philip.boyer@utsouthwestern.edu (phil

Your question appears focused on cell counts in the Hematology Lab.,
if I'm reading it correctly, rather than flow cytometry. The following
addresses stains - - although not a specific commercially available
stain - - and preparations.

First, I must define my goals in the evaluation. From the
neuropathologist perspective, I am interested in evaluating CSF from
several perspectives as I distinguish "normal" from "abnormal" cell
populations in the CSF. I find that a combination of preparations and
stains is optimal.

(1) Normal Hematopoietic Cells: cell counts of lymphs, monocytes, PMNs
/ bands, RBCs, etc.

(2) Normal "Lining" Cells: ependymal and choroid plexus

(3) Neoplastic Cells: leukemic or lymphomatous cells, medulloblastoma,
glioma, carcinoma, etc.

(4) Infectious Entities: cryptococcus, bacteria (sometimes tough to
distinguish from stain particles)

(5) Other: portions of neuropile obtained during insertion of
ventriculostomy tube; portions of tissues penetrated during lumbar
puncture (most commonly cartilage, dural fragments, arachnoid

To help me with evaluation for these various entities, I like to look
at BOTH (a) the Wright-Giemsa stained, air-dried cytospin preps from
the Hematology laboratory and (b) the PAP stained, alcohol-fixed
cytospin or ThinPrep preparations from the Cytology Laboratory. The
air dried Giemsa stained preparation is for me optimal for evaluating
both normal and abnormal hematopoietic cells . . . far superior to the
alcohol fixed preparations as the nuclei in the latter are tough to
"read." It turns out that the air dried preparations are often optimal
for me for evaluation for and classification of foreign entities. The
cells and cell clusters in the Cytology laboratory preparations that
I've seen in all hospitals in which I've trained and worked manifest
artifactual shrinkage which makes cytologic evaluation challenging. In
looking at all of the preparations from both laboratories, I also am
evaluating a greater number of cells and am increasing the sensitivity
of identification of abnormal cells (e.g. that one cluster of
pineocytoma) from a single CSF tap.

The following summary is taken from a work in progress, DIAGNOSTIC
CEREBROSPINAL FLUID EVALUATION, a mini-text book that will be
"published" / posted on the College of American Pathologists Web site
(http://www.cap.org) in January, 2004. This mini-text, including
multiple tables and illustrations, will be available as a free PDF
download.

++++++++++
PREPARATION FOR EVALUATION:
Ideal evaluation incorporates evaluation of a combination of both (a)
the Cytology Laboratory specimen, often prepared after a period of
delay, sometimes hours, and traditionally alcohol fixed and stained
with a PAP stain, and (b) the Hematology Laboratory specimen,
typically prepared quickly after collection, air dried, and stained
with Wright-Giemsa stain. The latter is typically a much better
preparation for assessment for hematopoietic neoplasm and distinction
from cytologically normal or reactive lymphocytes and monocytes. The
Cytology Laboratory specimen is often better for evaluation of primary
and metastatic neoplasms.

In general, because of the paucity of cells present in CSF specimens,
a method to optimally concentrate and minimally alter the cytology of
the cells is necessary. The cytocentrifuge technique is the most
commonly used (e.g. using the Shandon cytocentrifuge apparatus) to
concentrate the cells and affix them to, usually, two glass slides. In
the hematology laboratory, these slides are generally air dried and
then stained with a Wright-Giemsa type stain. In the cytology
laboratory, an additive that lyses red cells but maintains the
integrity of other cells with or without fixative is often added to
the CSF prior to cytocentrifugation. The prepared slides are then
typically fixed in alcohol and stained with a Papanicolaou method. If
the number of cells is large, which is only rarely the case,
additional unstained slides can be prepared by cytocentrifugation or a
cell block can be prepared, using either gelatin or thrombin mixed
with plasmin to capture the cells, with subsequent processing in the
histology laboratory to paraffin and sectioning like a typical
histology laboratory block. A hematoxylin and eosin stained slide is
produced and the cell block can subsequently be used for
immunophenotyping cells present in the specimen.

