Integrated Cytology of Cerebrospinal Fluid
Integrated Cytology of Cerebrospinal Fluid
Integrated Cytology of Cerebrospinal Fluid
Integrated Cytology
of Cerebrospinal Fluid
With 138 Figures
123
Michael Torzewski, MD Karl J. Lackner, MD
Institute of Clinical Chemistry and Institute of Clinical Chemistry
Laboratory Medicine and Department and Laboratory Medicine
of Neuropathology University of Mainz
University of Mainz 55101 Mainz
55101 Mainz Germany
Germany lackner@zentrallabor.klinik.uni-mainz.de
torzewski@zentrallabor.klinik.uni-mainz.de
DOI 10.1007/978-3-540-75885-3
This work is subject to copyright. All rights are reserved, wether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broad-
casting, reproduction on microfilm or any other way, and storage in data banks. Duplication of this
publication or parts thereof is permitted only under the provisions of the German Copyright Law
of September 9, 1965, in it current version, and permission for use must always be obtained from
Springer. Violations are liable to prosecution under the German Copyright Law.
springer.com
The use of general descriptive names, registered names, trademarks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant pro-
tective laws and regulations and therefore free for general use.
Product liability: the publishers cannot guarantee the accuracy of any information about dosage and
application contained in this book. In every individual case the user must check such information by
consulting the relevant literature.
543210
Preface
Histologic specimens are usually stained with hematoxylin and eosin (H&E) and
shown using a × 40 objective.
The authors would be glad to receive suggestions as to how the book might
be further improved. In this regard, submissions of additional illustrations or of
cytological preparations showing uncommon but diagnostically relevant find-
ings would be particularly helpful. We thank Springer-Verlag GmbH, Heidelberg,
Germany, and in particular Ellen Blasig and Gabriele Schröder for their extensive
help and advice during the preparation of this book. We are grateful to Andreas
Kreft, MD, Institute of Pathology, University of Mainz, for the critical reading of
Sect. 6.3. We also thank the technical staff of the Institute of Clinical Chemistry
and Laboratory Medicine and the Department of Neuropathology, University of
Mainz, for the preparation of the cytological and histologic specimens.
2 Common Artifacts
2.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Autolysis, High Cell Concentration, Unproper Loading, Adhesion . . . . 6
4 Inflammatory Conditions
4.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.2 Bacterial Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.3 Viral Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3.1 Viral Meningitis and Meningoencephalitis . . . . . . . . . . . . . . . . . . . . . . . 28
4.3.2 Rabies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.4 Fungal Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.5 Parasites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.5.1 Toxoplasma gondii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
VIII Contents
4.5.2 Acanthamoeba . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.5.3 Leishmaniasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5 Non-neoplastic Disorders
5.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.2 Infarction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.3 Hemorrhage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.4 Demyelinating Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.4.1 Multiple Sclerosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.4.2 Acute Disseminated Encephalomyelitis . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6 Neoplastic Disorders
6.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.2 Metastatic Tumors and Meningeal Carcinomatosis . . . . . . . . . . . . . . . . 54
6.2.1 Breast Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
6.2.2 Lung Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
6.2.3 Gastric Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
6.2.4 Malignant Melanoma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
6.3 Hematologic Neoplasms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
6.3.1 Leukemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
6.3.2 Lymphoma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.4 Primary Central Nervous System Neoplasms . . . . . . . . . . . . . . . . . . . . . 66
6.4.1 Astrocytic Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.4.2 Ependymal Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.4.3 Embryonal Tumors/Central Nervous System Primitive
Neuroectodermal Tumors (WHO Grade IV) . . . . . . . . . . . . . . . . . . . . . 72
6.4.4 Germ Cell Tumors; Germinoma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.4.5 Choroid Plexus Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
7 Contaminants
7.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.2 Bacterial and Fungal Contamination . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
7.3 Pollen and Starch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Chapter 1 Cerebrospinal Fluid
Cell Preparation
Since the cellular component of CSF obtained from the lumbar area is generally
scant, an efficient method of concentrating this material is necessary. Further-
more, considerations regarding the selection of cytopreparation techniques in-
clude the potential for cell loss, the clarity of cellular detail, and the spectrum of
stains offered. The most commonly utilized methods today are membrane filtra-
tion and cytocentrifugation.
Chapter 1
1 1.2 Cytocentrifugation
Of all the possible methods of transferring cells from CSF samples onto slides,
most laboratories now use a cytospin apparatus, such as the Shandon cytocen-
trifuge (Fig. 1A, reprinted by kind permission of Thermo Shandon), which is
efficient in terms of cell yield. The cytocentrifuge technique also allows use of
virtually all types of fixation and staining, including the preparation of cells for
immunocytochemistry, immunofluorescence and in situ hybridization. Because
of the centrifugal force when in the running position, the cytofunnel will raise
up in vertical position (right-hand side of the illustration). When the cytospin is
not running the cytofunnel is loaded at an angle to prevent the specimen com-
ing into contact with the filter card (left-hand side of the illustration; blue color,
specimen). To load the slide clip with a re-useable sample chamber and filter card
(Fig. 1B, reprinted by kind permission of Thermo Shandon) it is necessary to fit
the glass slide (1), to fit the filter card (2), to fit the re-useable sample chamber (3),
and finally to pull up the spring and press it into the two retaining hooks to hold
the chamber in place (4). After running the cytospin, cytospin sample chambers
are unloaded and samples are fixed as soon as possible to avoid autolysis. The
specimen can now be stained and examined microscopically.
