Introduction
Ionizing radiation includes x-rays and gamma rays (both of which transmit their energy via photons) and particulate radiation. The goal of XRT in treating tumors is to cause cell death or to stop cell replication. Photons impart critical energy to achieve this result by the photoelectric effect (at lower energies, < 0.05 MeV), by Compton scattering (at higher energies of 0.1-10 MeV, e.g. in linear accelerators and Gamma knives), or by pair-production (at the highest energies)1. In the Compton effect, the initial collision of the photon with an atom creates a free electron which then ionizes other atoms and breaks chemical bonds. The absorption of radiation by indirect ionization in the presence of water produces free radicals (containing an unpaired electron) which causes cellular injury (usually by damaging DNA) within the tumor.
For a discussion of radiation dosage and units, see page 126.
22.1. Conventional external beam radiation
Fractionation
The practice in which the total radiation dose is delivered in a series of smaller brief applications. This is one means of increasing the therapeutic ratio (the ratio of the effectiveness of XRT on tumor cells to that of normal cells). Radiation injury is a function of the dose, the exposure time, and the area exposed. Radiation oncologists refer to the four “R’s” of radiobiology2:
1. Repair of sublethal damage
2. Reoxygenation of tumor cells that were hypoxic before XRT: oxygenated cells are more sensitive than hypoxic cells because oxygen combines with unpaired electrons to form peroxides which are more stable and lethal than free radicals
3. Repopulation of tumor cells following treatment
4. Redistribution (or reassortment) of cells within the cell cycle: cells in the mitotic phase are the most sensitive
Dosing
The biologically effective dose of fractionated radiation is often modeled by the linear-quadratic equation (LQ-model) shown in Eq 22-1, where D = the total dose of radiation, d = dose per fraction, and the factors α & β are used to describe the cell response to radiation. A high α/β ratio ≥ 10 is designated as early-responding tissue such as tumor cells, and a ratio ≤ 3 is considered late-responding tissue (mitotically quiescent), such as normal brain and also AVMs.
22.1.1. Cranial radiation
Following surgery for tumor (craniotomy or spinal surgery), most surgeons wait ≈ 7-10 days before instituting XRT to the surgical site (allows initiation of healing).
Two CNS tumors that “melt away” with XRT but tend to recur later:
1. lymphoma
2. germ cell tumors
RADIATION INJURY AND NECROSIS
Radiation necrosis (RN) may mimic recurrent (or de novo) tumor both clinically and radiographically. Differences in prognosis and treatment make it important to distinguish between tumor and RN.
PATHOPHYSIOLOGY
As radiation is selectively toxic to more rapidly dividing cells, the two normal cell types within the CNS most vulnerable to RN are vascular endothelium (which have a turnover time of ≈ 6-10 mos) and oligodendroglial cells. Vascular injury may be the primary limiting factor to the tolerance of cranial XRT3. Injury from XRT occurs at lower doses when given concurrently with chemotherapy (especially methotrexate).
Radiation effects are divided into 3 phases4:
1. acute: occur during treatment. Rare. Usually an exacerbation of symptoms already present. Probably secondary to edema. Treat with ↑ steroids
2. early delayed: few weeks to 2-3 mos following completion of XRT. In spinal cord → Lhermitte’s sign. In brain → post-irradiation lethargy & memory difficulties
3. late delayed: 3 mos-12 yrs (most within 3 years). Due to small artery injury → thrombotic occlusion → white matter atrophy or frank coagulative necrosis
Manifestations of radiation effects:
1. decreased cognition
A. dementia may develop following XRT5 in as little as 1 year post-XRT. Incidence was higher when doses of 25-39 Gy were given in fractions > 300 cGy6
B. children: may attain lower IQ by ≈ 25 points, especially with > 40 Gy whole brain XRT. Measurable IQ differences occur in children radiated before age 7, but more subtle deficits occur even in older children7
2. radiation necrosis
3. injury to anterior optic pathways
4. injury to hypothalamic-pituitary axis → hypopituitarism → growth retardation in children (see page 655 for radiation injury to pituitary)
5. primary hypothyroidism (especially in children)
6. may induce formation of new tumor: tumors most commonly identified as having increased incidence following radiation treatment are gliomas (including glioblastoma8), meningiomas9, and nerve sheath tumors10. Skull base tumors have been reported following EBRT11
7. malignant transformation: e.g. after SRS for vestibular schwannomas (see page 633)
8. leukoencephalopathy: profound demyelinating/necrotizing reaction 4-12 mos after combined RXT and methotrexate, especially in children with acute lymphoblastic leukemia (ALL) and adults with primary CNS tumors
EVALUATION (DIFFERENTIATING RN FROM RECURRENT TUMOR)
|
IMHO |
Over the years many methods have been championed to differentiate radiation necrosis from recurrent high-grade glioma. Some are listed below. None have proven adequately reliable, and this may not even be a useful exercise. Tumor cells are frequently found on biopsy. The decision whether to reoperate is usually based on whether there is progressive mass effect (regardless of whether it is necrosis or tumor) taking into consideration the patient’s neurologic condition, projected longevity, patient desires… |
CT & MRI
Cannot reliably differentiate some cases of RN from tumor (especially astrocytoma; RN occasionally resembles glioblastoma).