Alternatively, use of a small-pore filters under pressure (e.g.
Millipore filter) to perform the function of concentration of the
cells is used by some cytopathology laboratories (Bigner, 1994). The
relatively recent introduction of the ThinPrep technology, an
automated membrane-based approach, has become an attractive method for
the preparation of glass slides onto which cells from gynecologic
smear collection (placed into a fixative buffered solution) or from
fluids are well distributed without overlap. It can be applied to CSF
samples.

In the Cytology Laboratory, many NON-CSF fluids and aspirates have a
fair number of cells and cells clusters present. After centrifugation
of the material from these specimens, a fairly large cell button is
present from which only a small proportion of the cells are sampled
for initial smear. Rather than discard the remainder of the button,
that material can be immobilized in a thrombin clot or in gelatin,
processed into a paraffin block in the Histology Laboratory; the
paraffin block is then treated like any biopsy specimen, with the
paraffin block sectioned and sections with hematoxylin and eosin (H&E)
and if desired special stains. The number of cells available for
evaluation in a CSF specimen is far too few to effect such a paraffin
block.

SPECIAL STAINS:
Special stains can be performed on cytocentrifuge preparations (with
destaining of one of the Papanicolaou stained slides if necessary) or
on sections from paraffin specimens. In general, filter preparations
are not amenable to special stains. The purposes of special stains
include (a) identification of microorganisms (a tissue Gram stain for
bacteria, a silver stain for fungus, or an acid fast stain for acid
fast bacilli), (b) histochemical identification of specific substances
like mucin (mucicarmine), glycogen or glycosylated proteins
(periodic-acid Schiff, PAS), collagenous material (trichrome or
reticulin), or elastin (elastin stain), and (c) immunoperoxidase
identification of antigens such as cluster designation proteins for
identification of white blood cells (common leukocyte antigen
(CD45rb), T cells (CD3), B lymphocytes (CD20), macrophages (CD68)),
cytokeratin for an epithelial cell, glial fibrillary acidic protein
for a glial cell, etc.

The difficulty in using any of these stains is that the cell or
microbial content in the fluid is often very limited, and much or all
of it is cytocentrifuged onto the original two slides stained with PAP
stains. If additional fluid is not available, then a PAP stained slide
can be decolorized and used. However, whereas with paraffin embedded
tissue you have the opportunity to pursue a panel of stains, with a
single slide or with a few slides options are considerably limited and
choice of stain must be judicious.

When abundant cells are present, flow cytometric evaluation can be
carried out. Most commonly, a limited panel of antibodies will be used
to establish B or T cell lineage and evaluate for monotypic expression
of kappa or lambda light chain in the former. In practice, an initial
lumbar puncture or ventriculostomy fluid collection both (a) an
atypical mononuclear cell population is present and (b) the abundance
of cells warrants an attempt at flow cytometric ****ysis. Actual
****ysis is then carried out on a subsequent specimen or specimens.
Ultimately, the number of cells may be too small to adequately
****yze.

EVALUATION OVERVIEW:
Cytologic evaluation of the cells in a CSF specimen is a screening
tool, as cytology is in other settings, such as evaluation of cells
present in effusions, gynecologic specimens, and fine needle
aspirations. In contrast to a biopsy specimen, where the pathologist
is afforded architectural context for appropriate classification of
atypical cells, in cytologic preparations, individual cells and small
clusters of cells are typically the substrate for evaluation. As a
result, non-definitive interpretations are sometimes necessary, e.g.
"atypical cells present," suspicious for neoplasm," and "malignant
cells present." Certainty of diagnosis can be bolstered by comparing,
for instance, the atypical cells in the CSF of a patient with a
history of acute leukemia with the leukemic blasts present in the
patient's bone marrow aspirate smear. In some situations, e.g. a CSF
specimen which shows tight clusters of atypical cells with
hyperchromatic nuclei, nuclear moulding, and minimal cytoplasm from a
patient with a history of medulloblastoma, the pathologist can be
definitive in her / his diagnosis. The typical paucity in number of
cells in a CSF specimen limits the use of special stains on CSF
preparations, but when they can be performed, useful data for the
differential diagnosis database can be obtained. Ultimately, as in all
areas of pathologic interpretation, clinicopathologic correlation is
required for ultimate appropriate interpretation of a result.