Cerebrospinal Fluid Cell Preparation
Chapter 2 Common Artifacts
Several preanalytical and analytical pitfalls may cause artifacts, making proper
assessment of the CSF cell preparation impossible or at least more difficult. First
of all, it is important to ensure that CSF samples reach the laboratory as soon as
possible after the puncture (within 2 h). If this is not possible, fixation of the CSF
sample with buffered formalin (1:1) is an option, but should not be routinely per-
formed. The storage temperature of native CSF should be between 5°C and 12°C to
minimize cell damage. Lower temperatures may lead to cold lysis, whereas higher
temperatures accelerate catabolic mechanisms [1].
Chapter 2
The subarachnoid space is a fluid-filled cavity that covers the brain and spinal cord
and communicates with the internal ventricles of the brain. It is bound externally
by a fine arachnoid membrane and internally by the pial surface of the brain, and
normally contains only a loose fibrovascular stroma. The arachnoid membrane
and pia mater are the two membranous layers that comprise the leptomeninges.
CSF is produced within the ventricular system by cells of the choroid plexus and
ependyma [2], and passes out of the ventricular system at the base of the brain
into the subarachnoid space by means of the foramina of Luschka and Magendie.
A specimen of CSF taken in the lumbar area and unaltered by diseases of the
central nervous system normally contains two types of cells, i.e., lymphocytic and
monocytic forms. The ratio of lymphocytes to monocytes is about 70 to 30 [3, 4].
Since the puncture needle penetrates the skin, adipose and fibrous tissue, and stri-
ated muscle, and since the cartilage and bones of the vertebral column may be also
in the path of the needle, non-neoplastic cellular elements, such as squamous cells,
adipose and fibrous tissue, and striated muscle, as well as chondrocytes and cells
of the hemopoietic system, are sometimes carried in the specimen. A specimen
taken from the ventricular area may also contain fragments of cerebral cortex,
white matter, and ependymal cells, as the needle traverses all of these structures.
10 Chapter 3
3
A specimen of CSF taken in the lumbar area and unaltered by diseases of the cen-
tral nervous system normally contains two types of cells: lymphocytic and mono-
cytic forms. The ratio of lymphocytes to monocytes is about 70 to 30 [3, 4]. There-
fore, small numbers of lymphocytes are seen in most CSF samples. These are small
(5–8 µm in diameter), relatively isomorphic cells. Usually, the nuclear chromatin
is dense and homogeneous (Fig. 3.2A). Occasionally, they show a discrete but ill-
defined paler structure within the nucleus, which is the nucleolus (Fig. 3.2B, ar-
rowhead). There is little cytoplasm, which is pale blue and clear (Fig. 3.2A,B, 1), or
sometimes slightly more intense (Fig. 3.2B, 2).
Immunophenotyping of lymphocytes by flow cytometry revealed that in nor-
mal CSF, CD3+ T lymphocytes constitute the vast majority of CSF lymphocytes
(about 93%), while the numbers of B lymphocytes and natural killer (NK) cells
are low [5].
Activated lymphocytic cell forms are part of the differentiation of normal B lym-
phocytes from plasma cells under inflammatory conditions. Compared with non-
activated lymphocytic cell forms, these are larger (up to 25 µm in diameter) or
dark cytoplasmic forms with a more pronounced cytoplasmic border (Fig. 3.2C,
1; D, 1) and/or more lightly stained and somewhat heterogeneous nuclear chro-
matin (Fig. 3.2C, 2). These cytologic features may sometimes lead to confusion
with the malignant hematopoietic cells of lymphoma or leukemia (see Sect. 6.3).
Immunophenotyping of these reactive cells is particularly helpful, because mix-
tures of reactive T and B lymphocytes are inevitably demonstrated, in contrast to
the monoclonal population that is present in the case of malignant lymphoma.
Plasma cells (Fig. 3.2C, 3) are never present in normal CSF, i.e., their presence al-
ways indicates an inflammation of the central nervous system (CNS). The nucleus
of mature plasma cells is excentrically located and contains a grainy chromatin
structure. A crescent-shaped area of light cytoplasm around the nucleus is typical,
but not present in every case. Due to enhanced proliferation, sometimes mitoses
(Fig. 3.2D, 2) or binuclear cells can be observed.
The Common Cell Types of Cerebrospinal Fluid 11
12 Chapter 3
The monocytic cell forms normally degenerate more rapidly in vitro than do the
lymphocytic cell forms. Monocytic cell forms have a hematogenic (monocyto-
3
poietic) origin [6]. The cells are large (15–20 µm in diameter) and have an ec-
centric, oval, kidney-shaped or horseshoe-shaped, blue–gray nucleus, which of-
ten contains large, pale nucleoli (Fig. 3.3A,B, arrowhead). The cytoplasm is pale
blue–gray, occasionally interspersed with vacuoles. Larger and/or numerous cyto-
plasmic vacuoles are already a sign of a higher activation state (Fig. 3.3C,D). Fur-
thermore, compared with non-activated monocytic cell forms, activated mono-
cytes are larger and their nucleus sometimes more rounded (Fig. 3.3C, D). Also in
Fig. 3.3C, 1 and D, 1, note the presence of erythrocytes.