MR spectroscopy (see page 133) was reliable in distinguishing pure tumor (elevated choline) from pure RN (low choline), but was less definitive with mixed tumor/necrosis12.
DWI: mean ADCs were lower with recurrence (1.18 ± 0.13 X 10-3 mm/s) vs. necrosis (1.4 ± 0.17 X 10-3 mm/s)13 (not all cases biopsy proven).
Nuclear brain scan
Some reports of success with thallium-201 and technetium-99m brain scans.
Computerized radionuclide studies
PET (positron emission tomography) scan: because positron emitting isotopes have short half lives, PET scanning requires a nearby cyclotron to generate the radiopharmaceuticals at great expense. Utilizing [18F]-fluorodeoxyglucose (FDG), regional glucose metabolism is imaged and is generally increased with recurrent tumor, and is decreased with RN. Specificity for distinguishing RN from tumor recurrence is > 90%, but sensitivity may be too low to make it reliable14. Amino acid tracers such as [11C]methionine and [18F]tyrosine are taken up by most brain tumors15, especially gliomas, and may also be used to help differentiate tumor from necrosis. Accuracy may be increased by fusing PET scan with MRI16.
SPECT (single positron emission computed tomography): “poor man’s PET scan”. Uses radiolabeled amphetamine. Uptake depends on presence of intact neurons and the condition of cerebral blood vessels (including blood brain barrier). Decreased radionuclide uptake indicates necrosis, whereas tumor recurrence has no decreased uptake.
TREATMENT
Symptoms from any form of radiation toxicity often respond initially to steroids.
Reoperation and excision is appropriate if there is deterioration from mass effect, regardless of whether the mass effect is from recurrent tumor or RN (the decision to reoperate should be based on the patient’s Karnofsky rating, see page 1182). Although some benefit has been shown, most reoperation studies are biased because they often select the patients who are doing better.
Other forms of therapy include: hyperbaric oxygen and anticoagulation.
Patients with documented tumor recurrence (as opposed to RN) may also be considered for additional radiation (external beam, interstitial brachytherapy, or stereotactic radiosurgery (SRS)) or chemotherapy.
PREVENTION
Injury is dependent on total radiation dose, number of treatments or fractions (less damage occurs with more frequent small treatments), and volume treated.
Various studies to determine the tolerance of normal brain to XRT have estimated that 65-75 Gy given over 6.5-8 wks in 5 fractions/week is usually tolerated (radiation necrosis will occur in ≈ 5% after 60 Gy fractionated in 30 treatments over 6 weeks). Other studies have shown tolerance to 45 Gy for 10 fractions, 60 Gy for 35 fractions, and 70 Gy for 60 fractions4.
22.1.2. Spinal radiation
SIDE EFFECTS
1. radiation myelopathy: see below
2. those due to overlap with GI tract: N/V, diarrhea
3. bone marrow suppression
4. growth retardation in children17
5. risk of developing cavernous malformations of the spinal cord (see page 1106)
RADIATION MYELOPATHY
Radiation myelopathy (RM) typically occurs in patients with spinal cord included in radiation therapy (XRT) ports used to treat cancer outside the spinal cord, includes breast, lung, thyroid, and epidural mets. Radiation neuropathy may occur with irradiation in the region of the axilla for carcinoma of the breast (see page 795). In the lower extremities, XRT for pelvic or bone tumors (e.g. of the femur) may produce lumbar plexopathy. In addition to permanent changes, radiation therapy may also produce spinal cord edema which may resolve after completion of radiation therapy.
EPIDEMIOLOGY
Incidence difficult to estimate due to the fact that the onset is typically delayed together with the poor survival of patients with malignant disease requiring XRT.
Most cases reported involve the cervical cord in spite of the higher frequency with which the thoracic cord is exposed to XRT (perhaps due to higher XRT doses to the head and neck and longer survival than with lung Ca)18. Delay between completion of XRT and onset of symptoms is usually ≈ 1 yr (reported range: 1 mos-5 yrs).
Important factors relating to the occurrence of RM include18:
1. rate of application (probably the most important factor)
2. total radiation dose
3. extent of cord shielding
4. individual susceptibility and variability
5. amount of tissue radiated
6. vascular supply to the region radiated
7. source of radiation
PATHOPHYSIOLOGY
Effects of XRT on the spinal cord that lead to RM are:
1. direct injury to cells (including neurons)
2. vascular changes, including endothelial cell proliferation → thrombosis
3. hyalinization of collagen fibers
CLINICAL
Clinical types of radiation myelopathy
Four clinical types have been described and are shown in Table 22-1.
Onset is usually insidious, but abruptness has also been described; the presentation often mimics epidural mets. First symptoms: usually paresthesias and hypesthesia of LEs, and Lhermitte’s sign. Then spastic weakness of LEs with hyperreflexia develops. A Brown-Séquard syndrome is not uncommon.
Approximately 50% of patients developing RM also have dysphagia from esophageal strictures requiring dilatations (the dysphagia often predates the myelopathy).
