B. Storch-Hagenlocher
Indications
Conventional cytology. Qualitative differences between the cells in the CSF make it possible to demonstrate morphological equivalents of various diseases. The main indications are:
• Acute, subacute, and chronic inflammation of the CNS.
• Tumors of the CNS and meninges.
• Detection of absorptive reactions in case of hemorrhage and irritation.
To fully exploit the possibilities of CSF diagnosis, differential cell counts should be performed with every cerebrospinal puncture.
The frequent practice of performing differential cell counts only when the total cell count is above 20 cells/μL may lead to an incorrect diagnosis:
• In the presence of meningeal tumor infiltration, the total cell count may be normal.
• An old, not too extensive subarachnoid hemorrhage will indeed always cause a phagocytic reaction, but the total cell count is often not increased.
Immunocytology. Immunocytochemical methods are a useful supplement to conventional cytological CSF diagnosis. Their main diagnostic indication lies in the detection of malignant cells, especially when:
• The cells cannot be adequately characterized morphologically.
• It is difficult to distinguish in the differential diagnosis between inflammatory transformation and malignant transformation of lymphocytes.
• The tumor cell fraction is very small.
The detection of intracellular immunoglobulins is also a sensitive inflammation marker and the object of diagnostic and scientific investigations.
Preanalytical Requirements
Conventional cytology. The CSF should be collected in a tightly closing sterile, transparent, colorless, and uncoated polystyrene tube. Conical tubes are best for subsequent processing.
Cell analysis must be carried out within 1–2 hours, since even at 4°C cell loss becomes unpredictable after 2 hours.
For protein studies, CSF without cells may be stored for 1–3 weeks at 4°C. The time and site of collection should be given on the laboratory requisition form.
Immunocytology. Immunocytological CSF diagnosis requires especially rapid and careful processing of the cell preparation. The emphasis should be on gentle sedimentation, high cell yield, and celerity. A low rotational speed of 200–300 g during precentrifugation and 750 rotations per minute (rpm) during cytocentrifugation usually ensure that cells are well preserved and surface structures intact. Coated multi-field adhesive slides are preferred, since they increase cell adhesion and make it possible to apply the cells obtained by gentle centrifugation directly onto the microscopic slide. The sediment must be prepared within 1–2 hours and air-dried, and staining is best done immediately thereafter. Unfortunately, this is often not possible for logistic reasons, but it has been shown that either freezing the sediment or storing the cells in preservative can damage surface structures and thus prevent optimal staining.
Procedures
Cell Counts
The cell count is obtained using untouched CSF. For this purpose, the CSF should be fresh and gently mixed. The cell counting is mainly done under the microscope, usually in a Fuchs-Rosenthal counting chamber, less often in a Neubauer counting chamber. Increasingly, automated hematological analyzers are being adapted for counting cells in CSF and are used routinely in diagnostics.
Counting in a Chamber
Structure of the chamber. The Fuchs-Rosenthal chamber differs from the usual chambers for counting blood cells in having a larger surface area (16 mm2) and larger chamber depth (0.2 mm), thus yielding a larger volume of 3.2 μL (the volume of the Neubauer chamber is 1 μL). The surface area is subdivided into 4 × 4 large squares, with 16 small squares in each; i. e., it consists of a total of 256 small squares.
How to Use the Counting Chamber
Gently mix the fresh CSF.
Mix the CSF with a staining solution (e. g., 5.0 mL glacial acetic acid + 0.2 mL methyl violet + 1.0 mL absolute ethanol + distilled water to make 100 mL) to stain the nuclei of leukocytes and lyse the erythrocytes.
Using a leukocyte pipette or a small tube, dilute the CSF with the staining solution at a ratio of 10:1 (corresponding to a CSF dilution of 9 in 10).
Thoroughly mix the diluted CSF.
Place one drop into the counting chamber.
Wait a few minutes to let the cells sediment.
Count the cells under a microscope.
