Vet Clin Small Anim 32 (2002) 1209–1235

Recognition of basic cell types

and criteria of malignancy
James H. Meinkoth, DVM, MS, PhD*,
Rick L. Cowell, DVM, MS
Department of Veterinary Pathobiology, Oklahoma State University College
of Veterinary Medicine, 250 McElroy Hall, Stillwater, OK 74078-2007, USA

A person examining cytologic preparations is often presented not only

with a confusing array of different cell types but, often, with an endless vari-
ety of cell debris and contaminants as well. With experience, most of the
common lesions are recognized quickly from memory. For the beginner,
or even the experienced cytologist when confronted with a lesion or sample
site not previously encountered, it is helpful to keep in mind certain funda-
mental questions that need to be considered when evaluating a sample. An
orderly approach to examining slides and consideration of certain questions
that can be answered based on cytology can reduce the chances of missing
important information or of misdiagnosing or overdiagnosing the sample.
Cells can usually be classified into one of a few basic categories based on
common features shared by different cells in that category. Recognizing the
common features of the different basic cell types makes it easier to identify
cells that are not obviously recognized at first.

Are sufficient numbers of well-stained intact cells present to evaluate?

A basic premise of cytology is that interpretations are generally based on
whole populations of cells rather than on individual cells. Any one cell or
few cells from a lesion may show features that are atypical or unusual. This
is especially true if cells are coming from a tissue in which the cells are not
well preserved or have been exposed to injurious stimuli. Cells coming from
inflammatory reactions or areas of tissue repair often show dysplastic or
other atypical features. Cells from urine samples often have significant aging
artifacts that could easily be misinterpreted.

* Corresponding author.
E-mail address: jhm@cvm.okstate.edu (J.H. Meinkoth).

0195-5616/02/$ - see front matter Ó 2002, Elsevier Science (USA). All rights reserved.
PII: S 0 1 9 5 - 5 6 1 6 ( 0 2 ) 0 0 0 4 8 - 7
1210 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

Interpretations based on inadequately cellular specimens may not contain

a representative sample of the lesion, giving a false impression of normalcy,
or, even worse, may result in a false impression of neoplasia that is not really
present. Although there is no easily defined limit to the question of how
many cells are enough, slides should have numerous cells per field across
a large portion of the slide. When a large lesion is being evaluated, several
smears of good cellularity collected from different areas of the lesion should
be evaluated to assess any variability that may be present within different
parts of the lesion. Large neoplasms may have areas of inflammation or
necrosis that yield markedly different cell populations than adjacent areas.
The cellularity of the smear is often evident the moment the slides are
stained. A slide containing high numbers of nucleated cells is visibly blue
after staining. Slides that are perfectly clear after staining probably have low
numbers of nucleated cells, although there may still be sufficient cells present
for a diagnosis. Therefore, cellularity must be confirmed by looking at the
slide under the microscope. Even for practitioners who do not have the time
or desire to evaluate cytology preparations themselves, it is beneficial to
stain one or two smears and determine that an adequately cellular specimen
is being sent off for evaluation.
When examined microscopically, the slides should first be scanned using
low power (10–20) to asses the cellularity of the slide and find the area(s)
containing the highest number of well-stained, well-spread, intact cells for
evaluation (Figs. 1 and 2). Cells are often unevenly distributed across the
slides, especially with impression smears or aspirates of solid tissue lesions.
Even with slides made from fluid samples, large cells may all be pulled out to
the feathered edge, where they can be easily overlooked if the whole slide is
not scanned first.
After locating the cellular areas of the slides, it is important to determine
if the cells are sufficiently spread for evaluation, intact, and well stained.
This can usually also be done using low power (10–20), which allows a

Fig. 1. Low-power (scanning) magnification of a slide showing cells that are well spread.
Although cell detail is difficult to evaluate at this power, individualized well-spread cells can be
recognized. Examination of this area at higher magnification should be rewarding. (Courtesy of
Oklahoma State University teaching files, Stillwater, OK.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1211

Fig. 2. Low-power magnification of another area from the same slide as in Figure 1. From
this magnification, the cells can be seen piling up on top of one another, and few well-spread
cells can be seen. Better areas of the slide should be found before progressing to higher
magnification. (Courtesy of Oklahoma State University teaching files, Stillwater, OK.)

greater portion of the slide to be evaluated in a short time. Inexperienced

cytologists often frustrate themselves by spending an inordinate amount
of time trying to identify cells that cannot be interpreted because of poor
staining or rupturing. Intact, well-stained, well-spread cells should have a
clearly evident demarcation between the nucleus and cytoplasm (Fig. 3). The
nucleus may be irregularly shaped (particularly in neoplastic cells), but the
nuclear outline should be smooth and distinct. A ‘‘fuzzy’’ appearance
around the outline of the nucleus generally indicates that the cell has been
minimally traumatized during sample collection and/or preparation. More

Fig. 3. Higher magnification showing many well-spread cells. A clear distinction can be seen
between the nucleus and the cytoplasm (arrowheads). Several traumatized cells are also present.
Note that their nuclear chromatin appears excessively fragmented (arrows). (Courtesy of
Oklahoma State University teaching files, Stillwater, OK.)
1212 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

significant cell trauma can result in the nucleus appearing fragmented or full
of holes (see Fig. 3). Severely traumatized cells often appear as strands of
light pink nuclear chromatin (Fig. 4). Some traumatized or ruptured cells
are present in virtually any cytologic specimen. The sample is usually still
interpretable if most of the cells are intact; however, the traumatized cells
are typically not evaluated, particularly when determining criteria of malig-
nancy. If most of the cells are traumatized or ruptured, additional samples
usually need to be collected.
Cells that are not well spread often stain diffusely dark, and it is difficult to
distinguish the line of demarcation between the cytoplasm and nucleus (Fig. 5).
Also, poor staining (even in well-spread cells) can result in a less distinct
demarcation between the nucleus and cytoplasm. Nucleoli often are more
prominent than usual in understained cells, and this can result in a false
impression of malignancy if this artifact is not recognized. The nuclei of
well-stained cells are usually stained a dark purple color, which is more intense
and deeper than that of the surrounding cytoplasm. There is some variation in
the intensity and pattern of nuclear staining between cell types. If you are
unsure whether light staining of nuclei is a characteristic of the cell or the result
of understaining, it may be helpful to evaluate the staining of more familiar
cells if any are present. Usually, some neutrophils are present as the result
of peripheral blood contamination or inflammation within the lesion.
In most specimens, there is significant variation in cellularity, degree of
cell spreading, and, hence, staining quality from area to area on the slide.
This is particularly true of extremely cellular specimens (ie, lymph node aspi-
rates), which may have areas that are of diagnostic quality even if most of
the slide is thick and understained. Diligent scanning of the slides on low
power is necessary to find these areas before attempting to evaluate the cells
at higher magnification. If the entire slide is found to be understained,
restaining the slide may improve the staining quality and result in a diag-
nostic sample.

