through the foramen of Lushka (bradycardia may prevent GTR).

2 weeks postoperatively, perform LP to look for “drop mets”: 10 cc of CSF is sent for cytology to quantitate (if any) number of malignant cells (may be used to follow treatment). If LP is positive, then by definition there are drop mets. If negative, it is not as helpful (sensitivity is not high). CSF from an EVD is not as sensitive as LP.

Lesions in fourth ventricle region are approached via midline suboccipital craniectomy.

Radiation therapy (XRT)

Ependymomas rank 2nd only to medulloblastomas in radiosensitivity. XRT is administered after surgical excision (survival is improved with post-op XRT: 50% survival time was 2 yrs longer with XRT than without⁵¹², and 5-year survival increased from 20-40% without XRT to 40-80% with XRT⁵¹⁶), however, for patients age < 3 years, see below.

1. cranial XRT

A. traditional therapy: 45-48 Gy to tumor bed⁵¹⁶ (recurrence treated with additional 15-20 Gy)^{515 (p 2797)}

B. recent recommendations: 3-D conformal XRT with higher doses (59.4 Gy delivered to tumor bed + 1 cm margins)⁵¹⁷

C. intensity modulated proton beam therapy appears equivalent in terms of local control, but may be better at sparing normal tissue⁵¹⁸

2. spinal XRT: most radiate only if drop mets or if positive CSF cytology (however, prophylactic spinal is controversial⁵¹⁹)

A. low dose XRT to entire spinal axis (median dose = 30 Gy⁵¹⁶)

B. boost to any regions showing drop mets

3. XRT is undesirable in age < 3 due to side effects. XRT was avoided in ≈ 30% of patients < 3 years age with comparable survival when XRT was reserved for treatment failures^{520, 521}. This concept of selective XRT may be applicable to older children as well⁵²²

Chemotherapy

Role is very limited.

1. has little impact on newly diagnosed cases. Adjuvant chemo after XRT in patients > 3 years showed no benefit

2. may reduce vascularity of ependymomas which may facilitate GTR (sometimes in a second stage operation)

3. may be considered for infants < ≈ 3 years age to delay use of XRT (see above)

4. chemo at the time of recurrence may arrest tumor progression for short periods

OUTCOME

Operative mortality^{515 (p 2797)}: 20-50% in early series; more recently: 5-8%.

Operative morbidity: advise patients/families pre-op of the likelihood need for post-op gastric feeding tube (G-tube) and tracheostomy (these may be temporary).

Age: peds vs. adults: 5-year survival is 20-30% in the pediatric group^{513, 523}, compared with up to 80% in adults. Patients 24-35 months old did better (5-year survival = 73%) than those younger than 24 months (26% 5-YS) or those older than 36 months (36% 5-YS)⁵²⁴.

Pathology: prognosis is worse with anaplastic ependymoma (WHO III) than with “standard” grade (WHO II)^{525, 526}. However, excluding WHO III tumors, malignant features in an ependymoma do not necessarily portend a worse prognosis⁵²⁷.

Extent of resection: the risk of recurrence is highest following subtotal resection. Gross total resection (GTR) (surgical) of primary intracranial tumor followed by craniospinal XRT as outlined above yields 41% 5-year survival.

Treatment failure: WHO Grade II tumors tend to recur initially at the site of origin⁵²⁵. However, primary failure in 9-25% of patients is via drop mets^{524, 528}.

SPINAL EPENDYMOMAS

The most common spinal cord glioma below the mid-thoracic region.

See Intramedullary spinal cord tumors on page 730.

21.2.19. Embryonal tumors

A few words about PNETs

Initially, the term primitive neuroectodermal tumor (PNET) encompassed a wide variety of previously individually named tumors which all seemed to share certain pathologic features suggesting origin from a common progenitor cell in the subependymal matrix (primitive neuroectodermal cells) (although the actual cell of origin is unknown). They are histologically indistinguishable but genetically distinct⁵²⁹. Now, the recommendation is to call these “embryonal tumors”⁵, but the term PNET is entrenched. These tumors include: retinoblastoma, pineoblastoma, neuroblastoma, esthesioneuroblastoma. Medulloblastoma (MB) is more than just a PNET of the posterior fossa (see below), as alterations involved in evolution of MBs such as beta-catenin and APC mutations are absent in pineoblastomas and supratentorial primitive PNETs (sPNETs). At least some MBs originate from the external granular layer (EGL) of the cerebellum.

Embryonal tumors

Location: Embryonal tumors most commonly arise in the cerebellar vermis (medulloblastoma), but also occur in cerebrum, pineal, brainstem or spinal cord. Primary spinal cord PNETs are extremely rare (approximately 30 cases reported by 2007530). sP-NETS have a worse prognosis than MB (see below).

Dissemination: Embryonal tumors (ETs) may disseminate via the CSF spontaneously⁵³¹, or iatrogenically (following surgery or shunting, the latter is a rare cause of tumor dissemination²⁵). Thus, all patients with ETs require spinal axis evaluation (gadolinium enhanced MRI is about as sensitive as water-soluble myelography) and cytologic examination of CSF. Prophylactic craniospinal XRT is indicated following surgical removal, but cranial XRT is avoided if at all possible before 3 years of age to avoid intellectual impairment and growth retardation (see Radiation injury and necrosis, page 771). Extraneural metastases can also occur.

21.2.19.1. Supratentorial primitive neuroectodermal tumors

Supratentorial primitive neuroectodermal tumor (sPNETs) are highly malignant lesions primarily affecting young children (65% occur in age < 5 years) and account for 2.5–6% of childhood brain tumors. Occur rarely in adults. No gender predilection. Histologically indistinguishable from medulloblastoma (MB), they have a distinct genetic pro-file, are more aggressive, and often respond poorly to MB-specific therapies (especially pineoblastomas). Overall survival rate for sPNETs is substantially lower than that for MBs, with an expected 3- year progression-free survival of approximately 50% for localized supratentorial PNETs^{532, 533}.

21.2.19.2. Medulloblastoma (MB)

Key concepts:

• a small-cell embryonal tumor of the cerebellum found predominantly in children (peak: 1st decade). The most common pediatric brain malignancy

• usually arises in the cerebellar vermis in the region of the apex of the roof of the 4th ventricle (fastigium), often producing hydrocephalus

• brainstem invasion usually limits complete surgical excision

• all patients must be evaluated for “drop mets”

Epidemiology

In children: MBs comprise 15-20% of intracranial tumors¹³⁸, 30-55% of p-fossa tumors. MB is the most common malignant pediatric brain tumor⁵³⁴. MBs comprise < 1% of adult brain neoplasms. Peak incidence: during 1st decade. Median age at diagnosis: 5-7 years (75% are diagnosed by age 15). Male:female ratio is 2:1. Familial cancer syndromes that include MB: Gorlin syndrome, Turcot syndrome (see page 588).

Clinical

Clinical history is typically brief (6-12 weeks). MBs usually arise in the cerebellar vermis, at the apex of the roof of the 4th ventricle (fastigium in the region of the posterior medullary velum), which predisposes to early obstructive hydrocephalus. Usual presenting symptoms: H/A, N/V, and truncal & appendicular ataxia. Infants with hydrocephalus may present with irritability, lethargy, or progressive macrocrania⁵³⁵. Spinal drop mets may produce back pain, urinary retention or leg weakness. Common signs: papilledema, ataxia, nystagmus, EOM palsies.

Seeding & metastases

≈ 10-35% have seeded the cranio-spinal axis at the time of diagnosis¹³⁸, and extra-neural mets occur in 5% of patients⁵³⁴, sometimes promoted by shunting⁵³⁶ (although this is uncommon²⁵).

EVALUATION

Usually appears as a solid, IV-contrast-enhancing lesion on CT or MRI (however, a rare diffuse variant in children < 3 yrs, medulloblastoma with extensive nodularity⁴⁴ (MBEN), has been described). Most are located in the midline in the region of the 4th ventricle (laterally situated tumors are more common in adults). Most have hydrocephalus. Ependymoma is the main entity to differentiate from on imaging (see page 1210.

CT: noncontrast → typically hyperdense (due to high cellularity), contrast → most enhance. 20% have calcifications.

MRI: T1WI → hypo- to isointense. T2WI → heterogeneous due to tumor cysts, vessels and calcifications⁵³⁷. Most enhance (including MBEN)

Spinal imaging: MRI with IV gadolinium or CT/myelography with water-soluble contrast should be done to rule-out “drop mets”. Staging is done either pre-op or within 2-3 weeks of surgery.

PATHOLOGY

All MB are WHO grade IV⁵³⁸.

Histologic subtypes⁵³⁸:

1. classic (90%): small, densely packed undifferentiated cells with hyperchromatic nuclei, scant cytoplasm (and inconstant cell clusters in Homer-Wright rosettes)⁵³⁷ (sometimes called “blue tumor”) (monotonous appearance)

2. desmoplastic (6%): similar to classic type with “glomeruli” AKA pale islands (collagen bundles and scattered, less cellular areas). Marked tendency for neuronal differentiation. More common in adults. Prognosis controversial: may be the same⁵³⁹ or less aggressive² than classic MB

3. large cell (4%⁵⁴⁰): large, round, and/or pleomorphic nucleoli, higher mitotic activity. In the few case reports, all were male. More aggressive than classic. Resembles atypical teratoid/rhabdoid tumors of cerebellum, but has different phenotype and cytogenic features

MOLECULAR BIOLOGY

The molecular genetic alterations in MBs can be divided into 3 groups:

1. non-random chromosomal abnormalities: (e.g. consistent deletion of 17p markers) has been shown in 35–40%

2. information from gene profiling:

A. ZIC and NSCL1 were the genes most closely correlated with MBs

B. certain genes were associated with more favorable outcome⁵²⁹

3. abnormalities in signal transduction pathways: e.g. neurotrophin signaling pathway (important in cerebellar development) or Sonic hedgehog (Shh)⁵⁴¹

TREATMENT

Stratification of patients into risk groups guides therapy (see Table 21-57 below).

MB are highly radiosensitive and moderately chemosensitive.

Treatment of choice: surgical debulking of as much tumor as possible (without causing neurological injury) followed by craniospinal XRT (radiation is necessary because of propensity to recur and to seed). Invasion of or attachment to the floor of the fourth ventricle (brainstem in the region of the facial colliculus) often limits excision. It is better to leave a small residual on the brain-stem (these patients do fairly well) than it is to chase every last remnant into the brain-stem (neurologic deficit is more likely with this).

Surgical exposure of midline cerebellar medulloblastomas requires opening of the foramen magnum, usually removal of the posterior arch of C1, and occasionally the arch of C2. Tumor spread with arachnoidal thickening (“sugar coating”) may occur.

XRT: optimal irradiation dose: 35-40 Gy to whole craniospinal axis + 10-15 Gy boost to tumor bed (usually posterior-fossa) and to any spinal mets seen, all fractionated over 6-7 wks^{542, 543}. Reduce dosages by 20-25% for age < 3 yrs, or use chemotherapy instead. Lower dose radiation (25 Gy) to the neuraxis may provide acceptable control when confirmed gross total excision is achieved⁵⁴⁴.

Chemotherapy: there is no standardized chemotherapy regimen. Lomustine (CCNU), cisplatin and vincristine (VCR) are primarily used, but are usually reserved for recurrence, for poor risk patients (see Prognosis below), or for children < 3 yrs age. Significant survival advantage was shown in poor-risk children with adjuvant chemotherapy (5-year actuarial disease-free survival rate = 87%) compared to those without (33%). No difference was observed among standard-risk patients⁵⁴⁵.

Shunts: 30-40% of children require permanent VP shunts following p-fossa resection. The risk of shunt-related seeding has been quoted as high as 10-20%¹³⁸, but this is probably overestimated²⁵. In the past, tumor filters were frequently used. They are less commonly used today because of the high incidence of obstruction.

Table 21-57 Risk stratification in medulloblastoma

Standard-risk patients

No residual tumor on post-op MRI and negative CSF results. 5-year survival is > 5%, and progression-free survival = 50%546, 547

Poor-risk patients

Bulky residual tumor > 1.5 cm2 post-op and dissemination in the brain, spine or CSF. Worse prognosis. 5-year disease-free survival is 35-50%548

Intermediate risk patients

An intermediate risk group probably exists, but has been poorly characterized

PROGNOSIS

Poor prognosticators⁵⁴⁹

• younger age (especially if < 3 yrs)

• disseminated (metastatic) disease

• inability to perform gross-total removal (especially if residual > 1.5 cm² in patient with localized disease)

• histological differentiation along glial, ependymal, or neuronal lines

One stratification scheme is shown in Table 21-57.

The sex of the child is an important predictor for survival of MB; girls had a much better outcome⁵⁵⁰. Gene expression profiling is highly predictive of response to therapy, predicting outcome with much greater accuracy than current staging criteria⁵²⁹. The ability of multiple biological and clinical markers to predict outcomes for patients with MB is currently under investigation^{551, 552}.

Long-term survivors of MB are at significant risk for permanent endocrinologic, cognitive, and psychological sequelae of treatments. Infants and very young children with MB remain a difficult therapeutic challenge because they have the most virulent form of the disease and are at highest risk for treatment-related sequelae.

Most common site of recurrence is p-fossa.

21.2.19.3. Ependymoblastoma

A highly cellular embryonal form of ependymal tumor⁵⁵³. Occurs most often in age < 5 yrs. Prognosis is poor, with median post-op survival ranging from 12-20 months, and almost 100% mortality rate at 3 yrs. As with other tumors in this category, there is a tendency for subarachnoid seeding.

21.2.19.4. Atypical teratoid/rhabdoid tumors (AT/RT)

A unique embryonal tumor of the CNS. Many of these tumors were probably previously misdiagnosed as MBs. Occurs primarily in infants and children (> 90% are < 5 years of age, with most age < 2 years). A minority are associated with primary renal rhabdoid tumor. 50% of AT/RTs occur in posterior fossa with a predilection for the cerebellopontine angle (CPA).

33% have CSF spread at presentation. Most patients die within 1 year of diagnosis.

Histopathology: some tumors are composed entirely of rhabdoid cells, others have a combination of rhabdoid and areas resembling PNET/MB. Other cell types include: malignant mesenchymal cells (usually spindle cells), malignant epithelial cells (glandular or squamous).

Molecular biology: AT/RT and the rhabdoid renal tumors have a deletion or monosomy of chromosome 22.

21.2.20. Epidermoid and dermoid tumors

AKA epidermoid or dermoid cysts.

Table 21-58 Comparison of epidermoids and dermoid

Feature	Epidermoid	Dermoid
frequency	0.5-1.5% of brain tumors	0.3% of brain tumors
lining	stratified squamous epithelium	also include dermal appendage organs (hair follicles and sebaceous glands)
contents	keratin, cellular debris, and cholesterol, occasional hair	same as epidermoids, plus hair and sebum
location	more common laterally (e.g. CP angle)	more commonly near midline
associated anomalies	tend to be isolated lesions	associated with other congenital anomalies in up to 50% of cases
meningitis	may have recurrent aseptic meningitis (including Mollarets meningitis, page 690)	may have repeated bouts of bacterial meningitis

Comparison of dermoids and epidermoids

Both are usually developmental, benign tumors that may arise when retained ectodermal implants are trapped by two fusing ectodermal surfaces. The growth rate of these tumors is linear, like skin (rather than exponential, as with neo-plastic tumors). Distinguishing features between the two tumors are shown in Table 21-58. They may occur in the following locations

1. calvaria: skull involvement occurs when ectodermal rests are included in the developing cranium (see page 699), epidural extension may occur with growth

2. intracranial: the most common sites include

A. suprasellar: commonly produce bitemporal hemianopsia and optic atrophy, and only occasionally pituitary (endocrine) symptoms (including DI)

B. sylvian fissure: may present with seizures

C. CPA: may produce trigeminal neuralgia, especially in young patient

D. basilar-posterior fossa: may produce lower cranial nerve findings, cerebellar dysfunction, and/or corticospinal tract abnormalities

E. within the ventricular system: occur within the 4th ventricle more commonly than any other

3. scalp

4. within the spinal canal:

A. most arise in the thoracic or upper lumbar spine

B. epidermoids of the lower lumbar spine may occur iatrogenically following LP (see Lumbar puncture, page 201)

C. dermoids of the spinal canal are usually associated with a dermal sinus tract (see page 252) and may produce recurrent bouts of spinal meningitis

EPIDERMOID CYSTS

Key concepts:

• usually arise from ectoderm trapped within or displaced into the CNS

• predilection for: CP angle, 4th ventricle, suprasellar region, spinal cord

• sometimes AKA cholesteatoma (not to be confused with cholesterol granuloma)

• grow at linear rate (unlike exponential rate of true neoplasms)

• imaging: CSF-like mass (hisignal on DWMRI is the best test to differentiate)

• may produce aseptic meningitis (Mollaret’s meningitis is one form)

• treatment: surgical excision. XRT has no role

AKA cholesteatoma (not cholesterol granuloma (see below)), AKA pearly tumor, AKA ectodermal inclusion cyst (see Table 21-58 above for comparison to dermoids). Although epidermoids and cholesteatomas are histologically identical (both arise from epithelium entrapped in an abnormal location, epidermoids are intradural, cholesteatomas are extradural), the term cholesteatoma is most often used to describe the lesion in the middle ear where the entrapped epithelium usually arises from chronic middle ear infections which lead to a retraction pocket (rarely, may instead be congenital).

May arise from any of the following⁵⁵⁴:

1. displaced dorsal midline ectodermal cell rests trapped during neural tube closure between gestational weeks 3-5

2. multipotential embryonic cell rests

3. epithelial cell rests carried to the CPA with the developing otic vesicle

4. epidermal cells displaced into CNS, e.g. by LP (see Lumbar puncture, page 201) or repeated percutaneous cranial subdural taps⁵⁵⁵

EPIDEMIOLOGY

Epidermoids comprise 1% of intracranial tumors⁵⁵⁶, and ≈ 7% of CPA tumors. Peak age of occurrence: 40 years. No gender difference.

HISTOLOGY

Epidermoids are lined by stratified squamous epithelium, and contain keratin (from desquamated epithelium), cellular debris, and cholesterol⁵⁵⁷. Growth occurs at a linear rate like normal skin, unlike the exponential growth of true neoplasms⁵⁵⁸. The cyst contents may be liquid or may have a flaky consistency. They tend to spread along normal cleavage planes and surround vital structures (cranial nerves, ICA…). Bony destruction occurs in a minority, usually with larger tumors. Rare degeneration to squamous cell cancer⁵⁵⁹ primarily in cases of repeated recurrences after multiple surgeries.

Distinction from cholesterol granuloma

Epidermoid cysts are sometimes mistakenly equated with cholesterol granulomas⁵⁶⁰, possibly because of the similarity between the terms cholesteatoma and cholesterol granuloma. However, these are distinct lesions⁵⁶¹. Cholesterol granulomas usually occur following chronic inflammation (usually in pneumatized portions of the temporal bone: petrous apex, mastoid air cells, middle ear space). Some differences are delineated in Table 21-59.

PRESENTATION

1. may present as any mass lesion in the same location

2. CPA lesions can produce V, VII or VIII neuropathies

3. recurrent episodes of aseptic meningitis caused by rupture of the cyst contents, which may also lead to hydrocephalus. Symptoms include fever and meningeal irritation. CSF shows pleocytosis, hypoglycorrhachia, elevated protein, and negative cultures. Cholesterol crystals may be seen and can be recognized by their amorphous birefringent appearance. Mollaret’s meningitis is a rare variant of aseptic meningitis which includes the finding of large cells in the CSF that resemble endothelial cells (which may be macrophages⁵⁶⁴) that may be seen in some patients with epidermoid cysts^{565, 566}

IMAGING

MRI: (see Figure 21-4) mimics CSF on T1WI (low signal, may be slightly > CSF) and T2WI (high signal). Tumors are usually also high signal on T2WI, but most enhance with contrast on T1WI (epidermoids do not enhance). An epidermoid may pass from the posterior fossa through the incisura to the middle fossa.

Diffusion weighted imaging (DWI) is the best test to differentiate epidermoids from CSF (e.g. as in similar appearing arachnoid cyst). Epidermoids show intense signal on DWI as a result of restriction of water movement.

Figure 21-4 MRI demonstrating left cerebellopontine angle epidermoid Note that the CSF is dark on the DWI

CT: low density, slightly greater than CSF. No enhancement⁵⁶⁷ (enhancement suggests a possible malignant epithelial component). Bone erosion is seen in 33%.

TREATMENT

Caution when removing epidermoid cysts to minimize spilling contents as they are quite irritating and may cause severe chemical meningitis (Mollaret’s meningitis, see above). Berger⁵⁵⁴ advocates intraoperative irrigation with hydrocortisone (100 mg/L of LR) to reduce the risk of post-op communicating hydrocephalus. Peri-operative IV steroids and copious saline irrigation during surgery may provide similar results. The tumor is cyst wall, and the surgical plan is generally to remove as much as possible but to leave capsule adherent to critical structures such as brainstem and blood vessels as the morbidity of removal is high and, a small residual is does not preclude satisfactory out-come.

In spite of adequate removal, it is not unusual to see persistent brainstem distortion on post-op imaging⁵⁶¹. Post-op radiation is not indicated as the tumor is benign and XRT does not prevent recurrence⁵⁶⁸.

21.2.21. Pineal region tumors

Key concepts:

• wide variety of pathology: germ cell tumors (mostly germinomas, teratomas), astrocytomas, & pineal tumors (mostly pineoblastomas) account for most tumors

• since tumors may be of mixed cell types, CSF tumor markers (ß-hCG, AFP…) are not as useful for diagnosis as they are for following response to treatment

• traditionally a test dose of XRT was employed, but there is a growing trend to obtain tissue diagnosis in all cases if possible before instituting treatment

Pineal region⁵⁶⁹: bounded dorsally by the splenium of the corpus callosum and the tela choroidea, ventrally by the quadrigeminal plate and midbrain tectum, rostrally by the posterior aspect of the 3rd ventricle, and caudally by the cerebellar vermis.

A striking feature is the diversity of lesions (neoplastic and nonneo-plastic) that may occur in this location due to the variety of tissues and conditions normally present, as shown in Table 21-60.

Table 21-60 Conditions giving rise to pineal region tumors

Substrate in pineal region	Tumor that may arise
pineal glandular tissue	pineocytomas and pineoblastomas
glial cells	astrocytomas (including pilocytic), oligodendrogliomas, glial cysts (AKA pineal cyst)
arachnoid cells	meningiomas, arachnoid cysts (non-neoplastic). Meningiomas characteristically displace the internal cerebral vein inferiorly
ependymal lining	ependymomas
sympathetic nerves	chemodectomas
rests of germ cells	germ cell tumors: choriocarcinoma, germinoma, embryonal carcinoma, endodermal sinus tumor (yolk sac tumor), and teratoma
absence of blood-brain barrier (BBB) in pineal gland	makes it a susceptible site for hematogenous metastases
remnants of ectoderm	epidermoid or dermoid cysts
Non neoplastic lesions that may mimic tumors
vascular	vein of Galen aneurysm (page 1112), AVM
infectious	cysticercosis (page 370)

PINEAL CYSTS (PCS)

Usually an incidental finding (i.e. not symptomatic), seen on ≈ 4% of MRIs⁵⁷⁰ or on 25-40% of autopsies⁵⁷¹ (many are microscopic). The most common ones are intra-pineal glial-lined cysts with diameter < 1 cm. Etiology is obscure, PCs are nonneoplastic, and may be due to ischemic glial degeneration or due to sequestration of the pineal diverticulum. They have been regarded as benign, but the natural history is not known with certainty⁵⁷². PCs may contain clear, slightly xanthochromic, or hemorrhagic fluid. Rarely, they may enlarge, and like other pineal region masses, may become symptomatic by causing hydrocephalus by aqueductal compression⁵⁷³, gaze paresis⁵⁷⁴ including Parinaud’s syndrome (see page 114), or hypothalamic symptoms.

Positional H/As have been attributed to PCs, the theory is that the cyst could intermittently compress the vein of Galen and/or sylvian aqueduct⁵⁷⁵. This remains unproven since asymptomatic compression of the vein of Galen and the quadrigeminal plate has been demonstrated on MRI⁵⁷⁶.

Imaging

May escape detection on CT because the cyst fluid density is often similar to CSF. MRI T1WI shows round or ovoid abnormality in region of pineal recess, signal varies with protein content (isointense or slightly hyperintense). T2WI occasionally show increased intensity⁵⁷². Gadolinium occasionally enhances the cyst wall with a maximum thickness of 2 mm; irregularities of the wall with nodular enhancement suggests the lesion is not benign.

Epidermoid-dermoid cysts may also occur in the pineal region, and are larger and have different signal characteristics on MRI.

Management

Asymptomatic PCs < 2 cm diameter with typical appearance should be followed clinically and with annual imaging studies. Surgery to relieve symptoms or to obtain a diagnosis is suggested for symptomatic lesions or for ones that show changes on MRI.

Surgery options for patients with hydrocephalus:

1. CSF shunt: may not relieve gaze disturbance (from pressure on tectal plate)

2. cyst excision: relieves symptoms and establishes diagnosis. Low morbidity

3. stereotactic or endoscopic aspiration: may not get enough tissue for diagnosis

4. endoscopic third ventriculostomy (ETV) (see page 212): useful only for typical PC as it does not obtain tissue for pathology. A few cases of regression of PCs after ETV have been reported⁵⁷⁷

PINEAL REGION NEOPLASMS

Tumors in this region are more common in children (3-8% of pediatric brain tumors) than in adults (≤ 1%)⁵⁷⁸. Over 17 tumor types occur in this region⁵⁷⁹. Germinoma is the most common tumor (21-44% in American/European population, 43-70% in Japan), followed by astrocytoma, teratoma and pineoblastoma⁵⁸⁰. Many tumors are of mixed cell type.

Germ cell tumors (GCT), ependymomas and pineal cell tumors metastasize easily through the CSF (“drop metastases”).

PINEAL GLAND TUMORS

Pineal cell tumors

A pineocytoma (AKA pinealcytoma) is a well differentiated neoplasm arising from pineal epithelium. Pineoblastoma (AKA pinealblastoma) is a malignant tumor that is considered a primitive neuroectodermal tumor (PNET)(see page 686). Both can metastasize through the CSF, and both are radiosensitive.

Germ cell tumors (GCT)

When they arise in the CNS, GCTs occur in the midline in the suprasellar and/or pineal region (simultaneous suprasellar and pineal region lesions is diagnostic of a GCT, so-called synchronous germ cell tumors, comprise 13% of GCTs, and are highly sensitive to XRT⁵⁸¹). In the pineal region, these tumors occur predominantly in males. In females, GCTs are more common in the suprasellar region⁵⁸². Aside from benign teratomas, all intracranial GCTs are malignant and may metastasize via CSF and systemically. Types of GCTs:

1. germinomas: malignant tumors of primitive germ cells that occur in the gonads (called testicular seminomas in males, dysgerminomas in females) or in the CNS. Survival with these is much better than with nongerminomatous tumors

2. non-germinomatous germ cell tumors (NGGCT) include:

A. embryonal carcinoma

B. choriocarcinoma

C. endodermal sinus tumor (EST) AKA yolk sac carcinoma: usually malignant

D. teratoma

1. mature

2. immature

Tumor markers

GCTs characteristically (but not always) give rise to tumor markers in the CSF (see Tumor markers used clinically, page 721).

Elevated CSF beta-human chorionic gonadotropin (ß-hCG) is classically associated with choriocarcinomas, but also occurs with up to 50% of germinomas (which are more common).

Alpha-fetoprotein (AFP) is elevated with endodermal sinus tumors, embryonal carcinoma and occasionally with teratomas. Elevated placental alkaline phosphatase (PLAP) in serum or CSF occurs with intracranial germinomas⁵⁸³. Table 21-61 summarizes these findings. When positive, tumor markers can be followed serially to assess treatment and to look for recurrence (they should be checked in serum and CSF). NB: tumor markers alone are not usually sufficient for making a definitive diagnosis of a pineal region tumor since many of these tumors are mixed cell type.

PEDIATRIC

A breakdown of pediatric pineal region tumors in one series is shown in Table 21-62 (series A).

In 36 patients < 18 yrs age, 17 distinct histological tumor types were identified: 11 germinomas (the most common tumor), 7 astrocytomas, and the remaining 18 had 15 different tumors⁵⁸⁴.

ADULT

GCTs and pineal cell tumors occur primarily in childhood and young adults. Thus, over the age of 40, a pi-neal region tumor is more likely to be a meningioma or a glioma. Series B in Table 21-62 includes both adult and pediatric patients.

CLINICAL

Almost all patients have hydrocephalus by the time of presentation, causing typical signs and symptoms of headache, vomiting, lethargy, memory disturbance, abnormally increasing head circumference in infants, and seizures. Parinaud’s syndrome (or the syndrome of the sylvian aqueduct) may be present (see page 114). Precocious puberty may occur only in boys with choriocarcinomas or germinomas with syncytiotrophoblastic cells, due to leuteinizing hormone-like effects of ß-hCG secreted in the CSF. Suprasellar GCT: triad of diabetes insipidus, visual deficit and panhypopituitarism⁵⁸².

Drop metastases from CSF seeding can produce radiculopathy and/or myelopathy.

Table 21-62 Pineal region tumors

Tumor	Series A* (%)	Series B† (%)
germinoma	30	27
astrocytoma	19	26
pineocytoma	6	12
malignant teratoma	6
unidentified germ-cell tumor	6
choriocarcinoma	3	1.1
malignant teratoma/embryonal cell tumor	3	1.6
glioblastoma	3
teratoma	3	4.3
germinoma/ectodermal sinus tumor	3
dermoid	3
embryonal cell tumor	3
pineoblastoma	3	12
pineocytoma/pineoblastoma	3
endodermal sinus tumor	3
glial cyst (pineal cyst)585	3	2.7
arachnoid cyst	3
metastases		2.7
meningioma		2.7
ependymoma		4.3
oligodendrogliomas		0.54
ganglioglioneuroma		2.7
lymphoma		2.7

* 36 children ≤ 18 yrs⁵⁸⁴

^† 370 tumors in patients 3-73 yrs old⁵⁷⁸

MANAGEMENT

The optimal management strategy for pineal region tumors has yet to be determined. “Test dose” radiation: Controversial. This is giving way to the doctrine of obtaining histology in most cases (e.g. by stereotactic biopsy) because of the harmful effects of XRT and because 36-50% of pineal tumors are benign or radioresistant⁵⁸⁶. The concept was that if a pineal region tumor enhanced uniformly and had the classic appearance of a germinoma on MRI, a test dose of 5 Gy was given, and if the tumor would shrink then the diagnosis of germinoma was virtually certain and XRT was continued without surgery. This may needlessly expose a patient with benign or radioresistant tumors to XRT, 578. “Trial XRT” should be avoided in tumors suspected of being teratomas or epidermoid cyst on MRI, and the response may be misleading in the relatively common situation of tumors with mixed cell types.

Management suggestions

1. get MRI of cervical, thoracic and lumbar spine to assess for drop mets

2. send for GCT markers (ß-hCG, AFP, PLAP (see page 692)) (somewhat helpful, but not adequate for diagnosis)^A:

A. serum

B. CSF (if able to safely obtain^B)

3. obtain histology in most cases. Most often this involves a biopsy, which should be generous (to avoid missing other histologies in mixed cell tumors)

A. if hydrocephalus: transventricular biopsy

B. if no hydrocephalus:

1. open biopsy or

2. stereotactic biopsy or

3. ? by CACE (see below)

4. based on markers and histology:

A. germinoma: XRT + chemo

B. all other tumors: one option is resection followed by adjuvant therapy (usually not very helpful) - see Indications:, page 695 for controversies

A. if negative (–) for GCT markers, it may be a pineal cell tumor, or it may be a GCT without markers (see Tumor markers, page 692). If positive, it can still be a mixed cell-type tumor

B. LP is contraindicated with large intracranial mass and/or obstructive hydrocephalus; CSF may be obtained from EVD if placed

Hydrocephalus

Patients presenting acutely due to hydrocephalus may be best treated with external ventricular drainage (EVD). This permits control over the amount of CSF drained, prevents peritoneal seeding with tumor (a rare event²⁵), and may avoid having a permanent shunt placed in the significant number of patients who will not need one after tumor removal (although ≈ 90% of patients with a pineal GCT require a shunt). Ventricular access (via EVD or Frazier burr hole, see page 156) in the post-op period is important in the event of acute hydrocephalus.

Stereotactic procedures

May be used to ascertain diagnosis (biopsy), or to treat symptomatic pineal region cysts^{587, 588}. Caution is advised since the pineal region has numerous vessels (vein of Galen, basal veins of Rosenthal, internal cerebral veins, posterior medial choroidal artery)⁵⁸⁹ which may be displaced from their normal position. The complication rate of stereotactic biopsy is: ≈ 1.3% mortality, ≈ 7% morbidity, and 1 case of seeding in 370 patients, and the diagnostic rate is ≈ 94%⁵⁷⁸. A shortcoming of stereotactic biopsy is that it may fail to disclose the histologic heterogeneity of some tumors.

Two main stereotactic trajectories: 1) anterolateral (low frontal) approach below the internal cerebral veins, and 2) posterolateral trans-parieto-occipital⁵⁷⁹. One study found that the trajectory correlated with complications, and they recommended the anterolateral approach⁵⁹⁰. However, the correlation of trajectory and complications was not born out in another study⁵⁷⁸, and they found that the complication rate was higher in firm tumors (pineocytomas, teratomas, and astrocytomas) and they recommend an open approach when the tumor appears difficult to penetrate on the first attempt at biopsy.

Stereotactic radiosurgery may be appropriate for treatment of some lesions.

Computer-assisted cisternal endoscopic approach (CACE)

Employs a supracerebellar infratentorial approach that permits visualization of neurovascular structures and avoids traversing brain parenchyma⁵⁷⁹.

Radiation treatment

For controversies regarding “test dose” XRT, see Management, page 693. Germinomas are very sensitive to radiation (and chemotherapy), and are probably best treated with these modalities and followed.

XRT is also utilized post-op for other malignant tumors. For highly malignant tumors or if there is evidence of CSF seeding, craniospinal XRT with a boost to the tumor bed is appropriate.

If possible, XRT is best avoided in the young child (see page 771). Chemotherapy may be used for age < 3 yrs until the child is older when XRT is better tolerated⁵⁸².

Surgical treatment of the tumor

Indications: Controversial. Some authors feel that most tumors (except germinomas, which are best treated with XRT) are amenable to open resection⁵⁹¹. Others feel that resection should be limited to ≈ 25% of tumors which are⁵⁷⁸:

1. radioresistant (e.g. malignant nongerminoma GCTs): 35-50% of pineal region tumors (larger numbers occur in series not limited to pediatric patients)

2. benign (e.g. meningioma, teratomas…)

3. well encapsulated

4. NB: malignant germ cell tumors should be without evidence of metastases (those with metastases do not benefit from surgery on the primary tumor)

5. pineocytoma: recommendation is for surgical excision + SRS for any residual

Options:

1. direct surgery: obtains generous tissue for biopsy. Curative for benign lesions. Not the optimum treatment for malignant tumors and germinomas without complications

2. biopsy followed by adjuvant therapy: the preferred management for malignancies and germinomatous germ cell tumors

Surgical approaches: Choice is aided by the pre-op MRI and includes:

1. most common approach: midline infratentorial-supracerebellar approach of Horsley and Krause as refined by Stein⁵⁹². Cannot be used if the angle of the tentorium is too steep (best assessed on MRI). May be done in the sitting position (risk of air embolism, see page 153) or in the Concorde position (see page 153)

2. occipital transtentorial: wide view. Risk of injury to occipital visual cortex or splenium of corpus callosum. Recommended for lesions centered at or superior to the tentorial edge or located above the vein of Galen or for rare cysts with superior extension. The occipital lobe is retracted laterally, and the tentorium is incised 1 cm lateral to the straight sinus

3. transventricular: indicated for large, eccentric lesions with ventricular dilatation. Usually via a cortical incision in the posterior portion of the superior temporal gyrus. Risks: visual defect, seizures, and on dominant side language dysfunction

4. lateral paramedian infratentorial

5. transcallosal: largely abandoned except for tumors extending into corpus callosum and third ventricle

6. paramedian infratentorial-supracerebellar approach may be used for cysts that do not extend superiorly or contralaterally⁵⁷²: avoids midline venous structures

Important surgical considerations:

The base of the pineal gland is the posterior wall of the 3rd ventricle. The splenium of the corpus callosum lies above, and the thalamus surrounds both sides. The pineal projects posteriorly and inferiorly into the quadrigeminal cistern. The deep cerebral veins are a major obstacle to operations in this region. Venous drainage of the pineal region must be preserved.

Surgical outcome:

Mortality rate: 5-10%⁵⁷⁸. Postoperative complications include: new visual field deficits, epidural fluid collection, infection, and cerebellar ataxia.

21.2.22. Choroid plexus tumors

Most are histologically benign (choroid plexus papilloma (CPP), WHO I), although intermediate (atypical choroid plexus papilloma, WHO II) and malignant tumors (choroid plexus carcinoma (CPC), WHO III) may occur. Malignant degeneration from WHO I or II to grade III was seen in 2 out of 124 patients with 59 months mean follow-up⁵⁹³. All may produce drop mets in the CSF, but WHO III do so more commonly. Although usually slow growing, they sometimes grow rapidly.

Atypical CPP have more mitotic figures than CPP without frank signs of malignancy seen in CPC⁵⁹⁴, and up to 2 of the following 4 features may be observed: increased cellularity, nuclear pleomorphism, blurring of the papillary pattern, areas of necrosis.

Epidemiology

Prevalence: 0.4-1% of all intracranial tumors. 1.5-6% of tumors in peds.

Although they may occur at any age, 70% of patients are < 2 yrs old⁵⁹⁵. Some tumors occur in neonates, supporting the hypothesis that some of these are congenital⁵⁹⁶.

Location: in adults these tumors are usually infratentorial, whereas in children they tend to occur supratentorially (a reverse from the situation for most other tumors) in the lateral ventricle⁵⁹⁶ with a predilection for the left side. See Intraventricular lesions on page 1224 for differential diagnosis. They can be located anywhere there is choroid plexus, with the most frequent locations: the lateral or fourth ventricles, the CPA (from extension of choroid plexus through the foramen of Luschka).

Presentation

Most present with symptoms of increased ICP from hydrocephalus (H/A, N/V, craniomegaly), others may present with seizures, subarachnoid hemorrhage (with meningismus), or focal neurologic deficit (hemiparesis, sensory deficits, cerebellar signs, or cranial nerve palsies of III, IV and VI).

Hydrocephalus, which may result from: overproduction of CSF (although total removal of these tumors does not always cure the hydrocephalus - especially in patients with high CSF protein, hemorrhage from tumor or surgery, or ependymitis), obstruction of CSF outflow, or communicating hydrocephalus from CSF borne particulates.

Imaging

Brain MRI or CT without and with contrast usually demonstrates a densely enhancing multilobulated intraventricular mass classically with projecting “fronds”. Hydrocephalus is common.

Treatment

There is no role for chemotherapy or radiation for WHO I lesions. For choroid plexus carcinoma, chemotherapy benefits a subset of patients⁵⁹⁷.

Surgical treatment:

Benign lesions may be cured surgically with total removal, and even the malignant tumors respond well to surgery. The operation may be difficult due to fragility of the tumor and bleeding from the choroidal arteries. However, persistence with a second and sometimes even third operation is recommended as 5-year survival rate of 84% can be achieved⁵⁹⁶. Post-operative subdural collections after transcortical tumor excision may occur, and may result from a persistent ventriculosubdural fistula, which may require subdural-peritoneal shunting⁵⁹⁵.

Recurrence

12 recurrences (6% of WHO I and 29% of WHO II patients) requiring neurosurgical intervention occurred in 124 complete resections with 59 months mean follow-up⁵⁹³.

21.2.23. Glial tumors of uncertain origin

1. astroblastoma

2. chordoid glioma of the 3rd ventricle⁵⁹⁸: rare, benign tumor of adulthood. Solid, enhancing mass of the 3rd ventricle. Female:male ratio = 3:1. Mitotic activity is absent in most tumors. GFAP immunostaining is common, S100 reactivity is variable. Histologically similar appearing to chordoid meningioma, which lacks GFAP staining. Attachment to wall of 3rd ventricle (hypothalamus) may prevent total removal

21.2.24. Tumors of peripheral nerves

PERINEURIOMA

A nerve sheath tumor. Variants:

1. intraneural perineurioma: usually solitary lesion of adolescence or young adulthood, affecting primarily peripheral nerves (cranial nerve involvement is rare). Pseudo-onion bulb formation with cylindrical enlargement of the nerve > 2-10 cm. Mitotic activity is rare, MIB-1 labeling index is low. Chromosome 22 loss is characteristic²¹⁹, no NF1 association. Treatment: conservative sampling of lesion, not resection

2. soft tissue perineurioma: uncommon. Only rarely can an associated nerve be identified. Almost exclusively benign, but malignant variety does occur. Female:male ratio = 4:1. In males, hands are often affected. Discrete, but not encapsulated, diameter = 1.5-20 cm. Treatment: gross total excision is curative

21.2.25. Miscellaneous primary brain tumors

PRIMARY CNS MELANOMA

Probably arises from melanocytes in the leptomeninges. May spread through CSF pathways. May occasionally metastasize outside the CNS to produce systemic metastases⁵⁹⁹.

The peak age for this tumor is in the 4th decade (compared to the 7th decade for primary cutaneous melanoma)⁶⁰⁰.

21.3. Pediatric brain tumors

Among all childhood cancers, brain tumors are the second only to leukemias in incidence (20%), and are the most common solid pediatric tumor⁵³⁴, comprising 40-50% of all tumors¹³⁸. Annual incidence: 2-5 cases per 100,000.

Types of tumors

The common pediatric brain tumors are gliomas (cerebellum, brain stem, and optic nerve), pineal tumors, craniopharyngiomas, teratomas, granulomas, and primitive neuroectodermal tumors (PNETs, primarily medulloblastoma).

Meningiomas: 1.5% of meningiomas occur in childhood and adolescence (usually between 10-20 years), comprising 0.4-4.6% of intracranial tumors^{127 (p 3263)} (see Meningiomas on page 613).

Table 21-63 Location of pediatric brain tumors by age

Age	% Infratentorial
0-6 mos	27%
6-12 mos	53%
12-24 mos	74%
2-16 yrs	42%

Infratentorial vs. supratentorial

It has traditionally been taught that most pediatric brain tumors (≈ 60%) are infratentorial, and that these are ≈ equally divided among brain stem gliomas, cerebellar astrocytomas, and medulloblastomas. In reality, the ratio of supratentorial to infratentorial tumors is dependent on the specific age group studied, as illustrated in Table 21-63. Table 21-64 shows the breakdown for pooled data from 1350 pediatric brain tumors.

Astrocytomas are the most common supratentorial tumor in pediatrics as in adulthood.

Table 21-64 Incidence of pediatric brain tumors*

Tumor type	Page	% of total
infratentorial tumors		54%
cerebellar astrocytomas	604	15%
medulloblastomas	686	14%
brain stem gliomas	607	12%
ependymomas513	682	9%
supratentorial benign astrocytomas	603	13%

* data from 1350 pediatric brain tumors^{126 (p 368)}

INTRACRANIAL NEOPLASMS DURING THE FIRST YEAR OF LIFE

Brain tumors presenting during the first year of life is a different subset of tumors than those presenting later in childhood. In a busy neurosurgical unit in a children’s hospital, they represented ≈ 8% of children admitted with brain tumors, an average of only ≈ 3 admissions per year⁶⁰¹.

90% of brain tumors in neonates are of neuroectodermal origin, teratoma being the most common. Some of these tumors may be congenital⁶⁰². Other supratentorial tumors include: astrocytoma, choroid plexus tumors, ependymomas, and craniopharyngiomas. Posterior fossa tumors include medulloblastoma and cerebellar astrocytoma.

Many of these tumors escape diagnosis until they are very large in size due to the elasticity of the infant skull, the adaptability of the developing nervous system to compensate for deficits, and the difficulty in examining a patient with limited neurologic repertoire and inability to cooperate. The most common presenting manifestations are vomiting, arrest or regression of psychomotor development, macrocrania, poor feeding/failure to thrive. They may also present with seizures.