Determination of cell count. If the number of cells is low, all the squares are counted. If there are more than 20 cells in one large square, it is enough just to count the cells in four large squares and then obtain the total cell count (CC) by multiplication. With the usual CSF dilution of 9/10 and a chamber volume of 3.2 μL, the number of cells per microliter of CSF is calculated as follows:
Assessment of cell count:
Normal: ≤ 4 cells/μL.
Pathological: ≥ 5 cells/μL.
If the CSF is contaminated with blood, the total cell count must be corrected; for this purpose, deduct 1 cell per 1000 erythrocytes.
Automated Cell Counting
For automated cell counting, different cytometric systems are available: Advia 120 (Beyer), CellDyn 4000 (Abott), and UF-100 Urine Flow Cytometer (Sysmex).
Examination. The amount of CSF required for examination is 100–300 μL; it is mixed with a stain-fixative solution according to the instructions. Measuring can be done after 4 minutes; the stained cells are stable for a period of 4 hours. To ensure that no cells transferred from previously processed samples or from blood samples are included, three blank measurements must be made before the CSF sample is measured.
Results. Comparative studies have shown varying degrees of agreement with the manual cell count (Strik et al., 2005) depending on the number of cells and any possible contamination with blood. The UF-100 and CellDyn 4000 systems are able to differentiate bacteria or fungi from cells, thus preventing falsely high cell counts (Ziebig et al., 2000; Hoffmann and Janssen, 2002).
Samples containing blood or a lot of cell debris pose a problem; they require dilution at higher than 1:10 to allow less than 1500 erythrocytes/μL to be obtained.
The Advia 120 system also allows distinction between different cells. Comparative studies with differential cell counts under the microscope showed good agreement in respect of neutrophilic granulocytes (van Acker et al., 2001; Aune et al., 2004; Mahieu et al., 2004). The agreement in respect of monocytes and eosinophilic granulocytes was less good. With regard to lymphocytes, reports regarding the reliability of automated cell counting are completely contradictory (Strik et al., 2005). No studies are available regarding differentiation between transformed lymphocytes, plasma cells, macrophages, and tumor cells.
|
Table 5.1 Advantages and disadvantages of cell-concentrating procedures |
||
|
Procedure |
Advantage |
Disadvantage |
|
Sedimentation chamber |
Good preservation of cell morphology, optimal for cytodiagnostic assessment |
• Requires relatively large volume of CSF • Cell loss when sucking off fluid with filter paper |
|
Cytocentrifugation |
• Practical to use • Centrifugation can be standardized • High yield of cells • Requires small volume of CSF |
Morphology of cell changes: • Elliptically stretched cytoplasm, irregular borders • Enlarged cytoplasmic borders of lymphocytes, prominent nucleoli (lymphocytes may appear transformed) |
For the time being, current methods of automated differential cell counts by no means replace conventional CSF cytology under the microscope.
Concentration of Cells
Since CSF contains few cells, differential cell counts can not be performed on a smear as they are for blood—the cells have to be concentrated first.
Sedimentation Chamber Method and Cytocentrifugation
Requirements. A procedure for cell concentration should neither change the cell composition nor denature the cells during processing, nor should it change their original morphology. The procedure must be simple, quick, and easy to perform in a reproducible manner. The amount of CSF required must be small.
Methods. Two methods are generally used:
• Sedimentation chamber procedure: The Sayk sedimentation chamber, which has undergone various modifications, uses forced natural sedimentation of the cells. CSF (1–5 mL, depending on the cell count) is placed in a cylinder standing on a microscopic slide. The cylinder is held down by a spring, and liquid is sucked up by filter paper surrounding the cylinder. The cells thus sediment within 20–30 minutes and are then available for further processing.
• Cytocentrifugation: Using a cytocentrifuge, cell preparations can be produced within 10 minutes. At least 200 μL of CSF is placed in a cylinder. The CSF is sucked up by filter paper, as in the Sayk chamber, and is therefore unavailable for further diagnostic tests. The centrifuge should be set at 800 rpm.