Fig. 4. Severely traumatized cells. The streaks of material (arrows) represent smeared out
nuclear chromatin from ruptured cells. (Courtesy of Oklahoma State University teaching files,
Stillwater, OK.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1213

Fig. 5. Poorly spread cells from a different area of the same slide as in Figure 3. The cells are
diffusely dark, and the demarcation between the nucleus and cytoplasm is difficult to determine.
(Courtesy of Oklahoma State University teaching files, Stillwater, OK.)

Are all the cells on the smear inflammatory cells?

A good initial decision, particularly for the beginning cytologist, is to
determine if the smear is composed entirely of inflammatory cells. Inflamma-
tory lesions are probably more commonly encountered in most practices
than neoplastic lesions. Also, most clinicians initially feel more comfortable
recognizing inflammatory cells because they are more familiar with the mor-
phology of these cells, having viewed them many times in peripheral blood
smears. Many clinicians choose to screen their cytology specimens, interpret-
ing inflammatory lesions in-house but submitting those composed of tissue
cells for outside evaluation. Although inflammatory cells in tissues often look
the same as they do in peripheral blood, some may appear different because
of morphologic changes induced by their presence in a focus of inflammation
or simply because they are not well spread. It is important to remember that
one may not be able to identify every cell present on a slide (or even in any
given field) and that the interpretation is based on the entire cell population
present. If a lesion is found to be composed entirely of inflammatory cells, the
relative percentages of the various types of inflammatory cells should be
noted, because this may provide clues to the etiology of the inflammation.
Finally, a search for infectious agents should be conducted. A discussion
of the various inflammatory patterns and morphology of common infectious
agents is covered in detail elsewhere in this issue. The basic types of inflam-
matory cells and their morphologic variations are discussed here.

Neutrophils
Neutrophils are commonly found in cytologic specimens. Their morphol-
ogy is often similar to that observed in peripheral blood smears (Fig. 6).
Normal neutrophil nuclei stain dark purple and contain one to multiple
1214 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

distinct segments or lobules. The neutrophil cytoplasm is typically clear.

Neutrophils are phagocytic cells and typically are the cells that phagocytize
pathogenic bacteria (see Fig. 6). Neutrophils contain intracytoplasmic gran-
ules; however, the neutrophil granules in most domestic animals generally
do not stain prominently with cytologic stains. Sometimes, however, these
granules are detectable as elongated and faintly eosinophilic structures, and
they must not be confused with bacteria.
In thick preparations or viscous fluids (eg, synovial fluid), neutrophils
may not spread well, and the segmented nature of their nucleus may be less
evident. Sometimes, the nucleus is essentially round, mimicking a lympho-
cyte, but the cell can still be identified as a neutrophil by the lobulated out-
line of the nucleus. More normal neutrophil morphology can be observed in
the thinner areas (often along the edges) of the smear.
Neutrophils may undergo several morphologic changes in tissues. Aging
change is a commonly encountered phenomenon. The initial change seen is
hypersegmentation of the nucleus (Fig. 7). Sometimes, an elongated thin
strand of nuclear material (Fig. 8) connects the nuclear lobes of aged neutro-
phils. This is especially common in neutrophils from fluid samples. The end
result of aging change is pyknosis of the nucleus. Pyknosis is condensation
of the nuclear chromatin into one or more small, discrete, densely staining
spheres lacking any nuclear chromatin pattern (see Fig. 7). Aging artifact
simply represents neutrophils dying of ‘‘old age’’ and must not be confused
with degenerative change.
Degenerative change occurs when neutrophils are present in an environ-
ment that is damaging to the cell. It is commonly seen in neutrophils from
lesions in which endotoxin-producing bacteria are present. The presence of
many neutrophils with marked degenerative change should prompt a dili-
gent search for bacteria. Degenerative change is an acquired change and
is distinct from the ‘‘toxic’’ changes noted in peripheral blood neutrophils.
Degenerative change occurs when the neutrophil is unable to control
water homeostasis and undergoes hydropic degeneration. The hallmark of

Fig. 6. Septic neutrophilic inflammation. Many neutrophils are present, some of which con-
tain phagocytized bacterial rods. (Courtesy of Oklahoma State University teaching files,
Stillwater, OK.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1215

Fig. 7. Sample from a nonseptic inflammatory lesion. Many of the neutrophils show nuclear
hypersegmentation (arrow), an aging artifact. This ultimately leads to pyknotic change
(arrowhead) in which the nuclear chromatin condenses and fragments into several discrete dense
spheres. (Courtesy of Oklahoma State University teaching files, Stillwater, OK.)

degenerative change is nuclear swelling (Fig. 9). The nucleus of the cell
swells, appearing thicker, staining more eosinophilic, and losing nuclear
lobulation. Degenerative neutrophils often resemble large band cells.

Macrophages
Macrophages in inflammatory lesions are derived from peripheral blood
monocytes. Many tissues have low numbers of fixed tissue macrophages as
normal resident cells (eg, Kupffer cells in the liver). Macrophages have an
extremely varied morphology in tissues, which can be somewhat confusing

Fig. 8. Hypersegmentation of neutrophils in which elongated thin filaments connect the nuclear
lobes. This is a common manifestation of aged neutrophils in fluid samples. (Courtesy of
Oklahoma State University teaching files, Stillwater, OK.)
1216 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

Fig. 9. Neutrophils showing degenerative change. Degenerative change is evidenced by nuclear

swelling, which leads to loss of distinct segmentation. Degenerative change is commonly
associated with the presence of endotoxin-producing bacteria. Many bacterial rods (arrow) were
present on this slide. (Courtesy of Oklahoma State University teaching files, Stillwater, OK.)

for the beginning cytologist. Initially, they may resemble peripheral blood
monocytes (Fig. 10). With time, the nucleus becomes round, and the cell
enlarges as the cytoplasm becomes greatly expanded and, sometimes,
extremely vacuolated. Macrophages are also phagocytic cells, typically phago-
cytizing larger structures like fungal organisms and other cells. Many times,
the cytoplasm of macrophages contains partially phagocytized debris that
cannot be identified but must not be misinterpreted as an infectious agent.
Binucleated or multinucleated macrophages are commonly encountered in
long-standing inflammatory lesions (Fig. 11). Binucleated or multinucleated
macrophages can get extremely large and are referred to as inflammatory
giant cells. In some chronic inflammatory lesions, ‘‘epithelioid’’ macrophages
may be encountered. This description is applied to macrophages that are
enlarged with expansive cytoplasm that stains uniformly basophilic, giving