21.4. Skull tumors

See Skull lesions on page 1219 for differential diagnosis and evaluation (including non-neoplastic lesions). Considering only tumors, the differential diagnosis includes:

1. benign tumors

A. osteoma: see below

B. hemangioma: see below

C. dermoid and epidermoid tumors: see below

D. chondroma: occur mainly in conjunction with the basal synchondroses

E. meningioma

F. aneurysmal bone cyst

2. malignant tumors: malignancy is suggested by a single large or multiple (> 6) small osteolytic lesions with margins that are ragged, undermined and lacking sclerosis⁶⁰³

A. bone metastases to the skull. Common ones include:

1. prostate

2. breast

3. lung

4. kidney

5. thyroid

6. lymphoma

7. multiple myeloma/plasmacytoma: see page 740

B. chondrosarcoma

C. osteogenic sarcoma

D. fibrosarcoma

21.4.1. Osteoma

Osteomas are the most common primary bone tumor of the calvaria. They are benign, slow-growing lesions, that occur commonly in the cranial vault, mastoid and para-nasal air sinuses, and the mandible. Lesions within air sinuses may present as recurrent sinusitis. More common in females, highest incidence is in 6th decade. Triad of Gardner’s syndrome: multiple cranial osteomas (of calvaria, sinuses, and mandible), colonic polyposis, and soft-tissue tumors.

See Localized increased density or hyperostosis of the calvaria on page 1222 for differential diagnosis.

Pathology

Consists of osteoid tissue within osteoblastic tissue, surrounded by reactive bone. Difficult to distinguish from fibrous dysplasia.

Radiographic evaluation

Skull x-ray: round, sclerotic, well demarcated, homogeneous dense projection. Usually arise from outer table of skull (inner table less common). May be compact or spongy (spongy osteoma may be radiolucent). Unlike meningiomas, diploë are preserved and vascular channels are not increased.

Osteomas are “hot” on nuclear bone scan.

Treatment

Asymptomatic lesions may simply be followed. Surgery may be considered for cosmetic reasons, or if pressure on adjacent tissues produces discomfort. Lesions involving only the outer table may be removed leaving the inner table intact.

21.4.2. Hemangioma

Comprise ≈ 7% of skull tumors⁶⁰³. These benign tumors commonly occur in the skull (discussed here) and spine (see page 738). Two types: cavernous (most common) and capillary (rare).

Radiographic evaluation

Skull x-ray: characteristically shows a circular lucency with honeycomb or trabecular pattern (seen in ≈ 50% of cases) or radial trabeculations producing a sunburst pattern (seen in ≈ 11% of cases)⁶⁰³. Sclerotic margins are evident in only ≈ 33%.

CT: hypodense lesion with sclerotic spaced trabeculations. Nonenhancing.

Bone scan: typically hot.

Treatment

Accessible lesions may be cured by en bloc excision or curettage. The gross appearance is of a hard, blue-domed mass beneath the pericranium. Radiation may be considered for inaccessible tumors.

21.4.3. Epidermoid and dermoid tumors of the skull

See also page 688 for epidermoids and dermoids in general. Skull involvement is rare and occurs when ectodermal rests are entrapped in the developing skull. Usually midline. Arise within the diploë and expand both inner and outer tables. Identical clinically and radiologically. These benign lesions may involve underlying dural venous structures or brain. They may become infected.

Radiographic evaluation

• skull x-ray: these osteolytic lesions have well-defined, sclerotic margins

• some imaging is required to evaluate possible intracranial involvement

CT: the lesions are hypodense (keratin contains fats), and non-enhancing

MRI: like CSF they are low intensity on T1WI and high signal on T2WI, but unlike CSF they are high signal on DWI MRI (see page 690)

Treatment

Treatment is surgical. Bone margins are curetted. Search must be made for a tract leading to the intracranial cavity which must be followed if found. Preparation for dural sinus repair must be made for lesions overlying the sagittal sinus (including torcular Herophili). Radiation and chemotherapy are not indicated.

21.4.4. Eosinophilic granuloma

A generally benign local disease of bone with mononuclear cells and eosinophils, most commonly found in the skull (43-80%). May also be seen in femur (14.5%), mandible, ribs, pelvis, and the spine (vertebra plana, see page 729). Classified as the mildest form of histiocytosis X which also includes multifocal eosinophilic granuloma (Hand-Schüller-Christian disease) and Letterer-Siwe syndrome (a fulminant, malignant lymphoma of infancy)⁶⁰⁴.

CLINICAL

Generally a condition of youth, 70% of patients are < 20 yrs age. In a series of 26 patients⁶⁰⁴, age range was 18 mos-49 yrs (mean: 16 yrs).

Most common presenting symptom: tender, enlarging skull mass (> 90%). May be asymptomatic and incidentally discovered on skull x-ray obtained for other reasons. Blood tests were normal in all except 1 who had eosinophilia of 23%.

Parietal bone was the most common site (42%), frontal bone next (31%)⁶⁰⁴ (some series show frontal bone was the most common).

EVALUATION

Skull x-rays

Classic radiographic finding: round or oval non-sclerotic punched out skull lesion with sharply defined margins, involving both inner and outer tables (the disease begins in diploic space), often with beveled edges. A central bone density is occasionally noted (rare, but diagnostic). No abnormal vascularity of adjacent bone. No periosteal reaction. Differentiate from hemangioma by absence of sunburst appearance.

CT scan

Characteristic appearance of a soft tissue mass within area of bony destruction having a central density⁶⁰⁵. Differentiate from epidermoid which has dense surrounding sclerosis.

PATHOLOGY

Gross: pinkish gray to purple lesion extending out of bone and involving pericranium. Dural involvement occurs in only 1 of 26 patients, but with no dural penetration.

Microscopic: numerous histiocytes, eosinophils, and multinucleated cells in a reticulin fiber network. No evidence that this is a result of an infection.

TREATMENT

Tendency toward spontaneous regression, however, most single lesions are treated by curettage. Multiple lesions are usually associated with extracalvarial bony involvement and are often treated with chemotherapy and/or low dose radiation therapy.

OUTCOME

After a mean 8 years follow-up, 8 patients (31%) developed additional lesions, 5 of these were ≤ 3 yrs age (all of 5 patients < 3 yrs age)⁶⁰⁴ (may suggest a form of Letterer-Siwe, thus young patients should be followed closely). Recurrences were local in one case, and in others involved other bones (including the skull, femur, lumbar spine) or brain (including the hypothalamus, presenting with diabetes insipidus and growth delay).

21.4.5. Non-neoplastic skull lesions

Includes:

1. osteopetrosis (see page 1204)

2. Paget’s disease of the skull (see page 498)

3. hyperostosis frontalis interna (see below)

4. fibrous dysplasia (see page 701)

HYPEROSTOSIS FRONTALIS INTERNA

See page 1222 for differential diagnosis. Hyperostosis frontalis interna (HFI) is a benign irregular nodular thickening of the inner table of the frontal bone that is almost always bilateral. The midline is spared at the insertion of the falx. Unilateral cases have been reported⁶⁰⁶, and in these cases one must R/O other etiologies such as meningioma, calcified epidural hematoma, osteoma, fibrous dysplasia, an epidural fibrous tumor⁶⁰⁷, or Paget’s disease.

The incidence of HFI in the general population is ≈ 1.4-5%⁶⁰⁶. HFI is more common in women (female:male ratio may be as high as 9:1) with an incidence of 15-72% in elderly women. A number of possible associated conditions have been described (most are unproven), the majority of which are metabolic, earning it the alias of metabolic craniopathy. Associated conditions include:

1. Morgagni’s syndrome (AKA Morgagni-Stewart-Morel syndrome): headache, obesity, virilism and neuropsychiatric disorders (including mental retardation)

2. endocrinologic abnormalities

A. acromegaly⁶⁰⁸ (elevated growth hormone levels): see page 639

B. hyperprolactinemia⁶⁰⁸

3. metabolic abnormalities

A. hyperphosphatemia

B. obesity

4. diffuse idiopathic skeletal hyperostosis (DISH): see page 506

CLINICAL

HFI may present without symptoms as an incidental finding on radiographic evaluation for other reasons. Many signs and symptoms have been attributed to HFI including: hypertension, seizures, headache, cranial nerve deficits, dementia, irritability, depression, hysteria, fatigability and mental dullness. The incidence of headache may be statistically higher in patients with HFI than in the general population⁶⁰⁹.

EVALUATION

Blood tests to R/O some of the above noted conditions may be indicated in appropriate cases: check growth hormone, prolactin, phosphate, alkaline phosphatase (to R/O Paget’s disease).

Plain skull x-ray shows thickening of the frontal bone with characteristic sparing of the midline. Spread to parietal and occipital bone occasionally occurs.

CT demonstrates the lesion which usually causes 5-10 mm of bone thickening, but as much as 4 cm has been reported.

Bone scan: usually shows moderate uptake in HFI (generally not as intense as with bone mets). Also, indium-111 leukocyte scan (commonly used to detect occult infection) will show accumulation in HFI (a false positive)^{610, 611}.

TREATMENT

Little has been written about treatment of cases where symptoms are suspected to be due to HFI. In one report, removal of the thickened bone was accomplished without evidence of dural adhesions, and with improvement in the presenting hysteria⁶⁰⁶.

Surgical technique

One technique described consists of using the craniotome to excise the thickened portion of the bone (a plain skull x-ray may be used to make a template), and then the thickened bone is thinned down with a high-speed drill, and the bone flap is then replaced. Alternatively, a cranioplasty with methylmethacrylate may be performed.

FIBROUS DYSPLASIA

Usually a benign condition in which normal bone is replaced by fibrous connective tissue (malignant transformation occurs in < 1%). Does not appear to be heritable. Most lesions occur in the ribs or craniofacial bones, especially the maxilla.

Patterns:

1. monostotic: most common

2. polyostotic: 25% with this form have > 50% of the skeleton involved with associated fractures and skeletal deformities

3. as part of McCune-Albright syndrome (endocrine dysfunction, café au lait spots which tend to occur on one side of the midline and tend to be more jagged than those seen in neurofibromatosis (see page 723), fibrous dysplasia, and precocious puberty primarily in females) and its variants

Clinical

Clinical manifestations of the fibrous dysplasia (FD) lesions include:

1. incidental finding (i.e. asymptomatic)

2. local pain

3. local swelling (rarely marked distortion resembling aneurysmal bone cyst may occur) or deformity

4. may predispose to pathologic fractures when they occur in long bones

5. cranial nerve involvement: including loss of hearing when the temporal bone is involved as a result of obliteration of the external auditory canal

6. seizures

7. serum alkaline phosphatase is elevated in about 33%, calcium levels are normal

8. darkened hair pigmentation overlying skull lesions

9. spontaneous scalp hemorrhages

10. rarely associated with Cushing’s syndrome, acromegaly

3 forms of the FD lesions:

1. cystic (the lesions are not actually cysts in the strict sense): widening of the diploë usually with thinning of the outer table and little involvement of the inner table. Typically occurs high in calvaria

2. sclerotic: usually involves skull base (especially sphenoid bone) and facial bones

3. mixed: appearance is similar to cystic type with patches of increased density within the lucent lesions

Ground glass appearance on x-rays is due to the thin spicules of woven bone.

Treatment

There is no cure for FD. Local procedures (mostly orthopedic) are used for deformities or bone pain that is refractory to other treatment. Neurosurgical involvement may be required for skull lesions producing refractory pain or neurologic symptoms. Calvarial lesions may be treated with curettage and cranioplasty. Calcitonin may be used for wide-spread lesions with bone pain and/or high serum alkaline phosphatase levels.

21.5. Cerebral metastases

Key concepts:

• brain metastases are the most common brain tumor seen clinically

• at the time of onset of neurologic symptoms, 70% will be multiple on MRI

• with solitary brain lesions in a patient with a history of cancer, biopsy should almost always be done since 11% of these lesions will not be mets

• although median survival with maximal treatment is only 8 months (similar to GBM), long-term survivors do occur

METASTASES TO THE BRAIN

Cerebral metastases are the most common brain tumor seen clinically, comprising slightly more than half of brain tumors (if one considers only imaging studies, they comprise ≈ 30%). In the U.S., the annual incidence of new cases of metastases is up to 170,000612, compared to 17,000 for primary brain tumors. 15-30% of patients with cancer (Ca) develop cerebral mets⁶¹³. In patients with no Ca history, a cerebral met was the presenting symptom in 15%; of these, 43-60% will have an abnormal chest x-ray (CXR)^{614, 615} (showing either a bronchogenic primary or other mets to lung).

In 9% of cases, a cerebral met is the only detectable site of spread. Cerebral mets occur in only 6% of pediatric cancers.

The route of metastatic spread to the brain is usually hematogenous, although local extension can also occur.

Solitary mets:

1. CT: at the time of neurologic diagnosis, 50% are solitary on CT^{616, 617} (see page 706)

2. MRI: if the same patients have an MRI, < 30% will be solitary⁶¹⁸

3. on autopsy: mets are solitary in one third of patients with brain mets, and 1-3% of solitary mets occur in the brain stem⁶¹⁹

Increasing incidence of cerebral mets: May be due to a number of factors:

1. increasing length of survival of cancer patients⁶²⁰ as a result of improvements in treatment of systemic cancer

2. enhanced ability to diagnose CNS tumors due to availability of CT and/or MRI

3. many chemotherapeutic agents used systemically do not cross the blood-brain barrier (BBB) well, providing a “haven” for tumor growth there

4. some chemotherapeutic agents may transiently weaken the BBB and allow CNS seeding with tumor

METASTASES OF PRIMARY CNS TUMORS

Spread via CSF pathways

CNS tumors that more commonly spread via CSF pathways include the following (when these tumors spread to the spinal cord, they are often called “drop mets”):

1. high grade gliomas (10-25%) (see page 597)

2. primitive neuroectodermal tumors (PNET), especially medulloblastoma (see page 686)

3. ependymoma (11%) (see page 682)

4. choroid plexus tumors (see page 695)

5. pineal region tumors

A. germ cell tumors (see page 692)

B. pineocytoma and pineoblastoma (see page 692)

6. rarely:

A. oligodendrogliomas (≈ 1%) (see page 609)

B. hemangioblastomas (see page 671)

C. primary CNS melanoma (see page 697)

Extraneural spread

Although most CNS tumors do not spread systemically, there is some potential for extraneural spread with the following tumors:

1. medulloblastoma (cerebellar-PNET): the most common primary responsible for extraneural spread. May spread to lung, bone marrow, lymph nodes, abdomen

2. meningioma: rarely goes to heart or lungs

3. malignant astrocytomas rarely metastasize systemically

4. ependymomas

5. pineoblastomas

6. meningeal sarcomas

7. choroid plexus tumors

8. tumors that spread through CSF pathways (see above) may spread via a CSF shunt (e.g. to peritoneum with VP shunt or hematogenously with a VA shunt), however, this risk is probably quite small²⁵

PRIMARY CANCERS IN PATIENTS WITH CEREBRAL METASTASES

In over 2,700 adults with a primary cancer undergoing autopsy at Sloan-Kettering, the sources of cerebral metastases are shown in Table 21-65. Sources of brain metastases in pediatrics is shown in Table 21-66.

In adults, lung and breast Ca together account for > 50% of cerebral mets.

In patients with a metastatic brain tumor as the initial presentation (i.e. undiagnosed primary) compared to patients with a known primary, there is about the same number of brain lesions, but there was an increased frequency of extracranial mets 622. In up to 26% of cases, the primary tumor was never identified⁶²².

Table 21-65 Sources of cerebral mets in adults (autopsy data)

Primary	%
lung Ca	44%
breast	10%
kidney (renal cell)*	7%
GI	6%
melanoma†	3%
undetermined	10%

* a rare tumor that metastasizes frequently to brain (in 20-25% of cases)

^† 16% in older series⁶²¹

LOCATION OF CEREBRAL METS

Intracranial metastases may be either parenchymal (≈ 75%) or may involve the leptomeninges in a carcinomatous meningitis (see page 711). 80% of solitary metastases are located in the cerebral hemispheres.

The highest incidence of parenchymal mets is posterior to the Sylvian fissure near the junction of temporal, parietal, and occipital lobes (presumably due to embolic spread to terminal MCA branches)⁶²³. Many tend to arise at the gray/white-matter interface.

The cerebellum is a common site of intracranial mets, and is the location in 16% of cases of solitary brain mets. It is the most common p-fossa tumor in adults, thus “a solitary lesion in the posterior fossa of an adult is considered a metastasis until proven otherwise”. Spread to the posterior fossa may be via the spinal epidural venous plexus (Batson’s plexus) and the vertebral veins.

Table 21-66 Sources of cerebral mets in peds

neuroblastoma

rhabdomyosarcoma

Wilm’s tumor

SPECIFIC TYPES OF CEREBRAL METS

The autopsy incidence of cerebral mets for various types of primary cancers at Sloan-Kettering Cancer Center is shown in Table 21-67.

LUNG CANCER

The lungs are the most common source of cerebral mets, and these are usually multiple. The lung primary may be so small as to render it occult.

Necropsy demonstrates cerebral mets in up to 50% of patients with small-cell lung Ca (SCLC) and non-squamous, non-small-cell lung Ca⁶²⁴.

Small-cell lung cancer (SCLC)

AKA “oat cell” Ca. A neuroendocrine tumor. 95% arise in proximal airways, usually in mainstem or lobar bronchi. Typically younger (27-66 years) than other lung Ca. Strongly associated with cigarette smoking. Median survival: 6-10 months. Staged in 1 of 2 categories:

• limited: confined to an area of the chest that can be encompassed by a single radiation port

• extensive: metastasis outside the thorax or intrathoracic disease that cannot be contained in a single radiation port

Although SCLC comprises only ≈ 20% of primary lung cancers, it is more likely to produce cerebral mets than other bronchogenic cell types (brain mets are found in 80% of patients who survive 2 yrs after diagnosis of SCLC)⁶²⁰.

Table 21-67 Autopsy incidence of cerebral mets for given primary cancers

Primary	% with cerebral mets
lung	21%
breast	9%
melanoma	40%
lymphoma	1%
Hodgkin’s	0
non-Hodgkin’s	2%
GI	3%
colon	5%
gastric	0
pancreatic	2%
GU	11%
kidney (renal)	21%
prostate*	0
testes	46%
cervix	5%
ovary	5%
osteosarcoma	10%
neuroblastoma	5%
head and neck	6%

* uncommon, but does occur

Treatment:

Very radiosensitive.

No identified brain mets: prophylactic cranial irradiation (PCI) with WBXRT reduces the incidence of symptomatic brain mets and increases survival (disease-free & overall)^{625, 626}. Typically 25 Gy in 10 fractions.

Brain mets: surgical resection considered for immediately life-threatening large lesions, XRT is used otherwise. Multiple SCLC brain lesions: XRT (initial treatment 30 Gy in 10 fractions) + chemotherapy.

Treatment of primary: usually not resected. Treated with chemotherapy ± XRT.

Recurrent brain mets after failure of initial treatment: 20 Gy in 10 fractions.

Non-small-cell lung cancer (NSCLC)

Includes: adenocarcinoma (the most common NSCLC), large cell, squamous cell, bronchoalveolar. Retrospective analysis of patients with NSCLC completely resected from lung found a 6.8% first recurrence rate in the brain⁶²⁴. Staged with typical TNM system. Prognosis better than SCLC.

Treatment of lung primary:

1. grades I, II, IIIA (i.e. no distal mets, excluding single brain met): resection

2. higher grades: XRT + chemotherapy

Staging studies for known lung primary

1. PET scan: can detect small malignancies. Useful in NSCLC to determine eligibility of resection of primary. Not useful in initial evaluation of SCLC

2. chest CT: usually includes adrenals and liver (thus abdomen and pelvis CT not necessary)

3. bone scan

4. brain: CT or MRI

When metastatic lung cancer is the suspected source of a newly diagnosed brain lesion, the lung lesion should be biopsied (if technically feasible) to rule out SCLC before obtaining tissue from the cerebral mass.

MELANOMA

Melanoma: the 5th most common cancer in men, 7th in women. Incidence is increasing. Most common sites of origin of melanoma metastases: skin, retina, brain (primary CNS melanoma, see page 697), nail bed. The primary site cannot be identified in up to ≈ 14% of cases⁶²⁷. Extremely difficult to locate primary sites: intraocular, GI mucosa.

Brain mets are found in 10-70% of patients with metastatic melanoma in clinical studies, and in 70-90% on autopsy of patients who died from melanoma. Patients with melanoma who have neurosurgical lesions typically presented 14 months after primary lesion was identified. Once cerebral mets of melanoma are detected, median survival is 113 days, and the mets contributed to the death in 94% of cases⁶²⁸. A small group with survival > 3 yrs had a single surgically treated met in the absence of other visceral lesions.

Evaluation

Metastatic melanoma to the brain classically causes pia/arachnoid involvement on imaging. Hemorrhagic involvement is common.

CT: lesions may be slightly hyperdense to brain on unenhanced CT due to melanin. Enhancement is less constant than for other mets (e.g. bronchogenic Ca).

MRI: decreased signal on T2WI surrounded by intense halo of edema. Enhancing T1WI lesions in a patient with melanoma is highly suggestive of melanoma metastases.

Systemic work-up: systemic disease determines ultimate survival after treatment of melanoma mets to the brain in 70% of patients. search for systemic mets should be done, including: CT of chest/abdomen/pelvis & bone scan. PET scan may be more sensitive for detecting metastatic spread than CT when there are clinical signs that the tumor has spread⁶²⁹; except for the brain, where brain MRI is more sensitive than CT or PET.

Treatment

Surgical indications:

1. patients with CNS metastases that can be completely resected with limited systemic disease: long-term survival is possible

2. patients with intracranial mets that cannot be completely removed or with uncontrolled systemic disease may be surgical candidates for the following:

A. for symptomatic relief: e.g. lesion causing painful pressure

B. life-threatening lesion: e.g. large p-fossa lesion with 4th ventricle compression

C. for hemorrhagic lesion causing symptoms by mass effect from the clot

Whole-brain radiation therapy (WBXRT):

Melanoma is typically radioresistant. WBXRT provides 2-3 month survival benefit and may be considered for palliation in patients with multiple mets that preclude complete excision or SRS.

Stereotactic radiosurgery (SRS):

Considered for ≤ 3 lesions all ≤ 3 cm diameter that are surgically inaccessible, with limited or quiescent systemic involvement. Relative contraindications: hemorrhagic lesions, lesions with significant mass effect surrounding edema.

Chemotherapy:

1. dacarbazine: the gold-standard treatment for melanoma. An alkylating agent which is about equally as effective as its newer orally administered analog temozolomide (Temodar®). Response rate: 10-20%

2. interferon and interleukin: interleukin-2 (IL-2) has been associated with significant cerebral edema in patients with melanoma mets in the brain, therefore therapy is withheld unless all mets can be removed prior to therapy 630

3. bevacizumab (Avastin®): monoclonal antibody to vascular endothelial growth factor (VGEF)

4. BAY 43-9006 (inhibits BRAF kinase - BRAF oncogene mutation is common in melanoma) used in conjunction with carboplatin and paclitaxel with a 50% response in Phase I trials

5. immunotherapy: Melacine® vaccine (melanoma tumor cell lysates + immunologic adjuvant Detox) is as effective as chemotherapy with fewer side effects. Used for treatment of documented melanoma, not as a preventative vaccine

Suggested algorithm for patients with metastatic melanoma to the brain⁶³¹:

1. Karnofsky performance scale (KPS) score < 70 (see page 1182): chemotherapy

2. KPS ≥ 70

A. active systemic disease:

1. life expectancy > 3 months: WBXRT

2. life expectancy ≤ 3 months: no treatment

B. no active systemic disease:

1. surgically accessible lesions: resection + WBXRT

2. surgically inaccessible lesions

a. ≤ 3 lesions all ≤ 3 cm diameter: SRS + WBXRT

b. > 3 lesions or any lesion > 3 cm diameter: WBXRT

Outcome

1. older studies with single CNS met and quiescent systemic disease: 8-10 months median survival

2. more recent study with active systemic disease and multiple intracranial lesions⁶³²: 18 months median survival, 20% 5-year survival

3. the presence of infratentorial lesions is a poor prognosticator

RENAL-CELL CARCINOMA

AKA hypernephroma. Usually associated with spread to lungs, lymph nodes, liver, bone (high affinity for bone), adrenals, and contralateral kidney before invading the CNS (thus, this tumor rarely presents as isolated cerebral metastases). Look for hematuria, abdominal pain, and/or abdominal mass on palpation or CT. Response to XRT is only ≈ 10%.

CLINICAL PRESENTATION

As with most brain tumors, signs and symptoms are usually slowly progressive compared to those from vascular events (ischemic or hemorrhagic infarcts) which tend to be sudden in onset and slowly resolve, or electrical events (seizures) which tend to be sudden in onset and rapidly resolve. There are no findings that would allow differentiation of a metastatic tumor from a primary neoplasm on clinical grounds.

Signs and symptoms include:

1. those due to increased ICP from mass effect and/or blockage of CSF drainage (hydrocephalus):

A. headache (H/A): the most common presenting symptom, occurs in ≈ 50%

B. nausea/vomiting

2. focal deficits:

A. due to compression of brain parenchyma by mass and/or peritumoral edema (e.g. monoparesis without sensory disturbance)

B. due to compression of cranial nerve

3. seizures: occur only in ≈ 15% of cases

4. mental status changes: depression, lethargy, apathy, confusion

5. symptoms suggestive of a TIA (dubbed “tumor TIA”) or stroke, may be due to:

A. occlusion of a vessel by tumor cells

B. hemorrhage into the tumor, especially common with metastatic melanoma, choriocarcinoma, and renal-cell carcinoma⁶³³ (see Hemorrhagic brain tumors, page 1123). May also occur due to decreased platelet count

EVALUATION

IMAGING STUDIES (CT OR MRI)

Metastases usually appear as “non-complicated” masses on CT (i.e. round, well circumscribed), often arising at the gray/white junction. Characteristically, profound white matter edema (“fingers of edema”) reach deep into brain from the tumor, usually more pronounced than that seen with primary (infiltrating) brain tumors. When multiple lesions are present (on CT or MRI of brains with multiple mets) Chamber’s rule applies: “Whoever counts the most mets is right.” Mets usually enhance, and must be considered in the differential diagnosis of a ring-enhancing lesion.

Solitary supratentorial lesion on CT⁶¹⁴

• brain mets from solid tumors are solitary on CT in 50-65% of cases

• with negative Ca history, negative CXR and negative IVP (presumably, this would also apply if a chest/abdominal/pelvic CT was negative): 7% of solitary lesions are mets, 87% are primary brain tumors, and 6% are nonneoplastic. Yield of further work-up to find primary is low (recommendation: follow serial CXRs)

• with history of treated Ca: 93% of solitary lesions are mets

MRI

More sensitive than CT, especially in the p-fossa (including brain stem). Detects multiple mets in ≈ 20% of single mets on CT⁶¹³. Multiple projections may also assist in surgical planning.

LUMBAR PUNCTURE

Relatively contraindicated when there is a cerebral mass (may be indicated once mass lesion has been ruled out). May be most useful in diagnosing carcinomatous meningitis (see Carcinomatous meningitis, page 711).

METASTATIC WORK-UP

Prior to obtaining tissue from brain lesion: When metastatic disease is suspected based on imaging or on surgical tissue, a search for a primary site and assessment for other lesions may be considered since it may provide alternative sites for tissue for histologic diagnosis, and it may guide treatment (e.g. widely disseminated metastases may preclude aggressive therapy). Metastatic work-up should include:

1. CXR: to rule out lung primary or other mets to lung

2. CT of the chest (more sensitive that CXR), abdomen and pelvis: to rule out renal or GI primary (second choice: IVP) or liver mets

3. test stool for occult blood

4. radionuclide bone scan: for patients with bone pain or for tumors that tend to produce osseous metastases (especially: prostate, breast, kidney, thyroid & lung)

5. mammogram in women

6. prostate specific antigen (PSA) in men

7. PET scan: can detect small malignancies

Cancer of unknown primary site (CUP): If the metastatic work-up (see above) is negative, the pathology of a metastatic brain lesion may implicate specific primary sites.

Small-cell carcinoma metastatic to the brain is most likely from the lung. Stains positive for neuroendocrine stains (see page 721).

Adenocarcinoma: lung is the most common primary. Other sources: GI (mostly colon), breast. The primary site may remain occult even after extensive evaluation in up to 88%⁶³⁴. Immunostaining has been tried to identify the primary site but has not been found to be widely useful.

MANAGEMENT

With optimal treatment, median survival of patients with cerebral mets is still only ≈ 26-32 weeks, therefore management is mostly palliative (also see Outcome, page 710 for comparison of various treatments).

Confirming the diagnosis

Caution: 11% of patients with abnormalities on brain CT or MRI with a history of cancer (within past 5 yrs) do not have cerebral metastases⁶³⁵. Differential diagnoses include: glioblastoma, low grade astrocytoma, abscess, and nonspecific inflammatory reaction. If non-surgical treatment (e.g. chemotherapy or RTX) is being contemplated, the diagnosis should be confirmed by biopsy in almost all cases.

MANAGEMENT DECISIONS

Prognostication

This is critical since many treatment decisions depend on overall prognosis.

RTOG RPA: Radiation Therapy Oncology Group recursive partitioning analysis classification⁶³⁶ (see Table 21-68) (from 1200 patients with brain metastases undergoing XRT). Conclusion: the specific tumor type, length of time since diagnosis, etc. are not as important prognostically as the Karnofsky Performance Scale (KPS) score (see page 1182).

RPA Class 3 patients have been shown to be unlikely to benefit from any of numerous treatment modalities studied. Class 1 are more likely to benefit. Most patients are Class 2, and benefit is unclear.

Table 21-68 RPA Classification for patients with brain mets

RPA class	Description	Median survival (mos)*
1	• KPS† ≥ 70 and • age < 65 years and • no systemic disease	7
2	• all others‡	4
3	• KPS < 70	2

* for patients undergoing XRT

^† KPS = Karnofsky Performance Scale score (see page 1182)

^‡ i.e. not RPA class 1 or 3

Management algorithm

Table 21-69 shows a summary of management suggestions (details appear in following sections).

Also, surgical excision may be considered for patients with completely resectable brain mets who are candidates for chemotherapy with interleukin-2 (IL-2) for systemic disease (e.g. for renal-cell Ca or melanoma) since this drug in some case reports produces significant cerebral edema if there are cerebral mets as well.

MEDICAL MANAGEMENT

Initial treatment:

1. anticonvulsants: e.g. phenytoin. Generally not needed for posterior fossa lesions

2. corticosteroids: many symptoms are due to peritumoral edema (which is primarily vasogenic), and respond to steroids within 24-48 hrs. This improvement is not permanent, and prolonged steroid administration may produce side effects (see Possible deleterious side effects of steroids, page 33).

Rx typical dose for a patient with significant symptoms who is not already on steroids: dexamethasone (Decadron®) 10-20 mg IV, followed by 6 mg IV q 6 hrs for 2-3 days, after which it is converted to ≈ 4 mg PO QID. Once symptoms are controlled, this is tapered to ≈ 2-4 mg PO TID as long as symptoms do not worsen

3. H₂ antagonists (e.g. ranitidine 150 mg PO q 12 hrs)

CHEMOTHERAPY

Limitations of chemotherapy in the brain are discussed on page 589. If multiple lesions of known small-cell Ca are detected on cerebral imaging, treatment of choice is radiation plus chemotherapy.

Table 21-69 Management suggestions for cerebral metastases*

Clinical situation		Management
unknown primary or unconfirmed diagnosis		stereotactic biopsy for ≈ all patients if surgical excision is not a consideration
uncontrolled widespread systemic cancer & obviously short life expectancy and/or poor performance status (Karnofsky ≤ 70, see page 1182)		(biopsy as indicated above) + WBXRT or no treatment
Stable systemic disease & KPS > 70
solitary met	symptomatic, large, or accessible lesion	surgical excision + WBXRT
	asymptomatic, small, or inaccessible lesion	WBXRT ± SRS boost
multiple mets	single large lesion that is life threatening or producing mass effect	surgery for the large lesion + WBXRT for the rest
	≤ 3 lesions: symptomatic & can all be removed	surgery + WBXRT or SRS + WBXRT
	≤ 3 lesions: cannot all be removed	WBXRT or SRS + WBXRT
	> 3 lesions: with no mass effect requiring surgery	WBXRT638

* adapted⁶³⁷. Abbreviations: WBXRT = whole brain radiation therapy, SRS = stereotactic radiosurgery

RADIATION THERAPY

Caution: not all brain lesions in cancer patients are mets (see above).

In patients not considered for surgery, steroids and radiation usually help H/A, and in ≈ 50% of cases symptoms improve or completely resolve⁶³⁹. This does not result in local control for the majority of these patients and they frequently succumb from progressive brain disease.

The usual dose is 30 Gy in 10 fractions given over 2 weeks. With this dose, 11% of 1-yr survivors and 50% of 2-yr survivors develop severe dementia.

“Radiosensitivity” of various metastatic tumors to whole brain radiation therapy (WBXRT) are shown in Table 21-70.

Prophylactic cranial irradiation

Prophylactic cranial irradiation after resection of small-cell lung carcinoma (SCLC) reduces relapses in brain, but does not affect survival⁶⁴⁰.

Table 21-70 “Radiosensitivity” of brain metastases to WBXRT

Radiosensitivity	Tumor
Radiosensitive635	• small-cell lung Ca • germ-cell tumors • lymphoma • leukemia • multiple myeloma
Moderately sensitive	• breast
Moderately resistant	• colon • non small-cell lung cancer
Highly resistant*	• thyroid • renal cell (10% respond) • malignant melanoma • sarcoma • adenocarcinoma

* SRS may be better than WBXRT for these

Post-op radiation therapy

WBXRT is usually recommended following craniotomy for metastatic disease⁶⁴¹, especially with SCLC where “micro-metastases” are presumed to be present throughout brain^A.

A. some centers do not routinely administer post-op WBXRT (except for very radiosensitive tumors such as SCLC) but instead follow patients with serial imaging studies and administer XRT only when metastases are documented

Optimal dose is controversial. Early reports recommended 30-39 Gy over 2-2.5 weeks (3 Gy fractions) with or without surgery⁶⁴². This is acceptable in patients not expected to live long enough to get long-term radiation effects. Recent recommendations are for smaller daily fractions of 1.8-2.0 Gy to reduce neurotoxicity⁶⁴³. These low doses are also associated with a higher rate of recurrent brain metastases⁶⁴⁴. Since 50 Gy are needed to achieve > 90% control of micrometastases, some use 45-50 Gy WBXRT, plus a boost to the tumor bed to bring the total treatment up to 55 Gy, all with low fractions of 1.80-2.0 Gy⁶⁴⁵.

Stereotactic radiosurgery

Inconsistent in its ability to reduce tumor size. Some retrospective studies show results comparable to surgery⁶⁴⁶. Other do not⁶⁴⁷. Does not obtain tissue for histological analysis, and cannot be used for lesions > 3 cm. Also, see page 710.

SURGICAL MANAGEMENT

SOLITARY LESIONS

Indications favoring surgical excision of a solitary lesion:

1. primary disease quiescent

2. lesion accessible

3. lesion is symptomatic or life-threatening

4. primary tumor known to be relatively radioresistant (excision is rarely indicated for untreated brain metastases from SCLC because of its radiosensitivity)

5. for recurrent SCLC following XRT

6. diagnosis unknown: alternatively consider biopsy, e.g. stereotactic biopsy

Surgical resection in patients with progressive systemic disease and/or significant neurologic deficit is probably unjustified⁶⁴⁸. Also, in newly diagnosed cancer patients, craniotomy may delay systemic treatment for weeks and the ramifications of this need to be considered.

MULTIPLE LESIONS

Patients with multiple metastases generally have much worse survival than those with solitary lesions⁶⁴³. Multiple metastases are usually treated with XRT without surgery. However, if total excision of all mets is feasible, then even multiple mets may be removed with survival similar to those having a single met removed⁶⁴⁹ (also see Table 21-69 for summary). If only incomplete excision is possible (i.e. cannot remove all mets, or portions of 1 or more must be left behind) then there is no improvement in survival with surgery, and XRT alone is recommended. The mortality of removing > 1 met at a single sitting is not statistically significantly higher than removing a single met.

Situations where surgery may be indicated for multiple mets⁶⁵⁰:

1. one particular and accessible lesion is clearly symptomatic and/or life threatening (life-threatening lesions include p-fossa and large temporal lobe lesions). This is palliative treatment to reduce the symptom/threat from that particular lesion

2. multiple lesions that can all be completely removed (see above)

3. no diagnosis (e.g. no identifiable primary): consider stereotactic biopsy

STEREOTACTIC BIOPSY

Considered for:

1. lesions not appropriate for surgery. Includes cases with no definite diagnosis and:

A. deep lesions

B. multiple small lesions

2. patients not candidates for surgical resection

A. poor medical condition

B. poor neurologic condition

C. active or widespread systemic disease

3. to ascertain a diagnosis

A. when another diagnosis is possible: e.g. no other sites of metastases, long interval between primary cancer and detection of brain mets…

B. especially if nonsurgical treatment modalities are planned (see above)

INTRA-OPERATIVE CONSIDERATIONS FOR SURGICAL REMOVAL

Most lesions present themselves on the surface of the brain or through the dura. For lesions not visible on the surface nor palpable immediately beneath the surface, intraoperative ultrasound or stereotactic techniques may be used to localize the lesion.

Metastases usually have a well defined border, thus a plane of separation from normal brain may be exploited, often allowing gross total removal.

OUTCOME

Table 21-71 lists factors associated with better survival regardless of treatment. Also, the prognosis gets worse as the number of mets increases⁶³⁷. Median survival even with best treatment in some studies is only ≈ 6 months.

Natural history

By the time that neurologic findings develop, median survival among untreated patients is ≈ 1 month⁶⁵¹.

Steroids

Using steroids alone (to control edema) doubles survival⁶⁵² to 2 mos^A.

A. NB: this is based largely on pre-CT era data, and the tumors were therefore probably larger than in current studies⁶⁵³

Table 21-71 Factors associated with better prognosis for brain mets (with any treatment)

• Karnofsky score* (KPS) > 70 • age < 60 yrs • metastases to brain only (no systemic mets) • absent or controlled primary disease • > 1 yr since diagnosis of primary • the fewer the number of brain mets • female gender

* the KPS (see page 1182) is probably the most important predictor; those with a score of 100 had median survival > 150 weeks

Whole brain radiation therapy (WBXRT)

WBXRT + steroids increases survival to 3-6 mos⁶⁴⁹. 50% of deaths are due to progression of intracranial disease.

Surgery ± WBXRT

Recurrence of tumor was significantly less frequent and more delayed with the use of post-op WBXRT⁶⁴¹. Length of survival was unchanged with supplemental use of WBXRT. There is also an additional loss of cognitive function in many cases, and patients are rarely independent after WBXRT.

In 33 patients treated with surgical resection of single mets and post-op WBXRT⁶⁵⁴: median survival was 8 months; with 44% 1-yr survival. If no evidence of systemic Ca, 1-yr survival is 81%. If systemic Ca is present (active or inactive), 1-yr survival is 20%. Patients with solitary mets and no evidence of active systemic tumor have the best prognosis^{639, 648}. With total removal, no recurrence nor new parenchymal mets occurred within 6 months, and the major cause of death was progression of Ca outside the CNS. A randomized trial verified the improved longevity and quality of survival of patients with solitary mets undergoing surgical excision plus WBXRT vs. WBXRT alone (40 weeks vs. 15 weeks median survival)⁶³⁵. The surgical mortality was 4% (≈ same as 30-day mortality in the RTX-only group). More patients treated with WBXRT alone die of their brain mets than those who underwent surgery. Following total removal and post-op WBXRT, 22% of patients will have recurrent brain tumor at 1 year⁶⁴³. This is better than surgery without XRT (with reported failure rates of 46%⁶⁴³ and 85%⁶⁴⁴).

Stereotactic radiosurgery (SRS)

Also, see page 778. There has not been a randomized study to compare surgery to SRS. Retrospective studies suggest that SRS may be comparable to surgery^{646, 655}. How-ever, a prospective (non-randomized, retrospectively matched) study⁶⁴⁷ found a median survival of 7.5 mos with SRS vs. 16.4 mos with surgery, and a higher mortality from cerebral disease in the SRS group (with the mortality due to the SRS treated lesions and not new lesions). A local control rate of ≈ 88% has been reported, with one study also recommending WBXRT following the SRS for better regional control⁶⁵⁶.

Actuarial control rates at 1 year following SRS + WBXRT were 75-80% and appear to be similar to surgery + WBXRT⁶³⁷. However, SRS was unreliable in reducing tumor size.

Multiple mets

Patients with multiple mets that were totally removed have a survival that is similar to those having single mets surgically removed⁶⁴⁹ (see above).

21.6. Carcinomatous meningitis

Carcinomatous meningitis (CM) AKA (lepto) meningeal carcinomatosis (LMC). Found in up to 8% of patients autopsied with systemic Ca. Up to 48% may present with CM before the presence of systemic Ca is known. Most common primaries: breast, lung, then melanoma^{267 (p 610-2)}. Always include lymphomatous meningitis in the differential diagnosis (see CNS lymphoma, page 672).

CLINICAL

Simultaneous onset of findings in multiple levels of neuraxis. Multiple cranial nerve findings are frequent (in up to 94%, most common: VII, III, V & VI), usually progressive. Most frequent symptoms: H/A, mental status changes, lethargy, seizure, ataxia. Non-obstructive hydrocephalus is also common. Painful radiculopathies can occur with “drop mets”.

DIAGNOSIS

Lumbar puncture

Perform only after mass lesion has been ruled out with cranial CT or MRI. Although the initial LP may be normal, CSF is eventually abnormal in > 95%.

CSF should be sent for:

1. cytology to look for malignant cells (requires ≈ 10 ml for adequate evaluation for CM). Repeat if negative (45% positive on first study, 81% eventually positive after up to 6 LPs). May need to pass CSF through a millipore filter

2. bacterial and fungal cultures (including unusual organisms, e.g. cryptococcus)

3. tumor markers: carcinoembryonic antigen, alpha-fetoprotein

4. protein/glucose: elevated protein is the most common abnormality. Glucose may be as low as ≈ 40 mg% in about a third of patients

MRI

Contrast enhanced MRI is more sensitive in showing meningeal enhancement⁶⁵⁷.

CT

May show (mild) ventricular dilatation, enhancement of basal cisterns. Sulcal enhancement may also occur with involvement of the convexities.

Myelography

Spinal seeding (“drop mets”) will produce filling defects on myelography.

SURVIVAL

Untreated: < 2 months. With radiation therapy + chemotherapy: median survival is 5.8 mos (range 1-29). Chemotherapy may be given intrathecally. About half of patients die of CNS involvement, and half die of systemic disease.

21.7. Foramen magnum tumors

DIFFERENTIAL DIAGNOSIS

See Foramen magnum lesions on page 1212 for nonneoplastic lesions. Most foramen magnum (FM) region tumors are extra-axial. This includes:

1. meningioma: the anterior lip of the foramen magnum is the second most common site of origin of posterior fossa meningiomas. Meningiomas comprise 38-46% of FM tumors^{192, 193} (see page 614) and most are intradural

2. chordoma

3. neurilemmoma

4. epidermoid

5. chondroma

6. chondrosarcoma

7. metastases

8. exophytic component of a brainstem tumor

PRESENTATION

In the pre-imaging era (i.e. before CT & MRI) these lesions were often diagnosed relatively late due to the unusual associated clinical syndromes and the rarity of visualizing this region on myelography.

CLINICAL FINDINGS

Symptoms:

1. sensory

A. craniocervical pain: usually an early symptom, commonly in neck and occiput. Aching in nature. ↑ with head movement

B. sensory findings: usually occur later. Numbness and tingling of the fingers

2. motor

A. spastic weakness of the extremities: weakness usually starts in the ipsilateral UE, then the ipsilateral LE, then contralateral LE, and finally contralateral UE (“rotating paralysis”)

Signs:

1. sensory

A. dissociated sensory loss: loss of pain and temperature contralateral to lesion with preservation of tactile sensation

B. loss of position and vibratory sense, greater in the upper than the lower extremities

2. motor

A. spastic weakness of the extremities

B. atrophy of the intrinsic hand muscles: a lower motor nerve finding

C. cerebellar findings may rarely be present with extensive intracranial extension

3. long tract findings

A. brisk muscle stretch reflexes (hyperreflexia, spasticity)

B. loss of abdominal cutaneous reflexes

C. neurogenic bladder: usually a very late finding

4. ipsilateral Horner’s syndrome: due to compression of cervical sympathetics

5. nystagmus: classically downbeat (see page 828), but other types can occur

It had been postulated that long tract findings were due to direct compression at the cervicomedullary junction, and that lower motor nerve findings in the upper extremities were due to central necrosis of the grey matter as a result of compression of arterial blood supply. Anatomic study suggests that it is actually venous infarction at lower cervical levels (C8-T1) that is responsible for the lower motor neuron findings.