Table 22-1 Types of radiation myelopathy
|
Type |
Description |
|
1 |
benign form; commonly several mos following XRT (reported as late as 1 yr). Usually resolves completely within several mos. Mild sensory symptoms (frequently limited to a Lhermitte’s sign) without objective neurological findings |
|
2 |
injury to anterior horn cells → lower motor neuron signs in arms or legs |
|
3 |
described only in experimental animals after doses larger than normal XRT. Complete cord lesion within hours due to injury to blood vessels |
|
4 |
the type commonly reported. Chronic, progressive myelopathy (see text) |
EVALUATION
Essentially a diagnosis of exclusion. Radiographic imaging (CT, myelography) will be normal. MRI may show spinal cord infarction. The history of previous radiation is key. The differential diagnosis is included in Acute paraplegia or quadriplegia on page 1190.
PROGNOSIS
Prognosis for Type 4 RM is poor. Usually progresses to complete (or near complete) cord lesion. Paraplegia and/or sphincter involvement are poor signs.
PREVENTION
Maximum recommended cord radiation dose depends on size of port, and varies with investigator. With large field techniques (> 10 cm of cord), the risk of RM is negligible with ≤ 3.3 Gy in 42 days (0.55 Gy/wk), and with small field techniques ≤ 4.3 Gy in 42 days (0.717 Gy/wk). Larger doses may possibly be given safely if fractionated over longer periods. Recommended upper limit: 0.2 Gy/fraction.
22.2. Stereotactic radiosurgery & radiotherapy
Key concepts:
• uses stereotactic localization to precisely focus therapeutic radiation on a lesion, a large dose given in a single treatment is called stereotactic radio surgery
• best accepted indication: AVM ≤ 3 cm diameter with compact nidus for which surgical removal is not appropriate (deep location, proximity to eloquent brain)
• advantage: low immediate procedural morbidity
• disadvantages: delayed complications of radiation. With AVM: long latency (1-3 years) to obliteration creates period with risk of hemorrhage
The term “stereotactic radiosurgery” (SRS) describes the use of stereotactic localization to administer large radiation doses via multiple noncoplanar ports or arcs (producing a very steep radiation gradient) to a precise locus while exposing normal structures to safely tolerated doses. Unlike conventional external beam radiation therapy (EBRT), the dose is usually administered in a single treatment session.
Using stereotactic techniques to administer fractionated radiation has been called stereotactic radio therapy (see page 775). Fractionation capitalizes on the differential response of normal tissue from tumor cells to radiation (see Fractionation, page 770).
Stereotactic radiosurgery is also increasingly being used for spine lesions19.
INDICATIONS
In general, SRS is useful for well circumscribed lesions less than approximately 2.5-3 cm diameter (the “classic” lesion for which SRS is used is for appropriate AVMs, see below). For larger lesions, the radiation dose must be reduced because of anatomic and radiobiological constraints, and the precision of the stereotactic technique is offset due to overlap.
Published uses of stereotactic radiosurgery include:
1. AVMs: see below
2. tumors: see below
A. vestibular schwannomas: see below
B. pituitary adenomas: conventional EBRT (fractionated over ≈ 5 wks) is generally preferred to SRS as the initial form of XRT
C. craniopharyngiomas
D. pineal tumors
E. metastases
F. high grade gliomas: see below
G. meningiomas of the cavernous sinus20
3. functional neurosurgery
A. for control of chronic pain21 including trigeminal neuralgia22, 23 (see page 555)
B. pallidotomies for Parkinson’s disease (see page 535): usually not a technique of choice because of inability to perform physiologic stimulation prior to lesioning to verify target location which may vary by several millimeters. May be a consideration for the rare patient who cannot undergo placement of a stimulating/lesioning needle (e.g. refractory coagulopathy)
4. for treating patients refusing open surgery for various conditions
AVMS
SRS is best accepted for the treatment of small (< 3 cm) AVMs that are deep or border on eloquent brain and have a “compact” (i.e. sharply demarcated) nidus24-26. This includes those incompletely excised with previous surgery. The radiation induces endothelial cell proliferation which produces thickening of the vascular wall and ultimately obliteration of the lumen over a period of ≈ 1-2 years. SRS is of no benefit for venous angiomas (see page 1104). For a comparison of treatment options for AVMs see page 1102.
Larger AVMs (up to 5 cm) have also been treated with SRS with some success. Tentorial dural AVMs (see page 1109) have also shown promising response to SRS28.
TUMORS
The use of SRS for tumors is controversial. It is not advisable for use on benign tumors in young patients because of possible delayed side-effects following radiation (see Delayed morbidity, page 778) (possible exception: vestibular schwannomas, see below).
Infiltrating tumors
Generally not indicated for infiltrating tumors, e.g gliomas (poorly defined tumor margins defeats the advantage of precisely localized radiation) although it has been used for recurrent lesions following traditional treatment (surgical excision and fractionated external beam radiation). One of the arguments for SRS in these tumors is the fact that 90% of recurrences are within the original radiographic solid tumor volume29. RTOG trial 9305 showed no benefit with upfront use of SRS added to EBRT and BCNU chemotherapy in treating glioblastoma (http://www.rtog.org/closedsummaries/9305.html).
Vestibular schwannoma
Possible indications for SRS for VS: poor operative candidates (due to poor medical condition and/or advanced age, some use > 65 or 70 yrs as a cutoff), patient refusing surgery, bilateral VSs, post-operative treatment of incompletely removed VSs that continue to grow on serial imaging, or recurrences following surgical removal (also see Vestibular schwannomas under Results below).