Fig. 10. Pyogranulomatous inflammatory response demonstrating many macrophages. Some

macrophages resemble peripheral blood monocytes (arrowhead). Other macrophages are
variably increased in size, resulting from increased amounts of cytoplasm, which is sometimes
vacuolated (arrows). (Courtesy of Oklahoma State University teaching files, Stillwater, OK.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1217

Fig. 11. Smear from a pyogranulomatous inflammatory reaction (Actinomyces infection). An

extremely large multinucleated macrophage is present. The macrophage has phagocytized
several neutrophils. (From Cowell RL, Tyler RD, Meinkoth JH, editors. Diagnostic cytology
and hematology of the dog and cat. 2nd edition. St. Louis: Mosby; 1999. p. 25; with permission.)

the cell the look of an epithelial cell. If such cells show variation in size and
multinucleation (both commonly seen in macrophages), they could poten-
tially be misinterpreted as neoplastic epithelial cells. Extreme caution should
be used when diagnosing malignancy in the face of inflammation because of
the potential for macrophages to display ‘‘atypical’’ criteria.

Lymphocytes (small, medium, and large)

Small lymphocytes are smaller than neutrophils, with rounded nuclei and
scant basophilic cytoplasm (Fig. 12). The nucleus is generally not perfectly
round but has a flattened or indented area on one side. The nuclear chroma-
tin has a ‘‘smudged’’ appearance. Nucleoli are not visible; however, darker
areas (heterochromatin) and lighter areas (euchromatin) are often visible.
Generally, the cytoplasm does not appear to completely encircle the nucleus;

Fig. 12. Smears made from an aspirate of a reactive lymph node. Many small lymphocytes are
present (arrowhead). These cells have scant amounts of cytoplasm, which does not appear to
encircle the nucleus. Two plasma cells (arrows) are also present. (Courtesy of Oklahoma State
University teaching files, Stillwater, OK.)
1218 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

it is visible for only a portion of the way around the nucleus. Medium-sized
lymphocytes may be present and are similar to small lymphocytes, but they
have moderately increased amounts of cytoplasm and may have nucleoli
visible. Lymphoblasts (large lymphocytes), commonly encountered in aspi-
rates of lymphoid tissue and lymphoid neoplasms, may be present in low
numbers in inflammatory lesions. Lymphoblasts are large cells with more
abundant cytoplasm, which typically stains basophilic. Nuclei can be vari-
ably shaped and have stippled nuclear chromatin. Distinct nucleoli are often
visible, and multiple nucleoli may be observed.
Reactive lymphocytes are lymphocytes that have been antigenically stimu-
lated. They have moderately increased amounts of basophilic cytoplasm.
Plasma cells are differentiated B lymphocytes stimulated to produce anti-
bodies. Plasma cells have a round and eccentrically placed nucleus, moderate
amounts of deeply basophilic cytoplasm, and usually a distinct clear area
located next to the nucleus (see Fig. 12). This clear area represents the Golgi
apparatus and is often located between the nucleus and the greatest volume of
cytoplasm. Some plasma cells (Mott cells) have numerous large clear vacuoles
filling their cytoplasm. These vacuoles represent retained immunoglobulin.

Eosinophils
Eosinophils are slightly larger than neutrophils. Their nuclei are seg-
mented but often divided into only two distinct lobes. Rarely, eosinophils
with perfectly round nuclei are identified in cytologic specimens. The cyto-
plasm of eosinophils contains prominent pink granules. These granules are
numerous, small, and rod shaped in cats. In dogs, eosinophil granules are
round and vary widely in size and number. The delicate densely packed gran-
ules of feline eosinophils are often less obvious than those of the dog, partic-
ularly in thick specimens that may not stain well (ie, transtracheal washes).
Also, neutrophils in exudates occasionally have mild eosinophilic stippling.
Care must be taken not to confuse neutrophils and poorly stained eosino-
phils when trying to differentiate eosinophilic from neutrophilic inflamma-
tory reactions in cats. Often, it is easier to identify feline eosinophils that
have been traumatized during slide preparation, because their granules
spread and become more obvious. If many eosinophils have been ruptured
during sample collection (eg, scraping of feline eosinophilic granuloma com-
plex lesions), high numbers of eosinophil granules are present throughout the
background of the smear and may be identified before intact cells are seen.

If a smear is composed of tissue cells rather than inflammatory cells,

what type of cells are present?
A wide variety of distinct cells types can be encountered from various
normal tissues and tumors sampled cytologically. With experience, many
of these cells can be easily recognized, particularly with the knowledge of
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1219

what structure is being sampled. Even if the cells are not immediately rec-
ognizable, however, they can generally be classified into one of three major
categories based on certain cytologic features:
1. discrete cells (or ‘‘round cells’’)
2. epithelial cells
3. mesenchymal cells
Categorizing cells into the major group to which they belong helps the
evaluator to identify the specific cell type present. Even if precise identifica-
tion cannot be made, relevant information may be gained, such as the pres-
ence of a cell type abnormal for the tissue sampled (eg, epithelial cells in a
lymph node aspirate).

Discrete cells (round cells)

Occurrence of discrete cells (what they are)
Discrete cells are a group of cells that share certain cytologic features
owing to the fact that they are present individually in tissues and not adhered
to other cells or connective tissue matrix. These are predominantly mobile
cells of hematogenous origin. Aspirates of normal lymphoid tissue, such as
spleen and lymph nodes, yield cell populations that have a discrete cell pat-
tern. Other than normal lymphoid tissue, a discrete cell pattern usually indi-
cates the presence of one of a group of tumors termed discrete cell tumors (or
round cell tumors). Recognition of discrete cell tumors is important, because
these are some of the more common neoplasms encountered in small animal
practice. Also, cells of most discrete cell tumors have cytologic characteristics
that are sufficiently distinct so as to allow for a specific diagnosis.

