Veterinary Clinical Pathology ISSN 0275-6382
CASE SERIES
Cytologic features of clear cell adnexal carcinoma in 3 dogs
Martina Piviani, Melissa D. Sanchez, Reema T. Patel
Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA, USA
Key Words
Canine, cytokeratin, cytology, epithelial
neoplasia, follicular stem cell carcinoma,
vimentin
Correspondence
Reema T. Patel, Department of Pathobiology,
School of Veterinary Medicine, University of
Pennsylvania, 3900 Delancey Street,
Philadelphia, PA 19104, USA
E-mail: rtpatel@vet.upenn.edu
DOI:10.1111/j.1939-165X.2012.00434.x
Introduction
Clear cell adnexal carcinoma is a rare primary cutane-
ous neoplasm of adnexa without definitive apocrine,
sebaceous, or follicular differentiation.1 The most dis-
tinctive histologic feature is the presence of clear
cells with clear or vacuolated cytoplasm. This tumor
has been described only in dogs and has been reported
previously as clear cell hidradenocarcinoma or follicu-
lar stem cell carcinoma.2,3 The likely cell of origin is a
cutaneous epithelial stem cell.1,3 Differential diagnoses
include sebaceous carcinoma, clear cell basal cell carci-
noma, liposarcoma, and balloon cell melanoma. Histo-
Vet Clin Pathol 0/0 (2012) 17 2012 American Society for Veterinary Clinical Pathology
Background: Clear cell adnexal carcinoma is a rare cutaneous adnexal
neoplasm without definitive apocrine, sebaceous, or follicular differentia-
tion. Differential diagnoses include sebaceous carcinoma, liposarcoma, and
balloon cell melanoma. Immunohistochemical analysis, with positive
immunoreactivity for pancytokeratin and vimentin, aids in the diagnosis.
Cytologic features of clear cell adnexal carcinoma have not been described
previously.
Objective: The aim of this study was to describe cytologic features of
canine clear cell adnexal carcinoma.
Methods: Fine-needle aspirates (FNA) obtained prior to biopsy of cutane-
ous neoplasms with a histologic diagnosis of clear cell adnexal carcinoma
confirmed by immunohistochemical analysis were reviewed retrospec-
tively. Slides prepared from FNA had been stained with modified Wright-
Giemsa or automated aqueous Romanowsky stain.
Results: Of 20 neoplasms diagnosed as clear cell adnexal carcinoma in
dogs, FNA of the mass had been performed in 3 cases. Cytologic features
were similar and included high cellularity, marked cellular pleomorphism,
loose arrangement of neoplastic cells, and a light blue to gray background
resulting from streaming of cytoplasm from ruptured cells. Neoplastic cells
were oval to polygonal to spindle-shaped with wispy cytoplasmic projec-
tions. Cytoplasmic eosinophilic stippling, globular deposits, or pink needle-
shaped inclusions were noted. Criteria of malignancy included marked
anisocytosis, anisokaryosis and anisonucleoleosis, multinucleation, karyo-
megaly, macronucleoli, and atypical mitotic figures.
Conclusions: Clear cell adnexal carcinoma should be included in the cyto-
logic differential diagnosis for a canine cutaneous mass even if an epithelial
origin is not readily identified owing to lack of characteristic epithelial fea-
tures, such as highly cohesive cell clusters, evident cell-to-cell junctions,
and distinct cytoplasmic edges.
pathologic and immunohistochemical features and
biological behavior, but not cytologic features, of clear
cell adnexal carcinoma have been previously
described.1 The objective of this study was to describe
the cytologic features of canine clear cell adnexal
carcinoma.
Materials and Methods
The pathology database of the Matthew J. Ryan Veteri-
nary Hospital of the University of Pennsylvania (MJR-
VHUP) was searched for all cases with a histologic
1
Cytology of canine clear cell adnexal carcinoma Piviani et al