CONTRAINDICATIONS
Compressive tumors of the spinal cord or medulla: even with the sharp isodose fall-off curves of SRS, there is still significant radiation delivered within a few millimeters of the margins of the lesion. This, together with the slight swelling of lesions that commonly follows SRS creates significant risk of neurologic injury, especially over the long term (and long survival is even more likely with benign lesions in young individuals).
COMPARISON OF SRS TECHNOLOGIES
Various methods are available, differing mostly on the source of the radiation and the technique for increasing the dose delivered to the lesion. A photon beam that is produced by electron acceleration is called an x-ray, whereas if it is produced by natural radioactive decay it is called a gamma ray. Although photons are identical regardless of how they are produced, gamma rays have a narrower distribution of energy than x-rays. The spatial accuracy of the gamma knife may be slightly better than linac systems, but the small difference does not seem critical because the error inherent in selecting the target margins exceeds the typical linac imprecision of ± 1 mm30. The linac has greater flexibility in dealing with non-spherical lesions and is much more economical than the gamma knife. For small lesions (< 3 cm diameter) both photon and charged particle beam sources have similar results.
Gamma knife
Different sized collimators and exposure times, using more than one isocenter, and plugging collimators that would pass radiation thorough sensitive structures are used to modify the treatment plan.
Linac
Standard linacs usually require modifications to provide the required precision (e.g. precision bearings, external collimators…).
Different sized collimators, different beam energies (arc weighting), and alterations of the arc paths and the number of arcs are used to modify treatments.
STEREOTACTIC RADIOTHERAPY (SRT)
AKA fractionated SRS. NB: there may be some confusion due to the similar names, and both use stereotactic localization, but stereotactic radio therapy is fractionated c.f. stereotactic radio surgery which is usually given in a single treatment. AVMs share some characteristics of what radiation oncologists call “late responding” lesions based on the linear quadratic model (LQ-model) (see page 770), and there is little rationale for fractionated protocols (although the LQ-model may not apply to SRS). Some slow growing tumors may also be similar to late responding tissue, but there may regions of hypoxic cells where XRT will be less effective, and where the reoxygenation phenomenon would improve response (see the Four “R’s” of radiobiology, page 770). Also, if there is some uncertainty regarding the tumor margins on CT or MRI and there is the possibility that some normal brain may be included in the treatment plan (or fear that constricting the treatment margins would exclude some tumor) this is again a situation where tissue repair may make fractionation more advantageous.
Accelerated fractionation (2-3 fractions/d x 1 week) are being investigated but are not appropriate in the vicinity of radiosensitive structures and may be inconvenient and expensive. Hypofractionation (1 fraction/d x 1 week) may be a better compromise.
For malignancies, fractionated schemes will almost always improve effectiveness of XRT. Research into SRT employs various methods to reposition the stereotactic frame, including masks, dental molds, etc. Displacement errors can be as high as 2-8 mm with mask systems, whereas recommended tolerances are 0.3 mm and 3°.
Although the optimal protocol has not yet been determined, SRT may have significant advantages for pituitary adenomas, peri-chiasmal lesions, in children (where it is even more desirable to minimize radiation of the normal brain), and in vestibular schw-annomas considered for XRT where there is useful hearing.
Vestibular schwannomas (VS)
Conventional EBRT is relatively effective in controlling residual or unresectable VSs (see page 633). SRT adds precision, with reported local control rates (LCR) of 94-100%31, comparable to that for SRS (with follow-up typically in 5-year range, which is short for these characteristically slow-growing tumors).
Rx: sample SRT protocol31: 6-MV Linac with a micromultileaf collimator used to deliver 54 Gy in 30 fractions of 1.8 Gy prescribed to the 90% isodose line via 7-22 noncoplanar static fields or 4-6 noncoplanar dynamic arcs to a target defined as the tumor volume plus a margin of 1-3 mm.
Cranial nerve dysfunction: SRT has not been compared head-to-head with SRS, but preliminary results suggest it may be superior to SRT (see page 632).
TREATMENT PLANNING
For a selected isocentric radiation dose to be delivered to a given volume, computer simulation programs help radiosurgeons select the number of arcs or beams, collimator width, etc., to keep exposure of nearby normal brain to acceptable limits, and limit radiation to particularly sensitive structures. Table 22-2 shows maximum recommended doses of various organs for a single fraction. In the brain, critical radiation sensitive structures include: optic vitreous, nerve, and chiasm, brain stem, and pituitary gland.
Cranial nerves: special sensory nerves (optic, vestibulocochlear) are the most radiosensitive. Somatic afferents (trigeminal), visceral efferents (facial), and somatic efferents (oculomotor, hypoglossal) are the next most radiosensitive33.
SRS treatment may also have a deleterious effect in structures sensitive to swelling, such as brain stem. Most radiosurgeons decline to use SRS for lesions in the region of the optic chiasm. However, in general it is not the radiosensitive structures located at a distance from the lesion that are at greatest risk. Rather, it is that tissue included in the higher dose isocenters immediately adjacent to the lesion.
For the linac, optimal dose drop-off usually occurs when ≥ 500° total degree-arc is used (e.g. 5 arcs of 100° each). Using more than 5 arcs rarely produces a significant difference out to the 20% isodose curve.