General cytologic characteristics of discrete cell populations

Because discrete cells are not adhered to other structures within the tis-
sues, they generally exfoliate readily during fine-needle biopsy (FNB).
Hence, the cellularity of the resulting smears is usually high. In addition, the
individual cells are usually evenly spread throughout the smear (Fig. 13).
Cell clusters or aggregates are not present; however, the extremely high cel-
lularity of the smears may result in cells being piled on top of each other in
thicker areas of the smears, resembling cell clusters. In the thinner areas of
the smears, the cells can be seen to be individually oriented.
The individual cells tend to be small to medium sized and round. If the
cells have not been traumatized during sample preparation, they have dis-
tinct cytoplasmic borders (ie, the boundary of the cells is well defined).

Specific discrete cell tumors

The discrete cell tumors are mast cell tumor, lymphosarcoma (lymph-
oma), histiocytoma, plasmacytoma, and transmissible venereal tumor
(TVT). In addition, melanomas are the great imitator, yielding cell popula-
tions that may appear discrete, epithelial, or mesenchymal.
1220 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

Fig. 13. This slide made from an aspirate of a transmissible venereal tumor shows a typical
discrete cell pattern. The slide is highly cellular, and the cells are evenly spread throughout the
smear. This pattern can usually be recognized from low-power magnification. (Courtesy of
Oklahoma State University teaching files, Stillwater, OK.)

Mast cell tumor. Mast cell tumors are the only lesions that yield highly cel-
lular smears consisting entirely (or predominantly) of mast cells. Mast cells
are recognized by their distinctive small red-purple intracytoplasmic gran-
ules (Fig. 14). Most mast cell tumors yield cells that contain a sufficient
number of granules to be easily recognized as mast cells. Sometimes, the
cells are so densely packed with granules that the cytoplasm appears dif-
fusely dark purple and the individual granules difficult or impossible to dis-
cern. In this situation, the granules are evident in cells that have been
ruptured. Because mast cell granules have such a high affinity for most cyto-
logic stains, the nucleus of a heavily granulated mast cell may appear pale or
even totally unstained, giving the cell a photographically negative look (dark
cytoplasm with pale nucleus) (Fig. 15). Some of the components of mast cell
granules are chemotactic for eosinophils. The number of eosinophils present
in smears from a mast cell tumor varies from only a few to many.

Fig. 14. Smears made from an aspirate of a mast cell tumor. Mast cells are recognized by the
variable number of distinct round intracytoplasmic granules (arrows). Rupturing of some cells
can lead to the presence of numerous free granules in the background (arrowheads). (Courtesy
of Oklahoma State University teaching files, Stillwater, OK.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1221

Fig. 15. Smears made from a heavily granulated mast cell tumor. Because the mast cell granules
take most cytologic stains so avidly, the cell nuclei are sometimes understained, appearing as a
clear area within the dark cytoplasmic granules. (Courtesy of Oklahoma State University
teaching files, Stillwater, OK.)

Cells from some mast cell tumors contain relatively few granules. If a
mast cell tumor has degranulated during or before aspiration, a percentage
of the cells may have relatively few granules. Generally, some granules are
still evident in most cells, and some cells remain heavily granulated. In addi-
tion, high numbers of free granules may be present in the background of the
slides. Aspirates of degranulated mast cell tumors may be of lower than nor-
mal cellularity because of resultant tissue edema. Anaplastic (poorly differ-
entiated) mast cell tumors may yield cells virtually devoid of granules
because the cells have not differentiated sufficiently to produce them. In this
case, the cells generally display high malignant potential (see section on eval-
uating cells for criteria of malignancy). Finally, Diff-Quik (Harleco, Gibbs-
town, NJ) sometimes fails to stain the granules of mast cell tumors. Slides
from the same tumor stained with Wright’s stain are heavily granulated.
This is an inconsistent event, occurring only in some mast cell tumors and
not in others. When it occurs, mast cells may resemble plasma cells or mac-
rophages. A diligent search of the slides usually reveals low numbers of iden-
tifiable, although poorly stained granules, in some cells. A person routinely
using Diff-Quik should always consider this possibility when evaluating a
discrete cell population.
Mast cell tumors should be evaluated for criteria of malignancy as
described later in this article. Evaluation is often limited by the fact that the
granules obscure nuclear detail. All mast cell tumors have the potential for
malignancy. A small percentage of tumors composed of heavily granulated
and well-differentiated cells are behaviorally malignant. Most tumors com-
posed of poorly granulated cells (stained with Wright’s stain) that display
marked cytologic criteria of malignancy are malignant. Examination of
buffy coat preparations, bone marrow aspirates, and aspirates of any en-
larged lymph nodes or abdominal organs (particularly the liver and spleen)
may be useful in detecting systemic spread of the mast cell tumor.
1222 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

Lymphosarcoma. Most lymphosarcomas of dogs and cats are high-grade

tumors composed predominantly of large blastic lymphoid cells. Detailed
morphologic classification of these cells is generally not necessary, and they
are typically referred to as ‘‘lymphoblastic’’ lymphosarcoma or lymphoma.
If lymphoblasts comprise greater than 50% of the cells in a highly cellular
smear containing mostly intact cells, a diagnosis of lymphosarcoma can reli-
ably be made. Lymphoblasts are usually easy to differentiate from the cells
of other discrete tumors based on their higher nuclear/cytoplasmic (N/C)
ratio and intensely basophilic cytoplasm. Also, in aspirates from lympho-
sarcoma, there are usually numerous small but variably sized basophilic
fragments of cytoplasm (‘‘lymphoglandular bodies’’) scattered among the
cells. Lymphoid cells, particularly lymphoblasts, are fragile cells and easily
ruptured during slide preparation. If an overwhelming majority of the cells
on the smears are ruptured, it is difficult to be confident that the remaining
cells accurately represent the cell population in the lymph node. In this sit-
uation, additional samples must be collected for a diagnosis.
Occasionally, low-grade lymphosarcoma is encountered. Such tumors
may yield predominantly small lymphocytes. Such tumors are difficult to dif-
ferentiate from normal or reactive lymphoid tissue based solely on cytology
and often require histologic biopsy for a definitive diagnosis.