Each method has its advantages and disadvantages (Table 5.1). Other procedures, such as the Kistler and Bischoff membrane filter method, and the use of prestained microscopic slides (test simplets) have not achieved general use, because they change the cell morphology or the preparations are unstable.
Microscopic slides. The slides for cell sedimentation are stored in 3% HCl-ethanol. Before use, they are thoroughly rinsed under tap water and dried.
Precentrifugation A widely used method for improving the yield of cells while maintaining cell morphology is precentrifugation.
Principle. CSF is centrifuged in a conical polystyrene tube for 8–10 minutes at 2200 rpm. The supernatant is pipetted off and stored for further diagnostic tests that can use cellfree CSF. The pellet is resuspended in at least 200 μL of a protein-containing medium (e. g., 9 mL medium 199 + 1 mL heat-inactivated fetal calf serum), mixed well, and pipetted into the cytocentrifugation chamber.
Advantages and disadvantages. The advantages of this method include an excellent cell yield, the possibility of preparing several sediments, the availability of the total supernatant for further protein diagnostics, and the cytoprotective effect of the medium. The method is disadvantageous when the CSF is mildly or moderately contaminated with blood and the cell image is obscured by overlying erythrocytes, making cell evaluation difficult. In addition, maintaining sterility is essential, since bacterial contamination of the medium leads to rapid growth of bacteria.
The medium needs to be stored in a refrigerator and should be prepared freshly every day.
Staining Methods
All staining methods used in hematology can be applied to CSF sediments. The cells become sufficiently fixed by airdrying; additional fixation is unnecessary and can change the cell morphology. Only Gram's stain requires the sediment (without any medium) to be heat-fixed by flaming.
May–Grünwald–Giemsa stain. Standard staining uses the May–Grünwald–Giemsa method. The nuclei appear in different hues of purple and the cytoplasm in different hues of blue, thus making it possible to differentiate individual structures (Kölmel, 1978).
Fig. 5.2 Immunocytochemical staining of B lymphoblasts with a CD 19 antibody and alkaline phosphatase anti-alkaline phosphatase (APAAP) detection reaction.
May–Grünwald–Giemsa Staining Procedure
Air-dry the cell preparation.
Cover the slide for 3 minutes with May–Grünwald solution (eosin + methylene blue).
Drain staining solution and rinse with tap water.
Cover slide for 15–20 minutes with freshly prepared, filtered Giemsa solution (standard solution diluted 1:10 with tap water).
Drain staining solution and rinse with tap water.
Air-dry the preparation.
Special stains. The Berlin blue reaction detects iron-containing material. Sudan black or Sudan red stains fat particles. Periodic-acid–Schiff stain (PAS reaction) demonstrates mucopolysaccharides and glycogen. PAS-positive reactions are found in lymphoblasts and carcinoma cells, and also in macrophages associated with Whipple's disease.
Immunocytochemistry
Principle. Antigenic structures on the cell surface or in the cytoplasm are labeled with specific antibodies. The antibody-labeled cells are visualized by coupling to a fluorescent dye, or they are stained by coupling to alkaline phosphatase (Fig. 5.2) or peroxidase. This makes accurate cell typing possible. Since usually only a few cells are available, microscopic methods are predominantly used. (Flow cytometric cell typing requires at least 10 000 cells and is available only in large centers.)
Procedure. To prevent nonspecific binding, CSF proteins have to be removed by washing with phosphate buffer; the preparation is then fixed with glutaraldehyde or methanol and then incubated with a target-specific monoclonal antibody. The optimal antibody concentrations vary; they have to be titrated for each individual antibody.
Selection of antibodies. The selection of antibodies is guided by the cell morphology, any known previous findings (such as a known primary tumor), and the clinical problem at hand. According to all of this, an antibody panel is assembled (Table 5.2). Double-staining with fluorescentlabeled antibodies makes it possible to assign antigens to certain cellular structures. However, fluorescent-stained preparations are perishable, so they have to be assessed immediately, and re-evaluation at a later time is not possible. Indirect staining of antibodies with alkaline phosphatase or immunoperoxidase allows long-term preservation of the preparations, thus ensuring correct assignment of antigenic structures to individual cell types.