The Common Cell Types of Cerebrospinal Fluid 13
14 Chapter 3
3.4 Macrophages
Cells from the choroid plexus (Fig. 3.5B: histologic specimen) and the cuboidal or
columnar ependymal cells that line the ventricles (Fig. 3.5D: histologic specimen,
3
oil immersion) are normal brain elements that have ready access to CSF. They
occasionally occur singularly or in papillary clusters in CSF obtained by lumbar
puncture from normal individuals. Cytologic differentiation between cells of the
ependyma and those of the choroid plexus is hardly possible. Plexus epithelial
cells usually appear as clusters of several cells with wide, coarse-grained cytoplasm
and uniform, round to slightly oval nuclei (Fig. 3.5A). Ependymal cells are rather
fragile cells with round, often pyknotic-appearing nuclei and a wide border of
pink or blue–gray cytoplasm, occasionally interspersed with vacuoles (Fig. 3.5C,
surrounded by many erythrocytes). Both cell types appear more frequently in
the CSF of small children and in cases of hydrocephalus, as well as in samples
obtained by ventricular puncture. Ependymal and choroidal cells also appear in
the CSF sample after intrathecal administration of drugs, especially of cytostatics.
They are generally of little diagnostic value. However, these cells may be morpho-
logically changed in reaction to an appropriate stimulus, and may therefore be a
source of error in diagnostic material.
The Common Cell Types of Cerebrospinal Fluid 17
18 Chapter 3
3.7.1 Chondrocytes
3
Cells that originate in the path of the aspiration needle are sometimes found in a
specimen. A frequent finding of lumbar puncture-derived CSF is the chondrocyte.
These cells enter the CSF as a result of an injury to the cartilage of a vertebra caused
by the puncture needle. Usually, they show a typical appearance with an intensely
red cytoplasm and uniform, round to slightly oval nuclei (Fig. 3.7.1A). Since these
cells originate from bradytrophic tissue, they are rather stable, even if all other
cell types are already autolytic due to inappropriate processing of the sample. In
CSF cell preparations obtained from patients with degenerative changes of the
cartilage (Fig. 3.7.1D: histologic specimen with chondrocyte “cloning” (1) and
granular change as evidence of degeneration and/or prolapse, objective × 20), they
sometimes appear in clusters (Fig. 3.7.1B) or even may be mistaken for neoplas-
tic cells (Fig. 3.7.1C). The presence of these cells in CSF, however, has no known
pathologic significance.
The Common Cell Types of Cerebrospinal Fluid 21
22 Chapter 3
atypical lymphoid or lymphoblastic cells in the CSF, perhaps making the differen-
tial diagnosis from a neoplastic process of the lymphatic system very difficult.
Intracellular inclusions, characteristic of a viral infection, are only rarely found
in the CSF. Definite identification of an infectious disease of the CNS must take
into consideration the results of all the different methods of clinical and labora-
tory diagnostics.
Many different bacteria are able to induce severe acute or subacute leptomeningi-
tis. Among the microorganisms that cause purulent phlegmonous inflammation
of the arachnoid are meningococci, streptococci, Haemophilus influenzae type b,
and others.
The laboratory investigation of the CSF reveals heavily increased cell numbers
(usually >1,000 leukocytes/µl). At the beginning, most of the inflammatory cells
are neutrophilic leukocytes. Very often they contain many microorganisms in
their cytoplasm.
Figure 4.2A shows typical diplococci in the leukocytic cytoplasm (arrowheads;
1, activated lymphocyte), sometimes even small packages of four bacteria; the
meningitis was caused by Neisseria meningitidis. In the later stages of bacterial
leptomeningitis the cellular population exhibits more and more activated mono-
cytes and macrophages, as shown in Fig. 4.2B and C. Figure 4.2B shows a macro-
phage (1) containing coagulase-negative staphylococci (arrowheads). In Fig. 4.2C,
large amounts of small lanceolate diplococci can be seen, sometimes surrounded
by a small unstained halo, corresponding to a capsule (arrowheads). These are
typical signs of a pneumococcal infection, caused by Streptococcus pneumoniae.
These bacteria prove to be Gram-positive (stained dark blue) after Gram staining
(Fig. 4.2D).
Exact identification of bacteria is only possible by microbiological investiga-
tion, which should include determination of resistance.
Inflammatory Conditions 27
28 Chapter 4
Acute viral infections of the CNS usually involve both the central nervous tissue
and the leptomeninges, e.g., infections with the herpes simplex virus, Coxsackie
4
virus or influenza virus.
At the beginning of the necrotizing process in the cerebral cortical areas and/
or in the brainstem nuclei there might be on the first day mainly polymorpho-
nuclear leukocytes in the CSF as a consequence of a severe necrotizing damage
of the brain tissue. After that, lymphocytic cells will predominate in the CSF
(Fig. 4.3.1A,B). In addition to normal lymphocytes (Fig. 4.3.1A, 1), there are ac-
tivated cells (Fig. 4.3.1A, 2), lymphoblastic elements (Fig. 4.3.1A, 3), and some-
times even some mitoses. Later on, the cytopathologic picture additionally shows
plasma cells (B, 1), activated monocytes (B, 2), and macrophages.
Figure 4.3.1C shows the macroscopic appearance of acute severe herpes sim-
plex encephalitis with hemorrhagic necrotizing alterations, mainly in the tempo-
ral lobe. The damage is not always totally symmetrical; in this case, the left side is
more severely involved.
In the brain tissue the inflammatory infiltrations are found around small blood
vessels and cuffing the capillaries and small veins (Fig. 4.3.1D, objective ×10). As
there is continuity between the perivascular Virchow–Robin spaces and the sub-
arachnoid compartment of the CSF, the inflammatory cells can easily move from
inside the brain tissue into the surrounding leptomeninges and, furthermore, into
the spinal channel.
Very often, acute generalized viral infections of the whole organism are ac-
companied by slight concomitant leptomeningitis with disseminated lymphocytic
infiltrations in the subarachnoid space and also in the Virchow–Robin spaces of
the cerebral parenchyma.