Histiocytoma. Histiocytomas are benign tumors of macrophage origin that

are common in young dogs (Fig. 16). Tumor cells are medium sized and
slightly larger than neutrophils. Nuclei are generally round to oval but may
be indented to irregular in shape. The nucleus has finely stippled chromatin
and may have indistinct nucleoli. The cells have a moderate amount of light
blue-gray cytoplasm. If there is a significant amount of protein-rich tissue
fluid present between cells, the cytoplasm of the cells may appear lighter
than the background, or cell borders may be indistinct.
Histiocytomas usually regress spontaneously within a few weeks to
months. Regression is associated with an infiltration of small lymphocytes

Fig. 16. Smears made from a histiocytoma. Histiocytomas have moderate amounts of light-
colored cytoplasm. Nuclei are usually round to oval but may be indented, kidney shaped, or
irregular (arrows). (Courtesy of Oklahoma State University teaching files, Stillwater, OK.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1223

into the tumor. Therefore, aspirates from these tumors sometimes contain a
mixture of tumor cells and small lymphocytes (Fig. 17). The presence of
small lymphocytes among the larger tumor cells sometimes leads to misiden-
tification of histiocytoma cells as lymphoblasts. The irregularly shaped
nuclei, light color and volume of the cytoplasm, and lack of lymphogland-
ular bodies help to differentiate histiocytoma cells from lymphoid cells.

Plasmacytoma. Tumors of plasma cell origin include multiple myeloma

(plasma cell myeloma), a systemic tumor arising primarily in the bone mar-
row, and extramedullary plasmacytomas. Extramedullary plasmacytomas
are commonly cutaneous tumors but have been described from other sites,
including the gastrointestinal tract. Cutaneous plasmacytomas are usually
benign, although malignant behavior has been reported in some tumors.
A higher percentage of plasmacytomas arising from the gastrointestinal
tract have been reported as having malignant behavior.
Well-differentiated plasmacytomas yield cells that resemble normal plasma
cells (Fig. 18). Distinguishing features include eccentrically placed, small,
round nuclei surrounded by a moderate amount of deeply basophilic cyto-
plasm with a distinct perinuclear clear zone. Poorly differentiated plasma-
cytomas may yield a less distinct population of discrete cells that demonstrate
significant cytologic criteria of malignancy. Binucleate and multinucleated
cells are common in both well-differentiated and poorly differentiated tumors
(Fig. 19). This and a lack of lymphoglandular bodies help to differentiate
these tumors from lymphosarcoma. Some plasma cells have a distinct red
color to the periphery of their cytoplasm and are referred to as ‘‘flame’’ cells.
Rarely, an eosinophilic extracellular matrix representing amyloid (composed
of immunoglobulin light chain) is seen among the neoplastic cells.

Transmissible venereal tumor. Except in certain geographic areas, TVTs

are less commonly encountered than the other discrete cell tumors. These
tumors are often present on the external genitalia but may occur in other
locations as well.

Fig. 17. Smears made from a histiocytoma. Regression of a histiocytoma is associated with an
infiltration of small lymphocytes (arrowheads), which may be more numerous than the his-
tiocytoma cells (arrows). (Courtesy of Oklahoma State University teaching files, Stillwater, OK.)
1224 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

Fig. 18. Smears made from an extramedullary plasmacytoma. Cells show a typical discrete cell
pattern. Many cells resemble mature plasma cells with eccentric nuclei and distinct perinuclear
clear areas (arrows). (Courtesy of Oklahoma State University teaching files, Stillwater, OK.)

The cells from a TVT are typically more pleomorphic than those from most
other discrete cell tumors (Fig. 20). They have moderate amounts of smoky to
light blue cytoplasm with sharply defined cytoplasmic boundaries. A promi-
nent characteristic of TVT cells, which can help to distinguish them from other
discrete cell tumors, is the presence of numerous, distinctly walled, cytoplas-
mic vacuoles. These vacuoles can also be found extracellularly, appearing as
clear areas against a proteinaceous background of tissue fluid. Nuclei show
moderate to marked anisokaryosis and have a coarse nuclear chromatin pat-
tern. Nucleoli may be prominent, and mitotic figures are common.

Epithelial cells
Occurrence of epithelial cells (what they are)
Normal epithelial cells are commonly encountered in many cytologic
preparations. Surface epithelium is present in most surface scrapings and

Fig. 19. Smears made from an extramedullary plasmacytoma. Binucleated and multinucleated
cells (arrows) are common in tumors of plasma cell origin. (Courtesy of Oklahoma State
University teaching files, Stillwater, OK.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1225

Fig. 20. Smears made from a transmissible venereal tumor (TVT). Cells from a TVT are char-
acterized by numerous clear vacuoles within the cytoplasm of the cells (arrows) and free in the
background (arrowheads). (Courtesy of Oklahoma State University teaching files, Stillwater,OK.)

swabs (eg, squamous cells from skin scrapings and nasal or vaginal swabs),
washings (eg, columnar cells from transtracheal washes), and as the result of
normal exfoliation (transitional cells from urine sediments). In addition, epi-
thelial cells are the major cellular component of smears made from FNB of
many parenchymal organs (eg, hepatocytes and bile duct epithelium from
liver aspirates, renal tubular cells from kidney aspirates) and glandular aspi-
rates (eg, mammary, prostate).
Epithelial cells may also originate from a hyperplastic proliferation or
neoplasm. Cells from benign epithelial tumors may be difficult or impossible
to differentiate from their normal or hyperplastic counterparts based solely
on cytology. Combining clinical and cytologic findings often allows the
diagnosis of a benign epithelial proliferation to be made. For example, a dis-
crete wart-like mass on an older dog that yields numerous clusters of ‘‘nor-
mal’’-appearing sebaceous epithelial cells suggests a sebaceous adenoma or
sebaceous gland hyperplasia.
Epithelial cells that display sufficient cytologic criteria of malignancy
indicate the presence of a carcinoma or adenocarcinoma. A specific diag-
nosis of cell type may or may not be possible based solely on cytology, de-
pending on how well differentiated and characteristic the cells are.
Histopathology may be required for a more specific diagnosis of cell type.
The ability to confirm the presence of a malignant epithelial tumor by cytol-
ogy is often sufficient to guide clinical management of a case, however.

General cytologic characteristics of epithelial cell populations

A main feature of epithelial cells is cell-to-cell adhesion (Fig. 21). Normal
epithelial cells are typically present in variably sized sheets or clumps. Some-
times, the area of adhesion between individual cells can be seen (Fig. 22).
True cell clustering from cell-to-cell adhesion must be differentiated from
crowding of cells in highly cellular aspirates of any cell type. This can usu-
ally be accomplished by looking at thinner areas of the smear. In thin areas,
1226 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

Fig. 21. Smears made from an aspirate of a feline kidney. Numerous renal tubular epithelial
cells are present. Epithelial cells tend to form cell clusters. Feline renal tubular cells may have
numerous lipid vacuoles within their cytoplasm. (From Cowell RL, Tyler RD, Meinkoth JH,
editors. Diagnostic cytology and hematology of the dog and cat. 2nd edition. St. Louis: Mosby;
1999. p. 204; with permission.)

if the cells are still present in clusters but are separated by acellular areas,
cell-to-cell adhesion is documented.
Mesenchymal cells are sometimes held together by an extracellular
matrix, resulting in large aggregates of cells that resemble cell adhesion.
Often, this extracellular matrix is apparent as brightly eosinophilic homoge-
nous material between the cells and can be used to identify the type of cell
present.
Mature squamous epithelial cells, often found in samples collected from
surface swabs or scrapings, tend to be more individually oriented and may
not show prominent cell clustering. Less well-differentiated squamous cells,
usually present in scrapings along with mature cells, are more cohesive and
have a greater tendency to be in clusters.