Table 22-2 Maximum recommended radiation dose of critical organs (delivered in a single fraction)
|
Structure |
Maximum dose (cGy) |
% of maximum (at a prescribed dose of 50 Gy) |
|
eye lens (cataract induction begins at 500 cGy) |
100 |
2% |
|
optic nerve32 |
100 |
2% |
|
skin in beam |
50 |
1% |
|
thyroid |
10 |
0.2% |
|
gonads |
1 |
0.02% |
|
breast |
3 |
0.06% |
Doses
Doses specify the amount of radiation delivered to the isocenter (or to a specified isodose curve, e.g. 18 Gy to the 50% isodose curve) and relating the isodose curve to a specific region of the lesion (e.g. at the edge of the AVM nidus). Dose-volume relation: the dose of radiation that can be tolerated is highly dependent on the volume being treated (larger treatment volumes require lower doses to avoid complications).
Dose selection is made based on known information or is estimated from dose-volume-relationship. If uncertain, err on the side of a lower dose. Previous XRT must also be taken into account by the radiation physicist. Adjacent structures within ≈ 2.5 mm of the target will receive injurious radiation and the total dose should reduced.
Target localization
CT: accuracy is never better than ≈ 0.6 mm which is the pixel size.
MRI: has 1-2 mm shift due to spatial distortion artifact from the magnet. If MRI is required to visualize the lesion, it may be better to use image fusion techniques with CT.
Stereotactic angiography: rarely required, and may even introduce errors in treatment planning. Stereotactic angiography should not be used alone because of problems including: the true geometry of the lesion cannot be fully appreciated, vessels may be obscured by other vessels or bone, etc.34-37. Digital subtraction angiography is even more problematic because it warps the image and requires an “unwarping” algorithm to be used for SRS.
Conformational planning
The shape of the treatment volume can be influenced by covering some sources (with gamma-knife units) or by choosing arcs with certain orientations (with linac based systems). Also, static and dynamic collimators have been developed.
Lesions that are not round or ellipsoid in shape may also be accommodated by using multiple isocenters. Lower total doses for each isocenter must then be used.
AVMs
If embolization is used before SRS, wait ≈ 30 days between the two procedures. DO NOT use radioopaque material in embolization mixture (see page 1103). Some experts find that target selection after embolization is extraordinarily difficult because of multiple small residual “nidi”.
A bolus-enhanced stereotactic CT is usually employed (except for those AVMs that are difficult to see on CT or when metal clips from previous surgery or radioopaque substance from embolization creates too much artifact). Caution with stereotactic angiography (see above).
A general consensus is that 15 Gy to the periphery of the AVM is optimal (range: 10-25). At McGill with linac SRS, they use 25-50 Gy delivered to 90% isodose curve at the edge of the nidus. With Bragg-peak, complications occurred less frequently with doses ≤ 19.2 Gy compared to doses above that (this may reduce the obliteration rate or increase the latency period)38.
Due to the fact that AVMs are benign lesions that are often treated in young patients, conformal planning is critical to avoid injury to nearby normal brain.
Tumors
Vestibular schwannomas (and meningiomas): For 1 isocenter: 10-15 Gy with the tumor at the 80% isodose line (current recommended maximum dose39, 40: 14 Gy) is associated with a lower incidence of cranial nerve palsies than higher doses41. For 2 isocenters: treat 10-15 Gy at the 70% isodose line.
Metastases: Median dose of 15 Gy (range: 9-25 Gy) at the center with the tumor contained in the 80% isodose curve has been recommended. One literature review found a reported range of 13-18 Gy at the center with good local control42.
RESULTS
AVMs
At 1 year, 46-61% of AVMs were completely obliterated on angiography, and at 2 years 86% were obliterated. There was no reduction in size in < 2% of cases. Smaller lesions have higher obliteration rates (with Bragg-peak in AVMs < 2 cm diameter, 94% thrombosed at 2 yrs, and 100% at 3 yrs38). AVMs > 25 mm diameter have only ≈ 50% chance of obliteration with 1 SRS treatment.
Although the immediate “procedural” mortality is 0%, Bragg-peak proton beam treatment of AVMs affords no protection against hemorrhage in the first 12-14 months following treatment26 (the so-called “incubation period”); this is similar to the 12-24 month latency for photon radiation24. Hemorrhages have occurred during the incubation period even in AVMs that had never bled before38, and the question has been raised whether a partially thrombosed AVM is more likely to bleed because of increased outflow resistance.
Factors associated with treatment failures include43: incomplete angiographic definition of the nidus (the most frequent factor, responsible for 57% of cases), recanalization of the nidus (7%), masking of nidus by hematoma, and a theorized “radiobiological resistance”. In some, no discernible reason for failure could be identified. In this series the complete obliteration rate was ≤ 64%, possibly because arteriography was heavily relied upon for treatment planning instead of emphasizing stereotactic CT.
If AVM persists 2-3 yrs after SRS, retreatment with SRS is an option43 (usually the residual is smaller).
Vestibular schwannomas
In 111 tumors ≤ 3 cm in size44, 44% decreased in size, 42% did not change, and 14% increased. Although retardation of growth is observed in the majority of cases, long-term results are not available to fully assess therapeutic efficacy and complication rate at this time45. Use in recurrent VSs following microsurgery is endorsed by some (see page 633). Also see Outcome & follow-up, page 631 for a comparison of outcomes (including cranial nerve dysfunction) with SRS versus microsurgery.