Inflammatory Conditions 29
30 Chapter 4
4.3.2 Rabies
4.5 Parasites
The activation of a latent infection with the protozoon Toxoplasma gondii (the
term “gondii” seems to be inappropriate; the germ was first found in small labora-
4
tory animals in the north of Africa, in a rodent named “gundi” and not “gondi”) is
a very common finding in immunosuppressed patients with severe malignancies
or in patients with HIV infections.
Figure 4.5.1A shows several tachyzoites of Toxoplasma gondii. Surrounded by
erythrocytes (1), supposedly artificial contamination, the protozoal microorgan-
isms (2) fill the cytoplasm of a phagocytic cell (3) and float freely in the CSF as
well (2). In Fig. 4.5.1B, the infectious agents are stained with a specific polyclonal
antibody. They are found as single tiny organisms (1) or as a pseudocystic accu-
mulation of tachyzoites (2).
These characteristic pseudocystic toxoplasmic bodies can also be found in his-
tologic specimens in many regions of the central nervous system (Fig. 4.5.1C, 1);
they represent resting forms of the infectious agent and may release the tachyzo-
ites during an acute reactivation of the disease. In areas of necrotizing encepha-
litic alterations the tissue contains many toxoplasmic protozoa, detected by im-
munohistochemistry (Fig. 4.5.1D, arrowheads, objective × 10). There is nearly no
inflammatory reaction as a consequence of the underlying immunosuppressed or
immunodeficient state of the patient.
Inflammatory Conditions 35
36 Chapter 4
4.5.2 Acanthamoeba
4.5.3 Leishmaniasis
Diseases caused by protozoa are very common all over the world, e.g., American
and African trypanosomiasis, cutaneous and visceral leishmaniasis, malaria and
toxoplasmosis, and furthermore pneumocystosis, babesiosis, amebiasis, balanti-
diasis, giardiasis, coccidiosis, and microsporidiosis.
It is surely a rare event that protozoa are found in the CSF obtained by lumbar
4
puncture; but sometimes they can be found in a sanguinolent sample of the CSF.
Figure 4.5.3A–D shows a specimen of a patient with Leishmaniasis (visceral form:
Kala-Azar, which is Hindi and means “black fever”). While entering the human
body the organisms in their promastigote form usually turn over into the smaller
amastigote type.
Figure 4.5.3A shows many intracellular (1) and also some extracellular (2)
leishmaniasis. The organisms appear larger in cytologic specimens than in tissue
specimens. Nuclei and kinetoplasts are distinct in these forms of leishmaniasis.
There is also one promastigote organism (3).
Figure 4.5.3B–D shows single forms of extracellular leishmaniasis (Fig. 4.5.3B, 1)
amidst damaged erythrocytes (Fig. 4.5.3C, 1) and single lymphocytes (Fig. 4.5.3B,
2; D, 1).
As a consequence of the increase in world-wide travel and climate change,
a higher frequency of tropical diseases will be probably be observed in the future,
even in temperate zones.
Inflammatory Conditions 39
Chapter 5 Non-neoplastic
Disorders
5.2 Infarction
5.3 Hemorrhage
Cerebral spinal fluid from patients suffering from multiple sclerosis (MS) is
characterized by mild lymphocytic pleocytosis. Cell counts rarely exceed 50/µl
and may be within the normal range, i.e., < 5/µl. The presence of granulocytes
should alert the cytologist to other potential diagnoses, e.g., Behçet’s disease with
CNS manifestations. Immunoglobulin-synthesizing B-lymphocytes are common
(Fig. 5.4.1A). They are indicative of the underlying inflammatory process, and are
rarely found in normal CSF. These cells show an excentric nucleus, and an in-
creased cytoplasm/nuclear ratio compared with non-activated, resting lympho-
cytes (inset, see also Sect. 3.2). Storage of immunoglobulins is indicated by a more
pronounced and basophilic cytoplasmic border (arrowheads). Even though im-
munoglobulins can also be shown by immunocytochemistry, this is usually not
needed. Besides activated B-lymphocytes, mature plasma cells are also found in
MS (Fig. 5.4.1B). They are characterized by more intense blue staining and are
slightly larger than activated B-cells.
A hallmark of inflammatory CNS diseases, including MS, is the intrathecal syn-
thesis of immunoglobulins, mostly IgG. This is reflected by a disproportionate in-
crease in the CSF/serum ratio of IgG as seen in the quotient diagram (Fig. 5.4.1C)
as well as the presence of oligoclonal IgG in CSF (Fig. 5.4.1D). These changes may
be accompanied by a mild or less common moderate disturbance of the blood–
CSF barrier (Fig. 5.4.1C). Again, a severe disturbance of the blood–CSF barrier
(CSF/serum ratio of albumin > 20 × 10-3) should raise the possibility of a different
diagnosis.
Non-neoplastic Disorders 47
48 Chapter 5
If unequivocally malignant cells are present, a few such atypical cells are sufficient
6
for the correct diagnosis of neoplastic meningitis. However, extreme caution is
advisable when analyzing CSF slides from infants and young children where lep-
tomeningeal cell clusters with a high proliferation rate may be present; these are
harmless and no indicator of a neoplasm [9]. Furthermore, blast-like tumor cell
clusters, thought to be of germinal matrix origin, are sometimes found in CSF of
neonates and young infants, particularly in connection with a history of prematu-
rity, hydrocephalus, or birth trauma, and must not be misinterpreted as neoplastic
meningitis [10–12].