diagnosis of clear cell adnexal carcinoma for which a
concurrent fine-needle aspirate (FNA) had been sub-
mitted for cytologic evaluation. A total of 20 cases, sub-
mitted from January 2005 to January 2011, had a
histologic diagnosis of clear cell adnexal carcinoma,
but cytologic evaluation was performed in only 3 dogs.
For all 3 dogs, the cytologic sample had been prepared
by gently smearing material obtained by aspiration on
a glass slide. For 2 of the dogs, the sample had been
submitted to the Cytopathology Service of MJR-VHUP,
prior to biopsy; for the 3rd dog, slides were obtained
from an external laboratory (Antech Diagnostics, Lake
Success, NY, USA).
Signalment and clinical information were
obtained from the submission forms or medical
records. All slides, including the histologic sections and
cytologic samples, were reviewed by board-certified
anatomic (MDS) and clinical (RTP) pathologists, in
addition to a clinical pathology resident (MP). Slides of
the FNAs had been air-dried and stained with modified
Wright-Giemsa (dogs 1 and 2) or aqueous Romanow-
sky stain (Aerospray 7120, Wescor Inc., Logan, UT,
USA) (dog 3). Excisional biopsies had been performed
in all 3 dogs, by 8-mm punch biopsy in dog 1 and by
elliptical incision with wide margins in dogs 2 and 3.
Tissue specimens were fixed in 10% neutral buffered
formalin, sectioned and stained with H&E and periodic
acid-Schiff (PAS) with and without pretreatment with
diastase. Immunohistochemical analysis for expression
of pancytokeratin AE1/AE3, vimentin, Melan A, PNL-2,
Cam 5.2, CD79a, calponin, S-100, and a-smooth mus-
cle actin (Table 1) was performed. All antibodies had
been validated for canine tissue in our laboratory. Posi-
tive controls for each antibody were analyzed using
sections of canine tissues expressing the relevant anti-
gen, including haired skin for cytokeratin, vimentin,
Cam 5.2, and calponin; malignant oral melanoma for
PNL-2; cutaneous melanocytoma for Melan A; lymph
node for CD79a; brain and peripheral nerves for S-100;
and a composite slide containing sections of lung and
urinary bladder, among other tissues, for a-smooth
muscle actin. Test samples with omission of primary
antibody were used as negative controls.
Results
Clinical and cytologic findings
Dog 1 was an 8-year-old intact male Boxer presented
for continued treatment and restaging of previously
diagnosed stage IVb lymphoma. During physical
examination a 34 mm pigmented, pedunculated
nodule was noted under the right eye. Dog 2 was a
4-year-old female spayed Pug presented for evaluation
of a slowly growing 34 cm cutaneous mass on the
dorsum that was first noticed 1 year earlier. The mass
had an irregular shape and no apparent deep fascial
connections. Dog 3 was a 4-year-old female spayed
mixed breed dog with a 3 9 4 cm haired, firm, slightly
lobulated cutaneous mass on the dorsum over the mid-
line. The mass had grown slowly during the past few
months. All dogs recovered uneventfully from the sur-
gical biopsy procedures, and metastasis and local recur-
rence were not reported over 79 months.
Slides made from all 3 masses were highly cellular
and contained large numbers of markedly pleomorphic
round to spindle-shaped to polygonal neoplastic cells
that were individualized or arranged in loose clusters
amid a background of many erythrocytes, proportional
numbers of leukocytes, frequent bare nuclei, many
clear vacuoles, and abundant light blue to gray amor-
phous material consistent with cytoplasmic debris
from ruptured cells (Figures 1A and 2A). Smears
from dogs 2 and 3 also contained low numbers of

Table 1. Antibodies used in immunohistochemical analysis of 3 canine clear cell adnexal carcinomas.

Primary antibody Monoclonal (M)/ Secondary

Primary antibody Primary antibody source Species dilution polyclonal (P) antibody

AE1/AE3 (pancytokeratin) Chemicon, Temecula, CA Mouse anti-human 1:500 M ENVision+*

Vimentin DAKO, Carpinteria, CA Mouse anti-human 1:50 M ENVision+
Melan A Novocastra, Bannockburn, KS Mouse anti-human 1:25 M ENVision+
PNL-2 DAKO Mouse anti-human Prediluted M ENVision+
Cam 5.2 Becton Dickinson, Franklin Lakes, NJ Mouse anti-human 1:20 M ENVision+
CD79a DAKO Mouse anti-human 1:50 M ENVision+
Calponin DAKO Mouse anti-human 1:50 M ENVision+
S-100 DAKO Rabbit anti-human 1:800 P LSAB2
a-smooth muscle actin DAKO Mouse anti-human 1:25 M LSAB2

*ENVision+, DAKO (dextran polymerhorseradish peroxidase [HRP], anti-mouse).

LSAB2, DAKO (streptavidin/biotinHRP, anti-mouse or anti-rabbit).

2 Vet Clin Pathol 0/0 (2012) 17 2012 American Society for Veterinary Clinical Pathology
Piviani et al Cytology of canine clear cell adnexal carcinoma

A B

C D

Figure 1. Fine-needle aspirate of a cutaneous nodule under the right eye of an 8-year-old intact male Boxer (Dog 1). Modified Wright-Giemsa stain. (A)
Note cells arranged individually or in loose clusters. Bar = 100 lm. (B) Neoplastic cells display multinucleation, marked anisokaryosis, and prominent
and sometimes large nucleoli. Bar = 50 lm. (C) Neoplastic cells have indistinct cytoplasmic borders. The light blue to gray background is the result
streaming of cytoplasm from ruptured cells. Bar = 20 lm. (D) Few neoplastic cells contain fibrillar to needle-shaped bright pink material (arrow).
Bar = 10 lm.