Gliomas
Median survival for large GBMs is so poor that SRS did not appear to have any benefit. Following SRS for gliomas, there is rarely reduction of enhancing volume (it is more common to have enlargement, sometimes with increased neurologic deficit).
Metastases
There has not been a randomized study to compare surgery to SRS. See page 710 for comparison of outcomes with cerebral mets with various treatments including SRS. Radiographic local control rate of ≈ 88% (reported range: 82-100%) has been cited42.
The advantage of SRS is that there is no risk from the treatment of hemorrhage, infection, or mechanical spread of tumor cells. Disadvantages include not obtaining tissue for diagnosis (11% of the time the lesions may not be mets, see page 707).
No significant difference has been found with SRS between tumors considered “radiosensitive” and those that are “radioresistant” as defined by standards developed for EBRT (see Table 21-70, page 708) (however, histology may affect the rate of response). The lack of significance of “radioresistance” may be due in part to the fact that the sharp dose drop-off with SRS allows higher doses to be delivered to tumors than would be used with EBRT.
Supratentorial control is better than infratentorial. Also, there is no significant difference in local control between single and dual mets. The RTOG has identified 3 or fewer mets as a more favorable prognosticator.
TREATMENT MORBIDITY AND MORTALITY
Immediate morbidity
Immediate mortality from the actual treatments themselves is probably zero. Morbidity: all but ≈ 2.5% of patients were discharged home within 24 hrs. Many centers do not admit patients overnight. Some immediate adverse reactions include46:
1. 16% of patients require analgesics for post-procedural headaches and antiemetics for nausea/vomiting
2. at least 10% of patients with subcortical AVMs had focal or generalized seizures within 24 hrs of treatment (only one was on subtherapeutic AEDs. All were controllable with additional AEDs)
Premedication: The Pittsburgh Gamma Knife group gives methylprednisolone 40 mg IV and phenobarbital 90 mg IV immediately after the radiation dose to patients with tumors or AVMs to reduce these adverse effects46.
Delayed morbidity
Long-term morbidity directly related to the radiation may occur, and just as with conventional XRT, is more frequent with larger doses and treatment volumes. Another risk particular to AVMs is that of hemorrhage during the latency period, which is 3-4% during the first year and is not higher following SRS. Radiation complications 47:
1. white matter changes: occurred 4-26 mos (mean 15.3) post-SRS. Seen on imaging (high intensity on MRI T2WI, or low density on CT) in ≈ 50% of patients, symptomatic in only 20% of patients38. Associated with radiation necrosis in ≈ 3% of cases
2. vasculopathy: diagnosed by narrowing seen on angiography or by ischemic changes on imaging in ≈ 5% of cases
3. cranial nerve deficits: occur in ≈ 1% of all cases. Incidence is higher with tumors of CPA or skull base
4. induced tumors:
A. malignant: only 6 reported malignant tumors in over 80,000 radiosurgical procedures for benign disease48. Estimated incidence: < 1 in 1000. Includes GBM, malignant degeneration of a schwannoma…
B. meningiomas: 0.7% chance of developing this within 10 years of XRT in a series with 2 cases identifiedA in 1333 patients contacted (out of 2500 treated)50 (risk was 1.9% in another series51)
A. meeting the criteria of Cahan et al.49 for XRT induced tumors
5. normal perfusion pressure breakthrough52: classically occurs following conventional microsurgery for AVMs (see page 1104), it has also been described following SRS53
22.3. Interstitial brachytherapy
Technique whereby radioactive implants are used to deliver locally high doses of radiation directly to tumors while exposing nearby normal brain to less toxic doses. At present, the numbers are too small and the follow-up too short to determine the efficacy of interstitial brachytherapy54. Controlled prospective studies have not yet been completed.
Interstitial brachytherapy (IB) may reduce the rate of tumor growth, but it rarely produces clinical improvement. Patients are generally not considered for IB unless their Karnofsky score is ≥ 70.
Techniques include:
1. insertion of high activity iodine-125 pellets which remain in place (either by conventional open surgery or by stereotactic technique)
2. insertion of catheters (so-called afterloading catheters) containing radioactive source (such as gold or I125) by stereotactic technique, which are then removed at a predetermined time (usually 1-7 days)
3. instillation of radioactive liquids (e.g. phosphorous isotope) into a cyst cavity
I125 has several characteristics that favor its use: it emits low-energy gamma rays which are absorbed by surrounding tissues minimizing radiation exposure of the normal brain, medical personnel and visitors. It is available as low-activity (< 5 mCi) or high-activity (5-40 mCi) seeds.
Treatment planning is devised to deliver 60 Gy to the edge of a volume that extends 1 cm beyond the contrast-enhancing tumor, with variations included to spare radiosensitive structures (e.g. optic chiasm). Usual delivery rates are 40-50 cGy/hr to the tumor margin (30 cGy/hr is the critical dose for cessation of human tumor growth) requiring that the seeds stay in the afterloading catheter ≈ 6 days.
RADIATION NECROSIS
Symptomatic radiation necrosis (RN) occurs in ≈ 40% of cases, and may occur as early as several months after IB. It may be impossible to differentiate from recurrent tumor in many cases. Symptomatic treatment is often achieved with increased corticosteroid dosages. Continued neurologic deterioration may require craniotomy.