Findings
Visual Assessment
Normal CSF is colorless and as clear as water.
Changes in color. If erythrocytes are present (> 1000/μL), the CSF is cloudy and pink; contamination with leukocytes (~ 1000/μL) results in an opaque-turbid to whitish-yellow color. In purulent meningitis, the CSF can be purulent-turbid due to an increased cell count, bacteria, and elevated protein content. The CSF can also be whitish-turbid when just the protein content is increased.
|
Table 5.2 Antibodies frequently used in immunocytochemistry |
|
|
Specific antibody against: |
Indication |
|
CD 34, CD 1, CD 2, CD 3, CD 4, CD 5, CD 7, CD 8, CD 10, CD 19, CD 20, CD 22, IgM, κ, λ, terminal deoxynucleotidyl transferase (TdT) |
Lymphoma, lymphatic leukemia |
|
CD 13, CD 14, CD 33, CD 34, CD 64, CD 117, myeloperoxidase (MPO), lysozyme |
Myelogenous leukemia |
|
Cytokeratin, carcinoembryonic antigen (CEA), epithelial membrane antigen (EMA) |
Carcinoma |
|
Neuron-specific enolase (NSE) |
Neuroendocrine tumors (e. g., small-cell bronchial carcinoma) |
|
Vimentin; in the case of melanoma, also protein S-100 and melanoma-specific antigens HMB-45 and HMB-50 |
Mesenchymal tumors (e. g., melanoma, sarcoma, and lymphoma) |
|
Vimentin, glial fibrillary acidic protein (GFAP), neurofilament (NF), NSE, protein S-100, cytokeratins |
Primary brain tumors |
Fig. 5.3 Normal lymphocytes in CSF.
Fibrin contamination. The sediment may contain delicate or, in purulent meningitis, coarse fibrin clots and sometimes also cells. Tuberculous meningitis may lead to a severe increase in fibrin, with coarse fibrin clots (“spiderweb clots”), often containing mycobacteria.
Xanthochromia. Xanthochromia, a diffuse yellow to yellowbrown discoloration of the CSF, is due to contamination with hemoglobin derivatives originating from a hemorrhage that is several days old. Xanthochromia is caused by the presence of bilirubin; it is also found in severe hyperbilirubinemia.
How to Tell Hemorrhage from Artificial Contamination with Blood
Bloody CSF should be gently centrifuged (at 200–300g) as soon as possible, preferably immediately after collection. If the CSF has been artificially contaminated with blood, the supernatant is colorless and clear; if an old hemorrhage is the cause, the supernatant is xanthochromic. If the CSF stands for more than 60–90 minutes after collection, or is centrifuged at higher speed, the erythrocytes lyse and the hemoglobin derivatives mimic an old hemorrhage.
CSF may be xanthochromic during the first weeks of life if there is marked hyperbilirubinemia; also when the protein content is very high.
Normal CSF
Lymphocytes and Monocytes
Normal lumbar CSF contains two cell types, lymphocytes and monocytes, in a ratio of 7:3. If more than 4–5 mL CSF is collected and the cells are allowed to differentiate further in stored aliquots, the proportion of monocytes becomes higher. The CSF occasionally contains solitary granulocytes and, more frequently, a few fresh erythrocytes artificially introduced during lumbar puncture.
Fig. 5.4 Lymphocyte and monocyte in CSF.
Lymphocytes. Lymphocytes circulating in the CSF originate partly from the blood, into which they can migrate back, and partly from the arachnoid tissue. They are predominantly T lymphocytes; B cells amount to only 1–2%. Lymphocytes occurring in normal CSF are small and isomorphic; they have a compact round or slightly oval nucleus and a narrow rim of pale cytoplasm (Fig. 5.3). Mild lymphocytic transformation may occur even in normal CSF, with enlargement of the nucleus and cytoplasmic rim and increased basophilia of the cytoplasm.