SURGICAL TREATMENT

Surgical approaches:

1. transoral approach: see page 176 for technique

A. disadvantage: cannot reach to > 1 cm to either side of midline

B. almost exclusively for extradural lesions (although some intra-axial lesions have been approached, the experience is extremely limited)

2. extreme lateral transcondylar approach

A. disadvantage: lack of familiarity of most neurosurgeons with this approach

B. advantage: excellent exposure of anterior foramen magnum with proximal control of vertebral artery

3. lateral posterior fossa approach: see Posterior fossa (suboccipital) craniectomy, page 152 for technique

A. disadvantage: cannot reach midline or contralateral component, however, some tumor in these regions may be pulled into the field as the tumor is debulked

21.8. Idiopathic intracranial hypertension (IIH)

Key concepts:

• papilledema and symptomatic ICP elevation > 25 cm H₂O in the absence of intracranial mass or infection. Often associated with dural sinus thrombosis

• a preventable cause of (often permanent) blindness from optic atrophy

• more common in obese females of childbearing age than general population

• recommended work-up:

preferred imaging studies: brain MRI (without & with contrast) and MRV. Imaging should be normal (allowed exception: slit-like ventricles)

LP. Findings: opening pressure (> 20 cm H₂O) & normal CSF analysis

ophthalmologic eval: test visual fields, acuity, and check for papilledema

• usually self-limited, recurrence is common, chronic in some patients

• risk of blindness is not reliably correlated to duration of symptoms, papilledema, H/A, Snellen visual acuity, or number of recurrences

• treatment for patients failing medical management (weight loss, Diamox…):

optic nerve sheath fenestration (ONSF) is best for visual loss without H/A

CSF shunt may be better than ONSF for H/A associated with visual loss

AKA pseudotumor cerebri, AKA benign intracranial hypertension, (plus numerous other obsolete terms⁶⁵⁸) is a heterogeneous group of conditions characterized by increased intracranial pressure with no evidence of intracranial mass, hydrocephalus, infection (e.g. chronic fungal meningitis), or hypertensive encephalopathy. Some, but not all, authors exclude patients with intracranial hypertension in the presence of dural sinus thrombosis. IIH is thus a diagnosis of exclusion. There is a juvenile and an adult form.

EPIDEMIOLOGY

1. female:male ratio reported from 2:1 to 8:1 (no gender difference in juvenile form)

2. obesity is reported in 11-90% of cases, and is not as prevalent in men⁶⁵⁹

3. incidence among obese women of childbearing years^{660, 661}: 19-21/100,000, (whereas incidence in general population⁶⁵⁸: 1-2/100,000)

4. peak incidence in 3rd decade (range: 1-55 years). 37% of cases are in children, 90% of these are age 5-15 years. Very rare in infancy

5. frequently self limited (recurrence rate: 9-43%)

6. severe visual deficits develop in 4-12%, unrelated to duration of symptoms, degree of papilledema, headache, visual obscuration, and number of recurrences⁶⁶². Perimetry is the best means to detect and follow visual loss

PATHOGENESIS

Not fully understood. Increased cerebral edema & brain water content, increased venous pressure & cerebral blood volume, and reduced CSF absorption have all been demonstrated. Theories that also explain the high prevalence in obese females:

1. mechanical theory: obesity → ↑ intraabdominal pressure → ↑ central venous pressure → ↓ CSF resorption → ↑ ICP^A

2. hormonal theory: adipocytes convert androstenedione → estrone → ↑ CSF production

A. however, other studies have indicated that elevated venous pressure may actually be an epiphenomenon to a primary increase in ICP⁶⁶³

Table 21-72 Modified Dandy’s criteria for IIH

• signs & symptoms of increased ICP • no localizing signs other than Cr. N VI palsy* in an otherwise awake and alert patient • increased CSF pressure without chemical or cytological abnormalities • normal to small ventricles and no intracranial mass

* may result from ↑ ICP (see page 836)

DIAGNOSTIC CRITERIA

Modified Dandy’s criteria are shown in Table 21-72.

More specifically, four diagnostic criteria⁶⁶⁴:

1. CSF pressure: > 25 cm H₂O (pressures > 40 are not uncommon)^A. Pressures 20-24.9 are nondiagnostic⁶⁶⁵. Pressure < 20 is normal

2. CSF composition: normal glucose and cell count. Protein is normal, or in ≈ two thirds of cases it is low (< 20 mg%)

3. symptoms & signs are those of elevated ICP alone (i.e. papilledema & H/A) with no focal findings (allowed exception: abducens nerve palsy which may be due to increased ICP, see page 836)

4. normal radiologic studies of the brain (CT or MRI) with the allowed exceptions of:

A. the occasionally seen slit ventricles (the incidence may be no higher in IIH than in age-matched controls⁶⁶⁶) or empty sella

B. infantile form may have generous ventricles and large fluid spaces over brain

C. intra-orbital abnormalities may be seen: see below

CLINICAL

PRESENTATION^{664, 667}

• symptoms

A. classic (major) symptoms

1. H/A (the most common symptom): 94-99%. Typically retro-ocular and pulsatile. May ↑ with eye movement. Severity does not correlate with degree of CSF pressure elevation. Occasionally worse in A.M.

2. nausea: 32% (actual vomiting is less common)

3. visual loss (see Visual loss in IIH below):

a. transient visual obscuration (TVO)

b. permanent afferent visual pathway injury

4. diplopia (more common in adult, usually due to VI nerve palsy): 30%

B. minor symptoms⁶⁶⁸

1. neck stiffness: 30-50%

2. tinnitus^B: up to 60%. Usually pulse synchronous. Described as rushing noise. May be unilateral (in these, may be reduced by ipsilateral jugular vein compression + ipsilateral head rotation)

3. ataxia: 4-11%

4. acral paresthesias: 25%

5. retrobulbar eye pain on eye movements

6. arthralgia^B: 11-18%

7. dizziness: 32%

8. fatigue

9. reduced olfactory acuity

• signs (generally restricted to visual system)

A. eye findings - also see Visual loss in IIH below

1. papilledema:

a. present in almost ≈ 100%

b. idiopathic intracranial hypertension without papilledema (IIHWOP)⁶⁶⁹: a variant of IIH. Visual loss tends not to occur

c. usually bilateral, occasionally unilateral⁶⁷⁰

d. may be mild (subtle nerve fiber elevation)

2. abducens nerve (Cr. N. VI) palsy: 20% (a false localizing sign, see page 836). The esotropia ranges from < 5 prism diopters dysconjugate angle in primary gaze to > 50671

3. visual acuity: relatively insensitive assessment of visual function

4. visual field defect: 9%.

a. early changes: peripheral fields & nasal quadrant defect

b. enlarged blind spot (66%) and concentric constriction of peripheral fields (blindness is very rare at presentation)

B. infantile form may have only enlarging OFC, frequently self limited, usually requires only follow-up without specific treatment

A. diurnal variations in CSF pressure may occasionally cause a falsely low (i.e normal) reading. if clinical suspicion is high, an LP at a different time of day or continuous ICP monitoring may be required

B. the causal relationship with IIH has been demonstrated by resolution of these symptoms with reduction of CSF pressure

C. conspicuous absence of altered level of consciousness in spite of high ICP

Worsening of any of the above symptoms with postural changes that increase ICP (bending over, Valsalva maneuver…) is characteristic in idiopathic intracranial hypertension.

Visual loss in IIH

Quoted range of occurrence in IIH: 48-68% (lower numbers generally come from population based samples). A prospective study found changes by Goldman perimetry in 96% of 50 patients⁶⁷². The only parameter associated with worsening vision is recent weight gain.

Pathomechanics: Increased ICP is transmitted along optic nerve sheath → circumferential compression of the retinal ganglion cell axons at the level of the lamina cribrosa⁶⁷¹.

Manifestations:

1. transient visual obscurations (TVO): graying or blacking out of vision. Lasts ≈ 1 second. Uni- or bilateral. Typically occur with eye movement, bending over or valsalva maneuver. Directly proportional to severity of papilledema. Frequency of TVOs parallels ICP elevation, but doesn’t correlate with permanent visual loss

2. visual loss in IIH may occur early or late, may be sudden or gradually progressive, and is not reliably correlated to duration of symptoms, papilledema, H/A, Snellen visual acuity, or number of recurrences. It may escape detection until profound.

A. early: usually constriction of fields and loss of color ( perimetry is the best test for following vision in IIH)

B. late: central vision is affected. Findings include: concentric constrictions, enlargement of the blind spot, inferior nasal defects, arcuate defects, cecocentral scotomas…

Table 21-73 Conditions that may be associated with IIH⁶⁷³

Proven association

Meets 4 criteria from Table 21-74

• obesity

Likely association

Meets 3 criteria from Table 21-74

• drugs: keprone, lindane • hypervitaminosis A

Probable association

Meets 2 criteria from Table 21-74

• steroid withdrawal* • thyroid replacement in children • ketoprofen & indomethacin in Bartter syndrome • hypoparathyroidism • Addison’s disease* • uremia • iron deficiency anemia • drugs: tetracycline, nalidixic acid, Danazol, lithium, amiodarone, phenytoin, nitrofurantoin, ciprofloxacin, nitroglycerin

Possible association

Meets 1 criterion from Table 21-74

• menstrual irregularity • oral contraceptive use† • Cushing’s syndrome • Vitamin A deficiency • minor head trauma • Behçet syndrome

Unlikely association

Meets none of the criteria in Table 21-74

• hyperthyroidism • steroid use • immunization

Unsupported association

• pregnancy • menarche

* may respond to steroids

^† may be associated with dural sinus thrombosis, see text

ASSOCIATED CONDITIONS

By definition, IIH is idiopathic. However, often what is considered “IIH” may actually be secondary to some other condition (e.g. transverse sinus thrombosis, see below). Many conditions cited as associations with IIH may be coincidental. Four criteria suggested to establish a cause-effect relationship are shown in Table 21-74⁶⁶⁷.

Table 21-74 Criteria for causality of IIH by another condition⁶⁶⁷

1. meets Dandy’s criteria (Table 21-72, page 713) 2. the condition should be proven to increase ICP 3. treatment of the condition should improve the IIH 4. properly controlled studies should show an association between the condition and IIH

Table 21-73 shows a scale⁶⁷³ to rank the likelihood of association between various conditions and IIH based on the number of the criteria met in Table 21-74.

Other conditions not included in this list that meet minimal criteria but are unconfirmed in case-control studies⁶⁵⁸ include:

1. other drugs: isotretinoin (Accutane®), trimethoprim-sulfamethoxazole, cimetidine, tamoxifen

2. systemic lupus erythematosus (SLE)

Conditions that may be related by virtue of increased pressure in the dural sinuses (see below) (some have called this “secondary IIH” which is an oxymoron):

1. otitis media with petrosal extension (so-called otitic hydrocephalus)

2. radical neck surgery with resection of the jugular vein

3. hypercoagulable states

VENOUS HYPERTENSION & SINOVENOUS ABNORMALITIES

Venous hypertension has often been proposed as a unifying underlying cause of IIH. Abnormalities of the dural sinuses, including thrombosis, stenosis⁶⁷⁴, obstruction, or elevated pressure have been demonstrated. While these findings may underlie a significant number of cases, they may in actuality be epiphenomena (e.g. venous hypertension may be due to compression of the transverse sinuses by elevated intracranial pressure⁶⁶³), and it is unlikely that such abnormalities will explain all cases.

DIFFERENTIAL DIAGNOSIS

1. true mass lesions: tumor, cerebral abscess, subdural hematomas, rarely gliomatosis cerebri may be undetectable on CT and will be misdiagnosed as IIH

2. cranial venous outflow impairment (some authors consider these as IIH)⁶⁷⁵

A. dural sinus thrombosis (see above and page 1166)

B. congestive heart failure

C. superior vena cava syndrome

D. unilateral or bilateral jugular vein or sigmoid sinus⁶⁷⁶ obstruction

E. hyperviscosity syndromes

F. Masson’s vegetant intravascular hemangioendothelioma⁶⁷⁷: an uncommon, usually benign lesion that may rarely involve the neuraxis (including intracranial occurrence)

3. Chiari I malformation (CIM): may produce findings similar to IIH. 6% of IIH patients have significant tonsillar ectopia, and ≈ 5% of patient with CIM have papilledema⁶⁷¹

4. infection (CSF will be abnormal in most of these): encephalitis, arachnoiditis, meningitis (especially basal meningitis or granulomatous infections, e.g. syphilitic meningitis, chronic cryptococcal meningitis), chronic brucellosis

5. inflammatory conditions: e.g. neurosarcoidosis (see page 71), SLE

6. vasculitis: e.g. Behçet’s syndrome

7. metabolic conditions: e.g. lead poisoning

8. pseudopapilledema (anomalous elevation of the optic nerve head) associated with hyperopia and drusen. Retinal venous pulsations are usually present. Especially deceptive when a patient with migraines has pseudopapilledema: treat the H/A

9. malignant hypertension: may produce H/A & bilateral optic disc edema which can be indistinguishable from papilledema. May also produce hypertensive encephalopathy (see page 73). Check BP in all IIH suspects

10. meningeal carcinomatosis

11. Guillain-Barré syndrome: CSF protein is usually elevated (see page 66)

12. following head trauma

EVALUATION RECOMMENDATIONS

Most tests are intended to rule out conditions that may mimic IIH:

1. cerebral imaging: cerebral CT or MRI (see below) scan without and with contrast

2. LP:

A. measure opening pressure (OP) with patient in lateral decubitus position

B. CSF analysis to rule-out infection (e.g. fungus, TB or Lyme disease), inflammation (e.g. sarcoidosis, SLE) or neoplasm (e.g. carcinomatous meningitis)

1. protein/glucose

2. cell count

3. routine & fungal cultures

4. cytology if suspicion of carcinomatous meningitis

3. routine labs: CBC, electrolytes, PT/PTT

4. W/U for sarcoidosis (see page 71) or SLE if other findings suggestive (e.g. cutaneous nodules, hypercoagulable state…)

5. neuro-ophthalmologic evaluation is recommended. Includes: visual field testing using quantitative perimetry, with evaluation of size of blind spot, slit-lamp examination ± fundus photographs

6. check BP to R/O malignant HTN → hypertensive encephalopathy (see page 716)

CT

CT without and with IV contrast is usually adequate to R/O intracranial mass, but may miss cases of dural sinus thrombosis. MRI & MRV are preferred.

MRI

Intracranial abnormalities are usually absent or minimal (slit ventricles, empty sella in 30-70%). However, intraorbital findings may be more substantial and include⁶⁷¹:

1. flattening of the posterior sclera: occurs in 80%

2. enhancement of the prelaminar optic nerve: in 50%

3. distention of the perioptic subarachnoid space: in 45%

4. vertical tortuosity of the orbital optic nerve: in 40%

5. intraocular protrusion of the prelaminar optic nerve: in 30%

Venography

Conventional venography or MR venography (MRV) to rule-out dural sinus or venous thrombosis.

TREATMENT AND MANAGEMENT

NATURAL HISTORY

Spontaneous resolution is common, sometimes within months, but usually after ≈ 1 year. Papilledema persists in ≈ 15%. Permanent visual loss occurs in 2-24%. Persistent H/A may occur in some. Recurs in ≈ 10% after initial resolution⁶⁷¹.

INTERVENTIONS

Studies are often difficult to interpret especially since spontaneous remission is common.

1. all patients must have repeated thorough ophthalmologic exams (see above)

2. stop possible offending drugs

3. weight loss: a weight loss of 6% usually results in complete resolution of papilledema⁶⁷⁸. However, resolution may be too slow for acutely threatened vision. Weight loss is also associated with reduction of other health risks of obesity. Symptoms recur if the weight is regained

A. dieting: 679rarely accomplished or sustained

B. bariatric surgery: gastric bypass, laparoscopic banding…

4. treatment of asymptomatic IIH patients is controversial as there is no reliable predictor for visual loss. Close follow up with serial formal visual field evaluation is necessary. Intervention is recommended in unreliable patients, or whenever visual fields deteriorate. It is possible to lose vision without H/A or papilledema

5. most cases remit by 6-15 weeks, however relapse is common

6. medical treatment

A. fluid and salt restriction

B. diuretics (slows CSF production)

1. carbonic anhydrase (CA) inhibitors:

a. acetazolamide (Diamox®): Rx start at 125-250 mg PO q 8-12 hrs, or long acting Diamox Sequels® 500 mg PO BID. Increase by 250 mg/day until symptoms improve, side effects occur, or 2 gm/day reached. SIDE EFFECTS: (in high doses): acral paresthesias, nausea, metabolic acidosis, altered taste, renal calculi, drowsiness. Rare: Stevens-Johnson syndrome, toxic epidermal necrolysis, agranulocytosis. ✖Contraindicated with allergy to sulfa or a history of renal calculi

b. methazolamide (Neptazane®): better tolerated but less effective. Rx 50-100 mg PO BID-TID. SIDE EFFECTS: similar to acetazolamide

c. topiramate (Topamax®): anticonvulsant with secondary inhibition of CA. Rx 200 mg PO BID. SIDE EFFECTS: Similar to acetazolamide, but can be used in sulfa allergic patients

2. furosemide (Lasix®)

a. start: 160 mg per day in adults, adjust per symptoms and eye exam (not to CSF pressure)

b. if ineffective, double (320 mg/day)

c. monitor K⁺ levels and supplement as needed

C. if ineffective, add steroids (options: dexamethasone (Decadron®) 12 mg/day, prednisone 40-60 mg/day, or methylprednisolone 250 mg IV q 6 hrs). May ↑ CSF resorption in cases of inflammation or venous thrombosis. Can be used as temporizing agents for patients awaiting surgery. A reduction in symptoms should occur by 2 weeks, after which time the steroid should be tapered over 2 weeks. Long-term use is not recommended due to, among other things, associated weight gain

7. surgical therapy^{267 (p 250-3)} only for cases refractory to above, or where visual loss is progressive or is severe initially or unreliable patient:

A. serial LPs until remission (25% remit after 1st LP⁶⁸⁰): remove up to 30 ml to halve OP, perform qod until OP < 20 cm H₂O, then decrease to q wk (no patient who had remission by 2nd LP had OP > 350 on 1st LP). Use a large gauge needle (e.g. 18 Ga) which may help promote a post-LP CSF leak into subcutaneous tissues. LPs may be difficult in obese patients. Revisions may be required in up to 50%. SIDE EFFECTS: include sciatica from nerve root irritation, acquired cerebellar tonsillar herniation (see page 317), spinal H/A (from intracranial hypotension)

B. shunts

1. lumbar shunt: usually lumboperitoneal (for insertion technique, see page 213). May be difficult in obese patient. May need a horizontal-vertical valve (see page 317) to prevent H/A from intracranial hypotension. Alternative: lumbopleural shunt

2. other shunts may be used, especially when arachnoiditis precludes use of lumbar subarachnoid space, e.g.:

• VP shunt: often difficult since the ventricles are frequently small or slit-like⁶⁸¹. Stereotactic techniques may make this more technically feasible

• cisterna magna shunt: may shunt to vascular system

C. optic nerve sheath fenestration: see below

D. obsolete treatment (presented for historical interest): subtemporal (advocated by Dandy) or suboccipital decompression. Usually bilateral silver-dollar size craniectomies under temporalis muscle to floor of middle fossa, open dura, cover brain with absorbable sponge, close fascia and muscle water-tight, anticonvulsants were started due to risk of post-op seizures

8. interventional procedures: venous sinus stenting may be considered for refractory cases⁶⁸²

9. patients should be followed at least two years (with repeat imaging, e.g. MRI) to R/O occult tumor

Optic nerve sheath fenestration (ONSF)^683-685

Generally better for protection of vision and reversal of papilledema than for other symptoms (e.g. H/A). Performed via medial or less commonly a lateral orbitotomy or transconjunctival medial approach. May reverse or stabilize visual deterioration⁶⁸⁶ and sometimes (but not always) lowers ICP (by continued CSF filtration) and may protect the contralateral eye (if not, contralateral ONSF must be performed). Has succeeded in cases where visual loss progressed after LP shunting⁶⁸⁷, possibly due to poor communication between orbital and intracranial subarachnoid spaces. SIDE EFFECTS: potential adverse include: pupillary dysfunction, peripapillary hemorrhage, chemosis, chorioretinal scarring⁶⁸⁸, diplopia (usually self-limited) from medial rectus disruption. Repeat fenestration is needed in 0-6%⁶⁷¹.

MANAGEMENT RECOMMENDATIONS FOR SPECIFIC SITUATIONS

Weight loss should be attempted in all.

1. IIH patients with H/A and no visual loss: medical therapy to control ↑ ICP and H/A. ONSF not recommended. Shunting is an option if medical management fails

2. IIH with visual loss without H/A:

A. mild visual loss: acetazolamide 500-1500 mg/d, follow-up q 2 weeks

B. moderate visual loss: acetazolamide 2000-3000 mg/d, follow-up q week

C. severe visual loss, moderate visual loss that doesn’t respond to acetazolamide, or optic disc at risk:

1. methylprednisolone 250 mg IV q 6 hrs + acetazolamide 1000 mg PO BID

2. if no improvement: ONSF. Consider shunt if ICP > 300 mm H₂O

3. IIH with visual loss AND H/A: for patients with surgical indications, either surgical procedure is appropriate. Shunting may relieve both problems simultaneously. ONSF may be more reliable to relieve the visual problems (the failure rate may be lower than the shunt malfunction rate) but is not as good for the H/A

4. IIHWOP: symptomatic treatment for H/A, diuretics

5. IIH in children and adolescents:

A. may be seen with withdrawal of steroids used for asthma

B. search for and correction of underlying etiology (offending drugs listed above, hypercalcemia, cancer…)

C. acetazolamide has been used with success

6. IIH in pregnancy:

A. women who first present with IIH during pregnancy: resolution of IIH following delivery is common

B. women who become pregnant during therapy:

1. 1st trimester: observation, limitation of weight gain, serial LPs. ✖ Acetazolamide should be avoided because of teratogenicity

2. 2nd & 3rd trimester: acetazolamide has been used safely, but involvement of high-risk obstetrician specialist is advised

7. pseudopapilledema (associated with drusen, etc., in the absence of IIH): no interventions⁶⁷¹. Reassurance and H/A management are employed

21.9. Empty sella syndrome

Empty sella syndrome (ESS) can be “primary” or “secondary”.

PRIMARY EMPTY SELLA SYNDROME

Occurs in the absence of prior treatment of a pituitary tumor (medical, surgical or XRT). Herniation of the arachnoid membrane into the sella turcica⁶⁸⁹ which can act as a mass, probably as a result of repeated CSF pulsation. The sella can become enlarged (see Sella turcica, page 138 for normal dimensions) and the pituitary gland may become compressed against the floor.

Frequent association: female sex (female:male ratio = 5:1), obesity and HTN. The frequency of intrasellar arachnoid herniation is higher in patients with pituitary tumors and in those with increased intracranial pressure for any reason (including idiopathic intracranial hypertension, see page 713) than in the general population.

These patients usually present with symptoms that do not suggest an intrasellar abnormality including: headache (the most common symptom), dizziness, seizures… Occasionally patients may develop CSF rhinorrhea⁶⁹⁰, deterioration of vision (acuity or field deficit resulting from kinking of optic chiasm due to herniation into the sella), or amenorrhea-galactorrhea syndrome.

Clinically evident endocrine disturbances are rare with primary ESS, however up to 30% have abnormal pituitary function tests, most commonly reduced growth hormone secretion following stimulation. Mild elevation of prolactin (PRL) and reduction of ADH may occur, probably from compression of the stalk. These patients show a normal PRL rise with TRH stimulation (whereas patients with prolactinomas do not).

Treatment: Surgical treatment is usually not indicated, except in the case of CSF rhinorrhea. In this setting, it is necessary to determine if there is increased ICP, and if so, if there is an identifiable cause. Simple shunting for hydrocephalus runs the risk of producing tension pneumocephalus from air drawn in through the former leak site. This may necessitate transsphenoidal repair with simultaneous external lumbar drainage, to be converted to a permanent shunt shortly thereafter. Hyperprolactinemia may be treated e.g. with bromocriptine (see page 651) if it interferes with gonadal function.

SECONDARY EMPTY SELLA SYNDROME

Entities associated with secondary empty sella syndrome:

1. following trauma⁶⁹¹

2. after successful transsphenoidal removal or XRT for a pituitary tumor⁶⁹¹

3. any cause of increased intracranial pressure, including: idiopathic intracranial hypertension (pseudotumor cerebri), Chiari malformation

Often presents with visual deterioration due to herniation of the optic chiasm into the empty sella. There may be hypopituitarism from the underlying cause.

Visual deterioration may be treated with chiasmopexy (propping up the chiasm) usually by transsphenoidal approach and packing the sella with fat, muscle or cartilage.

21.10. Tumor markers

TUMOR MARKERS USED HISTOLOGICALLY IN NEUROSURGERY

GLIAL FIBRILLARY ACIDIC PROTEI (GFAP)

Polypeptide, MW = 49,000 Daltons. Although the presence of GFAP usually indicates astroglial origin, it may occasionally be seen in oligodendrogliomas, ependymomas, and choroid plexus papillomas^{197 p (30-1)}. GFAP is only rarely found outside the CNS (in nonmyelinated Schwann cells, epithelium of the lens, hepatic Kupffer cells…). Thus, the presence of GFAP in a tumor found in the CNS is usually taken as good evidence for glial origin of the tumor. GFAP also occurs in normal brain parenchyma.

S-100 PROTEIN

A low molecular weight (21,000 Daltons) calcium-binding protein. Used on tissue microscopy for pathology. May participate in regulation of microtubule assembly. In CNS tumors, the distribution is similar to GFAP, but it is not as specific as GFAP (may be found in other cell types such as stellate cells of the adenohypophysis, chondrocytes)^{197 p (34-5)}, melanomas. In the peripheral nervous system, it is localized in Schwann cells. May be helpful in distinguishing Schwann cells from perineurial cells.

Clinically has been measured in serum (see below).

CYTOKERATIN (HIGH & LOW MOLECULAR WEIGHT)

Stains epithelial cells. Most primary brain tumors do not stain positive. Supports the diagnosis of carcinoma. May help distinguish metastatic tumors (when a positive stain occurs) from primary CNS tumors.

MIB-1 (AKA MONOCLONAL MOUSE ANTI-HUMAN KI-67ANTIBODY)

The Ki-67 antigen is expressed in all phases of the cell cycle except G0. A valuable marker of cell proliferation, but can only be used with fresh-frozen specimens. MIB-1 is a monoclonal antibody developed using recombinant parts of the Ki-67 protein as an immunogen, and can be used on paraffin-embedded sections of fixed tissue. Cells leaving the G₀/G₁-phase and entering the S-phase (performing DNA synthesis) stain positive with MIB-1 immunohistochemical stain. A high MIB-1 labeling index denotes high mitotic activity which often correlates with degree of malignancy. Most often used in lymphomas and breast cancer. For use in astrocytomas see page 596, for meningiomas see page 616.

NEUROENDOCRINE STAINS

Includes:

1. chromogranin: stains for neural crest derivatives, viz. pituitary adenomas, paragangliomas, neuroendocrine tumors

2. synaptophysin: stains neuronal and pineal tumors, PNET & medulloblastomas

3. neuron specific enolase (NSE): very sensitive but not specific for neuroendocrine

Metastases that are positive for neuroendocrine stains include: small-cell carcinoma of the lung, malignant pheochromocytoma, Merkel cell tumor. Metastatic small-cell tumors to the brain staining positive for neuroendocrine stains are almost all due to lung.

STAINING PATTERNS⁶⁹²

An individual tumor may lack a marker that is typically representative of its type. Therefore, a positive stain is more significant than a negative stain. General staining patterns are shown in Table 10-75.

TUMOR MARKERS USED CLINICALLY

HUMAN CHORIONIC GONADOTROPIN (HCG)

A glycoprotein, MW = 45,000. Secreted by placental trophoblastic epithelium. Beta chain (ß-hCG) is normally present only in the fetus or in gravid or postpartum females, otherwise it indicates disease. Classically associated with choriocarcinoma (uterine or testicular), also found in patients with embryonal cell tumors, teratocarcinoma of testis, and others.

CSF ß-hCG is 0.5-2% of serum ß-hCG in non-CNS tumors. Higher levels are diagnostic of cerebral mets from uterine or testicular choriocarcinoma, or primary choriocarcinoma or embryonal cell carcinoma of pineal (see page 691) or suprasellar region.

ALPHA-FETOPROTEIN

Alpha-fetoprotein (AFP) is a normal fetal glycoprotein (MW = 70,000) initially produced by the yolk sac, and later by the fetal liver. It is found in the fetal circulation throughout gestation, and drops rapidly during the first few weeks of life, reaching normal adult levels by age 1 yr. It is detectable only in trace amounts in normal adult males or nonpregnant females. It is present in amniotic fluid in normal pregnancies, and is detectable in maternal serum starting at ≈ 12-14 weeks gestation, increasing steadily throughout pregnancy until ≈ 32 weeks⁶⁹³.

Abnormally elevated serum AFP may occur in Ca of ovary, stomach, lung, colon, pancreas, as well as in cirrhosis or hepatitis and in the majority of gravid women carrying a fetus with an open neural tube defect (see Prenatal detection of neural tube defects, page 245). Serum AFP > 500 ng/ml usually means primary hepatic tumor.

CSF-AFP is elevated in some pineal region germ-cell tumors (see page 692). 16-25% of patients with testicular tumors get cerebral mets and elevated CSF AFP levels are reported in some.

CARCINOEMBRYONIC ANTIGEN (CEA)

A glycoprotein, MW = 200,000. Normally present in fetal endodermal cells. Originally described with colorectal adeno-Ca, now known to be elevated in many malignant and nonmalignant conditions (including cholecystitis, colitis, diverticulitis, hepatic involvement from any tumor, with 50-90% of terminal patients having elevation).

CSF CEA: levels > 1 ng/ml are reported with leptomeningeal spread of lung Ca (89%), breast Ca (60-67%), malignant melanoma (25-33%), and bladder Ca. May be normal even in CEA secreting cerebral mets if they don’t communicate with the subarachnoid space. Only carcinomatous meningitis from lung or breast Ca consistently elevates CSF CEA in the majority of patients.

S-100 PROTEIN

Serum S-100 protein levels rise after head trauma, and possibly after other insults to the brain. Levels may also be elevated in Creutzfeldt-Jakob disease (see page 363).

21.11. Neurocutaneous disorders

Formerly called phakomatoses. Neurocutaneous disorders (NCD) are a group of conditions, each with unique neurologic findings and benign cutaneous lesions (NB: both skin and the CNS derive embryologically from ectoderm), usually with dysplasia of other organ systems (often including the eyes). With the exception of ataxia-telangiectasia (not discussed here) all exhibit autosomal dominant inheritance. There is also a high rate of spontaneous mutations. These syndromes should be kept in mind in a pediatric patient with a tumor, and other stigmata of these syndromes should be sought.

NCDs that are more likely to come to the attention of the neurosurgeon:

1. neurofibromatosis: see below

2. tuberous sclerosis: see page 725

3. von Hippel-Lindau disease: see page 667

4. Sturge Weber syndrome: see page 726

5. racemose angioma (Wyburn-Mason syndrome): midbrain and retinal AVMs

21.11.1. Neurofibromatosis

Neurofibromatosis (NFT) is the most common of the NCDs. There are as many as 6 distinct types, the two most common of which (NF1 & NF2) are compared in Table 21-76 (variant forms also occur).

Schwannoma vs. neurofibroma: While similar in many ways, these tumors differ histologically. Schwannomas (nee: neurilemmomas) arise from schwann cells which produce myelin. Neurofibromasconsist of neurites (axons or dendrites of immature or developing neurons), Schwann’s cells, and fibroblasts within a collagenous or myxoid matrix. In contrast to shwannomas which displace axons (centrifugal), neurofibromas are unencapsulated and engulf the nerve of origin (centripetal). Neurofibromas may occur as solitary lesions, or, may be multiple as part of NF1 in the setting of which there is potential for malignant transformation. Both tumors have Antoni A (compact) and Antoni B (loose) fibers, but neurofibromas tend to have more Antoni B fibers. A patient ≤ 30 years age with a vestibular schwannoma is at increased risk of having NF2.

Table 21-76 Comparison of neurofibromatosis 1 &2⁶⁹⁴

current designation →	Neurofibromatosis 1 (NF1)	Neurofibromatosis 2 (NF2)
see page	723	724
alternate term	von Recklinghausen’s	bilateral acoustic NFT AKA MISME syndrome
obsolete term	peripheral NFT	central NFT
U.S. prevalence	100,000 people	≈ 3000 people
incidence	1/3000 births	1 in 40,000
inheritance	AD	AD
sporadic occurrence	30-50%	> 50%
gene locus	17 (17q11.2)	22 (22q12.2)
gene product	neurofibromin	schwannomin (merlin)
vestibular schwannomas (VS)	almost never bilateral	bilateral VSs are the hallmark
cutaneous schwannomas	no	70%
Lisch nodules	very common	not associated
cataracts	not associated	60-80%
skeletal anomalies	common	not associated
pheochromocytoma	occasional	not associated
MPNST*	≈ 2%	not associated
associated intramedullary spinal cord tumors	astrocytoma	ependymoma
intellectual impairment	associated	not associated

* malignant peripheral nerve sheath tumor

NEUROFIBROMATOSIS 1 (NF1 AKA VON RECKLINGHAUSEN’S DISEASE⁶⁹⁵)

CLINICAL FEATURES

More common than NF2, representing > 90% of cases of neurofibromatosis.

Diagnostic criteria: Are shown in Table 21-77⁶⁹⁶.

Associated conditions:

1. Schwann-cell tumors on any nerve (but bilateral VSs are virtually nonexistent)

2. spinal and/or peripheral-nerve neurofibromas

3. multiple skin neurofibromas

4. aqueductal stenosis: see page 241

5. macrocephaly: secondary to aqueductal stenosis and hydrocephalus, increased cerebral white matter

6. intracranial tumors: hemispheric astrocytomas are the most common, solitary or multicentric meningiomas (usually in adults). Gliomas associated with NF1 are usually pilocytic astrocytomas. Brain stem astrocytomas include both contrast-enhancing pilocytic lesions and those that are non-enhancing and radiologically diffuse

7. unilateral defect in superior orbit → pulsatile exophthalmos

8. neurologic or cognitive impairment: 30-60% have mild learning disabilities

9. kyphoscoliosis (seen in 2-10%, often progressive which then requires surgical stabilization)

10. visceral manifestations from involvement of autonomic nerves or ganglia within the organ. Up to 10% of patients have abnormal gastrointestinal motility/neuronal intestinal dysplasia related to neuronal hyperplasia within submucosal plexus

11. ≈ 20% develop plexiform neurofibromas: tumors from multiple nerve fascicles that grow along the length of the nerve. Almost pathognomonic for NF1⁶⁹⁷

12. syringomyelia

13. malignant tumors that have increased frequency in NFT: neuroblastoma, ganglioglioma, sarcoma, leukemia, Wilm’s tumor, breast cancer⁶⁹⁸

14. pheochromocytoma: is occasionally present

15. “unidentified bright objects” (UBOs) on brain or spinal MRI in 53-79% of patients (bright on T2WI, isointense on T1WI) that may be hamartomas, heterotopias, foci of abnormal myelination or low grade tumors⁶⁹⁹. Tend to resolve with age

Table 21-77 Diagnostic criteria for NF1⁶⁹⁶

Two or more of the following: • ≥ 6 café au lait spots*, each ≥ 5 mm in greatest diameter in prepubertal individuals, or ≥ 15 mm in greatest diameter in postpubertal patients • ≥ 2 neurofibromas of any type, or one plexiform neurofibroma (neurofibromas are usually not evident until age 10-15 yrs). May be painful • freckling (hyperpigmentation) in the axillary or intertriginous (inguinal) areas • optic glioma: see below• ≥ 2 Lisch nodules: pigmented iris hamartomas that appear as translucent yellow/brown elevations that tend to become more numerous with age • distinctive osseous abnormality, such as sphenoid dysplasia or thinning of long bone cortex with or without pseudarthrosis (e.g. of tibia or radius) • a first degree relative (parent, sibling or off-spring) with NF1 by above criteria

* café au lait spots: hyperpigmented oval light brown skin macules (flat). May be present at birth, increase in number and size during 1st decade. Are present in > 99% of NF1 cases. Rare on face

GENETICS

Simple autosomal dominant inheritance with variable expressivity but almost 100% penetrance after age 5 years. The NF1 gene is on chromosome 17q¹¹.2 which codes for neurofibromin⁷⁰⁰ (neurofibromin is a negative regulator of the Ras oncogene). Loss of neurofibromin as in NF1 results in elevation of growth-promoting signals. The spontaneous mutation rate is high, with 30-50% of cases representing new somatic mutations⁷⁰¹.

Counselling: prenatal diagnosis is possible by linkage analysis only if there are 2 or more affected family members⁷⁰⁰. 70% of NF1 gene mutations can be detected.

MANAGEMENT

• optic gliomas

A. unlike optic gliomas in the absence of NFT, these are rarely chiasmal (usually involving the nerve), are often multiple, and have a better prognosis

B. most are non progressive, and should be followed ophthalmologically and with serial imaging (MRI or CT)

C. surgical intervention probably does not alter visual impairment. Therefore, surgery is reserved for special situations (large disfiguring tumors, pressure on adjacent structures…)

• other neural tumors in patients with NF1 should be managed in the same manner as in the general population

A. focal, resectable, symptomatic lesions should be surgically removed

B. intracranial tumors in NF1 may often be unresectable, and in these cases chemotherapy and/or radiation therapy may be appropriate, with surgery reserved for cases with increasing ICP

C. when malignant degeneration is suspected (rare, but incidence of sarcomas and leukemias is increased), biopsy with or without internal decompression may be indicated

NEUROFIBROMATOSIS 2 (NF2 AKA BILATERAL ACOUSTIC NFT⁷⁰²)

AKA MISME syndrome (acronym for: Multiple Inherited Schwannomas, Meningiomas, and Ependymomas).

CLINICAL

Diagnostic criteria: Are shown in Table 21-78⁷⁰³.

Other clinical features:

1. seizures or other focal deficits

2. skin nodules, dermal neurofibromas, café au lait spots (less common than in NF1)

3. multiple intradural spinal tumors are common (less common in NF1)⁷⁰⁴: including intramedullary (especially ependymomas) and extramedullary (schwannomas, meningiomas…)

4. retinal hamartomas

5. antigenic nerve growth factor is increased (does not occur with NF1)

6. despite its name, is not associated with neurofibromas

Two subtypes⁷⁰³:

1. the more common, severe form with younger age of onset (2nd to 3rd decade), with rapid progression of hearing loss and multiple associated tumors

2. milder form, presents later in life, with slower deterioration in hearing and fewer associated tumors

Table 21-78 Diagnostic criteria for NF2⁷⁰³

Definite diagnosis if either: 1. bilateral vestibular schwannomas (VS) on imaging (MRI or CT) or 2. a first degree relative (parent, sibling or off-spring) with NF2 and either: A. unilateral VS at age < 30 years or B. any two of the following: meningioma, schwannoma (including spinal root), glioma (includes astrocytoma, ependymoma), posterior subcapsular lens opacity

Probable diagnosis if either: 1. unilateral VS at age < 30 and any of the following: meningioma, schwannoma, glioma, posterior subcapsular lens opacity or 2. multiple meningiomas and either of the following: schwannoma, glioma, or posterior lens opacity

GENETICS

Autosomal dominant inheritance. NF2 is due to a mutation at chromosome 22q¹².2 which results in the inactivation of schwannomin (AKA merlin, a semi-acronym for moesin-, ezrin-, and radixin-like proteins), a tumor suppression peptide.

MANAGEMENT CONSIDERATIONS

• bilateral vestibular schwannomas:

chance of preserving hearing is best when tumor is small. Thus, one should attempt to remove smaller tumor. If hearing is serviceable in that ear after surgery, then consider removing the second tumor, otherwise follow the second tumor as long as possible and perform a subtotal removal in an attempt to prevent total deafness

stereotactic radiosurgery therapy may be a treatment option

• most NF2 patients will become deaf at some time during their life

• prior to surgery, obtain MRI of cervical spine to R/O intraspinal tumors that may cause cord injuries during other operations

• NB: pregnancy may accelerate the growth of eighth nerve tumors

21.11.2. Tuberous sclerosis complex

Key concepts:

• autosomal dominant. Incidence: 1 in 6K-10K live births

• clinical triad: seizures, mental retardation and sebaceous adenomas

• typical CNS finding: subependymal nodule (“tuber”) - a hamartoma

• common associated neoplasm: subependymal giant cell astrocytoma

• 2 tumor suppressor genes: TSC1 (on chromosome 9q³⁴) codes for hamartin, and TSC2 (on chromosome 16p¹³) encodes tuberlin

• CT shows intracerebral calcifications (usually subependymal)

Tuberous sclerosis complex (TSC), AKA Bourneville’s disease, is a neurocutaneous disorder characterized by hamartomas of many organs including the skin, brain, eyes and kidneys. In the brain, the hamartomas may manifest as cortical tubers, glial nodules located subependymally or in deep white matter, or giant cell astrocytomas. Associated findings include pachygyria or microgyria.

EPIDEMIOLOGY/GENETICS

Incidence: 1 in 6,000-10,000 live births¹⁰. Point prevalence: 10.6 per 100,000 persons (from Rochester, MN⁷⁰⁵).

Autosomal dominant inheritance, however spontaneous mutation is common. Two distinct tumor suppressor genes have been identified: TSC1 (located on chromosome 9q³⁴) codes for hamartin, and TSC2 (on chromosome 16p¹³) codes for tuberlin.

Genetic counseling for parents with one affected child: 1-2% chance of recurrence,

DIAGNOSIS

Diagnostic criteria are shown in Table 21-79.

In the infant, the earliest finding is of “ash leaf” macules (hypomelanotic, leaf shaped) that are best seen with a Wood’s lamp. Infantile myoclonus may also occur.

In older children or adults, the myoclonus is often replaced by generalized tonic-clonic or partial complex seizures which occurs in 70-80%. Facial adenomas are not present at birth, but appear in > 90% by age 4 yrs (these are not really adenomas of the sebaceous glands, but are small hamartomas of cutaneous nerve elements that are yellowish-brown and glistening and tend to arise in a butterfly malar distribution usually sparing the upper lip).

Retinal hamartomas occur in ≈ 50% (central calcified hamartoma near the disc or a more subtle peripheral flat salmon-colored lesion). A distinctive depigmented iris lesion may also occur.

Plain skull x-rays

May show calcified cerebral nodules.

Table 21-79 Diagnostic criteria of tuberous sclerosis complex⁷⁰⁶

• TSC: diagnosis requires 2 major criteria, or 1 major and 2 minor criteria • Probable TSC: 1 major + 1 minor • Possible TSC: 1 major or 2 minor

Major criteria

• cutaneous manifestations: facial angiofibroma, ungual fibroma, > 3 hypomelanotic macules, shagreen patch • brain and eye lesions: cortical tuber, subependymal nodules, subependymal giant cell astrocytoma, multiple retinal nodular hamartomas • tumors in other organs: cardiac, rhabdomyoma, lymphangioleiomyomatosis, renal angiomyolipoma

Minor criteria

• rectal polyps • pits in dental enamel • bone cysts • migration abnormalities of cerebral white matter • gingival fibromas • nonrenal hamartomas • achromic retinal patches • confetti skin lesions • multiple renal cysts

CT scan⁷⁰⁷

Intracerebral calcifications are the most common (97% of cases) and characteristic finding. Primarily located subependymally along the lateral walls of the lateral ventricles or near the foramina of Monro.

Low density lesions that do not enhance are seen in 61%. Probably represent heterotopic tissue or defective myelination. Most common in occipital lobe.

Hydrocephalus (HCP) may occur even without obstruction. In the absence of tumor, HCP is usually mild. Moderate HCP usually occurs only in the presence of tumor.

Subependymal nodules are usually calcified, and protrude into the ventricle (“candle guttering” the appearance on pneumoencephalography).

Paraventricular tumors (mostly giant cell astrocytomas, see below) are essentially the only enhancing lesion in TSC.

PATHOLOGY

Subependymal nodules (“tubers”) are benign hamartomas that are almost always calcified, and protrude into the ventricles.

Giant cell astrocytoma: a transformation lesion. Almost always located at the foramen of Monro. Occurs in 5-15% of patients with TSC⁷⁰⁸. Histology shows fibrillary areas alternating with cells containing generous amounts of eosinophilic cytoplasm. Areas of necrosis and abundant mitotic figures may be seen, but are not associated with the typical malignant aggressiveness that these features usually denote⁷⁰⁹.

TREATMENT

Paraventricular tumors should be followed, and removed only if they are symptomatic. The transcallosal route is recommended by some.

Infantile myoclonus may respond to steroids. Seizures are treated with AEDs.