If there are only cells with suspected neoplasia, immunocytochemistry will be
very helpful – provided that the neoplastic cells have maintained their expression
characteristics. Unfortunately, however, the more malignant and dedifferentiated
a tumor, the stronger the likelihood of loss of expression of specific marker pro-
teins. Additional techniques, which in specific situations may substantially im-
prove information regarding the classification “neoplastic or non-neoplastic,” are
immunophenotyping by flow cytometry, in situ hybridization [13], or the use of
polymerase chain reaction. In the latter case, however, the genetic alteration of
the primary neoplasia must be known. These techniques may be particularly valu-
able in diagnosing neoplastic cells in chronic forms of leukemia where differen-
tiation from lymphocytic infiltration is extremely difficult and sometimes even
impossible.
Another significant problem in practice is how to deal with a negative cytol-
ogy, i.e., the absence of malignant cells or those with suspected neoplasia, de-
spite strong clinical or neuroradiological evidence. While some authors claim that
nearly every occurrence of neoplastic meningitis is detectable [9], it is reason-
able to assume that, due to the multifocal nature of neoplastic meningitis, CSF
obtained from sites distant from the pathological process may yield a negative
Neoplastic Disorders 53
cytology [14, 15]. Data from the literature suggest that up to 45% of cases will
be cytologically negative at first examination, but up to 90% will be identified by
means of a second CSF examination. Little benefit, however, will derive from ad-
ditional repetitive punctures [16]. In general, malignant cells will appear in CSF
most commonly when there is generalized seeding of the leptomeninges by tu-
mor cells, less often when there is only focal seeding and almost never when the
tumor is limited to the brain and the pial surface has not been breached [17]. If
neuroradiologically suspect leptomeningeal or, in particular, intraparenchymal le-
sions are present, either a leptomeningeal or a stereotactic brain biopsy should be
considered. Especially in the case of a suspected primary CNS lymphoma, with
the patient’s state of health being critical, no time should be wasted on repetitive
punctures for CSF analysis, aiming at a diagnosis.
Since cytology deals with single cells, one difficulty in arriving at the correct
diagnosis is lack of information concerning tumor architecture. Furthermore, due
to environmental differences, tumor cells frequently show secondary changes in the
CSF. Since, unfortunately, cytological diagnosis is frequently made by investigators
unfamiliar with the histopathology of tumors, we have added some figures with
typical histoarchitectural features, to give an impression of the primary tumor.
In summary, when neoplastic meningitis is suspected, the crucial question to
be answered is: “Neoplastic cells/suspected neoplasia cells or not?” Knowledge
consisting of clinical data, neuroradiologic findings, history, and location of the
tumor is frequently essential to arrive at a correct diagnosis. Clusters of cells in
CSF preparations from the ventricle may be confused with tumor cells when the
investigator takes it for a CSF sample from a lumbar puncture.
54 Chapter 6
Useful antibodies
• Anti-pan-cytokeratin (MNF-116) or antibodies against specific cytokeratins to
confirm the epithelial origin
• Anti-estrogen and/or progesterone receptor (if epithelial histogenesis has al-
ready been proven)
Neoplastic Disorders 55
56 Chapter 6
Useful antibodies
• Anti-pan-cytokeratin (MNF-116) or antibodies against specific cytokeratins
• Anti-TTF-1 (to confirm location of the primary tumor in the lung or thyroid)
• Anti-synaptophysin or other neuronal markers (as markers for SCLC)
Neoplastic Disorders 57
58 Chapter 6
Compared with carcinomas of the respiratory tract and breast, tumors of the gas-
trointestinal tract rarely lead to neoplastic meningitis (up to 0.25%) [18]. Some
specific entities show rather characteristic features in cytologic specimens. In
Fig. 6.2.3A, a cellular CSF specimen with densely packed atypical cells is shown.
One cell (Fig. 6.2.3A, 1) shows the typical appearance of a signet ring with nuclei
displaced to the border by mucous material in the center of the cell. Immunocyto-
chemistry with an antibody against pan-cytokeratin (MNF-116) labels numerous
cells and reveals multiple vacuoles within the cytoplasm (Fig. 6.2.3B, 1). In the
paraffin-embedded material, the typical histological aspect of a signet cell car-
6
cinoma is present (Fig. 6.2.3C), exhibiting strongly PAS-positive material in the
cytoplasm of the tumor cells (Fig. 6.2.3D).
Useful antibodies
• Anti-pan-cytokeratin (MNF-116) or antibodies against specific cytokeratins
Neoplastic Disorders 59
60 Chapter 6
Useful antibodies
• HMB-45 or anti-Melan A to confirm the melanotic nature
• Anti-S-100 as a helpful though unspecific marker
• Anti-pan-cytokeratin (MNF-116) to potentially identify an epithelial primary
6.3.1 Leukemia
eral blood count is essential. In chronic forms, the cytological criteria for ma-
lignancy may be less pronounced. Detailed information concerning the subtype
should be provided by the physician to facilitate correct diagnosis. CLL or CML
combined with pleocytosis is either due to additional inflammation or, if blasts
can unequivocally be identified, transformation into a high-grade neoplasm must
be assumed.