OUTCOME
IB is often used as a “last ditch” effort in a patient with a recurrent malignant tumor who has received maximal external beam irradiation and who is not a candidate for re-operation (as expected, the results in patients with such poor prognoses are not good). However, patients eligible for IB are usually better than those who are not candidates, and this may bias the results towards a better outcome55. Some studies with early (primary treatment) use have shown possible benefit56.
22.4. References
1. Thompson T P, Maitz A H, Kondziolka D, et al.: Radiation, radiobiology, and neurosurgery. Contemp Neurosurg 21 (13): 1-5, 1999.
2. Hall E J, Cox J D: Physical and biologic basis of radiation therapy. In Moss’ radiation oncology, Cox J D, (ed.). Mosby-Year Book, Inc., St. Louis, Missouri, 7th ed., 1994: pp 3-66.
3. O’Connor M M, Mayberg M R: Effects of radiation on cerebral vasculature: A review. Neurosurgery 46: 138-51, 2000.
4. Leibel S A, Sheline G E: Radiation therapy for neoplasms of the brain. J Neurosurg 66: 1-22, 1987.
5. Duffner P K, Cohen M E, Thomas P: Late effects of treatment on the intelligence of children with posterior fossa tumors. Cancer 51: 233-7, 1983.
6. DeAngelis L M, Delattre J Y, Posner J B: Radiation-induced dementia in patients cured of brain metastases. Neurology 39 (6): Neurology: 789-96, 1989.
7. Radcliffe J, Packer R J, Atkins T E, et al.: Three- and four-year cognitive outcome in children with non-cortical brain tumors treated with whole-brain radiotherapy. Ann Neurol 32: 551-4, 1992.
8. Zuccarello M, Sawaya R, deCourten-Myers: Glioblastoma occurring after radiation therapy for meningioma: Case report and review of literature. Neurosurgery 19: 114-9, 1986.
9. Mack E E, Wilson C B: Meningiomas induced by high-dose cranial irradiation. J Neurosurg 79: 28-31, 1993.
10. Ron E, Modan B, Boice J D, et al.: Tumors of the brain and nervous system after radiotherapy in childhood. N Engl J Med 319: 1033-9, 1988.
11. Lustig L R, Jackler R K, Lanser M J: Radiation-induced tumors of the temporal bone. Am J Otol 18: 230-5, 1997.
12. Rock J P, Hearshen D, Scarpace L, et al.: Correlations between magnetic resonance spectroscopy and image-guided histopathology, with special attention to radiation necrosis. Neurosurgery 51 (4): 912-9; discussion 919-20, 2002.
13. Hein P A, Eskey C J, Dunn J F, et al.: Diffusion-weighted imaging in the follow-up of treated high-grade gliomas: Tumor recurrence versus radiation injury. AJNR Am J Neuroradiol 25 (2): 201-9, 2004.
14. Thompson T P, Lunsford L D, Kondziolka D: Distinguishing recurrent tumor and radiation necrosis with positron emission tomography versus stereotactic biopsy. Stereotact Funct Neurosurg 73 (1-4): 9-14, 1999.
15. Ericson K, Lilja A, Bergstrom M, et al.: Positron emission tomography with ([11C]methyl)-l-methionine, [11C]d-glucose, and [68Ga]EDTA in supratentorial tumors. J Comput Assist Tomogr 9: 683-9, 1985.
16. Thiel A, Pietrzyk U, Sturm V, et al.: Enhanced accuracy in differential diagnosis of radiation necrosis by positron emission tomography-magnetic resonance imaging coregistration: Technical case report. Neurosurgery 46: 232-4, 2000.
17. Tomita T, McLone D G: Medulloblastoma in childhood: Results of radical resection and low-dose radiation therapy. J Neurosurg 64: 238-42, 1986.
18. Eyster E F, Wilson C B: Radiation myelopathy. J Neurosurg 32: 414-20, 1970.
19. 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.
20. Duma C M, Lunsford L D, Kondziolka D, et al.: Stereotactic radiosurgery of cavernous sinus meningiomas as an addition or alternative to microsurgery. Neurosurgery 32: 699-705, 1993.
21. Steiner L, Forster D, Leksell L, et al.: Gammathalamotomy in intractable pain. Acta Neurochir 52: 173-84, 1980.
22. Leksell L: Stereotactic radiosurgery in trigeminal neuralgia. Acta Chir Scand 137: 311-4, 1971.
23. Lunsford L D: Comment on Taha J M and Tew J M: Comparison of surgical treatments for trigeminal neuralgia: Reevaluation of radiofrequency rhizotomy. Neurosurgery 38: 871, 1996.
24. Saunders W M, Winston K R, Siddon R L, et al.: Radiosurgery for arteriovenous malformations of the brain using a standard linear accelerator: Rationale and technique. Int J Radiation Oncology Biol Phys 13: 441-7, 1988.
25. Poulsen M G: Arteriovenous malformations: A summary of 6 cases treated with radiation therapy. Int J Radiation Oncology Biol Phys 13: 1553-7, 1987.
26. Kjellberg R N, Hanamura T, Davis K R, et al.: Bragg-peak proton-beam therapy for arteriovenous malformations of the brain. N Engl J Med 309: 269-74, 1983.