Monocytes. Monocytes have either migrated into the CSF from the blood or stem from the leptomeninges or microglia. As mononuclear phagocytic leukocytes, they are slightly larger than lymphocytes (Fig. 5.4). They have an eccentric kidney- or horseshoe-shaped nucleus and a wider, more intensely stained cytoplasm which often contains vacuoles. Monocytes are unstable in vitro and degenerate rapidly.
In normal newborns, cell counts are often elevated until the 4th to 6th week of life, up to 10 cells/μL—in very premature newborns even up to 40 cells/μL. Cell counts above 40/μL after the 6th week are pathologic. The granulocyte fraction may be up to 50%. The monocyte fraction is often increased as well, with some of the monocytes showing considerable signs of activation. The erythrocyte count can also be markedly elevated as a result of birth-related manipulation or difficulties in the lumbar puncture. Erythrocyte counts of up to 1000/μL in mature newborns, and up to 20 000/μL in very premature newborns, are still considered normal (Hobusch, 2003).
Other Cells
Occasionally, cartilage cells get into the CSF as a result of lumbar puncture. They have a strongly acidophilic, cloudy cytoplasm, and a darkly stained nucleus (Fig. 5.5). Normal CSF may contain epithelial cells, and these are observed even more frequently after repeated punctures. They stem from the choroid plexus or the ependyma and are usually present in clusters. These cells are large with slightly basophilic or acidophilic cytoplasm, and the nuclei are usually marginal, round, pyknotic, and isomorphic.
Fig. 5.5 Cartilage cells in CSF.
Fig. 5.6 Phagocytes in a patient undergoing intrathecal chemotherapy.
Fig. 5.7 Signet-ring cell in a patient with breast cancer.
Fig. 5.8 Hemosiderophage with erythrocytes in a patient with subarachnoid hemorrhage.
Occasionally the bone marrow is injured during lumbar puncture, particularly in elderly patients and in infants. The cytological preparation then contains all stages of myelopoesis and erythropoesis (solitary or in cluster arrangements) and also plasma cells:
• Pronormoblasts (proerythroblasts) have a strongly basophilic cytoplasm and a large, round nucleus with nucleoli.
• Normoblasts (erythroblasts) show a small, coarse, and dark nucleus.
• Myeloblasts and promyelocytes, which are the largest cells; they have several prominent nucleoli which cease to be detectable during later maturation.
Phagocytes develop from the mononuclear phagocytic system; they remove foreign particles (cells, bacteria, viruses, fungi, pigments, lipids, and intrathecally administered substances). The phagocytic capability of these cells is extremely high; the phagocytosed material is deposited in the cytoplasm (Fig. 5.6) and often appears as an optically empty vacuole (signet ring cell, Fig. 5.7) or as numerous fine droplets, or vacuoles, as in lipophages.
Bloody CSF
Hemorrhage into the subarachnoid space triggers a violent leptomeningeal reaction leading to extensive pleocytosis of all cells of the hematopoietic system with transient predominance of granulocytes. After 4–6 hours, the mononuclear –phagocytic system is activated and the erythrocytes are phagocytosed (erythrophages). After 3–4 days, the hemoglobin derivatives are further degraded and dark-staining hemosiderin forms in the cells (hemosiderophages, Fig. 5.8). After 3–4 weeks, bright yellow hematoidin crystals develop that may be intra- or extracellular (Fig. 5.9).
Fig. 5.9 Hematoidin crystals in a patient with subarachnoid hemorrhage.
Pitfalls with Erythrophages in the Differential Cell Count
Even ex vivo, erythrocytes introduced by lumbar puncture and undergoing lysis after about 2 hours may still activate monocytes and initiate erythrocyte phagocytosis. Hence, the detection of one or two erythrophages in the differential cell count is not conclusive evidence of disease-related hemorrhage into the subarachnoid space.