Surgery for intractable seizures may be considered when a particular lesion is identified as a seizure focus. Better seizure control, not cure, is the goal in TSC.

21.11.3. Sturge–Weber syndrome

Key concepts:

• cardinal signs: 1) localized cerebral cortical atrophy and calcifications, 2) ipsilateral port-wine facial nevus (usually in distribution of V₁)

• contralateral seizures usually present

• plain skull films classically show “tram-tracking” (double parallel lines)

AKA encephalotrigeminal angiomatosis. A neurocutaneous disorder consisting of:

1. cardinal features:

A. localized cerebral cortical atrophy and calcifications (especially cortical layers 2 and 3, with a predilection for the occipital lobes):

1. calcifications appear as curvilinear double parallel lines (“tram-tracking”) on plain x-rays

2. cortical atrophy usually causes contralateral hemiparesis, hemiatrophy, and homonymous hemianopia (with occipital lobe involvement)

B. ipsilateral port-wine facial nevus (nevus flammeus) usually in distribution of 1st division of trigeminal nerve (rarely bilateral)

2. other findings that may be present:

A. ipsilateral exophthalmos and/or glaucoma, coloboma of the iris

B. oculomeningeal capillary hemangioma

C. convulsive seizures: contralateral to the facial nevus and cortical atrophy. Present in most patients starting in infancy

D. retinal angiomas

GENETICS

Most cases are sporadic. Other cases are suggestive of recessive inheritance, with chromosome 3 being implicated.

TREATMENT

Treatment is supportive. Anticonvulsants are used for seizures. Lobectomy or hemispherectomy may be needed for refractory seizures. XRT: complications are common and benefits are lacking. Laser surgery for the cutaneous nevus is disappointing; better results obtain from masking the nevus with a skin colored tattoo.

21.11.4. Neurocutaneous melanosis (NCM)

BACKGROUND

1. a rare, congenital, nonheritable phakomatosis in which large or numerous congenital melanocytic nevi are associated with benign and/or malignant melanocytic tumors of the leptomeninges⁷¹⁰

2. pathogenesis: neuroectodermal defect during morphogenesis involving melano-blasts of skin and pia mater originating from neural crest cells⁷¹⁰

CLINICAL FEATURES

1. two-thirds of patients with NCM have giant congenital melanocytic nevi^A710

2. one-third have numerous lesions^A without a single giant lesion⁷¹⁰

3. virtually all have large cutaneous melanocytic (pigmented) nevi located on the posterior torso⁷¹¹

4. neurologic manifestations: usually before age 2 years. Signs of intracranial hypertension (lethargy, vomiting…), focal seizures, motor deficits or aphasia⁷¹⁰

5. hydrocephalus: in almost 66%. Usually due to obstruction of CSF flow or reduced absorption as a result of thickened leptomeninges⁷¹⁰

A. pigmented nevi that are large, hairy, or both. The chances that nevi represents NCM is higher when the nevi are located on head, posterior neck or paravertebral)

CLINICAL DIAGNOSTIC CRITERIA⁷¹²

1. large or multiple congenital melanocytic nevi with meningeal melanosis or melanoma

2. absence of cutaneous melanoma, except in patients with benign meningeal lesions (i.e. must rule-out meningeal metastases from cutaneous melanoma)

3. no evidence of meningeal melanoma, except in patients with benign cutaneous lesions

ASSOCIATED CONDITIONS

NCM is sometimes associated with

1. neurocutaneous syndromes⁷¹⁰

A. Sturge-Weber syndrome (see page 726)

B. von Recklinghausen’s neurofibromatosis (NF1) (see page 723)

2. posterior fossa cystic malformations: (e.g. Dandy Walker malformation (see page 240)) occurs in up to 10%. These cases have worse prognosis due to malignant transformation⁷¹⁰

3. intraspinal lipoma and syringomyelia⁷¹⁰

DIAGNOSTIC TESTING

1. MRI: T₁ and T₂ signal shortening produced by melanin. IV gadolinium may demonstrate enhancement of tumor-infiltrated meninges⁷¹⁰

2. histological exam of CNS lesions shows leptomeningeal melanosis (benign) which develops from the melanocytes of the pia matter. Melanoma (malignant) occurs in 40-62% of cases but distinction has little prognostic significance because of the poor outcome of the symptomatic NCM patient even in the absence of melanoma⁷¹⁰

MANAGEMENT

The benefit of resecting skin lesions is questionable in the presence of leptomeningeal lesions⁷¹³. NCM appears refractory to radiation therapy and chemotherapy⁷¹³

Neurosurgical involvement is usually limited to⁷¹²:

1. shunting for hydrocephalus

2. palliative operative decompression if early in the course

3. biopsy for tissue diagnosis in questionable cases

PROGNOSIS

1. when neurological signs are present, prognosis is poor regardless of whether or not malignancy is present

2. > 50% of patients die within 3 years after the first neurologic manifestation⁷¹⁰

21.12. Tumors of the spine and spinal cord

15% of primary CNS tumors are intraspinal (the intracranial:spinal ratio for astrocytomas is 10:1; for ependymomas it’s 3-20:1)⁷¹⁴. There is disagreement over the prevalence, prognosis, and optimal treatment. Most primary CNS spinal tumors are benign (unlike the case with intracranial tumors). Most present by compression rather than invasion⁷¹⁵.

TYPES OF SPINAL TUMORS

May be classified in 3 groups. Although metastases may be found in each category, they are usually extradural. Frequencies quoted below are from a general hospital, extradural lesions are less common in neurosurgical clinics because of relative exclusion of extradural lymphomas, metastatic Ca, etc.:

1. extradural (ED) (55%): arise outside cord in vertebral bodies or epidural tissues

2. intradural extramedullary (ID-EM) (40%): arise in leptomeninges or roots. Primarily meningiomas and neurofibromas (together = 55% of ID-EM tumors)

3. intramedullary spinal cord tumors (IMSCT) (5%): arise in SC substance. Invade and destroy tracts and grey matter, see page 730

DIFFERENTIAL DIAGNOSIS: SPINE & SPINAL CORD TUMORS

See also Myelopathy on page 1185 for a list including nonneoplastic causes of spinal cord dysfunction (e.g. spinal meningeal cyst, epidural hematoma, transverse myelitis…).

1. extradural spinal cord tumors (55%): arise in vertebral bodies or epidural tissues

A. metastatic: comprise the majority of ED tumors

1. most are osteolytic (cause bony destruction): see Spinal epidural metastases, page 742. Common ones include:

a. lymphoma: most cases represent spread of systemic disease (secondary lymphoma); some cases may be primary (see below)

b. lung

c. breast

d. prostate

2. metastases that may be osteoblastic:

a. in men: prostate Ca is the most common

b. in women: breast Ca is the most common

B. primary spinal tumors (very rare)

1. chordomas: see page 675

2. osteoid osteoma: see page 736

3. osteoblastoma: see page 736

4. aneurysmal bone cyst (ABC): an expansile tumor-like osteolytic lesion consisting of a highly vascular honeycomb of blood-filled cavities separated by connective tissue septa, surrounded by a thin cortical bone shell which may expand. Comprise 15% of spine tumors⁷¹⁶. Etiology is controversial. May arise from preexisting tumor (including: osteoblastoma, giant cell tumor, fibrous dysplasia, chondrosarcoma) or following acute fracture. In spine, there is a tendency to involve primarily the posterior elements. Peak incidence is in 2nd decade of life. Treatment usually consists of intralesional curettage. High recurrence rate (25-50%) if not completely excised

5. chondrosarcoma: a malignant tumor of cartilage. Lobulated tumors with calcified areas

6. osteochondroma (chondroma): benign tumors of bone that arise from mature hyaline cartilage. Most common during adolescence. An enchondroma is a similar tumor arising within the medullary cavity

7. vertebral hemangioma: see page 738

8. giant cell tumors (GCT) of bone: AKA osteoclastoma (see page 742)

9. giant cell (reparative) granuloma: AKA solid variant of ABC⁷¹⁷. Related to GCT. Occurs primarily in mandible, maxilla, hands and feet, but there are case reports of spine involvement^{717, 718}. Not a true neoplasm - more of a reactive process. Treatment: curettage. Recurrence rate: 22-50%, treated with re-excision

10. brown tumor of hyperparathyroidism

11. osteogenic sarcoma: rare in spine

C. miscellaneous

1. plasmacytoma: see page 741

2. multiple myeloma: see page 740

3. eosinophilic granuloma (EG): osteolytic defect with progressive vertebral collapse (EG is one cause of vertebra plana - see page 1232). C-spine is the most commonly affected region. Isolated EG associated with systemic conditions (Letterer-Siwe or Hand-Schüller-Christian disease) are treated with biopsy and immobilization. Collapse or neurologic deficit from compression may require decompression and/or fusion. Low-dose RTX may also be effective^{719, 720}

4. Ewing’s sarcoma: aggressive malignant tumor with a peak incidence during 2nd decade of life. Spine mets are more common than primary spine lesions. Treatment is mostly palliative: radical excision followed by RTX (very radiosensitive) and chemotherapy⁷²¹

5. chloroma: focal infiltration of leukemic cells

6. angiolipoma: ≈ 60 cases reported in literature

7. neurofibromas: most are intradural, but some are extradural (see page 735), usually dilate neural foramen (dumbbell tumors)

8. Masson’s vegetant intravascular hemangioendothelioma⁷²² (see page 716)

2. intradural extramedullary spinal cord tumors (40%)

A. meningiomas: see below

B. neurofibromas

C. many lipomas are extramedullary with intramedullary extension

D. miscellaneous: only ≈ 4% of spinal metastases involve this compartment

3. tumors that are usually intradural, but may be partly or wholly extradural:

A. meningiomas: 15% of spinal meningiomas are extradural

B. neurofibromas

4. intramedullary spinal cord tumors (5%): see below

A. astrocytoma: 30% (see page 731)

B. ependymoma: 30% (see page 731) (including myxopapillary ependymoma, see page 731)

C. miscellaneous: 30%, includes:

1. malignant glioblastoma

2. dermoid. In addition to the general population, dermoids present in a delayed fashion following ≈ 16% of myelomeningocele (MM) closures⁷²³. An iatrogenic etiology has been debated⁷²⁴, however a case of a congenital dermoid in a newborn with MM⁷²⁵ indicates that the origin is not always from incompletely excised dermal elements at the time of MM closure

3. epidermoid

4. teratoma

5. lipoma

6. hemangioblastoma (see page 732)

7. neuroma (very rare intramedullary)

8. syringomyelia (not neoplastic)

9. extremely rare tumors

a. lymphoma

b. oligodendroglioma

c. cholesteatoma

d. intramedullary metastases: comprises only ≈ 2% of spinal mets

SPINAL MENINGIOMAS⁷²⁶

Epidemiology

Peak age: 40-70 years. Female:male ratio = 4:1 overall, but the ratio is 1:1 in the lumbar region. 82% thoracic, 15% cervical, 2% lumbar. 90% are completely intradural, 5% are extradural, and 5% both intra- and extra-dural. 68% are lateral to the spinal cord, 18% posterior, 15% anterior. Multiple spinal meningiomas occur rarely.

Clinical

Symptoms

	At onset	At time of first surgery
1. local or radicular pain:	42%	53%
2. motor deficits:	33%	92%
3. sensory symptoms:	25%	61%
4. sphincter disturbance:		50%

Signs prior to surgery (only 1 of 174 patients was intact)⁷²⁶

1. motor

A. pyramidal signs only: 26%

B. walks with aid: 41%

C. antigravity strength:17%

D. flexion-extension with gravity removed: 6%

E. paralysis: 9%

2. sensory

A. radicular: 7%

B. long tract: 90%

3. sphincter deficit: 51%

Outcome

Recurrence rate with complete excision is 7% with a minimum of 6 years follow-up (relapses occurred from 4 to 17 years post-op)⁷²⁶.

SPINAL LYMPHOMA

1. epidural

A. metastatic or secondary lymphoma: the most common form of spinal lymphoma. Spinal involvement occurs in 0.1-10% of patients with non-Hodgkin’s lymphoma

B. primary spinal epidural non-Hodgkin’s lymphoma: rare. Completely epidural with no bony involvement. The existence of this entity is controversial, and some investigators feel that it represents extension of undetected retroperitoneal or vertebral body lymphoma. May have a better prognosis than secondary lymphoma⁷²⁷

2. intramedullary

A. secondary: see page 732

B. primary: very rare (see below)

21.12.1. Intramedullary spinal cord tumors

21.12.1.1. Types of intramedullary spinal cord tumors

The following list excludes metastases (see below) and lipomas (of questionable neo-plastic origin⁷²⁸, and most are actually extramedullary intradural, see below).

1. astrocytoma (nonmalignant): 30%^A (the most common intramedullary spinal cord tumors (IMSCT) outside the filum terminale⁷¹⁵) tend to be eccentric

2. ependymoma: 30%A, tend to be more central, more uniform dense enhancement

3. miscellaneous: 30%, including:

A. malignant glioblastoma

B. dermoid

C. epidermoid (including iatrogenic from LP without stylet)^{729, 730}

D. teratoma

E. hemangioblastoma (see below)

F. hemangioma

G. neuroma (very rarely intramedullary)

H. extremely rare tumors

1. primary lymphoma (only 6 case reports, all non-Hodgkin type⁷³¹)

2. oligodendroglioma, only 38 cases in world literature⁷³²

3. cholesteatoma

4. paraganglioma

5. primarily spinal embryonal tumor (“spinal PNET”) (see page 685)⁵³⁰

6. pilomyxoid astrocytoma (see page 606)

7. metastasis

A. in pediatrics, astrocytoma and ependymoma constitute 90% of IMSCT

DIFFERENTIAL DIAGNOSIS

(also see DDx for Myelopathy on page 1185)

• neoplasm (tumor): (see above for list). Enhancement: 91% enhance⁷³³; of the 9% that do not, most were astrocytomas, 1 was a subependymoma; enhancement did not correlate with grade

• nonneoplastic lesions

vascular lesions (e.g. AVM): serpiginous linear flow-void. Spinal angiography may be useful⁷¹⁵

demyelinating disease (e.g. multiple sclerosis):

1. usually does not extend > 2 vertebral levels

2. cord lesions in MS are most common in the cervical region

inflammatory myelitis

paraneoplastic myelopathy

diseases causing pain over certain body segments (e.g. cholecystitis, pyelonephritis, intestinal pathology). To differentiate from these, look for dermatomal distribution, increase with Valsalva maneuver, and accompanying sensory and/or motor changes in LEs which suggest cord/radicular lesion. Radiographic studies are frequently required to differentiate

diseases of vertebral structures (e.g. Paget’s disease, giant cell tumors of bone (see page 742), etc.)

21.12.1.2. Specific intramedullary spinal cord tumors

EPENDYMOMA

Key concepts:

• the most common glioma of lower cord, conus and filum (most ependymomas in conus and filum are myxopapillary ependymomas). More common in adults

• evaluation: includes imaging the entire neuraxis (usually with enhanced MRI: cervical, thoracic, lumbar & brain) because of potential for seeding through CSF

• associated cysts are common

• treatment: surgical excision (most are encapsulated)

The most common glioma of the lower spinal cord, conus and filum (see Myxopapillary ependymoma below). Slow-growing. Benign. Slight male predominance; slight peak in 3rd to 6th decade. Over 50% in filum, next most common location is cervical. Histologically: papillary, cellular, epithelial, or mixed (in filum, myxopapillary ependymoma is most common, see below). Cystic degeneration in 46%. May expand spinal canal in filum⁷³⁴. Usually encapsulated and minimally vascular (papillary: may be highly vascular; may cause SAH). Symptoms present > 1 yr prior to diagnosis in 82% of cases⁵¹².

Myxopapillary ependymoma

Ependymomas of the conus medullaris and the filum terminale are usually of the myxopapillary subtype. WHO grade I. Usually solitary. Histology: papillary, with microcystic vacuoles, mucosubstance; connective tissue. No anaplasia, but CSF dissemination occurs rarely (can seed intracranially following removal of spinal tumor⁷³⁵). Denovo intracranial lesions also occur rarely. Rare reports of systemic mets⁷¹⁴. Outside the CNS, may occur in sacrococcygeal subcutaneous tissues from heterotopic rests of ependymal cells⁷³⁶.

Surgical removal of filum tumors consists of coagulating and dividing the filum terminale just above and below the lesion (see Distinguishing features of the filum terminale intraoperatively, page 255), and excising it in total. The filum is first cut above the lesion to prevent retraction upwards.

ASTROCYTOMA

Uncommon in first year. Peak: 3rd - 5th decade. Male:female = 1.5:1. The ratio of low-grade:high-grade = 3:1 in all ages⁷³⁴. Occurs at all levels, thoracic most common, then cervical. 38% are cystic; cyst fluid usually has high protein.

DERMOID AND EPIDERMOID

Epidermoids are rare before late childhood. Slight female predominance. Cervical and upper thoracic rare; conus common. Usually ID-EM, but conus/cauda equina may have IM component (completely IM lesions rare).

LIPOMA

May occur in conjunction with spinal dysraphism (see Lipomyeloschisis, page 251). The following considers lipomas that occur in the absence of spinal dysraphism.

Peak occurrence: 2nd, 3rd and 5th decade. Technically hamartomas. No sex predominance. Usually ID-EM (a sub-type is truly IM and essentially replaces the cord⁷³⁷), cervicothoracic region is the most common location. NB: unlike other IMSCT’s, most common symptom is ascending mono- or para-paresis (c.f. pain). Sphincter disturbance is common with low lesions. Local subcutaneous masses or dimples are frequent. Malis recommends early subtotal removal at about 1 year age in asymptomatic patient⁷³⁷. Superficial extrasacral removal is inadequate, as patients then develop dense scarring intraspinally leading to fairly rapid severe neurological damage with poor salvageability even after the definitive procedure.

HEMANGIOBLASTOMA

Usually non-infiltrating, well demarcated, may have cystic caps. 33% of patients with spinal hemangioblastoma will have von Hippel-Lindau disease (see page 667). Cannot incise nor core because of vascularity. Requires microsurgical approach similar to AVM, possibly with intraoperative hypotension.

METASTASES

Most spinal mets are extradural, only a few hundred case reports of IMSCT mets exist⁷³⁸, accounting for 3.4% of symptomatic metastatic spinal cord lesions⁷³⁹. Primaries include: small-cell lung Ca⁷⁴⁰, breast Ca, malignant melanoma, lymphoma and colon Ca^{739, 741}. Ca rarely presents first as an intramedullary spinal met.

PRESENTATION

1. pain: the most common complaint. Almost always present in filum tumors (exception: lipomas)⁷²⁹. Possible pain patterns:

• radicular: increases with Valsalva maneuver and spine movement. Suspect SCT if dermatome is unusual for disk herniation

• local: stiff neck or back, Valsalva maneuver increases. Pain during recumbency (“nocturnal pain”) is classic for SCT

• medullary (as in syrinx): oppressive, burning, dysesthetic, non-radicular, often bilateral, unaffected by Valsalva maneuver

2. motor disturbances

• weakness is 2nd or 3rd most common complaint. Usually follows sensory symptoms temporally

• children present most frequently with gait disturbances

• syringomyelic syndrome: suggests IMSCT. Findings: UE segmental weakness, decreased DTR, dissociative anesthesia (see below)

• long-tract involvement → clumsiness and ataxia (distinct from weakness)

• atrophy, muscle twitches, fasciculations

3. non-painful sensory disturbances

• dissociated sensory loss: decreased pain and temperature, preserved light touch (as in Brown-Séquard syndrome, see page 950). There is disagreement whether this is common⁷¹⁵ or uncommon⁷⁴² in IMSCT. ± non-radicular dysesthesias (early), with upward extension⁷⁴³

• paresthesias: either radicular or “medullary” distribution

4. sphincter disturbances

• usually urogenital (anal less common) → difficulty evacuating, retention, incontinence, and impotence. Early in conus/cauda equina lesions, especially lipomas (pain not prominent)

• sphincter dysfunction common in age < 1 yr due to frequency of lumbosacral lesions (dermoids, epidermoids, etc.)

5. miscellaneous symptoms:

• scoliosis or torticollis

• SAH

• visible mass over spine

TIME COURSE OF SYMPTOMS

Onset usually insidious, but abruptness occurs (benign lesions in children occasionally progress in hours). The onset is often erroneously attributed to coincidental injury. Temporal progression^A has been divided into 4 stages⁷⁴⁴:

1. pain only (neuralgic)

2. Brown-Séquard syndrome

3. incomplete transectional dysfunction

4. complete transectional dysfunction

A. 78% (of 23) ependymomas, 74% (of 42) gliomas, all 7 dermoids, and 50% (of 8) lipomas reached latter 2 stages before diagnosis (not affected by location in cross-sectional nor longitudinal dimension of SC (excludes conus lesions - more frequently diagnosed in 1st stage) (a pre-CT study)

DIAGNOSIS

It is usually difficult to distinguish IMSCT, ID-EM and ED on clinical grounds⁷¹⁵. Schwannomas often start with radicular symptoms that later progress to cord involvement. Most IMSCTs are located posteriorly in cord which may cause sensory findings to predominate early⁷²⁸.

DIAGNOSTIC STUDIES

Plain radiographs: vertebral body destruction, enlarged intervertebral foramina, or increases in interpedicular distances suggests ED SCT.

Lumbar puncture: Elevated protein is the most common abnormality⁷¹⁴ seen in ≈ 95%. The reported range with primary IMSCT’s is 50-2,240 mg%. Glucose is normal except with meningeal tumor. SCT can cause complete block, indicated by:

• Froin’s syndrome: clotting (due to fibrinogen) and xanthochromia of CSF

• Queckenstedt’s test (failure of jugular vein compression to increase CSF pressure, which it normally does in the absence of block)

• barrier to flow of myelographic contrast media

MRI: mainstay of diagnosis. Ependymomas enhance intensely and are often associated with hemorrhage and cysts. Cord edema may mimic a cyst.

Myelography: classically shows fusiform cord widening (may be normal early). Distinct from ED tumors which produce hourglass deformity (with incomplete block) or paintbrush effect (with complete block), or ID-EM tumors which produce a capping effect with a sharp cutoff (meniscus sign) (see page 746).

CT: some IMSCTs enhance with IV contrast. Metrizamide CT distinguishes IMSCT from ID-EM (poor in differentiating IMSCT subtypes).

Spinal angiography: rarely indicated, except in hemangioblastoma (may be suspected on myelography or MRI by linear serpiginous structures). MRI may obviate this test.

MANAGEMENT

Surgery should be performed as soon as possible (generally not as an emergency) after diagnosis since surgical results correlate with the preoperative neurologic condition, and it makes no sense to follow the patient as they develop progressive neurologic deficit⁷⁴⁵ (some of which may be irreversible).

Astrocytomas: For low grade lesions, if a plane can be developed between the tumor and spinal cord (when it can, it usually consists of a thin gliotic layer traversed by small blood vessels and adhesions⁷²⁸), an attempt at total excision is an option⁷⁴⁶. For high grade astrocytomas or for low-grade astrocytomas without a plane of separation, biopsy alone or biopsy plus limited excision is recommended⁷⁴⁶.

For high-grade lesions, post-op RTX (± chemotherapy) is recommended⁷⁴⁶. RTX is not supported following radical resection of low grade gliomas⁷⁴⁶.

Ependymomas: An attempt at gross total removal should be attempted. XRT is not recommended following gross total removal⁷⁴⁶.

TECHNICAL SURGICAL CONSIDERATIONS

• position: usually prone, well padded & securely taped to avoid movement if MEP monitoring is to be used. Other options include: lateral oblique, sitting

• if a cystic component is suspected, partial aspiration with a 25 Ga needle once the spinal cord is exposed will decrease the pressure (avoid total aspiration which makes it more difficult to locate the tumor)⁷⁴⁷. If the cyst forms a “cap” at either end of the tumor, the dura does not need to be opened over the cyst as drainage can be accomplished with removal of tumor

• adjunctive options include:

intraoperative spinal cord monitoring (SSEP, and motor evoked potentials (MEPs)⁷⁴⁸): SEPs almost always degrade with the initial myelotomy and do not correlate well with motor outcome⁷⁴⁹ (which is critical)^{750, 751} (e.g. it is not unusual for SEPs to be lost during the initial myelotomy without correlation with outcome) and postoperative motor deficit may occur in spite of unaltered intraoperative SEPs^748,749 and conversely SEPs may be lost without motor deficit. However, proof of improved outcomes with MEP monitoring is also lacking⁷⁵⁰

intraoperative ultrasound: also controversial⁷⁵¹, favored by some experts. Astrocytomas are usually isoechoic with spinal cord, whereas ependymomas are usually hyperechoic

• a myelotomy is performed either in the midline or just to one side of the dorsal midline to avoid the posteromedian vein. Alternatively, if the tumor is known to be very superficial off the midline (which may be confirmed by ultrasound), entry may be made there. Tumors may cause distortion and displacement of the midline - look for dorsal root entry zones on both sides to identify the midline as the mid-point between root entry zones

• 6-0 silk sutures are placed through the pial edge to gently retract the spinal cord open. Standard sized (i.e. non-micro) bayonet forceps can be used to gently spread tissues

• copious irrigation is used whenever bipolar cautery is employed on the tumor/spinal cord, to minimize transference of heat to the spinal cord. Monopolar cautery should not be used⁷⁴⁷

• either laser or ultrasonic aspiration (USA) are used to debulk tumor from within until the glial-tumor interface is reached. Charring from laser may make it more difficult to recognize the glial/tumor interface than USA, and the laser tends to be slower when debulking larger tumors

• watertight closure is critical

Table 21-80 Key concepts in surgical removal of IMSCT

• in almost all cases, IMSCTs should be debulked from within (to avoid manipulation of neural tissue) with ultrasonic aspirator or laser, and no attempt should be made initially to develop a plane between tumor and spinal cord (even for ependymomas, which of the 3 most common IMSCTs is the only one that actually has such a plane) • if MEPs are monitored: although it is arbitrary, it is suggested that tumor removal should be discontinued if the amplitudes drop to ≤ 50% of baseline

PROGNOSIS

No well designed studies give long term functional results with microsurgery, laser and radiotherapy. Better results occur with lesser initial deficits⁷²⁸. Recurrence depends on totality of removal, and on growth pattern of the specific tumor.

Ependymoma: total extirpation improves functional outcome, and myxopapillary ependymomas fare better than the “classic” type⁵¹². Best functional outcome occurs with modest initial deficits, symptoms < 2 years duration⁷⁵², and total removal. Survival is independent of extent of excision.

Astrocytomas: radical removal rarely possible (cleavage plane unusual even with microscope). Long term functional results poorer than ependymomas. There is 50% recurrence rate in 4-5 yrs.

21.12.2. Spinal schwannomas

Key concepts:

• slow growing benign tumors

• most (75%) arise from the dorsal (sensory) rootlets

• early symptoms are often radicular

• recurrence is rare after total excision (except in neurofibromatosis)

Incidence: 0.3-0.4/100,000/yr. Most occur sporadically and are solitary, but they may also be associated with neurofibromatosis (see page 722) primarily type 2 (NF2), but can occur with type 1.

Configurations

Most are entirely intradural, but 8-32% may be completely extradural^{753, 754}, 1-19% are a combination, 6-23% are dumbbell, and 1% are intramedullary.

Dumbbell tumors. Definition: tumors that develop an “hourglass” shape as a result of an anatomic barrier encountered during growth. Not all dumbbell tumors are Schwannomas (e.g. neuroblastoma, see page 748). Most have a contiguous intraspinal, foraminal (usually narrower) and extraforaminal components (widening of the neural foramen is a characteristic finding, can be recognized even on plain films, and speaks to the longstanding benign nature of the lesion). The waist may also be due to a dural constriction.

Asazuma et al.⁷⁵⁵ classification system for dumbbell spinal Schwannomas is shown in Figure 21-5.

Type I tumors are intradural and extradural and are restricted to the spinal canal. The constriction occurs at the dura.

Type II are all extradural, and are subclassified as: IIa do not extend beyond the neural foramen, IIb = inside spinal canal + paravertebral, IIc = foraminal + paravertebral.

Type IIIa are intradural and extradural foraminal, IIIb are intradural and extradural paravertebral.

Type IV are extradural and intravertebral. Type V are extradural and extralaminar with laminar invasion. Type VI show multidirectional bone erosion.

Craniocaudal spread: IF & TF designate the number of intervertebral foramina and transverse foramina involved, respectively (e.g. IF stage 2 = 2 foramens).

Figure 21-5 Classification of dumbbell spinal tumors Modified with permission from Asazuma T, Yoshiaki T, Hirofumi M, et al.: Surgical strategy for cervical dumbbell tumors based on a three-dimensional classification. Spine 29 (1): E10-4, 2003

Schwannomas involving C1 & C2: May involve vertebral arteries and require additional caution.

Clinical

Patients typically present with local pain. Neurologic deficits develop late.

Pathology

Composed of Antoni A (compact, interwoven bundles of long, spindly Schwann cells) and Antoni B tissue (sparse areas of Schwann cells in a loose eosinophilic matrix).

Treatment

Posterior approaches: Types I, IIa IIIa, some upper cervical IIIb and some VI are generally amenable to a posterior approach. IIa & IIIa usually require total facetectomy for complete removal⁷⁵⁵. Reconstruction with instrumentation may be needed if substantial posterior disruption occurs.

Anterior and combined anterior/posterior approaches: Asazuma et al.⁷⁵⁵ recommend a combined approach for Type IIb, IIc and IIIb lesions where the extraforaminal extension is large (viz. beyond the vertebral arteries). Reconstruction with instrumentation was required for some tumors (≈ 10% of all patients treated) which were type IV (2 patients), IIIb (1 pt) and VI (1 pt).

Nerve sacrifice: It is usually possible to preserve some fascicles of the nerve root, although sometimes section of the entire nerve root is required. New deficits may not occur since involved fascicles are often nonfunctional, and adjacent roots may compensate. The risk for motor deficit is higher for schwannomas than for neurofibromas, for cervical vs. lumbar tumors, and for cervical tumors with extradural extension.

Outcome

Recurrence is rare following gross total excision, except in the setting of NF2.

21.12.3. Bone tumors of the spine

For tumors that can affect the spine but are not necessarily of the spine, see page 728.

1. metastatic: the most common malignancy of spine

A. common osteolytic metastatic tumors include (see page 742):

1. lung

2. breast

3. prostate

4. lymphoma: most cases represent spread of systemic disease (secondary lymphoma), however some may be primary (see page 730)

5. plasmacytoma: see page 740

6. multiple myeloma: see page 740

7. eosinophilic granuloma: see page 729 for differentiating features

B. metastases that may be osteoblastic:

1. in men: prostate Ca is the most common

2. in women: breast Ca is the most common

C. Ewing’s sarcoma: see page 729

D. chloroma: focal infiltration of leukemic cells

2. primary spinal tumors (very rare)

A. benign

1. vertebral hemangioma: see page 738

2. osteoid osteoma: see page 736

3. osteoblastoma: see page 736

4. aneurysmal bone cyst: cavity of highly vascular honeycomb surrounded by a thin cortical shell which may expand (see page 728)

5. osteochondroma (chondroma): see page 728

6. giant cell tumors of bone: AKA osteoclastoma (see page 742). Almost always benign with pseudomalignant behavior

B. malignant

1. chondrosarcoma: see page 728

2. chordomas: see page 675

3. osteogenic sarcoma: rare in spine

21.12.3.1. Osteoid osteoma and osteoblastoma

Key concepts:

• both are benign bone tumors

• histologically identical, differentiation depends on size (≤1 cm = osteoid osteoma, > 1 cm = osteoblastoma)

• can occur in the spine and may cause neurologic symptoms (esp. osteoblastoma)

• high cure rate with complete excision

Two types of benign osteoblastic lesions of bone: osteoid osteoma (OO) and benign osteoblastoma (BOB) (see Table 21-81). They are indistinguishable histologically, and must be differentiated based on size and behavior.

Characteristically cause night pain and pain relieved by aspirin (see Clinical below).

Osteoblastoma is a rare, benign, locally recurrent tumor with a predilection for spine, that may rarely undergo sarcomatous change (to osteosarcoma⁷⁵⁷, only a handful of known cases of this). More vascular than OO⁷⁵⁸.

Differential diagnosis (for lesions with similar symptoms and increased uptake on radionuclide bone scan):

1. benign osteoblastoma

2. osteoid osteoma: more pronounced sclerosis of adjacent bone than BOB

3. osteogenic sarcoma: rare in spine

4. aneurysmal bone cyst: typically trabeculae in central, lucent region (see page 728)

5. unilateral pedicle/laminar necrosis

Table 21-81 Comparison of osteoid osteoma and benign osteoblastoma⁷⁵⁶

	Osteoid osteoma	Benign osteoblastoma
percent of primary bone tumors	3.2%
percent of primary vertebral tumors	1.4%
percent that occur in spine	10%	35%
size limitations	≤ 1 cm	> 1 cm
growth pattern	confined, self limiting	more extensive, may extend into spinal canal
potential for malignant change?	no	rare
location within spine (83 patients)
% in cervical spine	27%	25%
% in thoracic spine		35%
% in lumbar region	59%	35%
location within vertebra (81 patients)
lamina only	33%	16%
pedicle only	15%	32%
articular facet only	19%	0
vertebral body (VB) only	7%	5%
transverse process only	6%	8%
spinous process	5%	5%
> 1 element of neural arch	6%	19%
combined posterior elements & VB	0	11%

CLINICAL

See Table 21-82 for signs and symptoms. Tenderness confined to vicinity of the lesion occurs in ≈ 60%. 28% of patients with BOB presented with myelopathy. OO presented with neurologic deficit in only 22%.

EVALUATION

Bone scans are a very sensitive means for detecting these lesions. Once localized, CT or MRI may better define the lesion in that region.

Caution re needle biopsy: if the lesion turns out to be osteosarcoma, the contaminated needle tract can result in worse prognosis.

Osteoid osteoma

Radiolucent area with or without surrounding density, often isolated to pedicle or facet. May not show up on tomograms.

Osteoblastoma

Most are expansile, destructive lesions, with 17% having moderate sclerosis. 31% have areas of ↑ density, 20% surrounded by calcified shell. Often a contralateral spondylolysis⁷⁵⁷.

Table 21-82 Signs and symptoms in 82 patients⁷⁵⁶

Finding	Osteoid osteoma	Benign osteoblastoma
pain on presentation	100%	100%
pain increased by motion	49%	74%
pain increased by Valsalva	17%	36%
nocturnal pain	46%	36%
pain relieved by aspirin	40%	25%
radicular pain	50%	44%
scoliosis	66%	36%
neurologic abnormalities	22%	54%
myelopathy	0	28%
weakness	12%	51%
atrophy	9%	15%

TREATMENT

In order to obtain a cure, these lesions must be completely excised. The role of radiation therapy is poorly defined in these lesions, but is probably ineffective⁷⁵⁷.

Osteoid osteoma

Cortical bone may be hardened and thickened, with granulomatous mass in underlying cavity.

Osteoblastoma

Hemorrhagic, friable, red to purple mass well circumscribed from adjacent bone. Complete excision → complete pain relief in 93%. Curettage only → pain relief, with more likely recurrence. Recurrence rate with total excision is ≈ 10%.

21.12.3.2. Osteosarcoma

The most common primary bone cancer. More common in children, usually occurring near the ends of long bones, but also in the mandible, pelvis, and rarely in the spine⁷⁵⁹. Spinal osteosarcoma usually occurs in the lumbosacral region in males in their 40’s, sometimes arising from ares of osteoblastoma or Paget’s disease. If a percutaneous biopsy reveals osteosarcoma, the contaminated needle tract can increase the difficulty of subsequent surgery. Poor prognosis, median survival = 10 months⁷⁵⁹.

21.12.3.3. Vertebral hemangioma

Key concepts:

• the most common primary spine tumor. Benign

• rarely symptomatic (< 1.2%), symptoms more commonly from compression fracture, disc herniation, and rarely neural compression from bone expansion…

• MRI: small lesions are hyperintense on T1WI and T2WI. Larger ones may be hypointense. X-ray: striations (corduroy pattern) or “honeycomb” appearance. Bone scan: usually do not have increased uptake

• treatment: incidental lesions require no routine follow-up. Biopsy when mets are a strong consideration. Treatment options (when indicated): XRT, embolization, vertebroplasty (better than kyphoplasty), surgery

Vertebral hemangiomas (VH), AKA spinal hemangioma, cavernous hemangioma, or hemangiomatous angioma. Benign lesions of the spine. The most common primary tumor of the spine (10-12% of primary spinal bone tumors). Estimated incidence: 9-12%^{760, 761}. 70% are solitary, 30% are multiple (up to 5 levels may be involved, often noncontiguous). Lumbar and lower thoracic spine are the most common locations, cervical and sacral lesions are rare. Lesions involve only the vertebral bodies in ≈ 25%, posterior spinal arch in ≈ 25%, and both areas in ≈ 50%. Occasional cases of purely extradural lesions have been described⁷⁶². Intramedullary lesions are even less common⁷⁶³. Typically found in post-pubertal females.

Malignant degeneration has never been reported. Mature thin-walled blood vessels of varying sizes replace normal marrow, producing hypertrophic sclerotic bony trabeculations oriented in a rostral-caudal direction in one of two forms: cavernous (venous) or capillary (difference in subtype carries no prognostic significance).

PRESENTATION

1. incidental: most VH are asymptomatic, these require no follow-up (see below)

2. symptomatic: only 0.9-1.2% are symptomatic. There may be a hormonal influence (unproven) that may cause symptoms to increase with pregnancy (could also be due to increased blood volume and/or venous pressure)⁷⁶⁴ or to vary with the menstrual cycle and may explain why symptoms rarely occur before puberty

A. pain: occasionally VH may present with pain localized to the level of involvement with no radiculopathy. However, pain is more often due to other pathology (compression fracture, herniated disc, spinal stenosis…) rather than the VH itself

B. progressive neurologic deficit: this occurs rarely, and usually takes the form of thoracic myelopathy. Deficit may be caused by the following mechanisms

1. subperiosteal (epidural) growth of tumor into the spinal canal

2. expansion of the bone (cortical “blistering”) with widening of the pedicles and lamina producing a “bony” spinal stenosis

3. compression by vessels feeding or draining the lesion

4. compression fracture of the involved vertebra (very rare)⁷⁶⁵

5. spontaneous hemorrhage → spinal epidural hematoma⁷⁶⁶ (rare)

6. spinal cord ischemia due to “steal”

EVALUATION

Plain x-rays: classically show coarse vertically oriented striations (corduroy pattern) or a “honeycomb” appearance. At least ≈ one third of the VB must be involved to produce these findings on plain x-ray (see Figure 21-6).

Bone scan: VH are usually not hot (unless a compression fracture has occurred), which may help distinguish VH from metastatic disease (which usually light up).

CT: diagnostic procedure of choice. “Polka-dot sign”⁸¹⁷: multiple high density dots represents cross-sections through thickened trabeculae (see Figure 21-7).

MRI: small hemangiomas are focal, round, and hyperintense on T1WI and T2WI. More extensive lesions can be hypointense. Lesions that tend not to evolve (mottled increased signal on T1WI and T2WI, possibly due to adipose tissue) differ from those that tend to be symptomatic (isointense on T1WI,hyperintense on T2WI).

Spinal angiography: also may help distinguish nonevolutive (normal or slight increased vascularity compared to adjacent bone) from symptomatic (moderate to marked hypervascularity) lesions. Therapeutic: if the feeding artery does not also supply the anterior spinal artery, it may be embolized preoperatively or sacrificed at surgery.

Figure 21-6 Vertebral hemangioma. Sagittal CT reconstruction

Figure 21-7 Vertebral hemangioma. Axial CT showing polka dot sign

TREATMENT⁷⁶⁰

1. asymptomatic VH require no routine follow-up or evaluation unless pain or neurologic deficit develop, which are rare occurrences in incidentally discovered VH

2. biopsy: may be indicated in cases where diagnosis is uncertain (e.g. when metastases are a strong consideration). In spite of highly vascular nature, there have been no reported bleeding complications with CT guided biopsy

3. those presenting with pain or neurologic deficit

A. radiation therapy: may be used alone for painful lesions, preoperatively as a surgical adjunct, or post-op following incomplete removal. VH are radiosensitive and undergo sclerotic obliteration. Total dosage should be ≤ 40 Gy to reduce risk of radiation myelopathy. Improvement in pain may take months to years, and no radiographic evidence of response may occur

B. embolization: provides more rapid relief of pain than RTX, can also be used pre-op as surgical adjunct. Risks spinal cord infarction if major radicular artery (e.g. artery of Adamkiewicz, see page 96) is embolized

C. vertebroplasty: (see page 994) may be better than kyphoplasty for VH because kyphoplasty destroys the trabecular bone

D. surgery: for painful lesions that fail to respond to above measures, or for lesions with progressive neurologic deficit (see below)

Surgical treatment

For indications, see above. Recommended management is shown in Table 21-83.

Table 21-83 Recommendations for surgical management of VH*⁷⁶⁰

VH involvement	Approach	Post-op RTX?
posterior elements only	radical excision via posterior approach	not for total excision
VB involvement with anterior canal compression (with or without ST in canal)	anterior corpectomy with strut graft
VB involved but no expansion, ST in lateral canal	laminectomy with removal of soft-tissue	follow serial CT, give RTX if VB expansion or ST expansion
extensive involvement of anterior and posterior vertebral elements with circumferential bone expansion, no ST compression	laminectomy	either RTX, or close follow-up with CT and RTX for ST recurrence or progressive VB expansion
extensive anterior and posterior involvement with ST in anterior canal	anterior corpectomy with strut graft

* abbreviations: VB = vertebral body, ST = soft-tissue component of VH, RTX = radiation treatment

Major risks of surgery: blood loss, destabilization of the spine, neurologic deficit (during surgery, or post-op usually from epidural hematoma). Recurrence rate is 20-30% after subtotal resection, usually within 2 yrs. Patients with subtotal resection should have RTX which lowers recurrence rate to ≈ 7%.

21.12.3.4. Multiple myeloma

Multiple myeloma (MM) (or simply [myeloma) is a neoplasm of a single clone of plasma cells characterized by proliferation of plasma cells in bone marrow, infiltration of adjacent tissues with mature and immature plasma cells, and the production of an immunoglobulin, usually monoclonal IgG or IgA (referred to collectively as M-protein⁷⁶⁷). Circulating pre-myeloma cells lodge in appropriate microenvironments (e.g. in bone marrow) where they differentiate and expand. Although MM is often referred to in the context of “metastatic lesions” to bone, it is also sometimes considered a primary bone tumor. If only a single lesion is identified, then it is referred to as a plasmacytoma (see below).

MM presents as a result of the following (underscored items are characteristic for MM):

1. proliferation of plasma cells: interferes with normal immune system function → increased susceptibility to infection

2. bone involvement

A. bone marrow involvement → destruction of hematopoietic capacity → normocytic normochromic anemia, leukopenia, thrombocytopenia

B. bone resorption

1. → weakening of the bone → pathologic fractures (see below)

2. → hypercalcemia (present initially in 25% of MM patients, see below)

C. swelling or local tenderness of bone

D. bone pain: characteristically induced by movement, and absent at rest

E. spinal involvement

1. invasion of spinal canal in ≈ 10% of cases → spinal cord compression → myelopathy (see page 742)

2. nerve root compression (radiculopathy)

3. overproduction of certain proteins by plasma cells. May lead to:

A. hyperviscosity syndrome

B. cryoglobulinemia

C. amyloidosis

D. renal failure: multifactorial, but monoclonal light chains play a role

Skeletal disease

MM involvement is by definition multiple, and is usually restricted to sites of red marrow: ribs, sternum, spine, clavicles, skull, or proximal extremities. Lesions of the spine and/or skull are the usual reasons for presentation to the neurosurgeon.

Bone resorption in MM is not due simply to mechanical erosion by plasma cells. Increased osteoclastic activity has been observed.

Plasma cell tumors of the skull involving the cranial vault usually do not produce neurologic symptoms. Cranial nerve palsies can arise from skull base involvement. Orbital involvement may produce proptosis (exophthalmos).