64 Chapter 6
6.3.2 Lymphoma
While, in principle, all astrocytic tumors and, in particular, the anaplastic vari-
ants, have the potential for leptomeningeal spread and dissemination via the ce-
rebrospinal pathways, CSF samples for primary cytologic diagnosis are of very
limited value. Unfortunately, in the rare cases of primary leptomeningeal gliomas,
6
where diagnosis from cytologic specimens would be helpful, detection of tumor
cells in CSF samples is a rare exception [27]. Similarly, reports on spinal low-grade
gliomas with extensive leptomeningeal dissemination indicate that CSF findings
are largely restricted to elevated protein, while neoplastic cells are not detectable
[28]. Furthermore, while pilocytic astrocytomas (WHO grade I) frequently in-
filtrate the leptomeningeal space (Fig. 6.4.1.1A, 1) histologic specimen, objective
×5), true dissemination via the CSF is largely restricted to pilocytic astrocytomas
of the hypothalamic/chiasmatic region in infants and young children. Due to its
characteristic morphological features and a less favourable prognosis this tumor
has been listed as a variant in the current WHO classification, the so-called pilo-
myxoid astrocytoma (WHO grade II) [29, 30]. The classic situation in which the
neuropathologist faces the pivotal question of presence or absence of neoplastic
astrocytic cells in CSF samples, concerns either a known astrocytic tumor or a
check for tumor cells in the CSF for control purposes.
Morphology of astrocytic tumor cells in CSF samples is variable and fulfills the
general criteria for malignant cells. Figure 6.4.1.1B is a typical example of rather
large cells with broad, sometimes vacuolated cytoplasm (Fig. 6.4.1.1B, 1). Nuclei
are mildly pleomorphic, hyperchromatic, and possess distinct nucleoli. The cor-
responding histologic specimens (Fig. 6.4.1.1C, D) show an astrocytic tumor that
has developed multiple subependymal protrusions from which dissemination into
the ventricular CSF is easy to imagine. Glial histogenesis is confirmed by glial
fibrillary acidic protein (GFAP) immunohistochemistry (Fig. 6.4.1.1D). Due to its
cell pleomorphism and increased proliferation index, this tumor has been classi-
fied as an anaplastic astrocytoma (WHO grade III).
Neoplastic Disorders 67
Useful antibodies
• Anti-GFAP (to clarify glial histogenesis of tumor cells, although GFAP immu-
noreactivity does not automatically indicate a neoplastic or astrocytic cell; ana-
plastic glial cells, however, not infrequently lose their GFAP expression)
• Anti-MAP2 (normally expressed in neuronal cells, MAP2 can also be used as a
marker for glial histogenesis of tumor cells/progenitors)
68 Chapter 6
Glioblastoma multiforme (GBM) is the most common and most malignant gli-
oma (WHO grade IV) in adults. Despite its highly infiltrative growth manner,
metastasis via cerebrospinal pathways is rather rare. GBM may show the high-
est degree of anaplasia and pleomorphism. In Fig. 6.4.1.2A, a highly anaplastic
cell (1) can be seen with a hyperchromatic nucleus containing numerous nu-
cleoli. Furthermore, multiple bizarre cytoplasmic protrusions are detectable. In
addition, some activated lymphocytes (2), monocytes (3), and erythrocytes (4)
can be seen. In Fig. 6.4.1.2B, a neoplastic cell with prominent nucleoli and ba-
sophilic, plurivacuolated cytoplasm is present (1). Apart from GFAP immunos-
6
taining, immunohistochemistry with an antibody against the microtubule-associ-
ated protein-2 (MAP2) may be a useful tool for the detection of neoplastic glial
cells (Fig. 6.4.1.2C, 1). Although primarily regarded as a neuronal marker protein,
MAP2 is now a recognized marker for glial progenitors [31]. The corresponding
histologic specimen shows an anaplastic tumor of high cell density with palisad-
ing necrosis (Fig. 6.4.1.2D, 1), fulfilling all histological criteria for a diagnosis of
GBM.
Useful antibodies
• Anti-GFAP (to clarify glial histogenesis of tumor cells, although GFAP immu-
noreactivity does not automatically indicate a neoplastic or astrocytic cell)
• Anti-MAP2 (normally expressed in neuronal cells, MAP2 can also be used as a
marker for glial histogenesis of tumor cells/progenitors)
Neoplastic Disorders 69
70 Chapter 6
Useful antibodies
• Anti-GFAP (typically, ependymoma cells demonstrate strong GFAP immuno
staining; however, as already stated for astrocytomas, GFAP immunoreactivity
does not automatically indicate a neoplastic or ependymal cell)
Neoplastic Disorders 71
72 Chapter 6
6.4.3.1 Medulloblastoma
6.4.3.2 Retinoblastoma
The retinoblastoma (which is not listed in the current WHO classification of tu-
mors of the central nervous system) is a retinal embryonal neoplasm that either
extends directly from an ocular lesion or presents as a primary intracerebral le-
sion in the pineal or suprasellar region as part of the “trilateral retinoblastoma”
complex. Advanced tumors may break through the optic nerve and enter the sub-
arachnoid space. The cytological preparation shows neoplastic cells that are mostly
indistinguishable from other embryonal neoplasms (Fig. 6.4.3.1/6.4.3.2C; cf. A).
The same holds true for the histologic specimens. Again, a highly cellular embryo-
nal tumor is present (Fig. 6.4.3.1/6.4.3.2D; cf. B). Incidentally, scattered apoptotic
cells are visible (Fig. 6.4.3.1/6.4.3.2D, arrowheads). If distinctive features, such as
Flexner–Wintersteiner rosettes or “fleurettes” are lacking (Fig. 6.4.3.1/6.4.3.2D),
diagnosis cannot be made without additional clinical information or special im-
munohistochemical stains.
Neoplastic Disorders 73
Useful antibodies
• Anti-synaptophysin or other neuronal marker proteins may be helpful if the
tumor shows a tendency toward neuronal differentiation. In the context of
clinical data, however, immunocytochemistry is seldom necessary, due to the
characteristic cytology of the tumor cells
74 Chapter 6
Useful antibodies
• Anti-vimentin (however, specificity is very low)
Neoplastic Disorders 75
76 Chapter 6
Most CNS germ cell tumors develop in the midline structures of children and
adolescents with the pinealis region as the most common site of origin, followed
by the suprasellar region. Intracerebral germ cell tumors, however, may also be
metastases arising from the gonads. Like their extracranial counterparts, intra-
cranial germ cell tumors can be divided into germinomas and non-germinoma-
tous tumors, which can be further sub-classified as embryonal carcinoma, cho-
riocarcinoma, yolk sac tumor, and teratoma. Mixed germ cell tumors also occur.