Inflammation
Granulocytes
Granulocytes do not occur in normal CSF. They appear when the permeability of vascular walls in the subarachnoid space increases and toxins exert a chemotactic effect. The migration rate of granulocytes is then very high, and the CSF may be swamped with granulocytes within a few hours. The metabolism of granulocytes is based on anaerobic glycolysis; hence, granulocyte invasion is followed by elevated lactate levels.
Neutrophilic granulocytes. The nuclei of neutrophilic granulocytes are usually polymorphic, with several lobes strung together (Fig. 5.10); young granulocytes with rod-shaped nuclei are quite rare in CSF. The half-life of granulocytes is only a few hours in vivo and in vitro; it is therefore essential to determine the cell count immediately after CSF collection and to prepare a cell sediment. Neutrophilic granulocytes can invade the cerebrospinal space after any acute irritation, e. g., epileptic seizure, ischemic stroke, hemorrhage, and inflammation. They are therefore primarily a nonspecific sign of inflammation. The quantity and duration of granulocytes detected in the CSF permit conclusions on their genesis:
• Neutrophilic granulocytes are the real cells of acute defense against bacteria, and in bacterial infections can reach levels of several 10 000 cells/μL within a few hours. Even after successful treatment of bacterial inflammation they can still be detected for several more days (Fig. 5.11), but fall below 5% by about the 20th day of illness.
• In viral meningitis, granulocytes predominate during the first 3 days of illness, though rarely amount to more than 700/μL; after the 10th day of illness they are completely replaced by mononuclear cells (Fig. 5.12).
• In tuberculous meningitis and Listeria meningitis, granulocytes are still found after 30–40 days of treatment.
Fig. 5.10 Neutrophilic granulocytes.
Eosinophilic granulocytes. Eosinophilic granulocytes in CSF do not differ morphologically from those in the blood. Their nuclei have often the characteristic “spectacles” shape, but may have three segments. Typical are the eosinophilic granules in the cytoplasm, which are very resistant: even after cell lysis, isolated granules can still be detected in the CSF sediment. Small numbers of eosinophilic granulocytes appear during the healing phase of every inflammatory reaction (Fig. 5.13); they play an important role in allergic-hyperergic reactions and are frequently observed as a foreign body reaction after shunt operations. When eosinophilic granulocytes in the CSF persist over weeks and months, this is frequently due to zoonosis of the CNS (cysticercosis, hydatid disease, or other parasitic worm infestations). An eosinophilic tissue reaction detectable in the CSF is also observed with malignant tumors, particularly malignant lymphoma. The pathogenetic role of this eosinophilia is unknown and its prognostic relevance is the subject of much controversy (von Wasielewski et al., 2000; Axdorph et al., 2001; Dorta et al., 2002).
Basophilic granulocytes. Basophilic granulocytes are similar in size to neutrophilic and eosinophilic granulocytes, but their cytoplasm is packed with basophilic granules. Relatively little is known about the function of these cells; they appear during inflammation and interact with other inflammatory cells.
Fig. 5.11 Bacterial meningitis in the subsiding exudation phase: individual mononuclear cells.
Fig. 5.12 Viral meningitis: mononuclear pleocytosis, lymphocytic transformation, and a few granulocytes.
Fig. 5.13 Cryptococcal meningitis: neutrophilic, eosinophilic, and basophilic granulocytes.
Fig. 5.14 Transformed lymphocytes in neuroborreliosis.
Fig. 5.15 Plasma cells in the proliferative phase of bacterial meningitis.
Transformed Lymphocytes and Plasma Cells
Stimulated lymphocytes and plasma cells are characterized by an increase in the size of the cells, including the nuclei, and initially pale cytoplasm that becomes basophilic as the cells mature (Fig. 5.14). The nuclei may be slightly polymorphic and lie eccentrically, often surrounded by a bright perinuclear area. Plasma cells found in CSF have probably developed from small lymphocytes of the B cell system that have migrated from the blood into the CSF. The mature plasma cell has a round, eccentric nucleus with coarse chromatin, occasionally exhibiting a typical spoke structure and a bright perinuclear area (Fig. 5.15). Plasma cells are often binuclear (Fig. 5.16).