Neurologic involvement

Neurologic manifestations can occur as a result of:

1. tumor involvement of bone causing compression (see above)

A. tumor in spine with compression of spinal cord or nerve roots

B. tumor in skull with compression of brain or cranial nerves

2. deposition of amyloid within the flexor retinaculum of the wrist → carpal tunnel syndrome (the median nerve itself does not contain amyloid, and therefore responds well to surgical division of the transverse carpal ligament, see page 811)

3. diffuse progressive sensorimotor polyneuropathy: occurs in 3-5% of patients with MM (also, see page 799)

A. about half are due to amyloidosis (see page 800)

B. polyneuropathy can also occur without amyloidosis, especially in the rare osteosclerotic variant of MM

4. multifocal leukoencephalopathy has been described in MM⁷⁶⁹

5. hypercalcemia: may produce a dramatic encephalopathy with confusion, delerium or coma. Neurologic symptoms of hypercalcemia associated with MM are more common than in hypercalcemia of other etiologies

6. very rare: intraparenchymal metastases⁷⁷⁰

EPIDEMIOLOGY

Incidence in the U.S.: ≈ 1-2 per 100,000 in caucasians, and is ≈ twice that in blacks. MM accounts for 1% of malignancies, and 10% of hematologic cancers. Peak age of occurrence: 60-70 yrs of age, with < 2% of patients being < 40 yrs old. Slightly more common in males. Monoclonal gammopathy without MM occurs in ≈ 0.15% of the population, and in long-term follow-up 16% of these develop MM with an annual rate of 0.18%⁷⁶⁸.

EVALUATION

Diagnostic criteria for MM is shown in Table 21-84. Tests for MM include:

1. 24 hour urine for kappa Bence-Jones protein^A present in 75%

2. bloodwork: serum protein electrophoresis (SPEP) and immune electrophoresis (IEP) (looking for IgG kappa band)^A

3. skeletal radiologic survey. Characteristic x-ray finding: multiple, round, “punched-out” (sharply demarcated) lytic lesions in the bones typically involved (see above). Osteosclerotic lesions are seen in < 3% of patients with MM. Diffuse osteoporosis may also be seen

4. CBC: anemia eventually develops in most patients with MM. It is usually of moderate severity (Hgb ≈ 7-10 gm%) with a low reticulocyte count

5. technetium-99m nuclear bone scan is usually negative in untreated MM (due to rarity of spontaneous new bone formation) and is less sensitive than conventional radiographs. Therefore it is not usually helpful except perhaps to implicate etiologies other than MM to explain the observed findings. After treatment, bone scan may become positive as osteoblastic activity ensues (“flare” response)

6. serum creatinine: for prognostication

7. bone marrow biopsy: virtually all MM patients have “myeloma cells” (although sensitive, this is not specific and other diagnostic criteria should be sought)

A. monoclonal proteins cannot be detected in the urine or serum of ≈ 1% of MM patients. Two or more monoclonal bands are produced in ≈0.5-2.5% of patients with MM⁷⁷²

Table 21-84 Criteria for diagnosis of MM*

1. cytologic criteria A. marrow morphology: plasma cells and/or myeloma cells ≥ 10% of 1000 or more cells B. biopsy proven plasmacytoma

2. clinical and laboratory criteria A. myeloma protein (M-component) in serum (usually > 3 gm/dl) or urine IEP B. osteolytic lesions on x-ray (generalized osteoporosis qualifies if marrow contains > 30% plasma or myeloma cells) C. myeloma cells in ≥ 2 peripheral blood smears

* diagnosis requires⁷⁷¹: 1 A & B, or 1A or 1B and 2A, 2B, or 2C

TREATMENT

Many aspects of treatment fall into the purvey of the oncologist (see review⁷⁶⁸). Some aspects pertinent to neurosurgical care include:

1. XRT: MM Is very radiosensitive. Focal XRT for pain due to readily identifiable bone lesions, may allow pathologic fractures to heal and is effective in spinal cord compression (see page 708)

2. mobilization: immobilization due to pain and fear of pathologic compression fractures leads to further detrimental increases in serum calcium and weakness

3. pain control: mild pain often responds well to salicylates (contraindicated in thrombocytopenia). Local XRT is also effective (see below)

4. percutaneous kyphoplasty (see page 994) may be used for some spine lesions

5. therapy for hypercalcemia usually improves symptoms related to that

6. bisphosphonates inhibit bone resorption and rapidly reduces hypercalcemia (see page 500). Pamidronate is currently preferred over older agents

7. bortezomib (Velcade®): proteasome inhibitor, for treatment of refractory MM

PROGNOSIS

Untreated MM has a 6 month median survival. Solitary plasmacytoma has a 50% 10-year survival. If there is an solitary site of involvement but M-protein is present (i.e. essentially a plasmacytoma except for the M-protein), elimination of the M-protein following XRT indicates a 50-60% chance of remaining free of MM, if the M-protein doesn’t resolve, there is a high chance of developing MM.

PLASMACYTOMA

A neoplasm of a single clone of plasma cells similar to multiple myeloma (see above) but meeting the following criteria:

1. there must be no other lesions on complete skeletal survey (not bone scan)

2. bone marrow aspirate must show no evidence of myeloma

3. and serum and urine electrophoresis should show no M-protein

MM will develop in 55-60% of patients with a solitary plasmacytoma in 5 years, and in 70-80% by 10 yrs.

Treatment

1. local XRT provides good local control rates

2. percutaneous kyphoplasty (see page 994)

21.12.3.5. Giant cell tumors of bone

AKA osteoclastoma (cells arise from osteoclasts). In the same general category as aneurysmal bone cysts. Typically arise in adolescence. Most common in knees and wrists. May be seen by a neurosurgeon when they arise in the skull (especially the skull base, and in particular the sphenoid bone), or in the vertebral column (≈ 4% occur in sacrum).

Pathology

Lytic with bony collapse. Almost always benign with pseudomalignant behavior (recurrence is common, and pulmonary mets can occur).

Evaluation

Work-up includes chest CT because of possibility of pulmonary mets.

Treatment

Intratumoral curettage, possibly aided by pre-op embolization. Recurrence rate with this treatment (even if resection is subtotal) is only ≈ 20%. Role of RTX is controversial⁷¹⁹ because of the possibility of malignant degeneration (therefore use RTX only for non-resectable recurrence). Use of osteoclast inhibiting drugs (bisphosphonates e.g. pamidronate, see page 501) has met with some success following subtotal resection.

For gross residual disease after resection, re-resection is a consideration.

Cryosurgery with liquid nitrogen has been employed in long bones. Its use is limited in neurosurgical cases because of risk of injury to adjacent neural structures (brain, spinal cord) and cryotherapy induced fractures, although it has been described for use in the sacrum⁷⁷³.

Close follow-up is required due to propensity for recurrence. MRI or CT initially q 3 months is suggested.

21.12.4. Spinal epidural metastases

Key concepts:

• suspected in a cancer patient with back pain that persists in recumbency

• occurs in ≈ 10% of all cancer patients

• 80% of primary sites: lung, breast, GI, prostate, melanoma and lymphoma

• many treatments reduce pain. Surgery + XRT in selected cases increases chances of preserving ambulation & produces a modest improvement in survival

• if no neurologic compromise or bony instability, usual treatment: biopsy (CT- or fluoro-guided) followed by XRT (surgical indications: Table 21-88, page 745)

• surgery not helpful for: total paralysis > 8 hrs, loss of ambulation > 24 hrs, and not recommended for prognosis < 3-4 months survival, poor medical condition (poor PFTs…), or radiosensitive tumor

Spinal epidural metastases (SEM) occur in up to 10% of cancer patients at some time⁷⁷⁴, and are the most common spinal tumor. 5-10% of malignancies present initially with cord compression⁷⁷⁵. For other etiologies of spinal cord compression, see items marked with a dagger (^†) under Myelopathy on page 1185.

Routes of metastasis to spine:

1. arterial

2. venous: via spinal epidural veins (Batson’s plexus⁷⁷⁶)

3. perinervous (direct spread)

The usual route of spread is hematogenous dissemination to the VB with erosion back through pedicles and subsequent extension into the epidural space (i.e. anterior epicenter). Less commonly may initially metastasize to lateral or posterior aspect of canal. Most metastases (mets) are epidural, only 2-4% are intradural, and only 1-2% are intramedullary. Distribution between cervical, thoracic and lumbar spine is proportional to the length of the segment, thus the thoracic spine is the most common site (50-60%).

Primary sources of spine mets

Table 21-85 shows primary tumor types that give rise to SEM. The majority are common primaries that tend to metastasize to bone (lung, breast, prostate, renal-cell and thyroid). Rare tumors that may go to bone include the myxoid subtype of liposarcoma⁷⁷⁸ (17% of these patients develop bone mets, 5-year median survival is 16%).

Presentation

Pain: the most common initial symptom. Occurs in up to 95% of patients with SEM^{779, 780}. Types of pain:

• local pain: typically aching, experienced at the level of involvement. Increased pain with recumbency (especially at night) is characteristic

• radicular: tends to be sharp or shooting, referred into dermatome of the involved nerve root. Commonly bilateral in thoracic region

• mechanical: usually exacerbated by movement

Neck-flexion, straight-leg-raising, coughing, sneezing, or straining may also aggravate the pain.

Motor or autonomic dysfunction: the second most common presentation. Up to 85% of patients have weakness at the time of diagnosis. Leg stiffness may be an early symptom. Bladder dysfunction (urinary urgency, hesitancy or retention) is the most common autonomic manifestation; others include constipation or impotence.

Sensory dysfunction: anesthesia, hypesthesia, or paresthesias usually occur with motor dysfunction. Cervical or thoracic cord involvement may produce a sensory level.

Other presentations: pathologic fracture. Bone metastases can sometimes produce hypercalcemia (a medical emergency).

The greater the neurologic deficit when treatment is initiated, the worse the chances for recovery of lost function. 76% of patients have weakness by the time of diagnosis⁷⁷⁴. 15% are paraplegic on initial presentation, and < 5% of these can ambulate after treatment. Median time from onset of symptoms to diagnosis is 2 months⁷⁸¹.

Metastases to the upper cervical spine

For differential diagnosis, see Foramen magnum lesions, page 1212 and Axis (C2) vertebra lesions on page 1231.

Metastases to the C1-2 region comprise only ≈ 0.5% of spinal mets⁷⁸². They typically present initially with suboccipital and posterior cervical pain, and as the lesion progresses patients develop a characteristic pain that makes it difficult to sit up (some will hold their heads in their hands to stabilize it). Possibly as a result of the capacious spinal canal at this level, only ≈11-15% of patients present with neurologic symptoms. 15% develop spinal cord compression⁷⁸³, and quadriplegia from atlantoaxial subluxation occurred in ≈ 6%⁷⁸³.

Anterior approaches for stabilization at this location are difficult. Pathologic fractures due to osteoblastic types of tumors (e.g. prostate, some breast) may heal with radiation treatment and immobilization. For others, good pain relief and stabilization may be achieved with radiation followed by posterior fusion⁷⁸³.

EVALUATION AND MANAGEMENT OF EPIDURAL SPINAL METASTASES

There is no difference in outcome between lesions above or below the conus; thus spinal cord, conus medullaris, or cauda equina mets are considered together here as epidural spinal cord compression (ESCC). Features that help distinguish conus lesions from cauda equina are shown in Table 21-86.

Table 21-86 Features distinguishing conus lesions from cauda equina lesions⁷⁸⁴

	Conus medullaris lesions	Cauda equina lesions
spontaneous pain	rare; when present, is usually bilateral & symmetric in perineum or thighs	may be most prominent symptom; severe; radicular type; in perineum, thighs & legs, back or bladder
sensory deficit	saddle; bilateral; usually symmetric; sensory dissociation	saddle; no sensory dissociation; may be unilateral & asymmetric
motor loss	symmetric; not marked; fasciculations may be present	asymmetric; more marked; atrophy may occur; fasciculations rare
autonomic symptoms (including bladder dysfunction, impotency…)	prominent early	late
reflexes	only ankle jerk absent (preserved knee jerk)	ankle jerk & knee jerk may be absent
onset	sudden and bilateral	gradual and unilateral

GRADING FUNCTION

There is prognostic significance in the presenting neurologic condition. Grading scales such as that of Brice and McKissock (see Table 21-87) have been proposed.

Table 21-87 Grading spinal cord function with spinal metastases (Brice & McKissock)⁷⁸⁵

Group	Grade	Description
1	mild	patient able to walk
2	moderate	able to move legs, but not antigravity
3	severe	slight residual motor and sensory function
4	complete	no motor, sensory, or sphincter function below level of lesion

DIAGNOSTIC TESTS

PLAIN X-RAYS

Most spinal mets are osteolytic, but at least 50% of the bone must be eroded before plain x-rays will be abnormal⁷⁸⁶. Not very specific. Possible findings: pedicle erosion (defect in “owl’s eyes” AKA “winking owl sign” on LS or thoracic spine AP view) or widening, pathological compression fracture, vertebral body (VB) scalloping, VB sclerosis, osteoblastic changes (may occur with prostate Ca, Hodgkin’s disease, occasionally with breast Ca, and rarely with multiple myeloma)

MRI IN EVALUATING SEM

The test of choice in most situations. Advantages of MRI over myelography⁷⁷⁴:

1. non-invasive. Doesn’t require second procedure (C1-2 puncture) if complete block

2. no risk of neurologic deterioration from LP in patient with complete block

3. detects lesions that do not cause bony destruction or distortion of the spinal subarachnoid space

4. up to 20% of patients with SEM have at least two sites of cord compression, MRI can evaluate region between two complete blocks, myelography cannot

5. demonstrates paraspinal lesions

Disadvantages of MRI versus myelography:

1. does not obtain CSF for cytological study

2. contraindicated with cardiac pacemaker or internal defibrillator

MRI findings in spinal epidural metastases:

1. vertebral mets are slightly hypointense compared to normal bone marrow on T1WI, and are slightly hyperintense on T2WI

2. axial cuts typically show lesion involving the posterior vertebral body with invasion into one or both pedicles

3. when myelopathy or radiculopathy are present, there is usually tumor extension into the spinal canal (may not occur in lesions presenting only with local pain)

4. DWI images may help differentiate osteoporotic compression fracture from pathologic fracture⁷⁸⁷

CT IN EVALUATING SEM

Very good for bone detail. Often helpful for planning fusions. By itself, has low sensitivity for spinal cord compression by tumor. Sensitivity is increased with intrathecal contrast.

POSITRON EMISSION TOMOGRAPHY (PET)SCAN

PET scan using [18F]-fluorodeoxyglucose may be used for whole-body work-up for bone mets in patients with known cancer⁷⁸⁸. Sensitivity is high, but spatial resolution and specificity are low, so often must be used with CT and/or MRI.

MANAGEMENT

Patients are categorized into one of the three following groups below based on the rapidity and seriousness of the neurologic findings⁷⁸⁴.

A metastatic work-up is undertaken as time permits (see Metastatic work-up, page 747) (a preliminary work-up, e.g. CXR and physical exam, may be all that can be initially obtained for patients in Group I, whereas more complete work-up can be done in others).

GROUP I

Signs/symptoms of new or progressive (hours to days) cord compression (e.g. urinary urgency, ascending numbness). These patients have a high risk of rapid deterioration and require immediate evaluation.

Management

1. dexamethasone (DMZ) (Decadron®): reduces pain in 85%, may produce transient neurologic improvement. Optimal dose is not known. No difference was found comparing 100 mg IV bolus to 10 mg⁷⁸⁹. Suggestion: 10 mg IV or PO q 6 hrs x 72 hrs, followed by lower dose of 4-6 mg q 6 hrs

2. radiographic evaluation

A. plain x-rays of entire spine: 67-85% will be abnormal (see above)

B. emergency MRI (see MRI in evaluating SEM above)

C. emergency myelogram: if MRI cannot be done (include possible C1-2 puncture on the consent). Start with a so-called “blockogram” to R/O complete block: instill small volume of contrast (e.g. iohexol, see page 122) via LP and run the contrast all the way up the spinal column (CSF is usually xanthochromic with complete block, see Froin’s syndrome, page 733)

1. if there is not a complete block: with-draw 10 cc of CSF and send for cytology, protein & glucose. One may then inject more contrast to complete the study

2. if complete block: do not remove CSF (pressure shifts via LP caused neurologic deterioration in ≈ 14% of patients with complete block⁷⁹⁰, whereas there was no deterioration after C1-2 puncture)

a. in some cases, contrast can be “squeezed” past a “complete” block by injecting 5-10 ml of room air through a millipore filter⁷⁹¹

OR

b. perform a lateral C1-2 puncture (see page 205) and instill water soluble contrast to delineate the superior extent of the lesion

3. with myelography, epidural lesions classically produce hourglass deformity with smooth edges if block is incomplete, or paintbrush effect (feathered edges) if block is complete, unlike the sharp margins (capping or meniscus sign) of intradural extramedullary lesion, or fusiform cord widening of intramedullary tumors

D. bone scan if time permits. Abnormal in ≈ 66% of patients with spine mets

3. treatment based on results of radiographic evaluation

A. if no epidural mass: treat primary tumor (e.g. systemic chemotherapy). Local radiation therapy (XRT) to bony lesion if present. Analgesics for pain

B. if epidural lesion, either surgery or start XRT (usually 30-40 Gy in 10 treatments over 7-10 d with ports extending 2 levels above and below lesion). XRT is usually as effective as laminectomy with fewer complications (for further discussion see Treatment for SEM, page 747). Thus, surgery instead of XRT is considered only for the indications shown in Table 21-88

C. urgency of treatment (surgery or XRT) is based on degree of block and rapidity of deterioration:

1. if > 80% block or rapid progression of deficit: emergency treatment ASAP (if treating with XRT instead of surgery, continue DMZ next day at 24 mg IV q 6 hrs x 2 days, then taper during XRT over 2 wks)

2. if < 80% block: treatment on “routine” basis (for XRT, continue DMZ 4 mg IV q 6 hrs, taper during treatment as tolerated)

Table 21-88 Indications for surgery for spinal metastases

Indications

1. unknown primary and no tissue diagnosis (CT guided needle biopsy is an option for accessible lesions). NB: lesions such as spinal epidural abscess can be mistaken for metastases792 2. spinal instability 3. deficit due to spinal deformity or compression by bone rather than by tumor (e.g. due to compression fracture with collapse and retropulsed bone) 4. radio-resistant tumors (e.g. renal-cell carcinoma, melanoma…) or progression during XRT (usual trial: at least 48 hrs, unless significant or rapid deterioration) 5. recurrence after maximal XRT 6. rapid neurologic deterioration

Relative contraindications

1. very radiosensitive tumors (multiple myeloma, lymphoma…) not previously radiated 2. total paralysis (Brice and McKissock group 4) > 8 hours duration, or inability to walk (B&M group > 1) for > 24 hrs duration (after this, there is essentially no chance of recovery and surgery is not indicated) 3. expected survival: ≤ 3-4 months 4. multiple lesions at multiple levels 5. patient unable to tolerate surgery: for patients with lung lesions, check PFTs

GROUP II

Mild and stable signs/symptoms of cord compression (e.g. isolated Babinski), or either plexopathy or radiculopathy without evidence of cord compression. Admit and evaluate within 24 hrs.

Management

1. for suspected ESCC, manage as in Group I except on less emergent basis. Use low dose dexamethasone (DMZ) unless radiographic evaluation shows > 80% block

2. for radiculopathy alone (radicular pain, weakness or reflex changes in one myotome or sensory changes in one dermatome): if plain x-rays show bony lesion then 70-88% will have ESCC on myelography. If the plain film is normal, only 9-25% will have ESCC. Obtain MRI or myelogram and manage as for suspected ESCC

3. for plexopathy (brachial or lumbosacral): pain is the most common early symptom, distribution not limited to single dermatome, commonly referred to elbow or ankle. May mask coexistent radiculopathy, distinguish by EMG (denervation of paraspinal muscles occurs in radiculopathy) or presence of proximal signs and symptoms (Horner’s syndrome in cervical region, ureteral obstruction in lumbar region). Management:

A. MRI is initial diagnostic procedure (CT if MRI unavailable): C4 through T4 for brachial plexopathy, L1 through pelvis for lumbosacral

B. if CT shows bony lesion or paraspinal mass (with negative CT, plain films and bone scan are rarely helpful; however, if done, and plain x-ray shows malignant appearing bony lesion, or if bone scan shows vertebral abnormality, perform MRI or myelogram within 24 hrs) (give dexamethasone if ESCC suspected or MRI/myelogram delayed). Management as in Group I based on degree of block, XRT ports extended laterally to include any mass shown on CT

C. if no bony nor paraspinal lesion on MRI/CT, primary treatment of plexus tumor; analgesics for pain

GROUP III

Back pain without neuro signs/symptoms. Can be evaluated as outpatient over several days (modify based on ability of patient to travel, reliability, etc.).

Management

1. plain spine x-ray (AP, lat, oblique)

A. if focal bony pathology demonstrated: obtain MRI or myelogram (66% of patients with isolated LBP and X-ray abnormalities had SEM; in 81% with vertebral collapse > 50%, in 31% with pedicle erosion only, and in 7% with vertebral body lesion without collapse). Proceed as in Group I based on results of MRI/myelogram

B. if plain films normal: bone scan (bone scan positive in up to 20% of patients with normal plain films with ESCC). Consider MRI or myelogram if bone scan abnormal in absence of benign x-ray lesion

2. CT: if bony lesion or paraspinal mass, proceed to MRI or myelogram, otherwise primary tumor treatment and analgesics

METASTATIC WORK-UP

The appropriateness of the following tests depends on the amount of time available as well as clinical information that may rule some primaries in or out.

1. CXR: to rule out lung primary or other mets to lung

2. CT of chest, abdomen and pelvis

3. serum prostate specific antigen (PSA) in males

4. mammogram in females

5. for multiple myeloma: see page 741

6. careful physical exam of lymph nodes

TREATMENT FOR SEM

TREATMENT GOALS AND OUTCOME

No treatment for SEM significantly prolongs life. Treatment goals are palliative: pain control, preservation of spinal stability, and maintenance of sphincter control and ability to ambulate.

The most important factor affecting prognosis, regardless of treatment modality, is ability to walk at the time of initiation of therapy. Loss of sphincter control is a poor prognosticator and is usually irreversible.

The main decision is between surgery + post-op XRT, or XRT alone. As yet, no chemotherapy found useful for SEM (may help with primary). Surgery alone appears least effective for pain control (36%, compared to 67% for surgery + XRT, and 76% for XRT alone)⁷⁹³. Surgery has the attendant complications of anesthetic risk, post-op pain, wound problems in 11% (further complicated by radiation)⁷⁹³, and mortality in 5-6% after laminectomy and 10% after anterior approach with stabilization⁷⁹⁴. Therefore, surgery appears best reserved for situations described in Table 21-88, page 745.

Deterioration in one of the 3 major criteria (pain, continence, ambulation) occurred in 26% of patients treated with laminectomy alone, 20% of laminectomy + XRT, and 17% of XRT alone (roughly comparable). There is a 9% incidence of spinal instability⁷⁹³ following laminectomy without stabilization.

MEDICAL THERAPY

Chemotherapy is ineffective for SEM.

Bisphosphonates reduce the risk of vertebral compression fractures (VCF) by ≈ 50%, but the effect seems to abate after ≈ 2-3 years.

Promising agents undergoing trials include: denosumab, a RANK ligand (RANKL) inhibitor (see page 994) that may counteract RANKL which is overexpressed in response to lytic bony metastases⁷⁹⁵. The efficacy seems better than the bisphosphonate.

VERTEBROPLASTY/KYPHOPLASTY

Vertebroplasty/kyphoplasty (see page 994) reduces pain associated with pathologic fractures in up to 84%⁷⁹⁶ with an associated increase in functional outcome⁸¹⁶. Kyphoplasty appears to offer comparable pain relief to vertebroplasty with lower rates of cement leakage⁸¹⁶.

Relative contraindication: spinal cord compression. Unless the diagnosis has already been verified, a biopsy should be taken through one of the pedicles prior to injecting PMMA.

RADIATION THERAPY

Radiosensitive tumors: Table 21-70, page 708 lists radiosensitivity of metastatic tumors (to brain or spine). Other radiosensitive tumors that metastasize to the spine include: myxoid liposarcoma⁷⁹⁷.

Treatment⁷⁹⁸: Dose: range = 25-40 Gy. Typical plan: 30 Gy delivered in 3 Gy fractions over 10 days (2 working weeks) to ports extending at least 1 vertebral level above and below the extent o f the lesion. Timing: for initial treatment, try to start XRT within 24 hours of diagnosis; for post-op XRT, within about 14 days following surgery.

There is a theoretical risk of radiation induced edema causing or accelerating neurologic deterioration. This has not been born out by experimental studies with the usual small daily fractions utilized. Deterioration is more likely to be due to tumor progression⁷⁹⁹. The spinal cord is usually the dose limiting structure in treating SEM.

Increased doses are being made possible with the application of the added precision of stereotactic radiosurgery techniques to spinal metastases⁸⁰⁰.

SURGICAL TREATMENT

See Table 21-88, page 745 for indications for surgery.

Pre-op embolization by interventional radiologist may facilitate resection with less blood loss for highly vascular tumors such as: renal-cell, thyroid, and hepatocellular. Blood supply is through the intercostal arteries, and care must be taken to avoid embolizing vessels providing significant blood supply to the spinal cord (especially the artery of Adamkiewicz (see page 96)).

TECHNIQUE

Laminectomy alone is poor for spinal metastases when the pathology is anterior to the cord because of poor access to the tumor and the destabilizing effect of laminectomy when metastatic involvement of the vertebral body is significant^{801, 802}.

In a randomized controlled trial by Patchell et al.⁸⁰³, approaches directed at the location of the tumor (e.g. costotransversectomy, transthoracic approach…) with stabilization where necessary, produced better results than simple laminectomy, and surgery + XRT was superior to XRT alone (see Table 21-89). This study found a modest increase in survival, but more significant maintenance or regaining of lost ambulation. However, operative mortality with anterior decompression and stabilization was ≈ double (10%) that of laminectomy with (5%) or without (6%) stabilization in a literature review⁷⁹⁴.

Solitary spinal metastases with indolent tumors (e.g. renal cell Ca) may be candi-dates for attempted cure with en bloc resection (total spondylectomy)^{804, 805}.

Laminectomy is still appropriate with isolated involvement of the posterior elements. For anterior pathology, if the posterior elements are intact, a transthoracic approach with corpectomy and stabilization (e.g. with methylmethacrylate and Steinmann pins⁸⁰⁶, or with cage graft and lateral plate) followed by XRT improves neurologic function in ≈ 75% and pain in ≈ 85%. A posterolateral approach (e.g. costotransversectomy) may be used for anterolateral tumor⁸⁰⁷. Combining a corpectomy and removal of the pedicle and posterior elements destabilizes the spine, therefore posterior instrumentation prior to performing the corpectomy is required, followed by cage graft^808-814. To access a VB via a costotransversectomy, the rib of the like numbered VB and the one below need to be removed.

Table 21-89 Comparing surgery + XRT to XRT alone⁸⁰³

	XRT	Surgery + XRT
Ambulatory after treatment	57%	84%
Days ambulatory after treatment	13	122
Ambulatory after treatment when nonambulatory before treatment	19%	62%
Mean survival (days)	100	126

21.13. Neuroblastomas

Tumors arising from sympathetic ganglion⁸¹⁵. May occur anywhere in the sympathetic nervous system, most commonly from adrenal gland (40%), followed by sympathetic ganglia of thoracic (15%), cervical (5%) and pelvic regions (5%). Neoplasms under this rubric include:

1. neuroblastomas: the most undifferentiated and aggressive in this group. Olfactory neuroblastomas are called esthesioneuroblastomas (discussed on page 1230)

2. ganglioneuroblastomas

3. ganglioneuromas

Presentation

May present with abdominal mass, local or radicular pain, or (with high thoracic or cervical tumors) Horner’s syndrome. Spinal cord compression may occur from invasion through the neural foramen, and scoliosis may occur. Catecholamine precursors (homo-vanillic acid (HVA), vanillylmandelic acid (VMA) and dopamine) may be excreted and cause HTN (can be assayed in urine). Periorbital tumor metastases may produce raccoon’s eyes (usually unilateral ecchymosis and proptosis). Many of the low-grade tumors regress spontaneously and never present.

21.14. References

1. Kleihues P, Cavenee W K, (eds.): World Health Organization classification of tumors: Pathology and genetics of tumors of the nervous system. 3rd ed., International Agency for Research on Cancer Press, Lyon, 2000.

2. Kleihues P, Louis D N, Scheithauer B W, et al.: The WHO classification of tumors of the nervous system. J Neuropathol Exp Neurol 61 (3): 215-25; discussion 226-9, 2002.

3. Kleihues P, Burger P C, Scheithauer B W: The new WHO classification of brain tumors. Brain Pathol 3: 255-68, 1993.

4. Escourolle R, Poirier J: Manual of basic neuropathology. 2nd ed. W. B. Saunders, Philadelphia, 1971.

5. Louis D N, Ohgaki H, Wiestler O D, et al., (eds.): WHO classification of tumors of the central nervous system. 4th ed., World Health Organization classification of tumors, ed. Bosman F T, Jaffe E S, Lakhani S R, et al.International Agency for Research on Cancer, Lyon, 2007: pp 309.

6. Mahaley M S, Mettlin C, Natarajan N, et al.: National survey of patterns of care for brain-tumor patients. J Neurosurg 71: 826-36, 1989.

7. Whittle I R, Pringle A-M, Taylor R: Effects of resective surgery for left-sided intracranial tumors on language function: A prospective study. Lancet 351: 1014-8, 1998.

8. Forsyth P A, Posner J B: Headaches in patients with brain tumors: A study of 111 patients. Neurology 43: 1678-83, 1993.

9. Paraf F, Jothy S, Van Meir E G: Brain tumor-polyposis syndrome: Two genetic diseases. J Clin Oncol 15: 2744-58, 1997.

10. Hottinger A F, Khakoo Y: Neurooncology of familial cancer syndromes. J Child Neurol 24 (12): J Child Neurol: 1526-35, 2009.

11. Galicich J H, French L A: Use of dexamethasone in the treatment of cerebral edema resulting from brain tumors and brain surgery. Am Pract Dig Treat 12: 169-74, 1961.

12. French L A, Galicich J H: The use of steroids for control of cerebral edema. Clin Neurosurg 10: 212-23, 1964.

13. Glantz M J, Cole B F, Forsyth P A, et al.: Practice parameter: Anticonvulsant prophylaxis in patients with newly diagnosed brain tumors. Report of the quality standards subcommittee of the American Academy of Neurology. Neurology 54: 1886-93, 2000.

14. Chicoine M R, Silbergeld D L: Pharmacology for neurosurgeons. Part I: Anticonvulsants, chemotherapy, antibiotics. Contemp Neurosurg 18 (9): 1-6, 1996.

15. Prados M D, Berger M S, Wilson C B: Primary central nervous system tumors: Advances in knowledge and treatment. CA Cancer J Clin 48: 331-60, 1998.

16. Stewart D J: A critique of the role of the blood-brain barrier in the chemotherapy of human brain tumors. J Neurooncol (20): 121-39, 1994.

17. Salcman M, Broadwell R D: The blood brain barrier. In Neurobiology of brain tumors, Salcman M, (ed.). Williams and Wilkins, Baltimore, 1991, Vol. 4: pp 229-50.

18. Madajewicz S, Chowhan N, Tfayli A, et al.: Therapy for patients with high grade astrocytoma using intraarterial chemotherapy and radiation therapy. Cancer 88: 2350-6, 2000.

19. Barker F G, Prados M D, Chang S M, et al.: Radiation response and survival time in patients with glioblastoma multiforme. J Neurosurg 84: 442-8, 1996.

20. Laohaprasit V, Silbergeld D L, Ojemann G A, et al.: Postoperative CT contrast enhancement following lobectomy for epilepsy. J Neurosurg 73: 392-5, 1990.

21. Jeffries B F, Kishore P R, Singh K S, et al.: Contrast enhancement in the posoperative brain. Radiology 139: 409-13, 1981.

22. Gerber A M, Savolaine E R: Modification of tumor enhancement and brain edema in computerized tomography by corticosteroids: Case report. Neurosurgery 6: 282-4, 1980.

23. Hatam A, Bergström M, Yu Z Y, et al.: Effect of dexamethasone treatment in volume and contrast enhancement of intracranial neoplasms. J Comput Assist Tomogr 7: 295-300, 1983.

24. Albright L, Reigel D H: Management of hydrocephalus secondary to posterior fossa tumors. Preliminary report. J Neurosurg 46: 52-5, 1977.

25. Berger M S, Baumeister B, Geyer J R, et al.: The risks of metastases from shunting in children with primary central nervous system tumors. J Neurosurg 74: 872-7, 1991.

26. McLaurin R L, Venes J L, (eds.): Pediatric neurosurgery. 2nd ed., W. B. Saunders, Philadelphia, 1989.

27. Daumas-Duport C, Scheithauer B W, Kelly P J: A histologic and cytologic method for the spatial definition of gliomas. Mayo Clin Proc 62: 435-49, 1987.

28. Gunel M, Piepmeier J M: Management of low-grade gliomas. Contemp Neurosurg 19 (9): 1-6, 1997.

29. Packer R J, Lange B, Ater J, et al.: Carboplatin and vincristine for recurrent and newly diagnosed low-grade gliomas of childhood. J Clin Oncol 11: 850-6, 1993.

30. Daumas-Duport C, Scheithauer B W, Chodkiewicz J-P, et al.: Dysembryoplastic neuroepithelial tumor: A surgically curable tumor of young patients with intractable seizures. Neurosurgery 23: 545-56, 1988.

31. Daumas-Duport C, Varlet P, Bacha S, et al.: Dysembryoplastic neuroepithelial tumors: Nonspecific histological forms -- a study of 40 cases. J Neurooncol 41 (3): 267-80, 1999.

32. Adada B, Sayed K: Dysembryoplastic neuroepithelial tumors. Contemp Neurosurg 26 (23): 1-5, 2004.

33. Pahapill P A, Ramsay D A, Del Maestro R F: Pleomorphic xanthoastrocytoma: Case report and analysis of the literature concerning the efficacy of resection and the significance of necrosis. Neurosurgery38 (4): 822-8; discussion 828-9, 1996.

34. Fouladi M, Jenkins J, Burger P, et al.: Pleomorphic xanthoastrocytoma: Favorable outcome after complete surgical resection. Neuro-oncol 3 (3): 184-92, 2001.

35. Kepes J J, Rubinstein L J, Eng L F: Pleomorphic xanthoastrocytoma: A distinctive meningeal glioma of young subjects with relatively favorable prognosis. A study of 12 cases. Cancer 44: 1839-52, 1979.

36. Weldon-Linne C M, Victor T A, Groothuis D R, et al.: Pleomorphic xanthoastrocytoma: Ultrastructural and immunohistochemical study of a case with a rapidly fatal outcome following surgery. Cancer52: 2055-63, 1983.

37. Kumar S, Retnam T M, Menon G, et al.: Cerebellar hemisphere, an uncommon location for pleomorphic xanthoastrocytoma and lipidized glioblastoma multiformis. Neurol India 51 (2): 246-7, 2003.

38. Giannini C, Scheithauer B W, Burger P C, et al.: Pleomorphic xanthoastrocytoma: What do we really know about it? Cancer 85 (9): 2033-45, 1999.

39. Robinson S, Cohen A R: Cowden disease and Lhermitte-Duclos disease: An update. Case report and review of the literature. Neurosurg Focus 20 (1): E6, 2006.

40. Nelen M R, van Staveren W C, Peeters E A, et al.: Germline mutations in the PTEN/MMAC1 gene in patients with Cowden disease. Hum Mol Genet 6 (8): Hum Mol Genet: 1383-7, 1997.

41. Meltzer C C, Smirniotopoulos J G, Jones R V: The striated cerebellum: An MR imaging sign in Lhermitte-Duclos disease (dysplastic gangliocytoma). Radiology 194 (3): 699-703, 1995.

42. Chen K S, Hung P C, Wang H S, et al.: Medulloblastoma or cerebellar dysplastic gangliocytoma (Lhermitte-Duclos disease)? Pediatr Neurol 27 (5): 404-6, 2002.

43. Someshwar S, Hogg J P, Nield L S: Lhermitte-Duclos disease or neoplasm? Applied Neurology 3 (6): 37-9, 2007.

44. Suresh T N, Santosh V, Yasha T C, et al.: Medulloblastoma with extensive nodularity: A variant occurring in the very young-clinicopathological and immunohistochemical study of four cases. Childs Nerv Syst 20 (1): 55-60, 2004.

45. Capone Mori A, Hoeltzenbein M, Poetsch M, et al.: Lhermitte-Duclos disease in 3 children: A clinical long-term observation. Neuropediatrics 34 (1): 30-5, 2003.

46. Carlson J J, Milburn J M, Barre G M: Lhermitte-Duclos disease: Case report. J Neuroimaging 16 (2): 157-62, 2006.

47. Russell D S, Rubenstein L J: In Pathology of tumours of the nervous system. Williams and Wilkins, Baltimore, 5th ed., 1989: pp 83-161.

48. Kleihues P, Louis D N, Wiestler O D, et al.: WHO grading of tumors of the central nervous system. In WHO classification of tumors of the central nervous system, Louis D N, Ohgaki H, Wiestler O D, et al., (eds.). World Health Organization classification of tumors. Bosman F T, Jaffe E S, Lakhani S R, et al. International Agency for Research on Cancer, Lyon, 4th ed., 2007: pp 10-11.

49. Kernohan J W, Mabon R F, Svien H J, et al.: A simplified classification of the gliomas. Proc Staff Meet Mayo Clin 24: 71-5, 1949.

50. Daumas-Duport C, Scheithauer B, O’Fallon J, et al.: Grading of astrocytomas: A simple and reproducible method. Cancer 62: 2152-65, 1988.

51. Kim T S, Halliday A L, Hedley-Whyte T, et al.: Correlates of survival and the Daumas-Duport grading system for astrocytomas. J Neurosurg 74: 27-37, 1991.

52. Shafqat S, Hedley-Whyte E T, Henson J W: Age-dependent rate of anaplastic transformation in low-grade astrocytoma. Neurology 52: 867-9, 1999.

53. Berger M S: Role of surgery in diagnosis and management. In Benign cerebral glioma, Apuzzo M L J, (ed.). Neurosurgical topics. American Association of Neurological Surgeons, Park Ridge, Illinois, 1995, Vol. 2: pp 293-307.

54. Laws E R, Taylor W F, Clifton M B, et al.: Neurosurgical management of low-grade astrocytoma of the cerebral hemispheres. J Neurosurg 61: 665-73, 1984.

55. Kopelson G, Linggood R: Infratentorial glioblastoma: The role of neuraxis irradiation. Int J Radiation Oncology Biol Phys 8: 999-1003, 1982.

56. Kondziolka D, Lunsford L D, Martinez A J: Unreliability of contemporary neurodiagnostic imaging in evaluating suspected adult supratentorial (low-grade) astrocytoma. J Neurosurg 79: 533-6, 1993.

57. Zee C S, Conti P, Destian S, et al.: Imaging features of benign gliomas. In Benign cerebral glioma, Apuzzo M L J, (ed.). Neurosurgical topics. American Association of Neurological Surgeons, Park Ridge, Illinois, 1995, Vol. 2: pp 247-74.

58. Chang E F, Smith J S, Chang S M, et al.: Preoperative prognostic classification system for hemispheric low-grade gliomas in adults. J Neurosurg 109: 817-24, 2008.

59. Pignatti F, van den Bent M, Curran D, et al.: Prognostic factors for survival in adult patients with cerebral low-grade glioma. J Clin Oncol 20 (8): J Clin Oncol: 2076-84, 2002.

60. Narayan P, Olson J J: Management of anaplastic astrocytoma. Contemp Neurosurg 23 (24): 1-6, 2001.

61. Chamberlain M C, Murovic J, Levin V A: Absence of contrast enhancementon CT brain scans of patients with supratentorial malignant gliomas. Neurology 38: 1371-3, 1988.

62. Greene G M, Hitchon P W, Schelper R L, et al.: Diagnostic yield in CT-guided stereotactic biopsy of gliomas. J Neurosurg 71: 494-7, 1989.

63. Scherer H J: The forms of growth in gliomas and their practical significance. Brain 63: 1-35, 1940.

64. Choucair A K, Levin V A, Gutin P H, et al.: Development of multiple lesions during radiation therapy and chemotherapy. J Neurosurg 65: 654-8, 1986.

65. Erlich S S, Davis R L: Spinal subarachnoid metastasis from primary intracranial glioblastoma multi-forme. Cancer 42: 2854-64, 1978.

66. Artigas J, Cervos-Navarro J, Iglesias J R, et al.: Gliomatosis cerebri: Clinical and histological findings. Clin Neuropathol 4: 135-48, 1985.

67. Wilson N W, Symon L, Lantos P L: Gliomatosis cerebri: Report of a case presenting as a focal cerebral mass. J Neurol 234: 445-7, 1987.

68. Barnard R O, Geddes J F: The incidence of multifocal cerebral gliomas: A histological study of large hemisphere sections. Cancer 60: 1519-31, 1987.

69. van Tassel P, Lee Y-Y, Bruner J M: Synchronous and metachronous malignant gliomas: CT findings. AJNR 9: 725-32, 1988.

70. Harsh G R, Wilson C B: Nuroepithelial tumors of the adult brain. In Neurological surgery, Youmans J R, (ed.). W. B. Saunders, Philadelphia, 3rd ed., 1990, Vol. 3: pp 3040-136.

71. Salvati M, Caroli E, Orlando E R, et al.: Multicentric glioma: Our experience in 25 patients and critical review of the literature. Neurosurg Rev 26 (4): Neurosurg Rev: 275-9, 2003.

72. Cairncross J G, Laperriere N J: Low-grade glioma: To treat or not to treat? Arch Neurol 46: 1238-9, 1989.

73. Shaw E G: Low-grade gliomas: To treat or not to treat? A radiation oncologist’s viewpoint. Arch Neurol 47: 1138-9, 1990.

74. Karim A B M F, Maat B, Hatlevoll R, et al.: A randomized trial on dose-response in radiation therapy of low-grade cerebral glioma: European organization for research and treatment of cancer (EORTC) study 22844. Int J Radiation Oncology Biol Phys 36: 549-56, 1996.

75. Shibamoto Y, Kitakabu Y, Takahashi M, et al.: Supratentorial low-grade astrocytoma. Correlation of computed tomography findings with effect of radiation therapy and prognostic variables. Cancer 72: 190-5, 1993.

76. Kelly P J: Role of stereotaxis in the management of low-grade intracranial gliomas. In Benign cerebral glioma, Apuzzo M L J, (ed.). Neurosurgical topics. American Association of Neurological Surgeons, Park Ridge, Illinois, 1995, Vol. 2: pp 275-92.

77. Philippon J H, Clemenceau S H, Fauchon F H, et al.: Supratentorial low-grade astrocytomas in adults. Neurosurgery 32: 554-9, 1993.

78. van den Bent M J, Afra D, de Witte O, et al.: Long-term efficacy of early versus delayed radiotherapy for low-grade astrocytoma and oligodendroglioma in adults: The EORTC 22845 randomised trial. Lancet 366 (9490): Lancet: 985-90, 2005.

79. Shaw E G, Daumas-Duport C, Scheithauer B W, et al.: Radiation therapy in the management of low-grade supratentorial astrocytomas. J Neurosurg 70: 853-61, 1989.

80. Shaw E, Arusell R, Scheithauer B, et al.: Prospective randomized trial of low- versus high-dose radiation therapy in adults with supratentorial low-grade glioma: Initial report of a North Central Cancer Treatment Group/Radiation Therapy Oncology Group/Eastern Cooperative Oncology Group study. J Clin Oncol 20 (9): J Clin Oncol: 2267-76, 2002.

81. Laack N N, Brown P D, Ivnik R J, et al.: Cognitive function after radiotherapy for supratentorial low-grade glioma: A North Central Cancer Treatment Group prospective study. Int J Radiat Oncol Biol Phys 63 (4): Int J Radiat Oncol Biol Phys: 1175-83, 2005.

82. Morantz R A: Radiation therapy in the treatment of cerebral astrocytoma. Neurosurgery 20: 975-82, 1987.

83. Quinn J A, Reardon D A, Friedman A H, et al.: Phase II trial of temozolomide in patients with progressive low-grade glioma. J Clin Oncol 21 (4): 646-51, 2003.

84. Coffey R J, Lunsford L D, Taylor F H: Survival after stereotactic biopsy of malignant gliomas. Neurosurgery 22: 465-73, 1988.

85. Thomson A-M, Taylor R, Fraser D, et al.: Stereotactic biopsy of nonpolar tumors in the dominant hemisphere: A prospective study of effects on language functions. J Neurosurg 89: 923-6, 1997.

86. Leibel S A, Sheline G E: Radiation therapy for neoplasms of the brain. J Neurosurg 66: 1-22, 1987.

87. Kelly P J, Hunt C H: The limited value of cytoreductive surgery in elderly patients with malignant gliomas. Neurosurgery 33: 62-7, 1994.

88. Keles G E, Anderson B, Berger M S: The effect of extent of resection on time to tumor progression and survival in patients with glioblastoma multiforme of the cerebral hemirsphere. Surg Neurol 52: 371-9, 1999.