27. Pollock B E, Garces Y I, Stafford S L, et al.: Stereotactic radiosurgery for cavernous malformations. J Neurosurg 93 (6): 987-91, 2000.
28. Lewis A I, Tomsick T A, Tew J M: Management of tentorial dural arteriovenous malformations: Transarterial embolization combined with stereotactic radiation or surgery. J Neurosurg 81: 851-9, 1994.
29. 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.
30. Luxton G, Petrovich Z, Jozsef G, et al.: Stereotactic radiosurgery: Principles and comparison of treatment methods. Neurosurgery 32: 241-59, 1993.
31. Selch M T, Pedroso A, Lee S P, et al.: Stereotactic radiotherapy for the treatment of acoustic neuromas. J Neurosurg 101: 362-72, 2004.
32. Leber K A, Berglöff J, Pendi G: Dose-response tolerance of the visual pathways and cranial nerves of the cavernous sinus to stereotactic radiosurgery. J Neurosurg 88: 43-50, 1998.
33. Tishler R B, Loeffler J S, Lunsford L D, et al.: Tolerance of cranial nerves of the cavernous sinus to radiosurgery. Int J Radiation Oncology Biol Phys 27: 215-21, 1993.
34. Blatt D L, Friedman W A, Bova F J: Modifications based on computed tomographic imaging in planning the radiosurgical treatment of arteriovenous malformations. Neurosurgery 33: 588-96, 1993.
35. Bova F J, Friedman W A: Stereotactic angiography: Pitfalls and problems. Int J Radiation Oncology Biol Phys 20: 891-5, 1991.
36. Spiegelmann R, Friedman W A, Bova F J: Limitations of angiographic target localization in radiosurgical treatment planning. Neurosurgery 30: 619-24, 1992.
37. Friedman W A: Comment on Pollock B E, et al.: Repeat stereotactic radiosurgery of arteriovenous malformations: Factors associated with incomplete outcomes. Neurosurgery 38: 323, 1996.
38. Steinberg G K, Fabrikant J I, Marks M P, et al.: Stereotactic heavy-charged-particle Bragg-peak radiation for intracranial arteriovenous malformations. N Engl J Med 323: 96-101, 1990.
39. Kondziolka D, Lunsford L D, McLaughlin M R, et al.: Long-term outcomes after radiosurgery for acoustic neuromas. N Engl J Med 339: 1426-33, 1998.
40. Niranjan A, Lunsford L D, Flickinger J C, et al.: Dose reduction improves hearing preservation rates after intracanalicular acoustic tumor radiosurgery. Neurosurgery 45: 753-65, 1999.
41. Linskey M E, Lunsford L D, Flickinger J C: Radiosurgery for acoustic neuromas: Early experience. Neurosurgery 26: 736-45, 1990.
42. 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.
43. Pollock B E, Kondziolka D, Lunsford L D, et al.: Repeat stereotactic radiosurgery of arteriovenous malformations: Factors associated with incomplete outcomes. Neurosurgery 38: 318-24, 1996.
44. Hirsch A, Norén G: Audiological findings after stereotactic radiosurgery in acoustic neuromas. Acta Otolaryngol (Stockh) 106: 244-51, 1988.
45. 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.
46. Lunsford L D, Flickinger J, Coffey R J: Stereotactic gamma knife radiosurgery. Initial North American experience in 207 patients. Arch Neurol 47: 169-75, 1990.
47. Marks M P, Delapaz R L, Fabrikant J I, et al.: Intracranial vascular malformations: Imaging of charged-particle radiosurgery. Part II. Complications. Radiology 168: 457-62, 1988.
48. Loeffler J S, Niemierko A, Chapman P H: Second tumors after radiosurgery: Tip of the iceberg or a bump in the road? Neurosurgery 52 (6): 1436-40; discussion 1440-2, 2003.
49. Cahan W G: Radiation-induced sarcoma--50 years later. Cancer 82 (1): 6-7, 1998.
50. Sheehan J, Yen C P, Steiner L: Gamma knife surgery-induced meningioma: Report of two cases and review of the literature. J Neurosurg 105: 325-9, 2006.
51. Brada M, Rajan B, Traish D, et al.: The long-term efficacy of conservative surgery and radiotherapy in the control of pituitary adenomas. Clin Endocrinol (Oxf) 38 (6): 571-8, 1993.
52. Spetzler R F, Wilson C B, Weinstein P, et al.: Normal perfusion pressure breakthrough theory. Clin Neurosurg 25: 651-72, 1978.
53. Pollock B E: Occlusive hyperemia: A radiosurgical phenomenon? Neurosurgery 47: 1178-84, 2000.
54. Bernstein M, Laperriere N, Leung P, et al.: Interstitial brachytherapy for malignant brain tumors: Preliminary results. Neurosurgery 26: 371-80, 1990.
55. Florell R C, Macdonald D R, Irish W D, et al.: Selection bias, survival, and brachytherapy for glioma. J Neurosurg 76: 179-83, 1992.
56. Gutin P H, Prados M D, Phillips T L, et al.: External irradiation followed by an interstitial high activity iodine-125 implant “boost” in the initial treatment of malignant gliomas: NCOG study 6G-82-2. Int J Radiation Oncology Biol Phys 21: 601-6, 1991