Even normal CSF may contain a small number of transformed lymphocytes and solitary plasma cells. Apart from this, their presence can also be the result of a monomorphic cellular reaction to a variety of different stimuli. They can appear in the course of any acute, subacute, or chronic inflammation, whether caused by pathogens or of autoimmune origin; they can be a nonspecific reaction to hemorrhage, intrathecal treatment, or surgery of the CNS; or they may be an accompanying reaction to tumor cell infiltration.
Fig. 5.16 Binuclear plasma cell.
Fig. 5.17 Typical mitotic figure (arrow) and transformed lymphocytes in neuroborreliosis.
Fig. 5.18 a, b Mitotic figures in meningitis and malignant melanoma.
a Typical metaphase (arrow) in a patient with meningitis.
b Dividing cell in anaphase (arrow) in a patient with meningeal metastases of malignant melanoma.
Following the acute exudative phase of bacterial meningitis, mononuclear cells with prominent lymphocytic transformations usually predominate during the proliferative and reparative phases; in viral meningitis, this is already seen during the first days of the illness. In neuroborreliosis (Fig. 5.17), the presence of marked lymphocytic transformation and plasma cells is striking, as is also the case in Epstein-Barr virus infection, which induces pronounced B cell activity.
Tumor Cells
For tumor cell diagnosis in the CSF, the generally established criteria of malignancy apply. There are often signs of increased cell division (Fig. 5.18 a, b). However, the absence of mitotic cells does not exclude the diagnosis of tumor cells. Malignant cells occur solitarily or in clusters, and the cell size may vary considerably. The nucleocytoplasmic ratio has shifted toward the nucleus. The nuclei are often highly polymorphic with more prominent and more numerous nucleoli and irregular, loose chromatin structure (Fig. 5.19). Nucleus and cytoplasm stain intensively. The cytoplasmic rim is irregular and often pseudopodia-like. It is not usually possible to assign tumor cells to a particular organ on the basis of lumbar puncture cytology. Cytoplasmic inclusions are more common with tumors of epithelial origin, whereas cells of mesodermal origin have a more homogeneously stained cytoplasm. But these criteria are too vague to serve as the sole basis for diagnosis.
Fig. 5.19 Leptomeningeal metastases of breast cancer: polymorphic tumor cells (arrows) in CSF.
Special Features of Ventricular CSF
Ventricular CSF and lumbar CSF differ more in their protein compositions than in their cell populations.
Catheter Irritation. In patients with indwelling ventricular catheters, the CSF often contains eosinophilic granulocytes due to irritation and phagocytic reactions or to intrathecal treatment. Tissue fragments and giant cells are often seen, particularly when the ventricular catheter is not completely patent or adheres to the ventricular wall. Irritation-induced granulocytic pleocytosis after surgical intervention, or due to material intolerance, is often much more pronounced in ventricular CSF than in lumbar CSF and can attain 300–400 cells/μL (Kluge and Kalff, 2003).
Meningitis. In pathogen-induced meningitis, the pleocytosis is usually markedly less pronounced in the ventricular CSF, since the turnover of ventricular CSF is much higher. However, the granulocytic phase usually persists longer in the ventricular CSF.
Meningeal carcinoma. In meningeal blastoma or carcinoma, the ventricular CSF is not representative of the total CSF findings. There are markedly more tumor cells in the lumbar than in the ventricular CSF, so treatment cannot be monitored on the basis of the ventricular reservoir alone.