89. Stummer W, Reulen H-J, Meinel T, et al.: Extent of resection and survival in glioblastoma multiforme: Identification of and adjustment for bias. Neurosurgery 62: 564-76, 2008.

90. Friedman A H: Comment on Stummer W, et al.: Extent of resection and survival in glioblastoma multi-forme: Identification of and adjustment for bias. Neurosurgery 62: 574-5, 2008.

91. Vuorinen V, Hinkka S, Farkkila M, et al.: Debulking or biopsy of malignant glioma in elderly people - a randomized study. Acta Neurochir (Wien) 145: 5-10, 2003.

92. Piepmeier J M: Comment on Stummer W, et al., extent of resection and survival in glioblastoma multi-forme: Identification of and adjustment for bias. Neurosurgery 62: 576, 2008.

93. Apuzzo M L J: Comment on Coffey R J, et al.: Survival after stereotactic biopsy of malignant gliomas. Neurosurgery 22: 472-3, 1988.

94. Quigley M R, Maroon J C: The relationship between survival and the extent of resection in patients with supratentorial malignant gliomas. Neurosurgery 29: 385-9, 1991.

95. Shapiro W R, Green S B, Burger P C, et al.: Randomized trial of three chemotherapy regimens and two radiotherapy regimens in postoperative treatment of malignant glioma: Brain tumor cooperative group trial 8001. J Neurosurg 71: 1-9, 1989.

96. Sneed P K, McDermott M W, Gutin P H: Interstitial brachytherapy procedures for brain tumors. Semin Surg Oncol 13: 157-66, 1997.

97. Souhami L, Seiferheld W, Brachman D, et al.: Randomized comparison of stereotactic radiosurgery followed by conventional radiotherapy with carmustine to conventional radiotherapy with carmustine for patients with glioblastoma multiforme: Report of Radiation Therapy Oncology Group 93-05 protocol. Int J Radiat Oncol Biol Phys 60 (3): 853-60, 2004.

98. Kornblith P L: The role of cytotoxic chemotherapy in the treatment of malignant brain tumors. Surg Neurol 44: 551-2, 1995.

99. Medical Research Council Brain Tumour Working Party: Randomized trial of procarbazine, lomustine, and vincristine in the adjuvant treatment of high-grade astrocytoma: A Medical Research Council trial. J Clin Oncol 19 (2): 509-18, 2001.

100. Fine H A, Dear K B, Loeffler J S, et al.: Meta-analysis of radiation therapy with and without adjuvant chemotherapy for malignant gliomas in adults. Cancer 71: 2585-97, 1993.

101. Kornblith P L: Chemotherapy for malignant brain tumors. J Neurosurg 68: 1-17, 1988.

102. Esteller M, Garcia-Foncillas J, Andion E, et al.: Inactivation of the DNA-repair gene mgmt and the clinical response of gliomas to alkylating agents. N Engl J Med 343: 1350-4, 2000.

103. Bleehen N M, Stenning S P: A Medical Research Council trial of two radiotherapy doses in the treatment of grades 3 and 4 astrocytoma. The Medical Research Council Brain Tumour Working Party. Br J Cancer 64 (4): 769-74, 1991.

104. Stewart L A: Chemotherapy in adult high-grade glioma: A systematic review and meta-analysis of individual patient data from 12 randomised trials. Lancet 359 (9311): 1011-8, 2002.

105. Hochberg F H, Pruitt A A, Beck D O, et al.: The rationale and methodology for intraarterial chemo-therapy with BCNU as treatment for glioblastoma. J Neurosurg 63: 876-80, 1985.

106. Bashir R, Hochberg F H, Linggood R M, et al.: Preirradiation internal carotid artery BCNU in treatment of glioblastoma multiforme. J Neurosurg 68: 917-9, 1988.

107. Grossman S A, Reinhard C S, Colvin O M, et al.: The intracerebral distribution of BCNU delivery by surgically implanted biodegradable polymers. J Neurosurg 76: 640-7, 1992.

108. Westphal M, Ram Z, Riddle V, et al.: Gliadel wafer in initial surgery for malignant glioma: Long-term follow-up of a multicenter controlled trial. Acta Neurochir (Wien) 148 (3): 269-75; discussion 275, 2006.

109. Brem H, Piantadosi S, Burger P C, et al.: Placebocontrolled trial of safety and efficacy of intraoperative controlled delivery by biodegradable polymers of chemotherapy for recurrent gliomas. Lancet345: 1008-12, 1995.

110. Gliadel wafers for treatment of brain tumors. Med Letter 40 (1035): 92, 1998.

111. Friedman H S, McLendon R E, Kerby T, et al.: DNA mismatch repair and o⁶-alkylguanine-DNA alkyltransferase analysis and response to temodol in newly diagnosed malignant glioma. J Clin Oncol 16: 3851-7, 1998.

112. Yung W K, Prados M D, Yaya-Tur R, et al.: Multi-center phase II trial of temozolomide in patients with anaplastic astrocytoma or anaplastic oligoastrocytoma at first relapse. Temodal brain tumor group. J Clin Oncol 17: 2761-71, 1999.

113. Stupp R, Mason W P, van den Bent M J, et al.: European organisation for research and treatment of cancer brain tumor and radiotherapy groups; national cancer institute of Canada clinical trials group: Radiotherpy plus concommitant and adjuvant temozolamide for glioblastoma. N Engl J Med 352: 987-96, 2005.

114. Harsh G R, Levin V A, Gutin P H, et al.: Reoperation for recurrent glioblastoma and anaplastic astrocytoma. Neurosurgery 21: 615-21, 1987.

115. Ammirati M, Galicich J H, Arbit E, et al.: Reoperation in the treatment of recurrent intracranial malignant gliomas. Neurosurgery 21: 607-14, 1987.

116. Wen P Y, Fine H A, Black P M, et al.: High-grade astrocytomas. Neurol Clin 13: 875-96, 1995.

117. Burger P C, Vogel F S, Green S B, et al.: Glioblastoma multiforme and anaplastic astrocytoma: Pathologic criteria and prognostic implications. Cancer 56: 1106-11, 1985.

118. Burger P C, Scheithauer B W: Atlas of tumor pathology. Tumors of the central nervous system. Armed Forces Institute of Pathology, Washington, D.C., 1994.

119. Pollack I F, Hoffman H J, Humphreys R P, et al.: The long-term outcome after surgical treatment of dorsally exophytic brain-stem gliomas. J Neurosurg 78: 859-63, 1993.

120. Coakley K J, Huston J, Scheithauer B W, et al.: Pilocytic astrocytomas: Well-demarcated magnetic resonance appearance despite frequent infiltration histologically. Mayo Clin Proc 70: 747-51, 1995.

121. Hayostek C J, Shaw E G, Scheithauer B, et al.: Astrocytomas of the cerebellum: A comparative clinicopathologic study of pilocytic and diffuse astrocytomas. Cancer 72: 856-69, 1993.

122. Bernell W R, Kepes J J, Seitz E P: Late malignant recurrence of childhood cerebellar astrocytoma. J Neurosurg 37: 470-4, 1972.

123. Schwartz A M, Ghatak N R: Malignant transformation of benign cerebellar astrocytoma. Cancer 65: 333-6, 1990.

124. Wallner K E, Gonzales M F, Edwards M S B, et al.: Treatment of juvenile pilocytic astrocytoma. J Neurosurg 69: 171-6, 1988.

125. Zimmerman R A, Bilaniuk C T, Bruno L A, et al.: CT of cerebellar astrocytoma. Am J Roentgenol 130: 929-33, 1978.

126. Section of Pediatric Neurosurgery of the American Association of Neurological Surgeons, (ed.) Pediatric neurosurgery. 1st ed., Grune and Stratton, New York, 1982.

127. Youmans J R, (ed.) Neurological surgery. 3rd ed., W. B. Saunders, Philadelphia, 1990.

128. Ringertz N, Nordenstam H: Cerebellar astrocytoma. J Neuropathol Exp Neurol 10: 343-67, 1951.

129. Gol A: Cerebellar astrocytomas in children. Am J Dis Child 106: 21-4, 1963.

130. Winston K, Gilles F H, Leviton A, et al.: Cerebellar gliomas in children. J Natl Cancer Inst 58: 833-8, 1977.

131. Austin E J, Alvord E C: Recurrences of cerebellar astrocytomas: A violation of Collins’ law. J Neurosurg 68: 41-7, 1988.

132. Bucy P C, Thieman P W: Astrocytomas of the cerebellum. A study of patients operated upon over 28 years ago. Arch Neurol 18: 14-9, 1968.

133. Stein B M, Tenner M S, Fraser R A R: Hydrocephalus following removal of cerebellar astrocytomas in children. J Neurosurg 36: 763-8, 1972.

134. Tihan T, Fisher P G, Kepner J L, et al.: Pediatric astrocytomas with monomorphous pilomyxoid features and a less favorable outcome. J Neuropathol Exp Neurol 58 (10): J Neuropathol Exp Neurol: 1061-8, 1999.

135. Ceppa E P, Bouffet E, Griebel R, et al.: The pilomyxoid astrocytoma and its relationship to pilocytic astrocytoma: Report of a case and a critical review of the entity. J Neurooncol 81 (2): J Neurooncol: 191-6, 2007.

136. Arulrajah S, Huisman T A: Pilomyxoid astrocytoma of the spinal cord with cerebrospinal fluid and peritoneal metastasis. Neuropediatrics 39 (4): Neuro-pediatrics: 243-5, 2008.

137. Packer R J, Nicholson H S, Vezina L G, et al.: Brainstem gliomas. Neurosurg Clin N Am 3 (4): 863-79, 1992.

138. Laurent J P, Cheek W R: Brain tumors in children. J Pediatr Neurosci 1: 15-32, 1985.

139. Reigel D H, Scarff T B, Woodford J E: Biopsy of pediatric brain stem tumors. Childs Brain 5: 329-40, 1979.

140. Epstein F J, Farmaer J-P: Brain-stem glioma growth patterns. J Neurosurg 78: 408-12, 1993.

141. Epstein F, McCleary E L: Intrinsic brain-stem tumors of childhood: Surgical indications. J Neurosurg 64: 11-5, 1986.

142. Garcia C A, McGarry P A, Collada M: Ganglioglioma of the brain stem. Case report. J Neurosurg 60: 431-4, 1984.

143. Albright A L, Packer R J, Zimmerman R, et al.: Magnetic resonance scans should replace biopsies for the diagnosis of diffuse brain stem gliomas: A report from the Children’s Cancer Group. Neurosurgery 33: 1026-30, 1993.

144. Hoffman H J, Becker L, Craven M A: A clinically and pathologically distinct group of benign brainstem gliomas. Neurosurgery 7: 243-8, 1980.

145. Kyoshima K, Kobayashi S, Gibo H, et al.: A study of safe entry zones via the floor of the fourth ventricle for brain-stem lesions. J Neurosurg 78: 987-93, 1993.

146. Stark A M, Fritsch M J, Claviez A, et al.: Management of tectal glioma in childhood. Pediatr Neurol, 2005.

147. Bognar L, Turjman F, Villanyi E, et al.: Tectal plate gliomas. Part II: CT scans and MR imaging of tectal gliomas. Acta Neurochir (Wien) 127 (1-2): 48-54, 1994.

148. Pollack I F, Pang D, Albright A L: The long-term outcome in children with late-onset aqueductal stenosis resulting from benign intrinsic tectal tumors. J Neurosurg 80 (4): 681-8, 1994.

149. Grant G A, Avellino A M, Loeser J D, et al.: Management of intrinsic gliomas of the tectal plate in children. A ten-year review. Pediatr Neurosurg 31 (4): 170-6, 1999.

150. Oka K, Kin Y, Go Y, et al.: Neuroendoscopic approach to tectal tumors: A consecutive series. J Neurosurg 91 (6): 964-70, 1999.

151. Kihlstrom L, Lindquist C, Lindquist M, et al.: Stereotactic radiosurgery for tectal low-grade gliomas. Acta Neurochir Suppl 62: 55-7, 1994.

152. Mork S J, Lindegaard K F, Halvorsen T B, et al.: Oligodendroglioma: Incidence and biological behavior in a defined population. J Neurosurg 63: 881-9, 1985.

153. Daumas-Duport C, Tucker M-L, Kolles H, et al.: Oligodendrogliomas: Part II - a new grading system based on morphological and imaging criteria. J Neurooncol 34: 61-78, 1997.

154. Coons S W, Johnson P C, Scheithauer B W, et al.: Improving diagnostic accuracy and interobserver concordance in the classification and grading of primary gliomas. Cancer 79: 1381-93, 1997.

155. Daumas-Duport C, Varlet P, Tucker M-L, et al.: Oligodendrogliomas: Part I - patterns of growth, histological diagnosis, clinical and imaging correlations: A study of 153 cases. J Neurooncol 34: 37-59, 1997.

156. Chin H W, Hazel J J, Kim T H, et al.: Oligodendrogliomas. I. A clinical study of cerebral oligodendrogliomas. Cancer 45: 1458-66, 1980.

157. Roberts M, German W: A long term study of patients with oligodendrogliomas. J Neurosurg 24: 697-700, 1966.

158. Coons S W, Johnson P C, Pearl D K, et al.: The prognostic significance of Ki-67 labeling indices for oligodendrogliomas. Neurosurgery 41: 878-85, 1997.

159. Fortin D, Cairncross G J, Hammond R R: Oligodendroglioma: An appraisal of recent data pertaining to diagnosis and treatment. Neurosurgery 45: 1279-91, 1999.

160. Hart M N, Petito C K, Earle K M: Mixed gliomas. Cancer 33: 134-40, 1974.

161. Rutka J T, Murakami M, Dirks P B, et al.: Role of glial filaments in cells and tumors of glial origin: A review. J Neurosurg 87: 420-30, 1997.

162. Kros J M, Schouten W C D, Janssen P J A, et al.: Proliferation of gemistocytic cells and glial fibrillary acidic protein (GFAP)-positive oligodendroglial cells in gliomas: A MIB-1/GFAP double labeling study. Acta Neuropathol (Berl) 91: 99-103, 1996.

163. Cairncross J G, Macdonald D, Ludwin S, et al.: Chemotherapy for anaplastic oligodendroglioma. J Clin Oncol 12: 2013-21, 1994.

164. Cairncross J G, Ueki K, Zlatescu M C, et al.: Specific genetic predictors of chemotherapeutic response and survival in patients with anaplastic oligodendrogliomas. J Natl Cancer Inst 90: 1473-9, 1998.

165. Levin V A, Edwards M S, Wright D C, et al.: Modified procarbazine, CCNU and vincristine (PCV-3) combination chemotherapy in the treatment of malignant brain tumors. Cancer Treat Rep 64: 237-44, 1980.

166. Glass J, Hochberg F H, Gruber M L, et al.: The treatment of oligodendrogliomas and mixed oligodendroglioma-astrocytomas with PCV chemotherapy. J Neurosurg 76: 741-5, 1992.

167. Berger M S, Rostomily R C: Low grade gliomas: Functional mapping resection strategies, extent of resection, and outcome. J Neurooncol 34: 85-101, 1997.

168. Gonzales M, Sheline G E: Treatment of oligodendrogliomas with or without postoperative radiation. J Neurosurg 68: 684-8, 1988.

169. Reedy D P, Bay J W, Hahn J F: Role of radiation therapy in the treatment of cerebral oligodendroglioma. Neurosurgery 13: 499-503, 1983.

170. Taphoorn M J, Heimans J J, Snoek F J, et al.: Assessment of quality of life in patients treated for low-grade glioma: A preliminary report. J Neurol Neurosurg Psychiatry 55: 372-6, 1992.

171. Smith J S, Perry A, Borell T J, et al.: Alterations of chromosome arms 1p and 19q as predictors of survival in oligodendrogliomas, astrocytomas, and mixed oligoastrocytomas. J Clin Oncol 18 (3): 636-45, 2000.

172. Kraus J A, Lamszus K, Glesmann N, et al.: Molecular genetic alterations in glioblastomas with oligodendroglial component. Acta Neuropathol (Berl) 101 (4): 311-20, 2001.

173. Sgouros S, Carey M, Aluwihare N, et al.: Central neurocytoma: A correlative clinicopathologic and radiologic analysis. Surg Neurol 49 (2): 197-204, 1998.

174. Tong C Y, Ng H K, Pang J C, et al.: Central neurocytomas are genetically distinct from oligodendrogliomas and neuroblastomas. Histopathology 37 (2): 160-5, 2000.

175. Brat D J, Scheithauer B W, Eberhart C G, et al.: Extraventricular neurocytomas: Pathologic features and clinical outcome. Am J Surg Pathol 25 (10): 1252-60, 2001.

176. Jackson T R, Regine W F, Wilson D, et al.: Cerebellar liponeurocytoma. Case report and review of the literature. J Neurosurg 95 (4): 700-3, 2001.

177. George D H, Scheithauer B W: Central liponeurocytoma. Am J Surg Pathol 25 (12): 1551-5, 2001.

178. Bertalanffy A, Roessler K, Dietrich W, et al.: Gamma knife radiosurgery of recurrent central neurocytomas: A preliminary report. J Neurol Neurosurg Psychiatry 70 (4): 489-93, 2001.

179. Tyler-Kabara E, Kondziolka D, Flickinger J C, et al.: Stereotactic radiosurgery for residual neurocytoma. Report of four cases. J Neurosurg 95 (5): 879-82, 2001.

180. Brandes A A, Amista P, Gardiman M, et al.: Chemotherapy in patients with recurrent and progressive central neurocytoma. Cancer 88 (1): 169-74, 2000.

181. Mahmood A, Caccamo D V, Tomecek F J, et al.: Atypical and malignant meningiomas: A clinico-pathological review. Neurosurgery 33: 955-63, 1993.

182. Sheehy J P, Crockard H A: Multiple meningiomas: A long-term review. J Neurosurg 59: 1-5, 1983.

183. Kulali A, Ilcayto R, Rahmanli O: Primary calvarial ectopic meningiomas. Neurochirurgia (Stuttg) 34 (6): 174-7, 1991.

184. Nakasu S, Hirano A, Shimura T, et al.: Incidental meningiomas in autopsy studies. Surg Neurol 27: 319-22, 1987.

185. Wara W M, Sheline G E, Newman H, et al.: Radiation therapy of meningiomas. AJR 123: 453-8, 1975.

186. Yamashita J, Handa H, Iwaki K, et al.: Recurrence of intracranial meningiomas, with special reference to radiotherapy. Surg Neurol 14: 33-40, 1980.

187. Cushing H, Eisenhardt L: Mengiomas of the sphenoidal ridge. A. Those of the deep or clinoidal third. In Meningiomas: Their classification, regional behaviour, life history, and surgical end results. Charles C Thomas, Springfield, Illinois, 1938: pp 298-319.

188. Bonnal J, Brotchi J: Surgery of the superior sagittal sinus in parasagittal meningiomas. J Neurosurg 48 (6): J Neurosurg: 935-45, 1978.

189. Sindou M P, Alvernia J E: Results of attempted radical tumor removal and venous repair in 100 consecutive meningiomas involving the major dural sinuses. J Neurosurg 105 (4): 514-25, 2006.

190. Eskandary H, Hamzel A, Yasamy M T: Foot drop following brain lesion. Surg Neurol 43: 89-90, 1995.

191. Al-Mefty O: Tuberculum sella and olfactory groove meningiomas. In Surgery of cranial base tumors, Sekhar L N and Janecka I P, (eds.). Raven Press, New York, 1993: pp 507-19.

192. George B, Lot G, Velut S: Tumors of the foramen magnum. Neurochirurgie 39: 1-89, 1993.

193. George B, Lot G, Boissonnet H: Meningioma of the foramen magnum: A series of 40 cases. Surg Neurol 47: 371-9, 1997.

194. Vernooji M, Ikram A, Tanghe H L, et al.: Incidental findings on brain MRI in the general population. N Engl J Med 357: 1821-8, 2007.

195. Kuratsu J-I, Kochi M, Ushio Y: Incidence and clinical features of asymptomatic meningiomas. J Neurosurg 92: 766-70, 2000.

196. Zulch K J: Histologic typing or tumors of the central nervous system. International histological classification of tumors, no. 21. World Health Organization, Geneva, 1979: pp 17-57.

197. Russell D S, Rubenstein L J: Pathology of tumours of the nervous system. 5th ed. Williams and Wilkins, Baltimore, 1989.

198. Michaud J, Gagné F: Microcystic meningoma. Clinicopathologic report of eight cases. Arch Pathol Lab Med 107: 75-80, 1983.

199. Thomas H G, Dolman C L, Berry K: Malignant meningioma: Clinical and pathological features. J Neurosurg 55: 929-34, 1981.

200. Kolles H, Niedermayer I, Schmitt C, et al.: Triple approach for diagnosis and grading of meningiomas: Histology, morphometry of Ki-67/Feulgen stainings, and cytogenetics. Acta Neurochir (Wien) 137 (3-4): 174-81, 1995.

201. Perry A, Scheithauer B W, Stafford S L, et al.: “malignancy” in meningiomas: A clinicopathologic study of 116 patients with grading implications. Cancer 85: 2046-56, 1999.

202. Zimmerman R D, Fleming C A, Saint-Louis L A, et al.: Magnetic resonance of meningiomas. AJNR 6: 149-57, 1985.

203. Taylor S L, Barakos J A, Harsh G R, et al.: Magnetic resonance imaging of tuberculum sellae meningiomas: Preventing preoperative misdiagnosis as pituitary macroadenoma. Neurosurgery 31: 621-7, 1992.

204. Chun J Y, McDermott M W, Lamborn K R, et al.: Delayed surgical resection reduces intraoperative blood loss for embolized meningiomas. Neurosurgery 50 (6): 1231-5; discussion 1235-7, 2002.

205. Kai Y, Hamada J, Morioka M, et al.: Appropriate interval between embolization and surgery in patients with meningioma. AJNR Am J Neuroradiol 23 (1): 139-42, 2002.

206. Ojemann R G: Management of cranial and spinal meningiomas. Clin Neurosurg 40: 321-83, 1992.

207. Heros R C: Meningiomas involving the sinus. J Neurosurg 105 (4): 511-3, 2006 (editorial).

208. Bogaev C A, Sekhar L N: Ofactory groove and planum sphenoidale meningiomas. In Atlas of neurosurgical techniques, Sekhar L N and Fessler R G, (eds.). Thieme Medical Publishers, Inc., New York, 2006, Vol. “Brain”: pp 608-17.

209. Ojemann R G: Surgical management of olfactory groove meningiomas. In Operative neurosurgical techniques, Schmidek H H and Sweet W H, (eds.). W.B. Saunders, Philadelphia, 3rd ed., 1995, Vol. 1: pp 393-401.

210. Hakuba A, Tsujimoto T: Transcondyle approach for foramen magnum meningiomas. In Surgery of cranial base tumors, Sekhar L N and Janecka I P, (eds.). Raven Press, New York, 1993: pp 671-8.

211. David C A, Spetzler R: Foramen magnum meningiomas. Clin Neurosurg 44: 467-89, 1997.

212. Simpson D: The recurrence of intracranial meningiomas after surgical treatment. J Neurol Neurosurg Psychiatry 20: 22-39, 1957.

213. Mirimanoff R O, Dosoretz D E, Lingood R M, et al.: Meningioma: Analysis of recurrence and progression following neurosurgical resection. J Neurosurg 62: 18-24, 1985.

214. Adegbite A V, Khan M I, Paine K W E, et al.: The recurrence of intracranial meningiomas after surgical treatment. J Neurosurg 58: 51-6, 1983.

215. Jaaskelainen J: Seemingly complete removal of histologically benign intracranial meningioma: Late recurrence r. Surg Neurol 26: 461-9, 1986.

216. Barbaro N M, Gutin P H, Wilson C B, et al.: Radiation therapy in the treatment of partially resected meningiomas. Neurosurgery 20: 525-8, 1987.

217. Zuccarello M, Sawaya R, deCourten-Myers: Glioblastoma occurring after radiation therapy for meningioma: Case report and review of literature. Neurosurgery 19: 114-9, 1986.

218. Barba I, Moreno A, Martinez-Perez I, et al.: Magnetic resonance spectroscipy of brain hemangiopericytomas: High myoinositol concentrations and discrimination from meningiomas. J Neurosurg 94: 55-60, 2001.

219. Emory T S, Scheithauer B W, Hirose T, et al.: Intraneural perineurioma. A clonal neoplasm associated with abnormalities fo chromosome 22. Am J Clin Pathol 103: 696-704, 1995.

220. National Institutes of Health Consensus Development Conference: Acoustic neuroma: Consensus statement. NIH Consens Dev Conf Consens Statement. Bethesda, MD: Public Health Service, U.S. Department of Health and Human Services. Vol. 9, 1991.

221. Eldridge R, Parry D: Summary: Vestibular schwannoma (acoustic neuroma) consensus development conference. Neurosurgery 30: 962-4, 1992.

222. Harner S G, Laws E R: Clinical findings in patients with acoustic neuromas. Mayo Clin Proc 58: 721-8, 1983.

223. Stangerup S E, Caye-Thomasen P, Tos M, et al.: The natural history of vestibular schwannoma. Otol Neurotol 27 (4): Otol Neurotol: 547-52, 2006.

224. Jaffe B: Clinical studies in sudden deafness. Adv Otorhinolaryngol 20: 221-8, 1973.

225. Byl F: Seventy-six cases of presumed sudden hearing loss occurring in 1973: Prognosis and incidence. Laryngoscope 87: 817-24, 1977.

226. Berenholz L P, Eriksen C, Hirsh F A: Recovery from repeated sudden hearing loss with corticosteroid use in the presence of an acoustic neuroma. Ann Otol Rhinol Laryngol 101: 827-31, 1992.

227. Moskowitz D, Lee K J, Smith H W: Steroid use in idiopathic suden sensorineuroal hearing loss. Laryngoscope 94: 664-6, 1984.

228. Tarlov E C: Microsurgical vestibular nerve section for intractable Meniere’s disease. Clin Neurosurg 33: 667-84, 1985.

229. House W F, Brackmann D E: Facial nerve grading system. Otolaryngol Head Neck Surg 93: 184-93, 1985.

230. Hardy D G, Macfarlane R, Baguley D, et al.: Surgery for acoustic neurinoma: An analysis of 100 translabyrinthine operations. J Neurosurg 71: 799-804, 1989.

231. Daniels R L, Swallow C, Shelton C, et al.: Causes of unilateral sensorineural hearing loss screened by high-resolution fast spin echo magnetic resonance imaging: Review of 1,070 consecutive cases. Am J Otol 21 (2): Am J Otol: 173-80, 2000.

232. Stangerup S E, Caye-Thomasen P, Tos M, et al.: Change in hearing during ‘wait and scan’ management of patients with vestibular schwannoma. J Laryngol Otol 122 (7): J Laryngol Otol: 673-81, 2008.

233. Committee on Hearing and Equilibrium of the American Academy of Otolaryngology-Head and Neck Surgery Foundation: Guidelines for the evaluation of hearing preservation in acoustic neuroma (vestibular schwannoma). Otolaryngol Head Neck Surg 113 (3): 179-80, 1995.

234. Gardner G, Robertson J H: Hearing preservation in unilateral acoustic neuroma surgery. Ann Otol Rhinol Laryngol 97: 55-66, 1988.

235. Silverstein H, McDaniel A, Norrell H, et al.: Hearing preservation after acoustic neuroma surgery with intraoperative direct eighth cranial nerve monitoring: Part II. A classification of results. Otolaryngol Head Neck Surg 95, 1986.

236. Murofushi T, Matsuzaki M, Mizuno M: Vestibular evoked myogenic potentials in patients with acoustic neuromas. Arch Otolaryngol Head Neck Surg 124 (5): Arch Otolaryngol Head Neck Surg: 509-12, 1998.

237. Plotkin S R, Stemmer-Rachamimov A O, Barker F G, 2nd, et al.: Hearing improvement after bevacizumab in patients with neurofibromatosis type 2. N Engl J Med 361 (4): N Engl J Med: 358-67, 2009.

238. Bederson J B, von Ammon K, Wichmann W W, et al.: Conservative treatment of patients with acoustic tumors. Neurosurgery 28: 646-51, 1991.

239. Caye-Thomasen P, Dethloff T, Hansen S, et al.: Hearing in patients with intracanalicular vestibular schwannomas. Audiol Neurootol 12 (1): Audiol Neurootol: 1-12, 2007.

240. Asthagiri A R, Parry D M, Butman J A, et al.: Neurofibromatosis type 2. Lancet 373 (9679): Lancet: 1974-86, 2009.

241. Samii M, Matthies C: Management of 1000 vestibular schwannomas (acoustic neuromas): Surgical management with an emphasis on complications and how to avoid them. Neurosurgery 40: 11-23, 1997.

242. Brackmann D E, Fayad J N, Slattery W H, 3rd, et al.: Early proactive management of vestibular schwannomas in neurofibromatosis type 2. Neurosurgery 49 (2): Neurosurgery: 274-80; discussion 280-3, 2001.

243. Lobato-Polo J, Kondziolka D, Zorro O, et al.: Gamma knife radiosurgery in younger patients with vestibular schwannomas. Neurosurgery 65 (2): Neurosurgery: 294-300; discussion 300-1, 2009.

244. Timmer F C, Hanssens P E, van Haren A E, et al.: Gamma knife radiosurgery for vestibular schwannomas: Results of hearing preservation in relation to the cochlear radiation dose. Laryngoscope 119 (6): Laryngoscope: 1076-81, 2009.

245. Danner C, Mastrodimos B, Cueva R A: A comparison of direct eighth nerve monitoring and auditory brainstem response in hearing preservation surgery for vestibular schwannoma. Otol Neurotol 25 (5): Otol Neurotol: 826-32, 2004.

246. Yamakami I, Yoshinori H, Saeki N, et al.: Hearing preservation and intraoperative auditory brainstem response and cochlear nerve compound action potential monitoring in the removal of small acoustic neurinoma via the retrosigmoid approach. J Neurol Neurosurg Psychiatry 80 (2): J Neurol Neurosurg Psychiatry: 218-27, 2009.

247. Samii M, Gerganov V, Samii A: Improved preservation of hearing and facial nerve function in vestibular schwannoma surgery via the retrosigmoid approach in a series of 200 patients. J Neurosurg 105 (4): J Neurosurg: 527-35, 2006.

248. Patni A H, Kartush J M: Staged resection of large acoustic neuromas. Otolaryngol Head Neck Surg 132 (1): Otolaryngol Head Neck Surg: 11-9, 2005.

249. Roche P H, Ribeiro T, Khalil M, et al.: Recurrence of vestibular schwannomas after surgery. Prog Neurol Surg 21: Prog Neurol Surg: 89-92, 2008.

250. Lownie S P, Drake C G: Radical intracapsular removal of acoustic neurinomas. Long-term follow-up review of 11 cases. J Neurosurg 74: 422-5, 1991.

251. Wazen J, Silverstein H, Norrell H, et al.: Preoperative and postoperative growth rates in acoustic neuromas documented wtih CT scanning. Otolaryngol Head Neck Surg 93: 151-5, 1985.

252. Ojemann R G: Microsurgical suboccipital approach to cerebellopontine angle tumors. Clin Neurosurg 25: 461-79, 1978.

253. Sheptak P E, Jannetta P J: The two-stage excision of huge acoustic neuromas. J Neurosurg 51: 37-41, 1979.

254. Rhoton A L, Jr.: The cerebellopontine angle and posterior fossa cranial nerves by the retrosigmoid approach. Neurosurgery 47: S93-129, 2000.

255. Ebersold M J, Harner S G, Beatty C W, et al.: Current results of the retrosigmoid approach to acoustic neurinoma. J Neurosurg 76: 901-9, 1992.

256. Rhoton A L: Microsurgical anatomy of the brainstem surface facing an acoustic neuroma. Surg Neurol 25: 326-39, 1986.

257. Tatagiba M, Samii M, Matthies C, et al.: The significance for postoperative hearing of preserving the labyrinth in acoustic neurinoma surgery. J Neurosurg 77: 677-84, 1992.

258. Nutik S L, Korol H W: Cerebrospinal fluid leak after acoustic neuroma surgery. Surg Neurol 43: 553-7, 1995.

259. Symon L, Pell M F: Cerebrospinal fluid rhinorrhea following acoustic neurinoma surgery: Technical note. J Neurosurg 74: 152-3, 1991.

260. Ojemann R G: Management of acoustic neuromas (vestibular schwannomas). Clin Neurosurg 40: 498-539, 1993.

261. Malis L I: Comment on Nutik S l and Korol H W: Cerebrospinal fluid leak after acoustic neuroma surgery. Surg Neurol 43: 556-7, 1995.

262. Millen S J, Meyer G: Management of CSF otorhinorrhea following retrosigmoid removal of cerebellopontine angle tumors. Am J Otolaryngol 14: 585-9, 1993.

263. Coker N S, Jenkins H A, Fisch U: Obliteration of the middle ear and mastoid cleft in subtotal petrosectomy: Indications, technique and results. Ann Otol Rhinol Laryngol 995: 5-11, 1986.

264. Sekhar L N, Gormely W B, Wright D C: The best treatment for vestibular schwannoma (acoustic neuroma): Microsurgery or radiosurgery? Am J Otol 17: 676-89, 1996.

265. Wiegand D A, Fickel V: Acoustic neuromas. The patient’s perspective. Subjective assessment of symptoms, diagnosis, therapy, and outsome in 541 patients. Laryngoscope 99: 179-87, 1989.

266. Gormley W B, Sekhar L N, Wright D C, et al.: Acoustic neuroma: Results of current surgical management. Neurosurgery 41: 50-60, 1997.

267. Wilkins R H, Rengachary S S, (eds.): Neurosurgery. McGraw-Hill, New York, 1985.

268. Flickinger J C, Kondziolka D, Pollock B E, et al.: Evolution in technique for vestibular schwannoma radiosurgery and effect on outcome. Int J Radiation Oncology Biol Phys 36: 275-80, 1996.

269. Harner S G, Daube J R, Ebersold M J, et al.: Improved preservation of facial nerve function with use of electrical monitoring during removal of acoustic neuromas. Mayo Clin Proc 62: 92-102, 1987.

270. Moller A R, Jannetta P J: Preservation of facial function during removal of acoustic neuromas: Use of monopolar constant-voltage stimulation and EMG. J Neurosurg 61: 757-60, 1984.

271. Hirsch A, Norén G: Audiological findings after stereotactic radiosurgery in acoustic neuromas. Acta Otolaryngol (Stockh) 106: 244-51, 1988.

272. Pollock B E, Lunsford L D, Kondziolka D, et al.: Outcome analysis of acoustic neuroma management: A comparison of microsurgery and stereotactic radiosurgery. Neurosurgery 36: 215-29, 1995.

273. Ojemann R G: Comment on Pollock B E, et al.: Outcome analysis of acoustic neuroma management: A comparison of microsurgery and stereotactic radiosurgery. Neurosurgery 36: 225-6, 1995.

274. Selch M T, Pedroso A, Lee S P, et al.: Stereotactic radiotherapy for the treatment of acoustic neuromas. J Neurosurg 101: 362-72, 2004.

275. Glasscock M E, Hays J W, Minor L B, et al.: Preservation of hearing in surgery for acoustic neuromas. J Neurosurg 78: 864-70, 1993.

276. Ojemann R G, Levine R A, Montgomery W M, et al.: Use of intraoperative auditory evoked potentials to preserve hearing in unilateral acoustic neuroma removal. J Neurosurg 61: 938-48, 1984.

277. Brackmann D E, House J R I, Hitselberger W E: Technical modifications to the middle cranial fossa approach in removal of acoustic neuromas. Spring Scientific Meeting of the American Neurotolgy Society. Los Angeles, CA: 1993.

278. Pollock B E, Lunsford L D, Flickinger J C, et al.: Vestibular schwannoma management. Part I. Failed microsurgery and the role of delayed stereotactic radiosurgery. J Neurosurg 89: 944-8, 1998.

279. Shelton C, House W F: Hearing improvement after acoustic tumor removal. Otolaryngol Head Neck Surg 103: 963-5, 1990.

280. Flickinger J C, Lunsford L D, Coffey R J, et al.: Radiosurgery of acoustic neurinomas. Cancer 67: 345-53, 1991.

281. Wallner K E, Sheline G E, Pitts L H, et al.: Efficacy of irradiation for incompletely excised acoustic neurilemomas. J Neurosurg 67: 858-63, 1987.

282. Pitts L H, Jackler R K: Treatment of acoustic neuromas. N Engl J Med 339: 1471-3, 1998 (editorial).

283. Noren G, Hirsch A, Mosskin M: Long-term efficacy of gamma knife radiosurgery in vestibular schwannomas. Acta Neurochir 122: 164, 1993 (abstract).

284. Linskey M E, Lunsford L D, Flickinger J C: Neuroimaging of acoustic nerve sheath tumors after stereotactic radiosurgery. AJNR 12: 1165-75, 1991.

285. Pollock B E, Lunsford L D, Kondziolka D, et al.: Vestibular schwannoma management. Part II. Failed radiosurgery and the role of delayed micro-surgery. J Neurosurg 89: 949-55, 1998.

286. Slattery W H, Brackmann D E: Results of surgery following stereotactic irradiation for acoustic neuromas. Am J Otol 16: 315-21, 1995.

287. Wiet R J, Micco A G, Bauer G P: Complications of the gamma knife. Arch Otolaryngol Head Neck Surg 122: 414-6, 1996.

288. Yakulis R, Manack L, Murphy A I: Postradiation malignant triton tumor: A case report and review of the literature. Arch Pathol Lab Med 120: 541-8, 1996.

289. Comey C H, McLaughlin M R, Jho H D, et al.: Death from a malignant cerebellopontine angle triton tumor despite stereotactic radiosurgery. J Neurosurg 89: 653-8, 1998.

290. Lustig L R, Jackler R K, Lanser M J: Radiation-induced tumors of the temporal bone. Am J Otol 18: 230-5, 1997.

291. Beatty C W, Ebersold M J, Harner S G: Residual and recurrent acoustic neuromas. Laryngoscope 97: 1168-71, 1987.

292. Ragel B T, Couldwell W T: Pituitary carcinoma: A review of the literature. Neurosurgical Focus 16 (4): E7, 2004.

293. Ebersold M J, Quast L M, Laws E R, et al.: Long-term results in transsphenoidal removal of nonfunctioning pituitary adenomas. J Neurosurg 64: 713-9, 1986.

294. Biller B M, Swearingen B, Zervas N T: A decade of the Massachusetts General Hospital neuroendocrine clinical center. J Clin Endocrinol Metab 82: J Clin Endocrinol Metab: 1668-74, 1997.

295. Abe T, Matsumoto K, Sanno N, et al.: Lymphocytic hypophysitis: Case report. Neurosurgery 36: 1016-9, 1995.

296. Lieblich J M, Rogol A D, White B J, et al.: Syndrome of anosmia with hypogonadotropic hypogonadism (Kallmann syndrome): Clinical and laboratory studies in 23 cases. Am J Med 73 (4): Am J Med: 506-19, 1982.

297. Nutkiewicz A, DeFeo D R, Kohout R I, et al.: Cerebrospinal fluid rhinorrhea as a presentation of pituitary adenoma. Neurosurgery 6: 195-7, 1980.

298. Reid R L, Quigley M E, Yen S C: Pituitary apoplexy: A review. Arch Neurol 42: 712-9, 1985.

299. Cardoso E R, Peterson E W: Pituitary apoplexy: A review. Neurosurgery 14: 363-73, 1984.

300. Onesti S T, Wisniewski T, Post K D: Pituitary hemorrhage into a Rathke’s cleft cyst. Neurosurgery 27: 644-6, 1990.

301. Liu J K, Couldwell W: Pituitary apoplexy: Diagnosis and management. Contemp Neurosurg 25 (12): 1-5, 2003.

302. Yen M Y, Liu J H, Jaw S J: Ptosis as the early manifestation of pituitary tumour. Br J Ophthalmol 74 (3): Br J Ophthalmol: 188-91, 1990.

303. Telesca M, Santini F, Mazzucco A: Adenoma related pituitary apoplexy disclosed by ptosis after routine cardiac surgery: Occasional reappearance of a dismal complication. Intensive Care Med 35 (1): Intensive Care Med: 185-6, 2009.

304. Wakai S, Fukushima T, Teramoto A, et al.: Pituitary apoplexy: Its incidence and clinical significance. J Neurosurg 55: 187-93, 1981.

305. Rovit R L, Fein J M: Pituitary apoplexy, a review and reappraisal. J Neurosurg 37: 280-8, 1972.

306. Bills D C, Meyer F B, Laws E R, Jr., et al.: A retrospective analysis of pituitary apoplexy. Neurosurgery 33 (4): 602-8; discussion 608-9, 1993.

307. Wilson C B: Neurosurgical management of large and invasive pituitary tumors. In Clinical management of pituitary disorders, Tindall G T and Collins W F, (eds.). Raven Press, New York, 1979: pp 335-42.

308. Pei L, Melmed S, Scheithauer B, et al.: Frequent loss of heterozygosity at the retinoblastoma susceptibility gene (rb) locus in aggressive pituitary tumors: Evidence for a chromosome 13 tumor suppressor gene other than rb. Cancer Res 55: 1613-6, 1995.

309. Hardy J: Transsphenoidal surgery of hypersecreting pituitary tumors. In Diagnosis and treatment of pituitary tumors, Kohler P O and Ross G T, (eds.). Excerpta Medica/American Elsevier, New York, 1973: pp 179-94.

310. Hardy J: Transsphenoidal surgery of intracranial neoplasm. In Adv neurol, Thompson R A and Green R, (eds.). Raven Press, New York, 1976, Vol. 15: Neoplasia in the Central Nervous System: pp 261-74.

311. Krisht A F: Giant invasive pituitary adenomas. Contemp Neurosurg 21 (1): 1-6, 1999.

312. Laws E R: Comment on Knosp E, et al.: Pituitary adenomas with invasion of the cavernous sinus space: A magnetic resonance imaging classification compared with surgical findings. Neurosurgery 33: 617, 1993.

313. Scotti G, Yu C Y, Dillon W P, et al.: MR imaging of cavernous sinus involvement by pituitary adenomas. AJR 151: 799-806, 1988.

314. Barlas O, Bayindir C, Hepgul K, et al.: Bromocriptine-induced cerebrospinal fluid fistula in patients with macroprolactinomas: Report of three cases and a review of the literature. Surg Neurol 41 (6): 486-9, 1994.

315. Amar A P, Couldwell W T, Weiss M H: Prolactinomas: Focus on indications, outcomes, and management of recurrences. Contemp Neurosurg 21 (6): 1-6, 1989.

316. Banasiak M J, Malek A R: Nelson syndrome: Comprehensive review of pathophysiology, diagnosis, and management. Neurosurgical Focus 23 (3): E13, 2007.

317. Esposito F, Dusick J R, Cohan P, et al.: Early morning cortisol levels as a predictor of remission after transsphenoidal surgery for Cushing’s disease. J Clin Endocrinol Metab 91: 7-13, 2006.

318. Assie G, Bahurel H, Coste J, et al.: Corticotroph tumor progression after adrenalectomy in Cushing’s disease: A reappraisal of Nelson’s syndrome. J Clin Endocrinol Metab 49: 381-6, 2007.

319. Bertagna X, Raux-Demay M-C, Guilhaume B, et al.: Cushing’s disease. In The pituitary, Melmed S, (ed.). Blackwell Scientific, Malden, MA, 2nd ed., 2002: pp 496-560.

320. Kasperlik-Zaluska A A, Bonicki W, Jeske W, et al.: Nelson’s syndrome - 46 years later: Clinical experience with 37 patients. Zentralbl Neurochir 67: 14-20, 2006.

321. Cook D M: Aace medical guidelines for clinical practice for the diagnosis and treatment of acromegaly. Endocr Pract 10 (3): 213-25, 2004.

322. Melmed S: Acromegaly. N Engl J Med 322: 966-77, 1990.

323. Melmed S: Medical progress: Acromegaly. N Engl J Med 355 (24): 2558-73, 2006.

324. Renehan A, Shalet S: Acromegaly and colorectal cancer: Risk assessment should be based on population-based studies. J Clin Endocrinol Metab 87: 1909-, 2002.

325. Acromegaly Therapy Consensus Development Panel: Consensus statement: Benefits versus risks of medical therapy for acromegaly. Am J Med 97: 468-73, 1994.

326. Beck-Peccoz P, Brucker-Davis F, Persani L, et al.: Thyrotropin-secreting pituitary tumors. Endocr Rev 17 (6): 610-38, 1996.

327. Clarke M J, Erickson D, Castro M R, et al.: Thyroid-stimulating hormone pituitary adenomas. J Neurosurg 109 (1): 17-22, 2008.

328. Beck-Peccoz P, Persani L: Medical management of thyrotropin-secreting pituitary adenomas. Pituitary 5 (2): 83-8, 2002.

