16
Central Nervous System
(4)
A case of a foreign-body reaction in a postopera-
tively obtained sample erroneously thought to
contain tumor cells; and
(5)
A case of bacterial meningitis in a child with leuke-
mia incorrectly diagnosed as malignant.
In the series of 117 cases studied postmortem by Glass and asso-
ciates, 3 cases were falsely diagnosed as positive (2.6%).47 All
three cases involved patients with lymphoma and with infec-
tious meningitis erroneously interpreted as being neoplastic.
Borowitz and associates assessed the accuracy of the diagno-
sis of leukemia and lymphoma in CSF by reviewing the mor-
phology and clinical outcome of 45 patients.7 They reported
an overall false-positive rate of 10/72 or 14% of the samples.
Diagnostic accuracy was highest for acute leukemia, but all
patients with chronic lymphocytic leukemia had false-positive
diagnoses owing to misinterpretation of reactive lymphocytes
in cases of fungal or viral meningitis. Because these data were
acquired before the availability of lymphocyte marker studies,
one would anticipate that this source of false-positive diagnoses
could largely be eliminated by the application of immunohis-
tology to identify lymphocytic subsets.
These studies demonstrate that the rate of false-positive diag-
noses in CSF is low for metastases and primary brain tumors
but is a significant problem in CNS leukemia and lymphoma.
In general, the incidence of false-positive diagnoses in CSF can
be reduced by awareness of three areas of diagnostic pitfalls.
Because normal CSF is practically acellular, laboratory contami-
nation of these samples with benign or malignant cells from
other patients presents a major problem in CSF cytology, par-
ticularly when membrane filters are used. Contamination of
staining dishes or the instruments used in preparation with
malignant cells can lead to erroneous diagnoses of malignancy.
In addition, contamination of CSF with benign cells, such as
bronchial epithelium, squamous cells, or reactive mesothelium,
can cause diagnostic difficulties and can be misinterpreted as
representing a pathologic or neoplastic process.
A second pitfall is misinterpretation of normal brain ele-
ments. Thus, chondrocytes, ependymal-choroidal cells, and
neurons and glia, particularly in ventricular fluids, can cause a
problem if the observer is unaware of the cytologic appearance
of these elements.
A third major pitfall is misinterpretation of reactive lym-
phocytes as representing leukemia or lymphoma. This situ-
ation is frequently encountered both in patients who are
known to have leukemia or lymphoma and in patients who
Table 16.2 Supratentorial Masses in Adults
Low cellularity
High cellularity
Cytologically
benign
Cytologically
malignant
Normal brain
Abscess
Lymphoma
Reactive gliosis
Granuloma
Metastatic carcinoma
Low-grade
astrocytoma
Infarct
Melanoma
Hemorrhage
Sarcoma
Oligodendroglioma
Glioblastoma multiforme
Ependymoma
have meningeal symptoms and no prior history of neoplasia.
This problem can largely be eliminated by the application of
immunohistochemical markers as discussed previously.
Needle Aspirates of Intracranial Lesions
The most common application of fine-needle aspiration to
the CNS is in evaluation of a supratentorial mass in an adult.
Table 16.2 lists the lesions that should be considered in this
clinical setting. A general approach is to divide preparations into
those of low and high cellularity and within the latter category
to distinguish specimens containing cells with benign cytologic
features from those with malignant ones.
Because aspiration of brain lesions is performed to evaluate
mass lesions, samples of normal brain should theoretically not
be encountered. In reality, however, the lesion may be missed on
the first attempt, yielding only normal brain tissue containing
sparse neurons or glia (Fig. 16.13). In these cases, normal brain
must be distinguished from other processes yielding samples
of low cellularity, such as reactive gliosis and low-grade astro-
cytomas (Figs 16.14 and 16.15). In both gliosis and low-grade
tumors, the sample is more cellular than is normal brain. In
smear preparations, however, the thickness of the sample influ-
ences the cellularity; thus, this factor must be considered in
judging the cellular density. In smears of normal brain stained
with hematoxylin and eosin or Papanicolaou stain, only the
nuclei of glia are visible. Silver impregnation stains or immuno-
histochemistry using antibodies against GFAP is necessary to
demonstrate the cell bodies and processes. In gliosis, however,
the cytoplasm of the astrocytes is often dense and becomes vis-
ible in routine preparations. Low-grade astrocytomas may have
either inconspicuous or visible cytoplasm, but in addition to
increased cellularity, the nuclei become more hyperchromatic
and irregular in shape than do normal cells or those in gliosis
(see Fig. 16.15).
Aspiration of many mass lesions produces a hypercellular
smear. These processes include abscesses, granulomas, infarc-
tions, hemorrhages, lymphomas, metastases, and many types
of primary brain tumors (see Table 16.2). The distinctions
between these lesions depend to a large extent on individual
Fig. 16.13 Smear of normal white matter (H&E x MP).
449
previous page 444 ComprehensiveCytopathology 1104p 2008 read online next page 446 ComprehensiveCytopathology 1104p 2008 read online Home Toggle text on/off