Since prognosis and therapy differ significantly if additional non-germinomatous
components are present, careful examination is indispensable for an exact diagno-
6
sis. Analysis of the pattern of marker proteins in the CSF (AFP, CEA, beta-HCG,
PLAP) may be helpful in indicating the components present. Germ cell tumors
have a high tendency to disseminate via CSF pathways. While CSF cytological
analysis is well established for staging and control investigations, primary diagno-
sis of intracranial germ cell tumors is the domain of stereotactic brain biopsy.
Central nervous system germinomas are histologically identical to gonadal
seminomas and dysgerminomas of the ovary. Cytological preparations from pure
germinomas typically reveal large, mostly polygonal tumor cells with large vesicu-
lar nuclei and prominent nucleoli (Fig. 6.4.4.1A, 1). As in the histologic specimen
(Fig. 6.4.4.1C), a strikingly reactive lymphocytic component is frequently present
(Fig. 6.4.4.1A, 2; B, 1). Sometimes, even more pleomorphic cells are detectable
(Fig. 6.4.4.1B, 2). Very similar cytologic features are seen in the corresponding
histology. Again, large cells with abundant cytoplasm and vesicular nucleoli are
visible (Fig. 6.4.4.1C, 1). Immunohistochemistry against placental alkaline phos-
phatase (PLAP) confirms the diagnosis of a germinoma (Fig. 6.4.4.1, D) and may
also be of diagnostic use in cytologic specimens. Recently, OCT4 (also called
OCT3, OTF3, POU5F1), a 18-kDa POU-domain transcription factor, has been
shown to serve as a specific marker for CNS germinoma that is superior to the
usual PLAP immunohistochemistry [35–37].
Useful antibodies
• Anti-PLAP
• Anti-OCT4
Neoplastic Disorders 77
78 Chapter 6
While choroid plexus carcinomas (WHO grade III) frequently metastasize via ce-
rebrospinal pathways, this is a rare event in choroid plexus papillomas (WHO
grade I) [38]. Morphologically, normal choroid plexus cells (Fig. 6.4.5A: histo-
logic specimen) are indistinguishable from cells derived from a differentiated
choroid plexus papilloma (Fig. 6.4.5B: histologic specimen, objective × 10). This
may pose a diagnostic problem, particularly in CSF specimens from the ventricle,
where small pieces of normal choroid plexus are not infrequently present (see
Sect. 3.5). Single cells from a choroid plexus carcinoma, which is readily identi-
fied in the histologic specimen (Fig. 6.4.5C), may also be cytologically indistin-
6
guishable from normal choroid plexus or choroid plexus papilloma. Origin in the
choroid plexus can be demonstrated by use of some newly described antibodies
against an inward rectifier potassium channel Kir7.1 (Fig. 6.4.5D) or the glyco-
protein stanniocalcin-1, which seems to be highly specific for plexus tissue [39].
This may be particularly helpful, if a differential diagnosis of metastatic carcinoma
versus malignant dedifferentiated plexus tumor arises.
Useful antibodies
• Anti-Kir7.1 (seems to be very specific for plexus epithelium, both normal and
neoplastic)
When interpreting smears and specimens one can be misled by particles mim-
icking micro-organisms, especially parasites such as protozoa, mycotic agents or
helminths. Although some of these pitfalls are well-known, others can be prob-
lematic. For example, false protozoa parasites can correspond to exogenous agents
such as pollen (Fig. 7.3A,B, objective × 40).
Sometimes, the powder of gloves (starch) contaminating the specimen may be
misinterpreted as cryptococci (Fig. 7.3C, objective × 40). However, in this case
polarized light microscopy demonstrates the characteristic Maltese cross imaging
of starch (Fig. 7.3D, objective × 40). Furthermore, in the case of contamination
with pollen or starch the lack of inflammatory cells reveals these particles to be
7
contaminants.