329. Wilson C B: Endocrine-inactive pituitary adenomas. Clin Neurosurg 38: 10-31, 1992.

330. Cohen-Gadol A A, Pichelmann M A, Link M J, et al.: Granular cell tumor of the sellar and suprasellar region: Clinicopathologic study of 11 cases and literature review. Mayo Clin Proc 78 (5): 567-73, 2003.

331. Schaller B, Kirsch E, Tolnay M, et al.: Symptomatic granular cell tumor of the pituitary gland: Case report and review of the literature. Neurosurgery 42 (1): 166-70; discussion 170-1, 1998.

332. Markesbery W R, Duffy P E, Cowen D: Granular cell tumors of the central nervous system. J Neuropathol Exp Neurol 32 (1): 92-109, 1973.

333. Fuller G N, Wesseling P: Granular cell tumors of the neurohypophysis. In WHO classification of tumors of the central nervous system, Louis D N, Ohgaki H, Wiestler O D, et al., (eds.). World Health Organization classification of tumors. Bosman F T, Jaffe E S, Lakhani S R, et al. International Agency for Research on Cancer, Lyon, 4th ed., 2007: pp 241-2.

334. Gueguen B, Merland J J, Riche M C, et al.: Vascular malformations of the spinal cord: Intrathecal perimedullary arteriovanous fistulas fed by medullary arteries. Neurology 37: 969-79, 1987.

335. Wesseling P, Brat D J, Fuller G N: Pituicytoma. In WHO classification of tumors of the central nervous system, Louis D N, Ohgaki H, Wiestler O D, et al., (eds.). World Health Organization classification of tumors. Bosman F T, Jaffe E S, Lakhani S R, et al. International Agency for Research on Cancer, Lyon, 4th ed., 2007: pp 243-4.

336. Chandler W F: Treatment of disorders of the pituitary gland: Pearls and pitfalls from 30 years of experience. Clin Neurosurg 56: Clin Neurosurg: 18-22, 2008.

337. Walsh F B, Hoyt W F, (eds.): Clinical neuro-ophthalmology. 3rd ed., Williams and Wilkins, Baltimore, 1969.

338. Tindall G T, Barrow D L: Current management of pituitary tumors: Part I. Contemp Neurosurg 10: 1-6, 1988.

339. Watts N B: Cushing’s syndrome: An update. Contemp Neurosurg 17 (18): 1-7, 1995.

340. Bilaniuk L T, Moshang T, Cara J, et al.: Pituitary enlargement mimicking pituitary tumor. J Neurosurg 63: 39-42, 1985.

341. Atchison J A, Lee P A, Albright L: Reversible suprasellar pituitary mass secondary to hypothyroidism. JAMA 262: 3175-7, 1989.

342. Watanakunakorn C, Hodges R E, Evans T C: Myxedema. A study of 400 cases. Arch Intern Med 116: 183-90, 1965.

343. Gillam M P, Molitch M E, Lombardi G, et al.: Advances in the treatment of prolactinomas. Endocr Rev 27 (5): Endocr Rev: 485-534, 2006.

344. Kallenberg G A, Pesce C M, Norman B, et al.: Ectopic hyperprolactinemia resulting from an ovarian teratoma. JAMA 263: 2472-4, 1990.

345. Randall R V, Scheithauer B W, Laws E R, et al.: Pituitary adenomas associated with hyperprolactinemia. Mayo Clin Proc 60: 753-62, 1985.

346. Cowen P J, Sargent P A: Changes in plasma prolactin during ssri treatment: Evidence for a delayed increase in 5-HT neurotransmission. J Psychopharmacol 11 (4): 345-8, 1997.

347. Olukoga A O: Macroprolactinemia is clinically important. J Clin Endocrinol Metab 87 (10): 4833-4, 2002 (letter to the editor).

348. Tyrell J B, Aron D C, Forsham P H: Glucocorticoids and adrenal androgens. In Basic and clinical endocrinology, Greenspan F S, (ed.). Appleton and Lange, Norwalk, 3rd ed., 1991: pp 323-62.

349. Yanovski J A, Cutler G B, Chrousos G P, et al.: Corticotropin-releasing sormone stimulation following low-dose dexamethasone administration: A new test to distinguish Cushing’s syndrome from pseudo-Cushing’s states. JAMA 269: 2232-8, 1993.

350. McCutcheon I E, Oldfield E H: Cortisol: Regulation, disorders, and clinical evaluation. In Neuroendocrinology, Barrow D L and Selman W, (eds.). Concepts in neurosurgery. Williams and Wilkins, Baltimore, 1992, Vol. 5: pp 117-73.

351. Chrousos G P, Schulte H M, Oldfield E H, et al.: The corticotropin-releasing factor stimulation test: An aid in the evaluation of patients with Cushing’s syndrome. N Engl J Med 310: 622-6, 1984.

352. Watts N B, Tindall G T: Rapid assessment of corticotropin reserve after pituitary surgery. JAMA 259: 708-11, 1988.

353. Abboud C F: Laboratory diagnosis of hypopituitarism. Mayo Clin Proc 61: 35-48, 1986.

354. Dimaraki E V, Jaffe C A, DeMott-Friberg R, et al.: Acromegaly with apparently normal GH secretion: Implications for diagnosis and follow-up. J Clin Endocrinol Metab 87 (8): 3537-42, 2002.

355. Frohman L: Ectopic hormone production by tumors: Growth hormone-releasing factor. Neuroendocrine Perspect 3: 201-24, 1984.

356. Peyster R G, Hoover E D, Viscarello R R, et al.: CT appearance of the adolescent and preadolescent pituitary gland. AJNR 4: 411-4, 1983.

357. Watson J C, Shawker T H, Nieman L K, et al.: Localization of pituitary adenomas by using intraoperative ultrasound in patients with Cushing’s disease and no demonstrable pituitary tumor on magnetic resonance imaging. J Neurosurg 89: 927-32, 1998.

358. Kim L J, Lekovic G P, White W L, et al.: Preliminary experience with 3-tesla MRI and Cushing’s disease. Skull Base 17 (4): Skull Base: 273-7, 2007.

359. Kucharczyk W, Davis D O, Kelly W M, et al.: Pituitary adenomas: High-resolution MR imaging at 1.5 T. Radiology 161: Radiology: 761-5, 1986.

360. Barrow D L, Mizuno J, Tindall G T: Management of prolactinomas associated with very high serum prolactin levels. J Neurosurg 68: 554-8, 1988.

361. Blevins L S: Medical management of pituitary adenomas. Contemp Neurosurg 19 (11): 1-6, 1997.

362. Molitch M E: Pregnancy and the hyperprolactinemic woman. N Engl J Med 312 (21): N Engl J Med: 1364-70, 1985.

363. Landolt A M, Osterwalder V: Perivascular fibrosis in prolactinomas: Is it increased by bromocriptine? J Clin Endocrinol Metab 58: 1179-83, 1984.

364. Cabergoline for hyperprolactinemia. Med Letter 39: 58-9, 1997.

365. Webster J, Piscitelli G, Polli A, et al.: A comparison of cabergoline and bromocriptine in the treatment of hyperprolactinemic amenorrhea. N Engl J Med 331: 904-9, 1994.

366. Schade R, Andersohn F, Suissa S, et al.: Dopamine agonists and the risk of cardiac-valve regurgitation. N Engl J Med 356 (1): 29-38, 2007.

367. Johnston D G, Hall K, Kendall-Taylor P, et al.: Effect of dopamine agonist withdrawal after long-term therapy in prolactinomas. Studies with high-definition computerised tomography. Lancet 2 (8396): Lancet: 187-92, 1984.

368. Schlechte J A: Long-term management of prolactinomas. J Clin Endocrinol Metab 92 (8): J Clin Endocrinol Metab: 2861-5, 2007.

369. Colao A, Attanasio R, Pivonello R, et al.: Partial surgical removal of growth hormone-secreting pituitary tumors enhances the response to somatostatin analogs in acromegaly. J Clin Endocrinol Metab91: 85-92, 2006.

370. Turner H E, Thornton-Jones V A, Wass J A: Systematic dose-extension of octreotide LAR: The importance of individual tailoring of treatment in patients with acromegaly. Clin Endocrinol (Oxf) 61 (2): 224-31, 2004.

371. van der Lely A J, Hutson R K, Trainer P J, et al.: Long-term treatment of acromegaly with pegvisomant, a growth hormone receptor antagonist. Lancet 358 (9295): 1754-9, 2001.

372. Pegvisomant (Somavert) for acromegaly. Med Letter 45: 55-6, 2003.

373. Sheehan J M, Vance M L, Sheehan J P, et al.: Radiosurgery for Cushing’s disease after failed transsphenoidal surgery. J Neurosurg 93 (5): 738-42, 2000.

374. Sanno N, Teramoto A, Osamura R Y: Long-term surgical outcome in 16 patients with thyrotropin pituitary adenoma. J Neurosurg 93 (2): 194-200, 2000.

375. Ciric I, Mikhael M, Stafford T, et al.: Transsphenoidal microsurgery of pituitary macroadenomas with long-term follow-up results. J Neurosurg 59: 395-401, 1983.

376. Breen P, Flickinger J C, Kondziolka D, et al.: Radiotherapy for nonfunctional pituitary adenoma: Analysis of long-term tumor control. J Neurosurg 89: 933-8, 1998.

377. Schmidek H H, Sweet W H, (eds.): Operative neurosurgical techniques. 1st ed., Grune and Stratton, New York, 1982.

378. Powell M, Lightman S L, (eds.): Management of pituitary tumours: A handbook. Churchill Livingstone, New York, 1996.

379. Hardy J: Transsphenoidal hypophysectomy. J Neurosurg 34: 582-94, 1971.

380. Kern E B, Pearson B W, McDonald T J, et al.: The transseptal approach to lesions of the pituitary and parasellar region. Laryngoscope 89S: 1-34, 1979.

381. Spaziante R, de Divitiis E, Cappabianca P: Reconstruction of the pituitary fossa in transsphenoidal surgery: An experience of 140 cases. Neurosurgery 17: 453-8, 1985.

382. Zhang X, Fei Z, Zhang J, et al.: Management of nonfunctioning pituitary adenomas with suprasellar extension by transspehnoidal microsurgery. Surg Neurol 52: 380-5, 1999.

383. Fatemi N, Dusick J R, Mattozo C, et al.: Pituitary hormonal loss and recovery after transsphenoidal adenoma removal. Neurosurgery 63 (4): 709-18; discussion 718-9, 2008.

384. Decker R E, Chalif D J: Progressive coma after the transsphenoidal decompression of a pituitary adenoma with marked suprasellar extension: Report of two cases. Neurosurgery 28: 154-8, 1991.

385. Domingue J N, Wilson C B: Pituitary abscesses. J Neurosurg 46: 601-8, 1977.

386. Robinson B: Intrasellar abscess after transsphenoidal pituitary adenectomy. Neurosurgery 12: 684-6, 1983.

387. Verbalis J G, Robinson A G, Moses A M: Postoperative and post-traumatic diabetes insipidus. Front Horm Res 13: 247-65, 1985.

388. Teng M M H, Huang C I, Chang T: The pituitary mass after transsphenoidal hypophysectomy. AJNR 9: 23-6, 1988.

389. Cohen A R, Cooper P R, Kupersmith M J, et al.: Visual recovery after transsphenoidal removal of pituitary adenoma. Neurosurgery 17: 446-52, 1985.

390. Ayuk J, Clayton R, Holder G, et al.: Growth hormone and pituitary radiotherapy, but not serum insulin-like growth factor-I concentrations, predict excess mortality in patients with acromegaly. J Clin Endocrinol Metab 89: 1613-7, 2004.

391. Davis D H, Laws E R, Ilstrup D M, et al.: Results of surgical treatment for growth hormone-secreting pituitary adenomas. J Neurosurg 79: 70-5, 1993.

392. Chen J C, Amar A P, Choi S, et al.: Transsphenoidal microsurgical treatment of Cushing’s disease: Postoperative assessment of surgical efficacy by application of an overnight low-dose dexamethasone suppression test. J Neurosurg 98: 967-73, 2003.

393. Nishizawa S, Oki Y, Ohta S, et al.: What can predict postoperative “endocrinological cure” in Cushing’s disease? Neurosurgery 45: 239-44, 1999.

394. Trainer P J, Lawrie H S, Verhelst J, et al.: Transsphenoidal resection in Cushing’s disease: Undetectable serum cortisol as the definition of successful treatment. Clinical Endocrinology (Oxf) 38: 73-8, 1993.

395. Rees D A, Hanna F W, Davies J S, et al.: Long-term follow-up results of transsphenoidal surgery for Cushing’s disease in a single centre using strict criteria for remission. Clinical Endocrinology (Oxf)56: 541-51, 2002.

396. Yap L B, Turner H E, Adams C B, et al.: Undetectable postoperative cortisol does not always predict long-term remission in Cushing’s disease: A single centre audit. Clinical Endocrinology (Oxf) 56: 25-31, 2002.

397. Simmons N E, Alden T D, Thorner M O, et al.: Serum cortisol response to transsphenoidal surgery for cushing disease. J Neurosurg 95: 1-8, 2001.

398. Rollin G A, Ferreira N P, Junges M, et al.: Dynamics of serum cortisol levels after transsphenoidal surgery in a cohort of patients with Cushing’s disease. J Clin Endocrinol Metab 89: 1131-9, 2004.

399. Klein H J, Rath S A: Removal of tumors of the III ventricle using lamina terminalis approach: Three cases of isolated growth of craniopharyngiomas in the III ventricle. Childs Nerv Syst 5: 144-7, 1989.

400. Patterson R H, Denylevich A: Surgical removal of craniopharyngiomas by a transcranial approach through the lamina terminalis and sphenoid sinus. Neurosurgery 7: 111-7, 1980.

401. Manaka S, Teramoto A, Takakura K: The efficacy of radiotherapy for craniopharyngioma. J Neurosurg 62: 648-56, 1985.

402. Maggio W W, Cail W S, Brookeman J R, et al.: Rathke’s cleft cyst: Computed tomographic and magnetic resonance imaging appearances. Neurosurgery 21: 60-2, 1987.

403. Nishio S, Mizuno J, Barrow D L, et al.: Pituitary tumors composed of adenohypophysial adenoma and Rathke’s cleft cyst elements: A clinicopathological study. Neurosurgery 21: 371-7, 1987.

404. Voelker J L, Campbell R L, Muller J: Clinical, radiographic, and pathological features of symptomatic Rathke’s cleft cysts. J Neurosurg 74: 535-44, 1991.

405. Little J R, MacCarty C S: Colloid cysts of the third ventricle. J Neurosurg 39: 230-5, 1974.

406. Ciric I, Zivin I: Neuroepithelial (colloid) cysts of the septum pellucidum. J Neurosurg 43: 69-73, 1975.

407. Guner M, Shaw M D M, Turner J W, et al.: Computed tomography in the diagnosis of colloid cyst. Surg Neurol 6: 345-8, 1976.

408. Ryder J W, Kleinschmidt B K, Keller T S: Sudden deterioration and death in patients with benign tumors of the third ventricle area. J Neurosurg 64: 216-23, 1986.

409. Mamourian A C, Cromwell L D, Harbaugh R E: Colloid cyst of the third ventricle: Sometimes more conspicuous on CT than MR. AJNR 19: 875-8, 1998.

410. El Khoury C, Brugieres P, Decq P, et al.: Colloid cysts of the third ventricle: Are MR imaging patterns predictive of difficulty with percutaneous treatment? AJNR Am J Neuroradiol 21 (3): 489-92, 2000.

411. Torkildsen A: Should extirpation be attempted in cases of neoplasm in or near the third ventricle of the brain? Experiences with a palliative method. J Neurosurg 5: 249-75, 1948.

412. Bosch D A, Rahn T, Backlund E O: Treatment of colloid cyst of the third ventricle by stereotactic aspiration. Surg Neurol 9: 15-8, 1978.

413. Rivas J J, Lobato R D: CT-assisted stereotaxic aspiration of colloid cysts of the third ventricle. J Neurosurg 62: 238-42, 1985.

414. Mathiesen T, Grane P, Lindquist C, et al.: High recurrence rate following aspiration of colloid cysts in the third ventricle. J Neurosurg 78: 748-52, 1993.

415. Musolino A, Fosse S, Munari C, et al.: Diagnosis and treatment of colloid cysts of the third ventricle by stereotactic drainage. Report on eleven cases. Surg Neurol 32: 294-9, 1989.

416. Apuzzo M L J, Chandrasoma P T, Zelman V, et al.: Computed tomographic guidance stereotaxis in the management of lesions of the third ventricular region. Neurosurgery 15: 502-8, 1984.

417. Apuzzo M L J: Comment on Garrido E, et al.: Cerebral venous and sagittal sinus thrombosis after transcallosal removal of a colloid cyst of the third ventricle: Case report. Neurosurgery 26: 542, 1990.

418. Kondziolka D, Lunsford L D: Stereotactic management of colloid cysts: Factors predicting success. J Neurosurg 75: 45-51, 1991.

419. Hall W A, Lunsford L D: Changing concepts in the treatment of colloid cysts. An 11-year experience in the CT era. J Neurosurg 66: 186-91, 1987.

420. Glenn G M, Linehan W M, Hosoe S, et al.: Screening for von Hippel-Lindau disease by DNA polymorphism analysis. JAMA 267: 1226-31, 1992.

421. Catapano D, Muscarella L A, Guarnieri V, et al.: Hemangioblastomas of central nervous system: Molecular genetic analysis and clinical management. Neurosurgery 56 (6): 1215-21; discussion 1221, 2005.

422. Wanebo J E, Lonser R R, Glenn G M, et al.: The natural history of hemangioblastomas of the central nervous system in patients with von Hippel-Lindau disease. J Neurosurg 98 (1): J Neurosurg: 82-94, 2003.

423. Butman J A, Linehan W M, Lonser R R: Neurologic manifestations of von Hippel-Lindau disease. JAMA 300 (11): JAMA: 1334-42, 2008.

424. Go R C P, Lamiell J M, Hsia Y E, et al.: Segregation and linkage analysis of von Hippel-Lindau disease among 220 descendents from one kindred. Am J Human Genet 36: 131-42, 1984.

425. Friedrich C A: Genotype-phenotype correlation in von Hippel-Lindau syndrome. Hum Mol Genet 10 (7): 763-7, 2001.

426. Melmon K L, Rosen S W: Lindau’s disease. Review of the literature and study of a large kindred. Am J Med 36: 595-617, 1964.

427. Jagannathan J, Lonser R R, Smith R, et al.: Surgical management of cerebellar hemangioblastomas in patients with von Hippel-Lindau disease. J Neurosurg 108 (2): J Neurosurg: 210-22, 2008.

428. Manski T J, Heffner D K, Glenn G M, et al.: Endolymphatic sac tumors. A source of morbid hearing loss in von Hippel-Lindau disease. JAMA 277 (18): JAMA: 1461-6, 1997.

429. Chew E Y: Ocular manifestations of von Hippel-Lindau disease: Clinical and genetic investigations. Trans Am Ophthalmol Soc 103: Trans Am Ophthalmol Soc: 495-511, 2005.

430. Meyerle C B, Dahr S S, Wetjen N M, et al.: Clinical course of retrobulbar hemangioblastomas in von Hippel-Lindau disease. Ophthalmology 115 (8): Ophthalmology: 1382-9, 2008.

431. Niemela M, Lemeta S, Summanen P, et al.: Long-term prognosis of haemangioblastoma of the CNS: Impact of von Hippel-Lindau disease. Acta Neurochir (Wien) 141 (11): Acta Neurochir (Wien): 1147-56, 1999.

432. Choyke P L, Glenn G M, Walther M M, et al.: Hereditary renal cancers. Radiology 226 (1): Radiology: 33-46, 2003.

433. Meister M, Choyke P, Anderson C, et al.: Radiological evaluation, management, and surveillance of renal masses in von Hippel-Lindau disease. Clin Radiol 64 (6): Clin Radiol: 589-600, 2009.

434. Maher E R, Kaelin W G, Jr.: Von Hippel-Lindau disease. Medicine (Baltimore) 76 (6): Medicine (Baltimore): 381-91, 1997.

435. Hes F J, Feldberg M A: Von Hippel-Lindau disease: Strategies in early detection (renal-, adrenal-, pancreatic masses). Eur Radiol 9 (4): Eur Radiol: 598-610, 1999.

436. Bisceglia M, Galliani C A, Senger C, et al.: Renal cystic diseases: A review. Adv Anat Pathol 13 (1): Adv Anat Pathol: 26-56, 2006.

437. Truong L D, Choi Y J, Shen S S, et al.: Renal cystic neoplasms and renal neoplasms associated with cystic renal diseases: Pathogenetic and molecular links. Adv Anat Pathol 10 (3): Adv Anat Pathol: 135-59, 2003.

438. Bradley S, Dumas N, Ludman M, et al.: Hereditary renal cell carcinoma associated with von Hippel-Lindau disease: A description of a nova scotia cohort. Can Urol Assoc J 3 (1): Can Urol Assoc J: 32-6, 2009.

439. Moss J M, Choi C Y, Adler J R, Jr., et al.: Stereotactic radiosurgical treatment of cranial and spinal hemangioblastomas. Neurosurgery 65 (1): Neurosurgery: 79-85; discussion 85, 2009.

440. Chang S D, Meisel J A, Hancock S L, et al.: Treatment of hemangioblastomas in von Hippel-Lindau disease with linear accelerator-based radiosurgery. Neurosurgery 43 (1): Neurosurgery: 28-34; discussion 34-5, 1998.

441. VHL Family Alliance: VHL handbook. Section 5: Suggested screening guidelines. (December 10, 2009), 2009.

442. Constans J P, Meder F, Maiuri F, et al.: Posterior fossa hemangioblastomas. Surg Neurol 25 (3): 269-75, 1986.

443. Hes F J, van der Luijt R B: [von Hippel-Lindau disease: Protocols for diagnosis and periodical clinical monitoring. National von Hippel-Lindau disease working group]. Ned Tijdschr Geneeskd 144 (11): 505-9, 2000 (Dutch).

444. Poulsen M L, Budtz-Jorgensen E, Bisgaard M L: Surveillance in von Hippel-Lindau disease (vHL). Clin Genet: Clin Genet, 2009.

445. Ho V B, Smirniotopoulos J G, Murphy F M, et al.: Radiologic-pathologic correlation: Hemangioblastoma. AJNR 13: 1343-52, 1992.

446. Wakai S, Inoh S, Ueda Y, et al.: Hemangioblastoma presenting with intraparenchymatous hemorrhage. J Neurosurg 61: 956-60, 1984.

447. Silver M L, Hennigar G: Cerebellar hemangioma (hemangioblastoma). A clinicopathological review of 40 cases. J Neurosurg 9: 484-94, 1952.

448. Utsuki S, Oka H, Sato K, et al.: Fluorescence diagnosis of tumor cells in hemangioblastoma cysts with 5-aminolevulinic acid. J Neurosurg: J Neurosurg, 2009.

449. O’Neill B P, Illig J J: Primary central nervous system lymphoma. Mayo Clin Proc 64: 1005-20, 1989.

450. Kawakami Y, Tabuchi K, Ohnishi R, et al.: Primary central nervous system lymphoma. J Neurosurg 62: 522-7, 1985.

451. Jellinger K, Radaszkiewicz T: Involvement of the central nervous system in malignant lymphomas. Virchows Arch (Pathol Anat) 370: 345-62, 1976.

452. Helle T L, Britt R H, Colby T V: Primary lymphoma of the central nervous system. J Neurosurg 60: 94-103, 1984.

453. Alic L, Haid M: Primary lymphoma of the brain: A case report and review of the literature. J Surg Oncol 26: 115-21, 1984.

454. Eby N L, Grufferman S, Flannelly C M, et al.: Increasing incidence of primary brain lymphoma in the U.S. Cancer 62: 2461-5, 1988.

455. Murray K, Kun L, Cox J: Primary malignant lymphoma of the central nervous system: Results of treatment of 11 cases and review of the literature. J Neurosurg 65: 600-7, 1986.

456. Penn I: Development of cancer as a complication of clinical transplantation. Transplant Proc 9: 1121-7, 1977.

457. Levy R M, Bredesen D E, Rosenblum M L: Neurological manifestations of the acquired immunodeficiency syndrome (AIDS): Experience at ucsf and review of the literature. J Neurosurg 62: 475-95, 1985.

458. Jean W C, Hall W A: Management of cranial and spinal infections. Contemp Neurosurg 20 (9): 1-10, 1998.

459. Hochberg F H, Miller G, Schooley R T, et al.: Central-nervous-system lymphoma related to Epstein-Barr virus. N Engl J Med 309: 745-8, 1983.

460. MacMahon E M E, Glass J D, Hayward S D, et al.: Epstein-Barr virus in AIDS-related primary central nervous system lymphoma. Lancet 338: 969-73, 1991.

461. Burger P C, Scheithauer B W, Vogel F S: Surgical pathology of the nervous system and its coverings. 4th ed. Churchill Livingstone, New York, 2002.

462. Calamia K T, Miller A, Shuster E A, et al.: Intravascular lymphomatosis: A report of ten patients with central nervous system involvement and a review of the disease process. Adv Exp Med Biol 455: 249-65, 1999.

463. Glass J, Hochberg F H, Miller D C: Intravascular lymphomatosis. A systemic disease with neurologic manifestations. Cancer 71 (10): 3156-64, 1993.

464. So Y T, Beckstead J H, Davis R L: Primary central nervous system lymphoma in acquired immune deficiency syndrome: A clinical and pathological study. Ann Neurol 20: 566-72, 1986.

465. Poon T, Matoso I, Tchertkoff V, et al.: CT features of primary cerebral lymphoma in AIDS and non-AIDS patients. J Comput Assist Tomogr 13: 6-9, 1989.

466. DeAngelis L M: Cerebral lymphoma presenting as a nonenhancing lesion of computed tomographic/magnetic resonance scan. Ann Neurol 33: 308-11, 1993.

467. Enzmann D R, Krikorian J, Norman D, et al.: Computed tomography in primary reticulum cell sarcoma of the brain. Radiology 130: 165-70, 1979.

468. Vaquero J, Martinez R, Rossi E, et al.: Primary cerebral lymphoma: The ‘ghost tumor’. J Neurosurg 60: 174-6, 1984.

469. Gray R s, Abrahams J J, Hufnagel T J, et al.: Ghost-cell tumor of the optic chiasm; primary CNS lymphoma. J Clin Neuroophthalmol 9: 98-104, 1989.

470. DeAngelis L M: Cerebral lymphoma presenting as a nonenhancing lesion on computed tomographic/magnetic resonance scan. Ann Neurol 33 (3): 308-11, 1993.

471. O’Neill B P, Kelly P J, Earle J D, et al.: Computer-assisted stereotactic biopsy for the diagnosis of primary central nervous system lymphoma. Neurology 37: 1160-4, 1987.

472. DeAngelis L M, Yahalom J, Heinemann M-H, et al.: Primary central nervous system lymphomas: Combined treatment with chemotherapy and radiotherapy. Neurology 40: 80-6, 1990.

473. DeAngelis L M, Yahalom J, Thaler H T, et al.: Combined modality therapy for primary CNS lymphomas. J Clin Oncol 10: 635-43, 1992.

474. O’Marcaigh A S, Johnson C M, Smithson W A, et al.: Successful treatment of intrathecal methotrexate overdose by using ventriculolumbar perfusion and intrathecal instillation of carboxypeptidase g₂. Mayo Clin Proc 71: 161-5, 1996.

475. Hochberg F H, Miller D C: Primary central nervous system lymphoma. J Neurosurg 68: 835-53, 1988.

476. Baumgartner J E, Rachlin J R, Beckstead J H, et al.: Primary central nervous system lymphomas: Natural history and response to radiation therapy in 55 patients with acquied immunodeficiency syndrome. J Neurosurg 73: 206-11, 1990.

477. Formenti S C, Gill P S, Lean E, et al.: Primary central nervous system lymphoma in AIDS: Results of radiation therapy. Cancer 63: 1101-7, 1989.

478. O’Neill P, Bell B A, Miller J D, et al.: Fifty years of experience with chordomas in southeast scotland. Neurosurgery 16: 166-70, 1985.

479. Heffelfinger M J, Dahlin D C, MacCarty C S, et al.: Chordomas and cartilaginous tumors at the skull base. Cancer 32: 410-20, 1973.

480. Boriani S, Chevalley F, Weinstein J N, et al.: Chordoma of the spine above the sacrum. Treatment and outcome in 21 cases. Spine 21: 1569-77, 1996.

481. Hug E B, Loredo L N, Slater J D, et al.: Proton radiation therapy for chordomas and chondrosarcomas of the skull base. J Neurosurg 91: 432-9, 1999.

482. Meyer J E, Lepke R A, Lindfors K K, et al.: Chordomas: Their CT appearance in the cervical, thoracic and lumbar spine. Radiology 153: 693-6, 1984.

483. Schwarz S S, Fisher W S, Pulliam M W, et al.: Thoracic chordoma in a patient with paraparesis and ivory vertebral body. Neurosurgery 16: 100-2, 1985.

484. Wold L E, Laws E R: Cranial chordomas in children and young adults. J Neurosurg 59: 1043-7, 1983.

485. Azzarelli A, Quagliuolo V, Cerasoli S, et al.: Chordoma: Natural history and treatment results in 33 cases. J Surg Oncol 37: 185-91, 1988.

486. Mindell E R: Current concepts review. Chordoma. J Bone Joint Surg 63A: 501-5, 1981.

487. Jiang L, Liu Z J, Liu X G, et al.: Upper cervical spine chordoma of C2-C3. Eur Spine J 18 (3): Eur Spine J: 293-298; discussion 298-300, 2009.

488. Samson I R, Springfield D S, Suit H D, et al.: Operative treatment of sacrococcygeal chordoma. A review of twenty-one cases. J Bone Joint Surg 75: 1476-84, 1993.

489. Klekamp J, Samii M: Spinal chordomas - results of treatment over a 17-year period. Acta Neurochir (Wien) 138: 514-9, 1996.

490. Cheng E Y, Özerdemoglu R A, Transfeldt E E, et al.: Lumbosacral chordoma. Prognostic factors and treatment. Spine 24: 1639-45, 1999.

491. Suit H D, Goitein M, Munzenrider J, et al.: Definitive radiation therapy for chordoma and chondrosarcoma of base of skull and cervical spine. J Neurosurg 56: 377-85, 1982.

492. Rich T A, Schiller A, Mankin H J: Clinical and pathologic review of 48 cases of chordoma. Cancer 56: 182-7, 1985.

493. Magenau J M, Schuetze S M: New targets for therapy of sarcoma. Curr Opin Oncol 20 (4): Curr Opin Oncol: 400-6, 2008.

494. Rubinstein L J: Tumors of the central nervous system. Atlas of tumor pathology, second series, fascicle 6. Armed Forces Institute of Pathology, Washington, DC, 1972: pp 158-60.

495. Demierre B, Stichnoth F A, Hori A, et al.: Intracerebral ganglioglioma. J Neurosurg 65: 177-82, 1986.

496. Kalyan-Raman U P, Olivero W C: Ganglioglioma: A correlative clinicopathological and radiological study of ten surgically treated cases with follow-up. Neurosurgery 20: 428-33, 1987.

497. Sutton L N, Packer R J, Rorke L B, et al.: Cerebral gangliogliomas during childhood. Neurosurgery 13: 124-8, 1983.

498. Miller D C, Lang F F, Epstein F J: Central nervous system gangliogliomas. Part 1: Pathology. J Neurosurg 79: 859-66, 1993.

499. Lang F F, Epstein F J, Ransohoff J, et al.: Central nervous system gangliogliomas. Part 2: Clinical outcome. J Neurosurg 79: 867-73, 1993.

500. Russell D S, Rubenstein L J: Ganglioglioma: A case with a long history and malignant evolution. J Neuropathol Exp Neurol 21: 185-93, 1962.

501. Kudva Y C, Sawka A M, Young W F, Jr.: Clinical review 164: The laboratory diagnosis of adrenal pheochromocytoma: The Mayo Clinic experience. J Clin Endocrinol Metab 88 (10): J Clin Endocrinol Metab: 4533-9, 2003.

502. de Jong W H, Eisenhofer G, Post W J, et al.: Dietary influences on plasma and urinary metanephrines: Implications for diagnosis of catecholamine-producing tumors. J Clin Endocrinol Metab 94 (8): J Clin Endocrinol Metab: 2841-9, 2009.

503. Chretien P B, Engelman K, Hoye R C, et al.: Surgical management of intravascular glomus jugulare tumor. Am J Surg 122: 740-3, 1971.

504. Jackson C G, Harris P F, Glasscock M E I, et al.: Diagnosis and management of paragangliomas of the skull base. Am J Surg 159: 389-93, 1990.

505. Farrior J B, Hyams V J, Benke R, et al.: Carcinoid apudoma arising in a glomus jugulare tumor: Review of endocrine activity in glomus jugulare tumors. Laryngoscope 90: 110-9, 1980.

506. Jensen N F: Glomus tumors of the head and neck: Anesthetic considerations. Anesth Analg 78: 112-9, 1994.

507. Jackson C G, Glasscock M E, Nissen A J, et al.: Glomus tumor surgery: The approach, results, and problems. Otolaryngol Clin North Am 15: 897-916, 1982.

508. Kim J-A, Elkon D, Lim M-L, et al.: Optimum dose of radiotherapy for chemodectomas of the middle ear. Int J Radiation Oncology Biol Phys 6: 815-9, 1980.

509. Cummings B J, Beale F A, Garrett P G, et al.: The treatment of glomus tumors in the temporal bone by megavoltage radiation. Cancer 53: 2635-40, 1984.

510. Spector G J, Fierstein J, Ogura J H: A comparison of therapeutic modalities of glomus tumors in the temporal bone. Laryngoscope 86: 690-6, 1976.

511. Hatfield P M, James A E, Schulz M D: Chemodectomas of the glomus jugulare. Cancer 30: 1164-8, 1972.

512. Mork S J, Loken A C: Ependymoma: A follow-up study of 101 cases. Cancer 40: 907-15, 1977.

513. Duffner P K, Cohen M E, Freeman A I: Pediatric brain tumors: An overview. Ca 35: 287-301, 1985.

514. Taylor M D, Poppleton H, Fuller C, et al.: Radial glia cells are candidate stem cells of ependymoma. Cancer Cell 8 (4): Cancer Cell: 323-35, 2005.

515. Youmans J R, (ed.) Neurological surgery. 2nd ed., W. B. Saunders, Philadelphia, 1982.

516. Shaw E G, Evans R G, Scheithauer B W, et al.: Postoperative radiotherapy of intracranial ependymoma in pediatric and adult patients. Int J Radiation Oncology Biol Phys 13: 1457-62, 1987.

517. Merchant T E, Mulhern R K, Krasin M J, et al.: Preliminary results from a phase II trial of conformal radiation therapy and evaluation of radiation-related CNS effects for pediatric patients with localized ependymoma. J Clin Oncol 22 (15): J Clin Oncol: 3156-62, 2004.

518. MacDonald S M, Safai S, Trofimov A, et al.: Proton radiotherapy for childhood ependymoma: Initial clinical outcomes and dose comparisons. Int J Radiat Oncol Biol Phys 71 (4): Int J Radiat Oncol Biol Phys: 979-86, 2008.

519. Vanuytsel L, Brada M: The role of prophylactic apinal irradiation in localized intracranial ependymoma. Int J Radiation Oncology Biol Phys 21: 825-30, 1991.

520. van Veelen-Vincent M L, Pierre-Kahn A, Kalifa C, et al.: Ependymoma in childhood: Prognostic factors, extent of surgery, and adjuvant therapy. J Neurosurg 97 (4): J Neurosurg: 827-35, 2002.

521. Grundy R G, Wilne S A, Weston C L, et al.: Primary postoperative chemotherapy without radiotherapy for intracranial ependymoma in children: The UKCCSG/SIOP prospective study. Lancet Oncol 8 (8): Lancet Oncol: 696-705, 2007.

522. Little A S, Sheean T, Manoharan R, et al.: The management of completely resected childhood intracranial ependymoma: The argument for observation only. Childs Nerv Syst 25 (3): Childs Nerv Syst: 281-4, 2009.

523. Sutton L N, Goldwein J, Perilongo G, et al.: Prognostic factors in childhood ependymomas. Pediatr Neurosurg 16: 57-65, 1990.

524. Zacharoulis S, Ji L, Pollack I F, et al.: Metastatic ependymoma: A multi-institutional retrospective analysis of prognostic factors. Pediatric Blood and Cancer 50 (2): Pediatric Blood and Cancer: 231-5, 2008.

525. Kawabata Y, Takahashi J A, Arakawa Y, et al.: Long-term outcome in patients harboring intracranial ependymoma. J Neurosurg 103 (1): J Neurosurg: 31-7, 2005.

526. Tihan T, Zhou T, Holmes E, et al.: The prognostic value of histological grading of posterior fossa ependymomas in children: A Children’s Oncology Group study and a review of prognostic factors. Mod Pathol 21 (2): Mod Pathol: 165-77, 2008.

527. Ross G W, Rubinstein L J: Lack of histopathological correlation of malignant ependymomas with postoperative survival. J Neurosurg 70: 31-6, 1989.

528. Foreman N K, Love S, Thorne R: Intracranial ependymomas: Analysis of prognostic factors in a population-based series. Pediatr Neurosurg 24 (3): Pediatr Neurosurg: 119-25, 1996.

529. Pomeroy S L, Tamayo P, Gaasenbeek M, et al.: Prediction of central nervous system embryonal tumour outcome based on gene expression. Nature 415 (6870): 436-42, 2002.

530. Kumar R, Reddy S J, Wani A A, et al.: Primary spinal primitive neuroectodermal tumor: Case series and review of the literature. Pediatr Neurosurg 43 (1): 1-6, 2007.

531. Tomita T, McLone D G: Spontaneous seeding of medulloblastoma: Results of cerebrospinal fluid cytology and arachnoid biopsy from the cisterna magna. Neurosurgery 12: 265-7, 1983.

532. Reddy A T, Janss A J, Phillips P C, et al.: Outcome for children with supratentorial primitive neuroectodermal tumors treated with surgery, radiation, and chemotherapy. Cancer 88 (9): 2189-93, 2000.

533. Hong T S, Mehta M P, Boyett J M, et al.: Patterns of failure in supratentorial primitive neuroectodermal tumors treated in Children’s Cancer Group study 921, a phase III combined modality study. Int J Radiat Oncol Biol Phys60 (1): 204-13, 2004.

534. Allen J C: Childhood brain tumors: Current status of clinical trials in newly diagnosed and recurrent disease. Ped Clin N Am 32: 633-51, 1985.

535. Park T S, Hoffman H J, Hendrick E B, et al.: Medulloblastoma: Clinical presentation and management. J Neurosurg 58: 543-52, 1983.

536. Kessler L A, Dugan P, Concannon J P: Systemic metastases of medulloblastoma promoted by shunting. Surg Neurol 3: 147-52, 1975.

537. Blaser S I, Harwood-Nash D C: Neuroradiology of pediatric posterior fossa medulloblastoma. J Neurooncol 29 (1): 23-34, 1996.

538. Eberhart C G, Kepner J L, Goldthwaite P T, et al.: Histopathologic grading of medulloblastomas: A Pediatric Oncology Group study. Cancer 94 (2): 552-60, 2002.

539. Pramanik P, Sharma M C, Mukhopadhyay P, et al.: A comparative study of classical vs. Desmoplastic medulloblastomas. Neurol India 51 (1): 27-34, 2003.

540. Giangaspero F, Rigobello L, Badiali M, et al.: Large-cell medulloblastoma. Am J Surg Pathol 16: 687-93, 1992.

541. Corcoran R B, Scott M P: Oxysterols stimulate sonic hedgehog signal transduction and proliferation of medulloblastoma cells. Proc Natl Acad Sci U S A 103 (22): 8408-13, 2006.

542. Merchant T E, Wang M H, Haida T, et al.: Medulloblastoma: Long-term results for patients treated with definitive radiation therapy during the computed tomography era. Int J Radiat Oncol Biol Phys36 (1): 29-35, 1996.

543. Packer R J, Gajjar A, Vezina G, et al.: Phase III study of craniospinal radiation therapy followed by adjuvant chemotherapy for newly diagnosed average-risk medulloblastoma. J Clin Oncol 24 (25): 4202-8, 2006.

544. Tomita T, McLone D G: Medulloblastoma in childhood: Results of radical resection and low-dose radiation therapy. J Neurosurg 64: 238-42, 1986.

545. Packer R J, Sutton L N, Goldwein J W, et al.: Improved survival with the use of adjuvant chemotherapy in the treatment of medulloblastoma. J Neurosurg 74 (3): 433-40, 1991.

546. David K M, Casey A T, Hayward R D, et al.: Medulloblastoma: Is the 5-year survival rate improving? A review of 80 cases from a single institution. J Neurosurg 86 (1): 13-21, 1997.

547. Albright A L, Wisoff J H, Zeltzer P M, et al.: Effects of medulloblastoma resections on outcome in children: A report from the Children’s Cancer Group. Neurosurgery 38 (2): 265-71, 1996.

548. Evans A E, Jenkins R D, Sposto R, et al.: The treatment of medulloblastoma: Results of a prospective randomized trial of radiation therapy with and without CCNU, vincristine, and prednisone. J Neurosurg 72: 572-82, 1990.

549. Gilbertson R J: Medulloblastoma: Signalling a change in treatment. Lancet Oncol 5 (4): 209-18, 2004.

550. Weil M D, Lamborn K, Edwards M S, et al.: Influence of a child’s sex on medulloblastoma outcome. JAMA 279 (18): 1474-6, 1998.

551. Gajjar A, Hernan R, Kocak M, et al.: Clinical, histopathologic, and molecular markers of prognosis: Toward a new disease risk stratification system for medulloblastoma. J Clin Oncol 22 (6): 984-93, 2004.

552. Ray A, Ho M, Ma J, et al.: A clinicobiological model predicting survival in medulloblastoma. Clin Cancer Res 10 (22): 7613-20, 2004.

553. Mork S J, Rubinstein L J: Ependymoblastoma. A reappraisal of a rare embryonal tumor. Cancer 55: 1536-42, 1985.

554. Berger M S, Wilson C B: Epidermoid cysts of the posterior fossa. J Neurosurg 62: 214-9, 1985.

555. Gutin P H, Boehm J, Bank W O, et al.: Cerebral convexity epidermoid tumor subsequent to multiple percutaneous subdural aspirations. Case report. J Neurosurg 52 (4): 574-7, 1980.

556. Guidetti B, Gagliardi F M: Epidermoids and dermoid cysts. J Neurosurg 47: 12-8, 1977.

557. Fleming J F R, Botterell E H: Cranial dermoid and epidermoid tumors. Surg Gynecol Obstet 109: 57-79, 1959.

558. Alvord E C: Growth rates of epidermoid tumors. Ann Neurol 2: 367-70, 1977.

559. Link M J, Cohen P L, Breneman J C, et al.: Malignant squamous degeneration of a cerebellopontine angle epidermoid tumor. Case report. J Neurosurg 97 (5): 1237-43, 2002.

560. Sabin H I, Bardi L T, Symon L: Epidermoid cysts and cholesterol granulomas centered on the posterior fossa: Twenty years of diagnosis and management. Neurosurgery 21: 798-803, 1987.

561. Altschuler E M, Jungreis C A, Sekhar L N, et al.: Operative treatment of intracranial epidermoid cysts and cholesterol granulomas: Report of 21 cases. Neurosurgery 26: 606-14, 1990.

562. Friedman I: Epidermoid cholesteatoma and cholesterol granuloma: Experimental and human. Ann Otol Rhinol Laryngol 68: 57-79, 1959.

563. Chang P, Fagan P A, Atlas M D, et al.: Imaging destructive lesions of the petrous apex. Laryngoscope 108 (4 Pt 1): 599-604, 1998.

564. de Chadarevian J, Becker W J: Mollaret’s recurrent aseptic meningitis: Relationship and ultrastructural studies of the cerebrospinal fluid. J Neuropathol Exp Neurol 39: 661-9, 1980.

565. Abramson R C, Morawetz R B, Schlitt M: Multiple complications from an intracranial epidermoid cyst: Case report and literature review. Neurosurgery 24: 574-8, 1989.

566. Szabo M, Majtenyi C, Gusea A: Contribution to the background of Mollaret’s meningitis. Acta Neuropathol 59: 115-8, 1983.

567. Chambers A A, Lukin R R, Tomsick T A: Cranial epidermoid tumors: Diagnosis by computed tomography. Neurosurgery 1: 276-80, 1977.