References
Aune MW, Becker JL, Brugnara C, et al. Automated flow cytometric analysis of blood cells in cerebrospinal fluid: analytic performance. Am J Clin Pathol 2004;121:690–700
Axdorph U, Porwit-MacDonald A, Grimfors G, Björkholm M. Tissue eosinophilia in relation to immunopathological and clinical characteristics in Hodgkin's disease. Leuk Lymphoma 2001;42:1055–1065
Chamberlain MC, Kormanik PA, Glantz MJ. A comparison between ventricular and lumbar cerebrospinal fluid cytology in adult patients with leptomeningeal metastases. Neuro Oncol 2001;3:42–45
Dorta RG, Landman G, Kowalski LP, et al. Tumour-associated tissue eosinophilia as a prognostic factor in oral squamous cell carcinomas. Histopathology 2002;41:152–157
Ellis MJ, Kazina CJ, Del Bigio MR, McDonald PJ. Treatment of recurrent ventriculoperitoneal shunt failure associated with persistent cerebrospinal fluid eosinophilia and latex allergy by use of an “extracted” shunt. J Neurosurg Pediatr 2008;1:237–239
Fulkerson DH, Boaz JC. Cerebrospinal fluid eosinophilia in children with ventricular shunts. J Neurosurg Pediatr 2008;1:88–295
Hobusch D. Besonderheiten der Liquordiagnostik im Kindesalter. In: Zettl UK, Lehmitz R, Mix E, eds. Klinische Liquordiagnostik. Berlin: de Gruyter; 2003
Hoffmann JJ, Janssen WC. Automated counting of cells in cerebrospinal fluid using the CellDyn-4000 haematology analyser. Clin Chem Lab Med 2002;40:1168–1173
Kluge H, Kalff R. Besonderheiten des Ventrikelliquors. In: Zettl UK, Lehmitz R, Mix E, eds. Klinische Liquordiagnostik. 2nd ed. Berlin: de Gruyter; 2005
Kölmel HW. Liquorzytologie. Berlin, Heidelberg: Springer; 1978
Mahieu S, Vertessen F, Van der Planken M. Evaluation of ADVIA 120 CSF assay (Bayer) vs. chamber counting of cerebrospinal fluid specimens. Clin Lab Haematol 2004;26:195–199
Strik H, Luthe H, Nagel I, et al. Automated cerebrospinal fluid cytology: limitations and reasonable applications. Anal Quant Cytol Histol 2005;27:167–173
Van Acker JT, Delanghe JR, Langlois MR, et al. Automated flow cytometric analysis of cerebrospinal fluid. Clin Chem 2001;47:556–560
von Wasielewski R, Seth S, Franklin J, et al. Tissue eosinophilia correlates strongly with poor prognosis in nodular sclerosing Hodgkin's disease, allowing for known prognostic factors. Blood 2000;95:1207–1213
Ziebig R, Lun A, Sinha P. Leukocyte counts in cerebrospinal fluid with the automated hematology analyzer CellDyn 3500 and the urine flow cytometer UF-100. Clin Chem 2000;46:242–247
Further Reading
Hohlfeld R, Schwendemann G, Schwartz A, et al. Typisierung von Liquorzellpopulationen mit monoklonalen Antikörpern. In: Kölmel HW, ed. Zytologie des Liquor cerebrospinalis. Weinheim: VCH; 1986
Kluge H, Wieczorek V, Linke E, Zimmermann K, Isenmann S, Witte OW (eds.) Atlas of CSF Cytology. Stuttgart, New York: Thieme; 2006
Kölmel HW. Liquorzytologie. In: Zettl UK, Lehmitz R, Mix E, eds. Klinische Liquordiagnostik. 2nd ed. Berlin: de Gruyter; 2005
Kranz BR, Thiel E, Thierfelder S. ZNS-Befall bei lymphohämopoetischen Neoplasien: Inzidenz und immunzytologischer Nachweis im Liquor. In: Kölmel HW, ed. Zytologie des Liquor cerebrospinalis. Weinheim: VCH; 1986
Wick M. Immunzytologie. In: Zettl UK, Lehmitz R, Mix E, eds. Klinische Liquordiagnostik. 2nd ed. Berlin: de Gruyter; 2005