Fig. 22. Aspirate from a perianal adenoma. The areas of cell adhesion can be seen in some cells
(arrows). (From Cowell RL, Tyler RD, Meinkoth JH, editors. Diagnostic cytology and
hematology of the dog and cat. 2nd edition. St. Louis: Mosby; 1999. p. 44; with permission.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1227

Epithelial cells can be extremely large cells and have abundant cytoplasm.
Whereas normal epithelial cells vary in size from small (basal cells) to large
depending on the specific type, the presence of extremely large cells on a
cytologic smear suggests an epithelial origin. Epithelial cells are round to
columnar to caudate and have distinct and sharply defined cytoplasmic bor-
ders. Cytoplasmic borders within cell clusters may be difficult to discern;
however, the outer edges of the cells in the clusters are typically clearly
demarcated. Nuclei of epithelial cells are generally round to somewhat oval.

Epithelial criteria specific to certain cell types

Mature squamous epithelial cells are easily recognized by their angular
rather than round outline. Cells become progressively angular with matura-
tion. The nuclei of mature squamous cells become progressively smaller and
pyknotic and eventually disappear.
Normal epithelial cells from the respiratory tract and gastrointestinal
tract are distinctly columnar. Cell clusters may show long rows of cells, with
nuclei lined up at the basal end of the cell. Sometimes, cilia can be seen at the
apical surface.
Although most tissue architecture is lost during fine-needle aspiration,
epithelial cells of glandular origin sometimes show evidence of tubule or aci-
nus formation. Secretory epithelial cells may show cytoplasmic vacuolation,
which is often confined to the area immediately surrounding the nucleus
(perinuclear vacuolation).

Mesenchymal cells
Occurrence of mesenchymal cells (what they are)
Mesenchymal cells are cells of connective tissue origin. Because blood is a
connective tissue, hematopoietic cells (including many of the cells described
as ‘‘discrete cells’’) are included as mesenchymal tissue. Because these hema-
topoietic cells have a cytologic appearance so distinct from the other con-
nective tissues, they are typically classified separately. Most often, the
discussion of ‘‘mesenchymal cells’’ in cytology texts implies ‘‘stromal’’ con-
nective tissue cells.
Most normal connective tissues exfoliate no cells when sampled by FNB.
Fibroblasts and fibrocytes are the only normal mesenchymal cells commonly
encountered. Clusters of normal stromal cells can be seen in aspirates of
internal organs, particularly the spleen. Scattered individual fibroblasts may
be seen in aspirates from virtually any tissue. Reactive fibroblasts may be
present in significant numbers in aspirates from areas of inflammation or tis-
sue repair (eg, surgical scars). Reactive fibroblasts may show many of the
cytologic criteria of malignancy, so caution should be taken in evaluating
mesenchymal cells when a significant inflammatory response is present.
Reactive fibroblasts should be suspected when scattered mesenchymal cells
are present along with a population of inflammatory cells.
1228 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

Highly cellular smears that predominantly contain a pure population of

mesenchymal cells are likely to indicate a mesenchymal neoplasia. Malignant
tumors of mesenchymal origin are, by definition, sarcomas, although some
tumors are inappropriately named (eg, malignant fibrous histiocytoma).

General cytologic characteristics of mesenchymal cell populations

As previously mentioned, aspirates of normal mesenchymal tissue are
usually sparsely cellular because of the cohesive nature of connective tissue.
Benign mesenchymal tumors tend to exfoliate few cells, and samples of diag-
nostic quality may be difficult to obtain. In contrast, malignant mesenchy-
mal tumors may yield highly cellular aspirates. The cells are usually
individually oriented, although large clusters may be present, particularly
if held together by an extracellular matrix (see below).
Mesenchymal cells are often elongated cells with cytoplasm that tapers in
one or more directions (Fig. 23). These are commonly referred to as ‘‘spindle
cells.’’ Typically, mature mesenchymal cells are extremely elongated
fusiform cells, whereas less-differentiated cells are round and plump. Nuclei
of mesenchymal cells may be round or rod-like. Aspirates from malignant
mesenchymal tumors often show a mixture of cells of various shapes, so the
entire population must be examined to determine the cell type.
In contrast to the other cell types (discrete and epithelial), the cytoplasmic
borders of mesenchymal cells are often indistinct (Fig. 24), and the cytoplasm
may blend imperceptibly with the background, making it nearly impossible
to distinguish the limits of the cell membrane. Ruptured cells may also have
indistinct cytoplasmic borders, but the nuclear membrane is usually dis-
rupted in these traumatized cells, whereas the nuclear outline of intact mes-
enchymal cells is well defined. Another characteristic of mesenchymal cells is
production of an extracellular matrix. This is seen as variably eosinophilic
material present between cells, often holding them in large aggregates.

Fig. 23. A ‘‘spindle cell’’ from a malignant tumor of mesenchymal origin (myxosarcoma). Note
the elongated appearance with tapered cytoplasm. (From Cowell RL, Tyler RD, Meinkoth JH,
editors. Diagnostic cytology and hematology of the dog and cat. 2nd edition. St. Louis: Mosby;
1999. p. 49; with permission.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1229

Fig. 24. Cells from a tumor of mesenchymal origin. The cytoplasmic boundary of these cells is
extremely indistinct, and the cytoplasm seems to fade gradually into the background. This is
another common feature of cells of mesenchymal origin. (Courtesy of Oklahoma State
University teaching files, Stillwater, OK.)

These descriptions outline general characteristics. The entire cell popula-

tion present should be carefully evaluated, because no single criterion can
definitively identify a cell population. Sometimes, particularly with poorly
differentiated malignant tumors, the cells show criteria of more than one cat-
egory. In these cases, it may be impossible to accurately classify the type of
cell present. If the cell type cannot be categorized, the cells should be eval-
uated for criteria of malignancy, because the identification of a malignant
tumor may be sufficient information to direct management of the case. Sur-
gical biopsy and histopathology may be able to provide a more specific diag-
nosis as to tissue of origin, if needed.