568. Keville F J, Wise B L: Intracranial epidermoid and dermoid tumors. J Neurosurg 16: 564-9, 1959.

569. Ringertz N, Nordenstam H, Flyger G: Tumors of the pineal region. J Neuropathol Exp Neurol 13: 540-61, 1954.

570. Di Costanzo A, Tedeschi G, Di Salle F, et al.: Pineal cysts: An incidental MRI finding? J Neurol Neurosurg Psychiatry 56: 207-8, 1993.

571. Hasegawa A, Ohtsubo K, Mori W: Pineal gland in old age: Quantitative and qualitative morphological study of 168 human autopsy cases. Brain Res 409: 343-9, 1987.

572. Torres A, Krisht A F, Akouri S: Current management of pineal cysts. Contemp Neurosurg 27 (7): 1-5, 2005.

573. Maurer P K, Ecklund J, Parisi J E, et al.: Symptomatic pineal cysts: Case report. Neurosurgery 27: 451-4, 1990.

574. Wisoff J H, Epstein F: Surgical management of symptomatic pineal cysts. J Neurosurg 77: 896-900, 1992.

575. Klein P, Rubinstein L J: Benign symptomatic glial cysts of the pineal gland: A report of seven cases and review of the literature. J Neurol Neurosurg Psychiatry 52: 991-5, 1989.

576. Mamourian A C, Towfighi J: Pineal cysts: MR imaging. AJNR 7: 1081-6, 1986.

577. Di Chirico A, Di Rocco F, Velardi F: Spontaneous regression of a symptomatic pineal cyst after endoscopic third-ventriculostomy. Childs Nerv Syst 17 (1-2): 42-6, 2001.

578. Regis J, Bouillot P, Rouby-Volot F, et al.: Pineal region tumors and the role of stereotactic biopsy: Review of the mortality, morbidity, and diagnostic rates in 370 cases. Neurosurgery 39: 907-14, 1996.

579. Youssef A S, Keller J T, van Loveren H R: Novel application of computer-assisted cisternal endoscopy for the biopsy of pineal region tumors: Cadaveric study. Acta Neurochir (Wien) 149 (4): 399-406, 2007.

580. Oi S, Matsumoto S: Controversy pertaining to therapeutic modalities for tumors of the pineal region: A worldwide survey of different patient populations. Childs Nerv Syst 8 (6): 332-6, 1992.

581. Sugiyama K, Uozumi T, Kiya K, et al.: Intracranial germ-cell tumor with synchronous lesions in the pi-neal and suprasellar regions: Report of six cases and review of the literature. Surg Neurol 38 (2): 114-20, 1992.

582. Hoffman H J, Ostubo H, Hendrick E B, et al.: Intracranial germ-cell tumors in children. J Neurosurg 74: 545-51, 1991.

583. Shinoda J, Yamada H, Sakai N, et al.: Placental alkaline phosphatase as a tumor marker for primary intracranial germinoma. J Neurosurg 68 (5): 710-20, 1988.

584. Edwards M S B, Hudgins R J, Wilson C B, et al.: Pi-neal region tumors in children. J Neurosurg 68: 689-97, 1988.

585. Todo T, Kondo T, Shinoura N, et al.: Large cysts of the pineal gland: Report of two cases. Neurosurgery 29: 101-6, 1991.

586. Oi S, Matsuzawa K, Choi J U, et al.: Identical characteristics of the patient populations with pineal region tumors in Japan and in korea and therapeutic modalities. Childs Nerv Syst 14 (1-2): 36-40, 1998.

587. Stern J D, Ross D A: Stereotactic management of benign pineal region cysts: Report of two cases. Neurosurgery 32: 310-4, 1993.

588. Musolino A, Cambria S, Rizzo G, et al.: Symptomatic cysts of the pineal gland: Stereotactic diagnosis and treatment of two cases and review of the literature. Neurosurgery 32: 315-21, 1993.

589. Kelly P J: Comment on Musolino A, et al.: Symptomatic cysts of the pineal gland: Stereotactic diagnosis and treatment of two cases and review of the literature. Neurosurgery 32: 320-1, 1993.

590. Dempsey P K, Kondziolka D, Lunsford L D: Stereotactic diagnosis and treatment of pineal region tumors and vascular malformations. Acta Neurochir 116: 14-22, 1992.

591. Kelly P J: Comment on Regis J, et al.: Pineal region tumors and the role of stereotactic biopsy: Review of the mortality, morbidity, and diagnostic rates in 370 cases. Neurosurgery 39: 912-3, 1996.

592. Stein B M: The infratentorial supracerebellar approach to pineal lesions. J Neurosurg 35 (2): 197-202, 1971.

593. Jeibmann A, Wrede B, Peters O, et al.: Malignant progression in choroid plexus papillomas. J Neurosurg 107 (3 Suppl): J Neurosurg: 199-202, 2007.

594. Jeibmann A, Hasselblatt M, Gerss J, et al.: Prognostic implications of atypical histologic features in choroid plexus papilloma. J Neuropathol Exp Neurol 65 (11): J Neuropathol Exp Neurol: 1069-73, 2006.

595. Boyd M C, Steinbok P: Choroid plexus tumors: Problems in diagnosis and management. J Neurosurg 66: 800-5, 1987.

596. Ellenbogen R G, Winston K R, Kupsky W J: Tumors of the choroid plexus in children. Neurosurgery 25: 327-35, 1989.

597. Wrede B, Liu P, Wolff J E: Chemotherapy improves the survival of patients with choroid plexus carcinoma: A metaanalysis of individual cases with choroid plexus tumors. J Neurooncol 85 (3): J Neurooncol: 345-51, 2007.

598. Brat D J, Scheithauer B W, Staugaitis S M, et al.: Third ventricular chordoid glioma: A distinct clinicopathologic entity. J Neuropathol Exp Neurol 57 (3): 283-90, 1998.

599. Savitz M H, Anderson P J: Primary melanoma of the leptomeninges: A review. Mt Sinai J Med 41: 774-91, 1974.

600. Gibson J B, Burrows D, Weir W P: Primary melanoma of the meninges. J Pathol Bacteriol 74: 419-38, 1957.

601. Jooma R, Hayward R D, Grant D N: Intracranial neoplasms during the first year of life: Analysis of one hundred consecutive cases. Neurosurgery 14: 31-41, 1984.

602. Wakai S, Arai T, Nagai M: Congenital brain tumors. Surg Neurol 21: 597-609, 1984.

603. Thomas J E, Baker H L: Assessment of roentgeno-graphic lucencies of the skull: A systematic approach. Neurology 25: 99-106, 1975.

604. Rawlings C E, Wilkins R H: Solitary eosinophilic granuloma of the skull. Neurosurgery 15: 155-61, 1984.

605. Mitnick J S, Pinto R S: CT in the diagnosis of eosinophilic granuloma. J Comput Assist Tomogr 4: 791-3, 1980.

606. Hasegawa T, Ito H, Yamamoto S, et al.: Unilateral hyperostosis frontalis interna: Case report. J Neurosurg 59: 710-3, 1983.

607. Willison C D, Schochet S S, Voelker J L: Cranial epidural fibrous tumor associated with hyperostosis: A case report. Surg Neurol 40: 508-11, 1993.

608. Fulton J D, Shand J, Ritchie D, et al.: Hyperostosis frontalis interna, acromegaly and hyperprolactinemia. Postgrad Med J 66: 16-9, 1990.

609. Bavazzano A, Del Bianco P L, Del Bene E, et al.: A statistical evaluation of the relationships between headache and internal frontal hyperostosis. Res Clin Stud Headache 3: 191-7, 1970.

610. Floyd J L, Jackson D E, Carretta R: Appearance of hyperostosis frontalis interna on indium-111 leukocyte scans: Potential diagnostic pitfall. J Nucl Med 27: 495-7, 1986.

611. Oates E: Spectrum of appearance of hyperostosis frontalis interna on in-111 leukocyte scans. Clin Nucl Med 13: 922-3, 1988.

612. Johnson J D, Young B: Demographics of brain metastasis. Neurosurg Clin N Am 7 (3): 337-44, 1996.

613. Mintz A P, Cairncross J G: Treatment of a single brain metastasis. The role of radiation following surgical excision. JAMA 280: 1527-9, 1998 (editorial).

614. Voorhies R M, Sundaresan N, Thaler H T: The single supratentorial lesion: An evaluation of preoperative diagnosis. J Neurosurg 53: 364-8, 1980.

615. Patchell R A, Posner J B: Neurologic complications of systemic cancer. Neurol Clin 3: 729-50, 1985.

616. Zimm S, Galen L, Wampler G L, et al.: Intracerebral metastases in solid-tumor patients: Natural history and results of treatment. Cancer 48: 384-94, 1981.

617. DeAngelis L M: Management of brain metastases. Cancer Invest 12: 156-65, 1994.

618. Davis P C, Hudgins P A, Peterman S B, et al.: Diagnosis of cerebral metastases: Double-dose delayed CT versus contrast-enhanced MR imaging. AJNR 12: 293-300, 1991.

619. Weiss H D, Richardson E P: Solitary brainstem metastasis. Neurology 28: 562-6, 1978.

620. Nugent J L, Bunn P A, Matthews M J, et al.: CNS metastses in small-cell bronchogenic carcinoma: Increasing frequency and changing pattern with lengthening survival. Cancer 44: 1885-93, 1979.

621. Vieth R G, Odom G L: Intracranial metastases and their neurosurgical treatment. J Neurosurg 23: 375-83, 1965.

622. Agazzi S, Pampallona S, Pica A, et al.: The origin of brain metastases in patients with an undiagnosed primary tumour. Acta Neurochir (Wien) 146 (2): 153-7, 2004.

623. Kindt G W: The pattern of location of cerebral metastatic tumors. J Neurosurg 21: 54-7, 1964.

624. Figlin R A, Piantadosi S, Feld R, et al.: Intracranial recurrence of carcinoma after complete resection of stage I, II, and III non-small-cell lung cancer. N Engl J Med 318: 1300-5, 1988.

625. Auperin A, Arriagada R, Pignon J P, et al.: Prophylactic cranial irradiation for patients with small-cell lung cancer in complete remission. Prophylactic cranial irradiation overview collaborative group. N Engl J Med 341 (7): 476-84, 1999.

626. Slotman B, Faivre-Finn C, Kramer G, et al.: Prophylactic cranial irradiation in extensive small-cell lung cancer. N Engl J Med 357 (7): 664-72, 2007.

627. Solis O J, Davis K R, Adair L B, et al.: Intracerebral metastatic melanoma: CT evaluation. Comput Tomogr 1: 135-43, 1977.

628. Sampson J H, Carter J H, Friedman A H, et al.: Demographics, prognosis, and therapy in 702 patients with brain metastases from malignant melanoma. J Neurosurg 88: 11-20, 1998.

629. Swetter S M, Carroll L A, Johnson D L, et al.: Positron emission tomography is superior to computed tomography for metastatic detection in melanoma patients. Ann Surg Oncol 9 (7): 646-53, 2002.

630. Lochead R, McKhann G, Hankinson T, et al.: High dose systemic interleukin-2 for metastatic melanoma in patients with treated brain metastases. J Immunother 27 (6): J Immunother: S15, 2004.

631. Nabhani T, Liau L M: Neurosurgical and other treatment options for metastatic melanoma in the central nervous system: Part II. Contemp Neurosurg 27 (13): 1-5, 2005.

632. Essner R, Lee J H, Wanek L A, et al.: Contemporary surgical treatment of advanced-stage melanoma. Arch Surg 139 (9): 961-6; discussion 966-7, 2004.

633. Kondziolka D, Bernstein M, Resch L, et al.: Significance of hemorrhage into brain tumors: Clinico-pathological study. J Neurosurg 67: 852-7, 1987.

634. Shildt R A, Kennedy P S, Chen T T, et al.: Management of patients with metastatic adenocarcinoma of unknown origin: A Southwest Oncology Group study. Cancer Treat Rep 67 (1): Cancer Treat Rep: 77-9, 1983.

635. Patchell R A, Tibbs P A, Walsh J W, et al.: A randomized trial of surgery in the treatment of single metastases to the brain. N Engl J Med 322: 494-500, 1990.

636. Gaspar L, Scott C, Rotman M, et al.: Recursive partitioning analysis (RPA) of prognostic factors in three Radiation Therapy Oncology Group (RTOG) brain metastases trials. Int J Radiat Oncol Biol Phys 37 (4): 745-51, 1997.

637. Pollock B E: Management of patients with multiple brain metastases. Contemp Neurosurg 21 (18): 1-6, 1999.

638. Nieder C, Andratschke N, Grosu A L, et al.: Recursive partitioning analysis (RPA) class does not predict survival in patients with four or more brain metastases. Strahlenther Onkol 179 (1): 16-20, 2003.

639. Horton J: Treatment of metastases to the brain. Current Concept Oncology. 1984, pp 18-22.

640. Jackson D V, Richards F, Cooper M R, et al.: Prophylactic cranial irradiation in small cell carcinoma of the lung: A randomized study. JAMA 237: 2730-3, 1977.

641. Patchell R A, Tibbs P A, Regine W F, et al.: Postoperative radiotherapy in the treatment of single metastases to the brain: A randomized trial. JAMA 280: 1485-9, 1998.

642. Kramer S, Hendrickson F, Zelen M, et al.: Therapeutic trials in the management of metastatic brain tumors by different time/dose fraction schemes. Natl Cancer Inst Monogr 46: 213-21, 1977.

643. DeAngelis L M, Mandell L R, Thaler H T, et al.: The role of postoperative radiotherapy after resection of single brain metastases. Neurosurgery 24: 798-804, 1989.

644. Smalley S R, Schray M F, Laws E R, et al.: Adjuvant radiation therapy after surgical resection of solitary brain metastasis: Association with pattern of failure and survival. Int J Radiation Oncology Biol Phys 13: 1611-6, 1987.

645. Shaw E: Comment on DeAngelis L M, et al.: The role of postoperative radiotherapy after resection of single brain metastases. Neurosurgery 24: 804-5, 1989.

646. Sills A K: Current treatment approaches to surgery for brain metastases. Neurosurgery 57 (5 Suppl): S24-32; discusssion S1-4, 2005.

647. Bindal A K, Bindal R K, Hess K R, et al.: Surgery versus radiosurgery in the treatment of brain metastasis. J Neurosurg 84: 748-54, 1996.

648. Smalley S R, Laws E R, O’Fallon J R, et al.: Resection for solitary brain metastasis: Role of adjuvant radiation and prognostic variables in 229 patients. J Neurosurg 77: 531-40, 1992.

649. Bindal R K, Sawaya R, Leavens M E, et al.: Surgical treatment of multiple brain metastases. J Neurosurg 79: 210-6, 1993.

650. Tobler W D, Sawaya R, Tew J M: Successful laser-assisted excision of a metastatic midbrain tumor. Neurosurgery 18: 795-7, 1986.

651. Markesbery W R, Brooks W H, Gupta G D, et al.: Treatment for patients with cerebral metastases. Arch Neurol 35: 754-6, 1978.

652. Ruderman N B, Hall T C: Use of glucocorticoids in the palliative treatment of metastatic brain tumors. Cancer 18: 298-306, 1965.

653. Posner J B: Surgery for metastases to the brain. N Engl J Med 322: 544-5, 1990 (editorial).

654. Galicich J H, Sundaresan N, Thaler H T: Surgical treatment of single brain metastasis: Evaluation of results by CT scanning. J Neurosurg 53: 63-7, 1980.

655. Alexander E, Moriarty T M, Davis R B, et al.: Stereotactic radiosurgery for the definitive noninvasive treatment of brain metastases. J Natl Cancer Inst 87: 34-40, 1995.

656. Fuller B G, Kaplan I D, Adler J, et al.: Stereotactic radiosurgery for brain metastases: The importance of adjuvant whole brain irradiation. Int J Radiation Oncology Biol Phys 23: 413-8, 1992.

657. Sze G, Soletsky S, Bronen R, et al.: MR imaging of the cranial meninges with emphasis on contrast enhancement and meningeal carcinomatosis. AJNR 10: 965-75, 1989.

658. Radhakrishnan K, Ahlskog J E, Garrity J A, et al.: Idiopathic intracranial hypertension. Mayo Clin Proc 69: 169-80, 1994.

659. Digre K B, Corbett J J: Pseudotumor cerebri in men. Arch Neurol 45: 866-72, 1988.

660. Durcan F J, Corbett J J, Wall M: The incidence of pseudotumor cerebri: Population studies in Iowa and Louisiana. Arch Neurol 45: 875-7, 1988.

661. Radhakrishnan K, Ahlskog J E, Cross S A, et al.: Idiopathic intracranial hypertension (pseudotumor cerebri): Descriptive epidemiology in Rochester, Minn, 1976 to 1990. Arch Neurol 50: 78-80, 1993.

662. Rush J A: Pseudotumor cerebri: Clinical profile and visual outcome in 63 patients. Mayo Clin Proc 55: 541-6, 1980.

663. King J O, Mitchell P J, Thomson K R, et al.: Manometry combined with cervical puncture in idiopathic intracranial hypertension. Neurology 58 (1): 26-30, 2002.

664. Ahlskog J E, O’Neill B P: Pseudotumor cerebri. Ann Int Med 97: 249-56, 1982.

665. Corbett J J, Mehta M P: Cerebrospinal fluid pressure in normal obese subjects and patients with pseudotumor cerebri. Neurology 33: 1386-8, 1983.

666. Jacobson D M, Karanjia P N, Olson K A, et al.: Computed tomography ventricular size has no predictive value in diagnosing pseudotumor cerebri. Neurology 40: 1454-5, 1990.

667. Giuseffi V, Wall M, Siegel P Z, et al.: Symptoms and disease associations in idiopathic intracranial hypertension (pseudotumor cerebri): A case-control study. Neurology 41 (2 (Pt 1)): 239-44, 1991.

668. Round R, Keane J R: The minor symptoms of increased intracranial hypertension: 101 patients with benign intracranial hypertension. Neurology 38 (9): 1461-4, 1988.

669. Wang S J, Silberstein S D, Patterson S, et al.: Idiopathic intracranial hypertension without papilledema: A case control study in a headache center. Neurology 51: 245-9, 1998.

670. Sher N A, Wirtschafter J, Shapiro S K, et al.: Unilateral papilledema in ‘benign’ intracranial hypertension (pseudotumor cerebri). JAMA 250: 2346-7, 1983.

671. Bejjani G K, Cockerham K P, Pless M, et al.: Idiopathic intracranial hypertension. Contemp Neurosurg 24 (9): 1-8, 2002.

672. Wall M, George D: Idiopathic intracranial hypertension: A prospective study of 50 patients. Brain 114: 155-80, 1991.

673. Digre K B: Epidemioligy of idiopathic intracranial hypertension. Annual meeting of the North American Neuro-Ophthalmoligical Society (NANOS). 1992.

674. Farb R I, Vanek I, Scott J N, et al.: Idiopathic intracranial hypertension: The prevalence and morphology of sinovenous stenosis. Neurology 60: 1418-24, 2003.

675. Johnston I, Hawke S, Halmagyi M, et al.: The pseudotumor syndrome: Disorders of cerebrospinal fluid circulation causing intracranial hypertension without ventriculomegaly. Arch Neurol 48: 740-7, 1991.

676. Powers J M, Schnur J A, Baldree M E: Pseudotumor cerebri due to partial obstruction of the sigmoid sinus by a cholesteatoma. Arch Neurol 43: 519-21, 1986.

677. Wen D Y, Hardten D R, Wirtschafter J D, et al.: Elevated intracranial pressure from cerebral venous obstruction by Masson’s vegetant intravascular hemangioendothelioma. J Neurosurg 75: 787-90, 1991.

678. Johnson L N, Krohel G B, Madsen R W, et al.: The role of weight loss and acetazolamide in the treatment of idiopathic intracranial hypertension (pseudotumor cerebri). Ophthalmology 105 (12): 2313-7, 1998.

679. Newberg B: Pseudotumor cerebri treated by rice/reduction diet. Arch Intern Med 133: 802-7, 1974.

680. Weisberg L A: Benign intracranial hypertension. Medicine (Baltimore) 54: 197-207, 1975.

681. Hahn F J, McWilliams F E: The small ventricle in pseudotumor cerebri: Demonstration of the small ventricle in benign intracranial hypertension. CT 2: 249-53, 1978.

682. Higgins J N, Owler B K, Cousins C, et al.: Venous sinus stenting for refractory benign intracranial hypertension. Lancet 359 (9302): 228-30, 2002.

683. Brourman N D, Spoor T C, Ramocki J M: Optic nerve sheath decompression for pseudotumor cerebri. Arch Ophthalmol 106: 1384-90, 1988.

684. Sergott R C, Savino P J, Bosley T M: Modified optic nerve sheath decompression provides long-term visual improvement for pseudotumor cerebri. Arch Ophthalmol 106: 1384-90, 1988.

685. Corbett J J, Nerad J A, Tse D, et al.: Optic nerve sheath fenestration for pseudotumor cerebri: The lateral orbitotomy approach. Arch Ophthalmol 106: 1391-7, 1988.

686. Kelman S E, Heaps R, Wolf A, et al.: Optic nerve decompression surgery improves visual function in patients with pseudotumor cerebri. Neurosurgery 30: 391-5, 1992.

687. Kelman S E, Sergott R C, Cioffi G A, et al.: Modified optic nerve decompression in patients with functioning lumboperitoneal shunts and progressive visual loss. Ophthalmology 98: 1449-53, 1991.

688. Spoor T C, Ramocki J M, Madion M P, et al.: Treatment of pseudotumor cerebri by primary and secondary optic nerve sheath decompression. Am J Ophthalmol 112: 177-85, 1991.

689. Kaufman B: The “empty” sella turcica - a manifestation of the intrasellar subarachnoid space. Radiology 90: 931-41, 1968.

690. Perani D, Scotti G, Colombo N, et al.: Spontaneous CSF rhinorrhea through the lamina cribrosa associated with primary empty sella. Ital J Neurol Sci 5 (2): Ital J Neurol Sci: 167-72, 1984.

691. Lee W M, Adams J E: The empty sella syndrome. J Neurosurg 28: 351-6, 1968.

692. McKeever P E: Immunohistochemistry of the nervous system. In Diagnostic immunohistochemistry, Dabbs D J, (ed.). Churchill Livingstone, New York, 2002: pp 559-624.

693. Burton B K: Alpha-fetoprotein screening. Adv Pediatr 33: 181-96, 1986.

694. Burger P C, Scheithauer B W: AFIP atlas of tumor pathology. Fourth series. Fascicle 7: Tumors of the central nervous system. Armed Forces Institute of Pathology, Washington, D.C., 2007.

695. Riccardi V M: Von recklinghausen neurofibromatosis. N Engl J Med 305: 1617-27, 1981.

696. National Institutes of Health Consensus Development Conference: Neurofibromatosis: Conference statement. Arch Neurol 45: 575-8, 1988.

697. Packer R J, Gutmann D H, Rubenstein A, et al.: Plexiform neurofibromas in NF1: Toward biologic-based therapy. Neurology 58 (10): Neurology: 1461-70, 2002.

698. Sharif S, Moran A, Huson S M, et al.: Women with neurofibromatosis 1 are at a moderately increased risk of developing breast cancer and should be considered for early screening. J Med Genet 44 (8): J Med Genet: 481-4, 2007.

699. Sevick R J, Barkovich A J, Edwards M S, et al.: Evolution of white matter lesions in neurofibromatosis type 1: MR findings. AJR Am J Roentgenol 159 (1): AJR Am J Roentgenol: 171-5, 1992.

700. Karnes P S: Neurofibromatosis: A common neuro-cutaneous disorder. Mayo Clin Proc 73: 1071-6, 1998.

701. Walker L, Thompson D, Easton D, et al.: A prospective study of neurofibromatosis type 1 cancer incidence in the UK. Br J Cancer 95 (2): Br J Cancer: 233-8, 2006.

702. Martuza R L, Eldridge R: Neurofibromatosis 2: (bilateral acoustic neurofibromatosis). N Engl J Med 318: 684-8, 1988.

703. Parry D M, Eldridge R, Kaiser-Kupfer M I, et al.: Neurofibromatosis 2 (NF2): Clinical characteristics of 63 affected individuals and clinical evidence for heterogeneity. Am J Med Genet 52 (4): Am J Med Genet: 450-61, 1994.

704. Egelhoff J C, Bates D J, Ross J S, et al.: Spinal MR findings in neurofibromatosis types 1 and 2. AJNR 13: 1071-7, 1992.

705. Wiederholt W C, Gomez M R, Kurland L T: Incidence and prevalence of tuberous sclerosis in Rochester, Minnesota, 1950 through 1982. Neurology 35: 600-3, 1985.

706. Roach E S, Gomez M R, Northrup H: Tuberous sclerosis complex consensus conference: Revised clinical diagnostic criteria. J Child Neurol 13 (12): J Child Neurol: 624-8, 1998.

707. McLaurin R L, Towbin R B: Tuberous sclerosis: Diagnostic and surgical considerations. Pediat Neurosci 12: 43-8, 1985.

708. Thiele E A: Managing epilepsy in tuberous sclerosis complex. J Child Neurol 19 (9): J Child Neurol: 680-6, 2004.

709. Chow C W, Klug G L, Lewis E A: Subependymal giant-cell astrocytoma in children: An unusual discrepancy between histological and clinical features. J Neurosurg 68: 880-3, 1988.

710. Di Rocco F, Sabatino G, Koutzoglou M, et al.: Neurocutaneous melanosis. Childs Nerv Syst 20 (1): Childs Nerv Syst: 23-8, 2004.

711. DeDavid M, Orlow S J, Provost N, et al.: Neurocutaneous melanosis: Clinical features of large congenital melanocytic nevi in patients with manifest central nervous system melanosis. J Am Acad Dermatol 35 (4): J Am Acad Dermatol: 529-38, 1996.

712. McClelland S, 3rd, Charnas L R, SantaCruz K S, et al.: Progressive brainstem compression in an infant with neurocutaneous melanosis and Dandy-Walker complex following ventriculoperitoneal shunt placement for hydrocephalus. Case report. J Neurosurg 107 (6 Suppl): J Neurosurg: 500-3, 2007.

713. Mena-Cedillos C A, Valencia-Herrera A M, Arroyo-Pineda A I, et al.: Neurocutaneous melanosis in association with the Dandy-Walker complex, complicated by melanoma: Report of a case and literature review. Pediatr Dermatol 19 (3): Pediatr Dermatol: 237-42, 2002.

714. Kopelson G, Linggood R M, Kleinman G M, et al.: Management of intramedullary spinal cord tumors. Radiology 135: 473-9, 1980.

715. Adams R D, Victor M: Intraspinal tumors. In Principles of neurology. McGraw-Hill, New York, 2nd ed., 1981: pp 638-41.

716. Liu J K, Brockmeyer D L, Dailey A T, et al.: Surgical management of aneurysmal bone cysts of the spine. Neurosurg Focus 15 (5): Neurosurg Focus: E4, 2003.

717. Suzuki M, Satoh T, Nishida J, et al.: Solid variant of aneurysmal bone cyst of the cervical spine. Spine 29 (17): Spine: E376-81, 2004.

718. Neviaser J S, Eisenberg S H: Giant cell reparative granuloma of the cervical spine; case report. Bull Hosp Joint Dis 15 (1): Bull Hosp Joint Dis: 73-8, 1954.

719. Dunn E J, Davidson R I, Desai S: Diagnosis and management of tumors of the cervical spine. In The cervical spine, The Cervical Spine Research Society Editorial Committee, (ed.). JB Lippincott, Philadelphia, 2nd ed., 1989: pp 693-722.

720. Menezes A H, Sato Y: Primary tumors of the spine in children - natural history and management. Concepts Pediatr Neurosurg 10: 30-53, 1990.

721. Grubb M R, Currier B L, Pritchard D J, et al.: Primary Ewing’s sarcoma of the spine. Spine 19: 309-13, 1994.

722. Porter D G, Martin A J, Mallucci C L, et al.: Spinal cord compression due to Masson’s vegetant intravascular hemangioendothelioma: Case report. J Neurosurg 82: 125-7, 1995.

723. Scott R M, Wolpert S M, Bartoshesky L E, et al.: Dermoid tumors occurring at the site of previous myelomeningocele repair. J Neurosurg 65 (6): J Neurosurg: 779-83, 1986.

724. Storrs B B: Are dermoid and epidermoid tumors preventable complications of myelomeningocele repair? Pediatr Neurosurg 20 (2): Pediatr Neurosurg: 160-2, 1994.

725. Ramos E, Marlin A E, Gaskill S J: Congenital dermoid tumor in a child at initial myelomeningocele closure: An etiological discussion. J Neurosurg Pediatrics 2 (6): J Neurosurg Pediatrics: 414-5, 2008.

726. Solero C L, Fornari M, Giombini S, et al.: Spinal meningiomas: Review of 174 operated cases. Neurosurgery 25 (2): 153-60, 1989.

727. Lyons M K, O’Neill B P, Kurtin P J, et al.: Diagnosis and management of primary spinal epidural nonhodgkin’s lymphoma. Mayo Clin Proc 71: 453-7, 1996.

728. Stein B: Intramedullary spinal cord tumors. Clin Neurosurg 30: 717-41, 1983.

729. Stern W E: Localization and diagnosis of spinal cord tumors. Clin Neurosurg 25: 480-94, 1977.

730. DeSousa A L, Kalsbeck J E, Mealey J, et al.: Intraspinal tumors in children. A review of 81 cases. J Neurosurg 51: 437-45, 1979.

731. Hautzer N W, Aiyesimoju A, Robitaille Y: Primary spinal intramedullary lymphomas: A review. Ann Neurol 14: 62-6, 1983.

732. Alvisi C, Cerisoli M, Giuloni M: Intramedullary spinal gliomas: Long term results of surgical treatment. Acta Neurochir 70: 169-79, 1984.

733. White J B, Miller G M, Layton K F, et al.: Nonenhancing tumors of the spinal cord. J Neurosurg Spine 7 (4): 403-7, 2007.

734. Dorwart R H, LaMasters D L, Watanabe T J: Tumors. In Computed tomography of the spine and spinal cord, Newton T H and Potts D G, (eds.). Clavadal Press, San Anselmo, 1983: pp 115-31.

735. Tzerakis N, Georgakoulias N, Kontogeorgos G, et al.: Intraparenchymal myxopapillary ependymoma: Case report. Neurosurgery 55 (4): 981, 2004.

736. Helwig E B, Stern J B: Subcutaneous sacrococcygeal myxopapillary ependymoma. A clinicopathologic study of 32 cases. Am J Clin Pathol 81 (2): 156-61, 1984.

737. Malis L I: Intramedullary spinal cord tumors. Clin Neurosurg 25: 512-39, 1978.

738. Smaltino F, Bernini F P, Santoro S: Computerized tomography in the diagnosis of intramedullary metastases. Acta Neurochir 52: 299-303, 1980.

739. Edelson R N, Deck M D F, Posner J B: Intramedullary spinal cord metastases. Neurology 22: 1222-31, 1972.

740. Murphy K C, Feld R, Evans W K, et al.: Intramedullary spinal cord metastases from small cell carcinoma of the lung. J Clin Onc 1: 99-106, 1983.

741. Jellinger K, Kothbauer P, Sunder-Plassmann, et al.: Intramedullary spinal cord metastases. J Neurol 220: 31-41, 1979.

742. Stein B: Surgery of intramedullary spinal cord tumors. Clin Neurosurg 26: 473-9, 1979.

743. Sebastian P R, Fisher M, Smith T W, et al.: Intramedullary spinal cord metastasis. Surg Neurol 16: 336-9, 1981.

744. Nittner K: In Handbuch der neurochirurgie, Olivecrona H and Tonnis W, (eds.). Springer-Verlag, New York, 1972, Vol. VII 2: pp 1-606.

745. Post K D, Stein B M: Surgical management of spinal cord tumors and arteriovenous malformations. In Operative neurosurgical techniques, Schmidek H H and Sweet W H, (eds.). W.B. Saunders, Philadelphia, 3rd ed., 1995, Vol. 2: pp 2027-48.

746. Nadkarni T D, Rekate H L: Pediatric intramedullary spinal cord tumors: Critical review of the literature. Childs Nerv Syst 15: 17-28, 1999.

747. Greenwood J: Surgical removal of intramedullary tumors. J Neurosurg 26: 276-82, 1967.

748. Morota N, Deletis V, Constantini S, et al.: The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery 41: 1327-36, 1997.

749. Kothbauer P, Deletis V, Epstein F J: Intraoperative spinal cord monitoring for intramedullary surgery: An essential adjunct. Pediatric Neurosurgery 26: 247-54, 1997.

750. Albright A L: Intraoperative spinal cord monitoring for intramedullary surgery: An essential adjunct? Pediatric Neurosurgery 29: 112, 1998 (letter).

751. Albright A L: Pediatric intramedullary spinal cord tumors. Childs Nerv Syst 15: 436-7, 1999 (letter).

752. Guidetti B, Mercuri S, Vagnozzi R: Long-term results of the surgical treatment of 129 intramedullary spinal gliomas. J Neurosurg 54: 323-30, 1981.

753. Seppala M T, Haltia M J, Sankila R J, et al.: Long-term outcome after removal of spinal schwannoma: A clinicopathological study of 187 cases. J Neurosurg 83 (4): 621-6, 1995.

754. Conti P, Pansini G, Mouchaty H, et al.: Spinal neurinomas: Retrospective analysis and long-term outcome of 179 consecutively operated cases and review of the literature. Surg Neurol 61 (1): 34-43; discussion 44, 2004.

755. Asazuma T, Toyama Y, Maruiwa H, et al.: Surgical strategy for cervical dumbbell tumors based on a three-dimensional classification. Spine 29 (1): E10-4, 2004.

756. Janin Y, Epstein J A, Carras R, et al.: Osteoid osteomas and osteoblastomas of the spine. Neurosurgery 8: 31-8, 1981.

757. Amacher A L, Eltomey A: Spinal osteoblastoma in children and adolescents. Childs Nerv Syst 1: 29-32, 1985.

758. Lichtenstein L, Sawyer W R: Benign osteoblastoma. J Bone Joint Surg 46A: 755-65, 1964.

759. Shives T C, Dahlin D C, Sim F H, et al.: Osteosarcoma of the spine. J Bone Joint Surg Am 68 (5): 660-8, 1986.

760. Fox M W, Onofrio B M: The natural history and management of symptomatic and asymptomatic vertebral hemangiomas. J Neurosurg 78: 36-45, 1993.

761. Healy M, Herz D A, Pearl L: Spinal hemangiomas. Neurosurgery 13: 689-91, 1983.

762. Richardson R R, Cerullo L J: Spinal epidural cavernous hemangioma. Surg Neurol 12: 266-8, 1979.

763. Cosgrove G R, Bertrand G, Fontaine S, et al.: Cavernous angiomas of the spinal cord. J Neurosurg 68: 31-6, 1988.

764. Tekkök I H, Açikgöz B, Saglam A, et al.: Vertebral hemangioma symptomatic during pregnancy - report of a case and review of the literature. Neurosurgery 32: 302-6, 1993.

765. Graham J J, Yang W C: Vertebral hemangioma with compression fracture and paraparesis treated with preoperative embolization and vertebral resection. Spine 9: 97-101, 1984.

766. Kosary I A, Braham J, Shacked I, et al.: Spinal epidural hematoma due to hemangioma of vertebra. Surg Neurol 7: 61-2, 1977.

767. Keren D F, Alexanian R, Goeken J A, et al.: Guidelines for clinical and laboratory evaluation of patients with monoclonal gammopathies. Arch Pathol Lab Med 123: 106-7, 1999.

768. Bataille R, Harousseau J-L: Multiple myeloma. N Engl J Med 336: 1657-64, 1997.

769. McCarthy J, Proctor S J: Cerebral involvement in multiple myeloma. Case report. J Clin Pathol 31: 259-64, 1978.

770. Norum J, Wist E, Dahil I M: Cerebral metastases from multiple myeloma. Acta Oncol 30: 868-9, 1991 (letter).

771. Costa G, Engle R L, Schilling A, et al.: Melphalan and prednisone: An effective combination for the treatment of multiple myeloma. Am J Med 54: 589-99, 1973.

772. Foerster J: Multiple myeloma. In Wintrobe’s clinical hematology, Lee G R, Bithell T C, Foerster J, et al., (eds.). Lea and Febiger, Philadelphia, 9th ed., 1993, Vol. 2: pp 2219-49.

773. Marcove R C, Sheth D S, Brien E W, et al.: Conservative surgery for giant cell tumors of the sacrum. The role of cryosurgery as a supplement to curettage and partial excision. Cancer 74 (4): 1253-60, 1994.

774. Godersky J C, Smoker W R K, Knutzon R: Use of MRI in the evaluation of metastatic spinal disease. Neurosurgery 21: 676-80, 1987.

775. Livingston K E, Perrin R G: The neurosurgical management of spinal metastases causing cord and cauda equina compression. J Neurosurg 49: 839-43, 1978.

776. Batson O V: The function of the vertebral veins and their role in the spread of metastases. Ann Surg 112: 138, 1940.

777. Rodichok L D, Ruckdeschel J C, Harper G R, et al.: Early detection and treatment of spinal epidural metastases: The role of myelography. Ann Neurol 20: 696-702, 1986.

778. Schwab J H, Boland P, Guo T, et al.: Skeletal metastases in myxoid liposarcoma: An unusual pattern of distant spread. Ann Surg Oncol 14 (4): 1507-14, 2007.

779. Bach F, Larsen B H, Rhode K, et al.: Metastatic spinal cord compression. Occurrence, symptoms, clinical presentations and prognosis in 398 patients with spinal cord compression. Acta Neurochir (Wien) 107: 37-43, 1990.

780. Helwig-Larsen S, Sorensen P S: Symptoms and signs in metastatic spinal cord compression: A study from first symptom until diagnosis in 153 patients. Eur J Cancer 30A: 396-8, 1994.

781. Levack P, Graham J, Collie D, et al.: Don’t wait for a sensory level: Listen to the symptoms: A prospective audit of the delays in diagnosis of malignant cord compression. Clin Oncol (R Coll Radiol)14: 472-80, 2002.

782. Sherk H H: Lesions of the atlas and axis. Clin Orthop 109: 33-41, 1975.

783. Nakamura M, Toyama Y, Suzuki N, et al.: Metastases to the upper cervical spine. J Spinal Disord 9: 195-201, 1996.

784. Portenoy R K, Lipton R B, Foley K M: Back pain in the cancer patient: An algorithm for evaluation and management. Neurology 37: 134-8, 1987.

785. Brice J, McKissock W: Surgical treatment of malignant extradural spinal tumors. Br Med J 1: 1341-4, 1965.

786. Gabriel K, Schi D: Metastatic spinal cord compression by solid tumors. Semin Neurol 24: 375-83, 2004.

787. Li K, Poon P: Sensitivity and specificity of MRI in detecting spinal cord compression and in distinguishing malignant from benign compression fractures of vertebrae. Magn Reson Imaging 6: 547-56, 1988.

788. Francken A B, Hong A M, Fulham M J, et al.: Detection of unsuspected spinal cord compression in melanoma patients by 18F-fluorodeoxyglucose-positron emission tomography. Eur J Surg Oncol 31: 197-204, 2005.

789. Vecht C J, Haaxma-Reiche H, van Putten W L, et al.: Initial bolus of conventional versus high-dose dexamethasone in metastatic spinal cord compression. Neurology 39: 1255-7, 1989.

790. Hollis P H, Malis L I, Zappulla R A: Neurological deterioration after lumbar puncture below complete spinal subarachnoid block. J Neurosurg 64: 253-6, 1986.

791. Lee Y-Y, Glass J P, Wallace S: Myelography in cancer patients: Modified technique. AJR 145: 791-5, 1985.

792. Danner R L, Hartman B J: Update of spinal epidural abscess: 35 cases and review of the literature. Rev Infect Dis 9: 265-74, 1987.

793. Findlay G F G: Adverse effects of the management of malignant spinal cord compression. J Neurol Neurosurg Psychiatry 47: 761-8, 1984.

794. Witham T F, Khavkin Y A, Gallia G L, et al.: Surgery insight: Current management of epidural spinal cord compression from metastatic spine disease. Nat Clin Pract Neurol 2: 87-94, 2006.

795. Mundy G R: Metastasis to bone: Causes, consequences and therapeutic opportunities. Nat Rev Cancer 2 (8): Nat Rev Cancer: 584-93, 2002.

796. Fourney D R, Schomer D F, Nader R, et al.: Percutaneous vertebroplasty and kyphoplasty for painful vertebral body fractures in cancer patients. J Neurosurg 98: 21-30, 2003.

797. Reitan J B, Kaalhus O: Radiotherapy of liposarcomas. Br J Radiol 53 (634): 969-75, 1980.

798. Faul C M, Flickinger J C: The use of radiation in the management of spinal metastases. J Neurooncol 23: 149-61, 1995.

799. Rubin P: Extradural spinal cord compression by tumor: Part I. Experimental production and treatment trials. Radiology 93: 1243-8, 1969.

800. Rock J P, Ryu S, Yin F F, et al.: The evolving role of stereotactic radiosurgery and stereotactic radiation therapy for patients with spine tumors. J Neurooncol 69 (1-3): 319-34, 2004.

801. Onimus M, Schraub S, Bertin D, et al.: Surgical treatment of vertebral metastasis. Spine 11: 883-91, 1986.

802. Cooper P R, Errico T J, Martin R, et al.: A systematic approach to spinal reconstruction after anterior decompression for neoplastic disease of the thoracic and lumbar spine. Neurosurgery 32: 1-8, 1993.

803. Patchell R A, Tibbs P A, Regine W F, et al.: Direct decompressive surgical resection in the treatment of spinal cord compression caused by metastatic cancer: A randomized trial. Lancet 366 (9486): 643-8, 2005.

804. Fourney D R, Abi-Said D, Rhines L D, et al.: Simultaneous anterior-posterior approach to the thoracic and lumbar spine for the radical resection of tumors followed by reconstruction and stabilization. J Neurosurg 94: 232-44, 2001.

805. Sakaura H, Hosono N, Mukai Y, et al.: Outcome of total en bloc spondylectomy for solitary metastasis of the thoracolumbar spine. J Spinal Disord 17: 297-300, 2004.

806. Sundaresan N, Galicich J H, Lane J M, et al.: Treatment of neoplastic epidural cord compression by vertebral body resection and stabilization. J Neurosurg 63: 676-84, 1985.

807. Overby M C, Rothman A S: Anterolateral decompression for metastatic epidural spinal cord tumors: Results of a modified costrotransversectomy approach. J Neurosurg 62: 344-8, 1985.

808. Shaw B, Mansfield F L, Borges L: One-stage posterolateral decompression and stabilization for primary and metastatic vertebral tumors in the thoracic and lumbar spine. J Neurosurg 70: 405-10, 1989.

809. Akeyson E W, McCutcheon I E: Single-stage posterior vertebrectomy and replacement combined with posterior instrumentation for spinal metastasis. J Neurosurg 85: 211-20, 1996.

810. Fourney D R, Abi-Said D, Lang F F, et al.: Use of pedicle screw fixation in the management of malignant spinal disease: Experience in 100 consecutive cases. J Neurosurg 94: 25-37, 2001.

811. Wang J C, Boland P, Mitra N, et al.: Single-stage posterolateral transpedicular approach for resection of epidural metastatic spine tumors involving the vertebral body with circumferential reconstruction: Results in 140 patients. J Neurosurg Spine 1: 287-98, 2004.

812. Hunt T, Shen F H, Arlet V: Expandable cage placement via a posterolateral approach in lumbar spine reconstructions: Technical note. J Neurosurg Spine 5: 271-4, 2006.

813. Snell B E, Nasr F F, Wolfla C E: Single-stage thoracolumbar vertebrectomy with circumferential reconstruction and arthrodesis: Surgical technique and results in 15 patients. Neurosurgery (Operative Neurosurgery) 58: 263-9, 2006.

814. Sciubba D M, Gallia G L, McGirt M J, et al.: Thoracic kyphotic deformity reduction with a distractible titanium cage via an entirely posterior approach. Neurosurgery 60 (4 Suppl 2): 223-31, 2007.

815. Brodeur G, Pritchard J, Berthold F, et al.: Revisions of the international criteria for neuroblastoma diagnosis, staging, and response to treatment. J Clin Oncol 11: 1466-77, 1993.

816. Bouza C, Lopez-Cuadrado T, Cediel P, et al.: Balloon kyphoplasty in malignant spinal fractures: A systematic review and meta-analysis. BMC Palliat Care 8: 12, 2009.

817. Persaud T: The polka-dot sign. Radiology 246 (3): 980-1, 2008.

818. Lamborn K R, Chang S M, Prados M D: Prognostic factors for survival of patients with glioblastoma: Recursive partitioning analysis. Neuro Oncol 6 (3): 227-35, 2004.