Contaminants 85
References
17. Glass JP, Melamed M, Chernik NL, et al. (1979) Malignant cells in cerebrospinal fluid
(CSF): the meaning of a positive CSF cytology. Neurology 29:1369–1375
18. Gleissner B, Chamberlain MC (2006) Neoplastic meningitis. Lancet Neurol
5:443–452
19. Chamberlain MC (2005) Neoplastic meningitis. J Clin Oncol 23:3605–3613
20. Srodon M, Westra WH (2002) Immunohistochemical staining for thyroid transcrip-
tion factor-1: a helpful aid in discerning primary site of tumor origin in patients with
brain metastases. Hum Pathol 33:642–645
21. Brat DJ, Giannini C, Scheithauer BW, et al. (1999) Primary melanocytic neoplasms of
the central nervous systems. Am J Surg Pathol 23:745–754
22. Katz KA, Jonasch E, Hodi FS, et al. (2005) Melanoma of unknown primary: experience
at Massachusetts General Hospital and Dana-Farber Cancer Institute. Melanoma Res
15:77–82
23. Jaffe ES, Harris NL, Stein H, Vardiman JW (eds) (2001) WHO classification of tu-
mours. Pathology and genetics of tumours of haematopoietic and lymphoid tissues,
2nd edn. IARC, Lyon
24. Chamberlain MC, Nolan C, Abrey LE (2005) Leukemic and lymphomatous meningi-
tis: incidence, prognosis and treatment. J Neurooncol 75:71–83
25. Balmaceda C, Gaynor JJ, Sun M, et al. (1995) Leptomeningeal tumor in primary cen-
tral nervous system lymphoma: recognition, significance, and implications. Ann Neu-
rol 38:202–209
26. Batchelor T, Loeffler JS (2006) Primary CNS lymphoma. J Clin Oncol 24:1281–1288
27. Riva M, Bacigaluppi S, Galli C, et al. (2005) Primary leptomeningeal gliomatosis: case
report and review of the literature. Neurol Sci 26:129–134
28. Perilongo G, Gardiman M, Bisaglia L, et al. (2002) Spinal low-grade neoplasms with
extensive leptomeningeal dissemination in children. Childs Nerv Syst 18:505–512
29. Tihan T, Fisher PG, Kepner JL, et al. (1999) Pediatric astrocytomas with monomor-
phous pilomyxoid features and a less favorable outcome. J Neuropathol Exp Neurol
58:1061–1068
30. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (eds) (2007) WHO classification of
tumours of the central nervous system, 4th edn. IARC, Lyon
31. Blümcke I, Muller S, Buslei R, et al. (2004) Microtubule-associated protein-2 immu-
noreactivity: a useful tool in the differential diagnosis of low-grade neuroepithelial
tumors. Acta Neuropathol (Berl) 108:89–96
32. Seol HJ, Hwang SK, Choi YL, et al. (2003) A case of recurrent subependymoma with
subependymal seeding: case report. J Neurooncol 62:315–320
33. Lassaletta A, Perez-Olleros P, Scaglione C, et al. (2007) Successful treatment of in-
tracranial ependymoma with leptomeningeal spread with systemic chemotherapy and
intrathecal liposomal cytarabine in a two-year-old child. J Neurooncol 24:24
34. Merchant TE, Fouladi M (2005) Ependymoma: new therapeutic approaches including
radiation and chemotherapy. J Neurooncol 75:287–299
35. Hattab EM, Tu PH, Wilson JD, et al. (2005) OCT4 immunohistochemistry is superior
to placental alkaline phosphatase (PLAP) in the diagnosis of central nervous system
germinoma. Am J Surg Pathol 29:368–371
References 89
36. Cheng L, Sung MT, Cossu-Rocca P, et al. (2007) OCT4: biological functions and clin
ical applications as a marker of germ cell neoplasia. J Pathol 211:1–9
37. Takei H, Bhattacharjee MB, Rivera A, et al. (2007) New immunohistochemical mark-
ers in the evaluation of central nervous system tumors: a review of 7 selected adult and
pediatric brain tumors. Arch Pathol Lab Med 131:234–241
38. McEvoy AW, Galloway M, Revesz T, et al. (2002) Metastatic choroid plexus papilloma:
a case report. J Neurooncol 56:241–246
39. Hasselblatt M, Bohm C, Tatenhorst L, et al. (2006) Identification of novel diagnostic
markers for choroid plexus tumors: a microarray-based approach. Am J Surg Pathol
30:66–74
Subject Index
A D
accessory nuclei 52 diluent 6
adenocarcinoma 56
adhesion 6 E
albumin 42, 46 eosinophilic granulocyte 25
apoptosis 64, 72 ependyma 9
arachnoid granulations 18 erythrocyte 6, 14
arachnoid membrane 18 erythrophage 14
Aspergillus 32 erythrophagocytosis 6
atypical cell 52 estrogen 54
autolysis 2, 20
F
B filum terminale 70
bacteria 14 fixation 2, 5
bilirubin 44 flow cytometry 10, 64
blood–CSF barrier 46, 48 free-living amoebae (FLA) 36
borreliosis 25
budding 32 G
glial fibrillary acidic protein
C (GFAP) 66 – 68, 70
cartilage 20 glucose 82
cell number 6 granulocyte 30, 32, 41, 42, 44, 46
choroid plexus 9, 16, 78 granulomatous amebic encephalitis
concentration 6 (GAE) 36
CSF/serum quotient diagram 46
CSF–serum ratio 42 H
cytocentrifugation 1 hematoidin 14, 44
cytokeratin 54, 56, 58, 60 hemoglobin 14, 44
cytostatic therapy 64 hemosiderin 14, 44, 62
92 Subject Index
K P
Kir7.1 78 parasite 84
periodic acid Schiff (PAS) 32, 58
L phagocytosis 14, 42, 44
lactate 82 pilocytic astrocytoma 66
leptomeningitis 26 placental alkaline phosphatase
lumbar puncture 16, 20 (PLAP) 76
lymphocyte 9, 28, 41, 48, 76 plasma cell 10, 25, 28, 41, 46
– B lymphocyte 10, 46 polarized light microscopy 84
– T lymphocyte 10 primary tumor 51, 53, 56, 72
progesterone 54
M protozoa 38
macrophage 36, 41, 44, 48 protrusions 60, 62, 66, 68
Maltese cross 84 – cytoplasmic 60, 68
marrow 22 – nuclear 62
Melan A 60 – subependymal 66
melanin 44, 60 Prussian blue 44
membrane filtration 1 pseudomeningitis 82
meningioma 18
microtubule-associated protein-2 S
(MAP2) 67, 68 S-100 60
mitosis 10, 28, 52, 74 siderophage 14
mold fungal infection 32 signet cell carcinoma 58
monocyte 6, 9, 28, 41, 42, 48, 82 skin 22
multiple sclerosis (MS) 46, 48 staining 2
Subject Index 93