Do the tissue cells present display significant criteria of malignancy?

Tissue cells should be evaluated for cytologic criteria of malignancy. If
sufficient criteria are present, a diagnosis of malignant neoplasia can be
made. Cells from normal tissue, hyperplastic tissue, and benign neoplasia
generally do not contain significant criteria of malignancy.

Cytologic criteria of malignancy

Although some assessment of the arrangement of cells within cell clusters
can be made, cytologic samples lack the architectural information that is
available on histologic sections. Therefore, factors like disruption of normal
architecture and invasion of suspect cells into adjacent normal tissue or lym-
phatics usually cannot be detected. Evaluation of malignant potential in
cytology specimens revolves around evaluating cell populations for lack of
differentiation, rapid cell division, and cellular atypia. In general, benign
lesions yield morphologically uniform populations of well-differentiated
cells, whereas malignant tumors are characterized by variability of cell fea-
tures. Cytologic criteria are divided into general criteria of malignancy and
1230 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

nuclear criteria of malignancy. Nuclear criteria of malignancy are more reli-

able because they are less likely to be induced by nonneoplastic processes,
such as inflammation-induced dysplasia. No single criterion indicates the
presence of malignancy, and any of the features described below may be
seen in certain cells or cell populations.

General criteria of malignancy

Anisocytosis and macrocytosis. Anisocytosis (Fig. 25) refers to variation in
cell size, whereas macrocytosis (Fig. 26) refers to exceptionally large cells.
Both are atypical findings in most cell populations, although there are
exceptions. In samples of normal or reactive lymphoid tissue, variation in
cell size is an expected finding because of the variety of different cell types
present (eg, mature lymphocytes, lymphoblasts, plasma cells). In contrast,
lymphoid malignancy (lymphosarcoma) yields a uniform monotonous pop-
ulation of lymphoblasts. In scrapings from skin surfaces and some vaginal
swabs, there can be moderate to marked anisocytosis of the squamous
epithelial cells. This relates to the fact that such samples can collect squa-
mous cells of varying degrees of maturation, ranging from small and
immature basal or parabasal cells to mature, fully keratinized, superficial
squamous cells. Transitional epithelial cells also show moderate anisocytosis
as a normal feature of that cell type.
Some degree of anisocytosis is normal in any cell population. The ten-
dency of many beginning cytologists is to overinterpret normal variability
in cell size rather than to ignore significant variation when it is present; cau-
tion is thus warranted in evaluating subjective parameters. ‘‘Significant’’
variation in cell size is when one cell is multiple times the size of other cells
from the same population. There are no easily defined objective criteria of
macrocytosis, although this usually indicates cells that are decidedly larger
than what is normal for the cell population. Obviously, this requires having

Fig. 25. Aspirate from a transitional cell carcinoma. The cells show significant anisocytosis and
anisokaryosis. Some cells (arrowhead) are several times larger than other cells in the population
(arrows). Several binucleated cells and a multinucleated cell are also present. (Courtesy of
Oklahoma State University teaching files, Stillwater, OK.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1231

Fig. 26. Low-power photomicrograph of smears made from thoracic fluid of a cat with a
metastatic carcinoma. Atypically macrocytic cells with macronuclei are present. A neutrophil
(arrow) is present for size comparison. (Courtesy of Oklahoma State University teaching files,
Stillwater, OK.)

sufficient experience to recognize the limits of normal. Macrocytic cells are

most commonly observed in tumors of epithelial origin.

Hypercellularity. Malignant tumors tend to exfoliate high numbers of cells,

even when arising from tissues that would not normally exfoliate any cells.
A classic example of this is primary bone tumors, such as osteosarcoma and
chondrosarcoma. Normal bone obviously exfoliates few, if any, cells. Aspi-
rates from primary bone tumors are often highly cellular, however. Cells in
malignant tumors are often anaplastic and have not differentiated to the
point where they develop cell receptors or produce the extracellular matrix
that makes them adhesive to other tissues in the body. Therefore, these cells
exfoliate well by FNB. Similarly, these cells often demonstrate a loss of
cohesion. Highly malignant epithelial tumors yield cells that distribute in
more of a discrete cell pattern, with fewer cell clusters.
Obviously, hypercellularity as a criterion of malignancy must be viewed
in terms of the tissue sampled. Inflammatory lesions, lymphoid tissue, and
some other tissues normally yield high numbers of cells. In these instances,
hypercellularity cannot be considered a criterion of malignancy. Highly cel-
lular slides containing a single population of mesenchymal cells is not a nor-
mal finding, however. Hypercellularity is also important, because the sample
is more likely to be representative of the lesion than if relatively few cells are
present. A definitive diagnosis of malignancy should be made with extreme
caution if the sample is of low cellularity.

Pleomorphism. This term refers to variability in the shape of cells. Pleomor-

phism may be normal if more than one cell type is present on a smear. Also, pleo-
morphism among cells of a single cell type is seen in some normal tissue, such as
transitional cells from the urinary tract and samples containing squamous cells
of varying degrees of maturation (skin scrapings and vaginal smears).
1232 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

Nuclear criteria of malignancy

Anisokaryosis and macrokaryosis. These terms refer to variation in nuclear
size (see Fig. 25) and excessively large nuclei (see Fig. 26), respectively.
Nuclei that are multiple times the size of those in other cells within the same
population represent significant anisokaryosis. In some malignant tumors,
particularly carcinomas, macronuclei, which may be larger than some entire
cells of the same population, may be present.
Anisokaryosis is a normal finding in samples containing squamous epi-
thelial cells. As squamous cells mature, the nucleus becomes small and
pyknotic, eventually disappearing from the cell.

Multinucleation. Cells with multiple nuclei may be seen in malignant tumors

of any cell type. Multinucleation is particularly important when anisokaryo-
sis is present among nuclei within a single cell. Multinucleation in neoplastic
cells results from nuclear division without cell division. Usually, even num-
bers of nuclei are present. Odd numbers of nuclei indicate atypical nuclear
division and are an important finding (Fig. 27).
Inflammatory lesions may have macrophages that are multinucleated
(‘‘multinucleated inflammatory giant cells’’). Osteoclasts are normally multi-
nucleated. Megakaryocytes, which are commonly present in the spleen as a
reflection of extramedullary hematopoiesis, may have multiple nuclear lobes
that may sometimes appear to be multiple individual nuclei.