In the end, after all evaluation is carried out, the combined efforts
of the hematology laboratory technicians, cytology technicians, and
cytopathologists may end up with un-satisfying designations such as
(1) positive for malignant cell or (2) suspicious cells present,
favor neoplasm or reactive process. However, such limitations are
inherent in the realm of cytology.

DIAGNOSTIC PITFALLS:
A variety of pitfalls are present in the cytologic interpretation of
CSF. Placement of a ventriculostomy tube requires passage of the tube
through cerebral parenchyma into ventricular space. As a result, small
portions of cerebral parenchyma and ependymal surface are commonly
noted in the initial CSF sample sent after tube placement, and the
specimen requisition sheet may not note that the source of the
specimen was from a ventriculostomy tube. Most worrisome are fragments
of germinal matrix that can be noted in specimens from premature
infants. In addition, small clusters of "lining cells," choroid
plexus, ependymal, or arachnoid in origin, can be noted in the CSF.
Interpretation can be extremely difficult if slide preparation is not
optimal. Incorporation of all clinical data is necessary for
appropriate interpretation. Further, blasts present in peripheral
blood can contaminate a CSF preparation as the result of a bloody tap,
and this possibility must be carefully considered.

All cytopathology preparations are eminently susceptible to cross
contamination from another specimen that is concomitantly or had
previously been fixed or stained in solutions used for the CSF
preparation. As a single cell is, in reality, enough to call "positive
for malignancy," scrupulous care must be exercised to avoid
contamination.

Critical to accurate evaluation of CSF, cells present in the CSF must
be distinguished from cells present in the blood. Contamination of CSF
by blood containing a large number of polymorphonuclear cells or
leukemic cells could spuriously lead to consideration of an infectious
or leukemic CSF process.

++++++++++++++

Philip J. Boyer, M.D., Ph.D.
Department of Pathology, Division of Neuropathology
University of Texas Southwestern Medical School
5323 Harry Hines Blvd. H2.132
Dallas, Texas 75390-9073
Philip.Boyer@utsouthwestern.edu

robert

nice response and very interesting, I think what the original question was
referring to a wet chamber count of csf specimens. We use new methylene
blue stain, same as in retic for doing these counts. Microcrit tubes are
coated and allowed to dry. Csf is then placed in these crit tubes and
placed on the chamber. I think that was the original question.

kuhnfucius

"The
air dried Giemsa stained preparation is for me optimal for evaluating
both normal and abnormal hematopoietic cells . . . far superior to the
alcohol fixed preparations as the nuclei in the latter are tough to
"read." It turns out that the air dried preparations are often optimal
for me for evaluation for and classification of foreign entities. "

I agree. In addition, I clear any stain precipitate and lighten the stain
(after staining) body fluid preps (particular when Wrights / Giemsa is used)
by dipping them in a week solution of alcohol. When hematological stains
are applied to body fluids preps the usual absence of large numbers of cells
tends to create are much darker stain in the cells. Thanks for the
addition info.

kuhnfucius

A very interesting solution! Sounds great. I want to try this. We no
longer have micro hematocrit tubes due to the sharps hazard, but I can order
a some. My only concern is whether there will be a significant amount of
stain precipitate. Since I switched to totally automated retics I have no
long stock new methylene blue stain (any irregular patterned retics are sent
to the reference lab). Perhaps the instrument stain pack has some. I do
not miss the inspection difficulties or time involved with manual retic
evaluation. I have digressed here from the original thread to invite a
critic such a policy.