Abnormal nuclear/cytoplasmic ratio. The N/C ratio refers to the relative

areas occupied by the nucleus and cytoplasm of the cell. A low N/C ratio
indicates a cell with a relatively small nucleus and vast amounts of cyto-
plasm (Fig. 28). In contrast, cells with minimal amounts of cytoplasm have

Fig. 27. A multinucleated cell with an odd number of nuclei. Prominent large nucleoli of
varying size are also present. The erythrocytes and neutrophil can be used for size comparison.
(Courtesy of Oklahoma State University teaching files, Stillwater, OK.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1233

Fig. 28. Smears from an aspirate of a normal salivary gland (accidentally aspirated instead of
the submandibular lymph node). The cells have a relatively small mature nucleus and abundant
cytoplasm, yielding a low nuclear/cytoplasmic ratio. (Courtesy of Oklahoma State University
teaching files, Stillwater, OK.)

a high N/C ratio (Fig. 29). Epithelial and mesenchymal cells having a high
or variable N/C ratio are suggestive of malignancy. A high N/C ratio is a
particularly important finding in extremely large cells, because some small
cells (eg, mature lymphocytes, basal epithelial cells) normally have a high
N/C ratio. A high N/C ratio in a large cell generally indicates a poorly differ-
entiated cell.

Abnormal nucleoli. Nucleoli are areas within the nucleus that are respon-
sible for production of ribosomal RNA. All cells have nucleoli, but they
are usually small and often not readily visible. Nucleoli that are large

Fig. 29. Smears from an aspirate of an oral melanoma. These relatively large cells (compare
with size of erythrocytes) have a high nuclear/cytoplasmic ratio. (Courtesy of Oklahoma State
University teaching files, Stillwater, OK.)
1234 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235

(macronucleoli), atypically shaped (angular nucleoli), or vary in size (aniso-

nucleoliosis) are strong indicators of malignancy (see Fig. 27). Nucleoli in
normal cells are small, approximately 1 to 2 lm in diameter. Nucleoli greater
than 5 lm in diameter are suggestive of malignancy. Erythrocytes can be used
as a reference for evaluating the size of nucleoli. Canine erythrocytes are 7 to
8 lm in diameter, whereas feline erythrocytes are approximately 5 to 6 lm
in diameter (if well spread). Normal cells have round nucleoli. Fusiform,
pleomorphic, or angular nucleoli are indicative of malignancy. Diff-Quik
often stains nucleoli more prominently than other cytologic stains, so this
must be considered when evaluating cells for malignant potential.

Abnormal mitosis. Mitotic figures are rare in samples from most normal
tissue (except lymphoid tissue and bone marrow). Increased numbers of
mitoses or mitotic figures showing abnormal alignment of chromosomes
are suggestive of malignancy (Fig. 30).

Coarse nuclear chromatin. Nuclear chromatin patterns are not as distinctive

and evident in cells stained with Romanowsky type stains as they are when
cells are wet fixed and stained with Papanicolaou type stains. Still, an abnor-
mally coarse nuclear chromatin pattern is often visible in malignant cells.

Nuclear molding. Sometimes, the nucleus of one cell can be seen to deform
around the nucleus of another cell (or another nucleus within a multi-
nucleated cell). This is referred to as nuclear molding and indicates rapid
growth and loss of contact inhibition.

General cautions regarding evaluating cytologic criteria of malignancy

As stated previously, there is no single cellular feature that reliably distin-
guishes malignant from benign cells. A reliable diagnosis of malignancy can
usually be made if three or more nuclear criteria of malignancy are present

Fig. 30. Smears from a malignant tumor in the lung of a cat. One abnormal mitotic figure is
present. (From Cowell RL, Tyler RD, Meinkoth JH, editors. Diagnostic cytology and
hematology of the dog and cat. 2nd edition. St. Louis: Mosby; 1999. p. 182; with permission.)
 J.H. Meinkoth, R.L. Cowell / Vet Clin Small Anim 32 (2002) 1209–1235 1235

in most of the cells present in the smear. If cytologic features of malignancy

are not unambiguous, the diagnosis should be confirmed with a biopsy and
histologic evaluation. It is imperative that a representative sample (highly
cellular) is available and that only intact (nontraumatized), well-spread, well
stained cells are evaluated. Nucleoli are often more distinct in understained
cells present in thick areas of the smears. When cells are partially ruptured,
the nuclear chromatin spreads and uncoils. This results in the nucleus
appearing larger than it really is and also makes nucleoli normally obscured
by the condensed chromatin more visible.
Also, caution should be used in diagnosing neoplasia in the face of
inflammation. Inflammation can induce dysplastic changes in tissue cells
that can mimic neoplasia. Inflammatory lesions may also contain large epi-
thelioid macrophages and proliferating fibroblasts, both of which may have
some features that are similar to those of malignant cells.
Conversely, not every malignant tumor shows marked cellular atypia and
variability. Some tumors may yield relatively uniform populations of cells
yet exhibit aggressive biologic behavior. This finding is frequently encoun-
tered in endocrine tumors. Most thyroid tumors in dogs are malignant, yet
samples from some thyroid carcinomas contain relatively uniform cells with-
out marked criteria of malignancy. The same situation is described in other
tumors of endocrine and neuroendocrine origin. Many other well-differenti-
ated carcinomas (eg, perianal gland tumors) are difficult to distinguish from
a benign proliferation based solely on cytologic examination. In some cases,
this differentiation is also difficult to make on histologic examination. In
these cases, experience with the peculiarities of each tumor type is helpful
in forming a correct interpretation of the sample.

Suggested readings
Tyler RD, Cowell RL, Baldwin CJ, et al. Introduction. In: Cowell RL, Tyler RD, Meinkoth JH,
editors. Diagnostic cytology and hematology of the dog and cat. 2nd edition. St. Louis:
Mosby; 1999. p. 1–19.
Tyler RD, Cowell RL, Meinkoth JH. Cutaneous and subcutaneous lesions. In: Cowell RL,
Tyler RD, Meinkoth JH, editors. Diagnostic cytology and hematology of the dog and cat.
2nd edition. St. Louis: Mosby; 1999. p. 20–51.
Raskin RE. General categories of cytologic interpretation. In: Raskin RE, Meyer DJ, editors.
Atlas of canine and feline cytology. Philadelphia: WB Saunders; 2001. p. 19–33.
Lumsden JH, Baker R. Cytopathology techniques and interpretation. In: Baker R, Lumsden
JH, editors. Color atlas of cytology of the dog and cat. St. Louis: Mosby; 2000. p. 7–20.