A B

C D

Figure 2. Fine-needle aspirate of a cutaneous mass on the dorsum of a 4-year-old female spayed mixed breed dog (Dog 3). Automated aqueous stain,
Aerospray 7120. (A) Note the presence of fine dark pigment consistent with melanin within the cytoplasm of the neoplastic cell in the center (arrow).
Bar = 100 lm. (B) Bizarre mitotic figures are present (arrow). Bar = 50 lm. (C) Neoplastic cells are oval to fusiform with wispy cytoplasmic projections.
Bar = 20 lm. (D) Note the eosinophilic stippling condensing in globular aggregates of pink material in the paranuclear area of neoplastic cells.
Bar = 10 lm.

Vet Clin Pathol 0/0 (2012) 17 2012 American Society for Veterinary Clinical Pathology 3
Cytology of canine clear cell adnexal carcinoma
inflammatory cells, which were predominantly macro-
phages, small lymphocytes, and plasma cells admixed
with few neutrophils. Macrophages appeared foamy
and contained erythrocytes or granular, blue to black
or golden pigment, interpreted as hemosiderin and he-
matoidin, respectively. Neoplastic cells had variable N:
C ratios, with large round to oval nuclei, granular to
finely reticular chromatin, and 13 prominent, round,
basophilic nucleoli (Figure 1B). Cytoplasmic bound-
aries appeared indistinct within cell clusters and cyto-
plasmic streaming was frequently noted (Figures 1C
and 2C). Cytoplasm was moderate to abundant in
amount, was light blue, and occasionally formed wispy
projections (Figures 1C, 2B, and 2C). Fine cytoplasmic
eosinophilic stippling, occasionally condensing in glob-
ular aggregates in the paranuclear area, was present in
many neoplastic cells (Figure 2D). Cells from the
masses of dogs 1 and 2 frequently contained needle-
shaped pink structures, 13 lm long, that were also
noted free in the background (Figure 1D). Few indis-
tinct intracytoplasmic clear vacuoles were present pro-
ducing peripheral indistinct clearing in the cytoplasm.
Rare cells in the mass of dog 3 contained granular
green-black pigment, suggestive of melanin (Figure 2A).
Rarely cells were surrounded by small amounts of
bright pink, glassy, amorphous material. Criteria of
malignancy included marked anisocytosis and aniso-
karyosis, multinucleation, karyomegaly, macronucle-
oli, and bizarre-shaped nucleoli (Figure 1B). Mitotic
figures were also noted occasionally and often
appeared atypical (Figure 2B). The original cytologic
diagnosis for the 3 tumors had been sarcoma for dog 1,
carcinoma, likely clear cell adnexal carcinoma, for dog
2, and spindle-cell tumor or melanoma for dog 3.
Histologic and immunohistochemical findings
H&E sections of haired skin revealed dermal or
subcutaneous neoplasms with similar morphologic
appearances in all 3 dogs. The neoplasms consisted of
well-demarcated, multinodular proliferations of cells
arranged in nests and lobules, supported by variably
thick bands of fibrous stroma (Figures 3A and 3C).
Cells ranged from polygonal to elongate with variably
distinct cell borders and moderate to abundant
amounts of eosinophilic cytoplasm; extensive cyto-
plasmic clearing (Figures 3B and 3D) was frequently
observed. Neoplastic cells rarely contained cytoplasmic
melanin. Nuclei were round to oval with coarsely
stippled chromatin and 12 prominent nucleoli. There
was moderate to marked pleomorphism, and
occasionally cells exhibited karyomegaly and multinu-
cleation. Cellular and nuclear pleomorphism was more
4 Vet Clin Pathol 0/0 (2012) 17 2012 American Society for Veterinary Clinical Pathology
Piviani et al
prominent in the mass of dog 3 (Figure 3D). Mitoses
ranged from 1 to 4 per 10 4009 fields. Foci of minerali-
zation, hemorrhage, necrosis, and aggregates of hemo-
siderin-laden macrophages were scattered throughout
the neoplasms of all 3 dogs. Aggregates of lymphocytes
and plasma cells were occasionally present at the
periphery of the neoplasm. The margins of the exci-
sional biopsies were free of neoplastic cells. The diag-
nosis in all 3 cases was cutaneous clear cell adnexal
carcinoma with complete surgical excision.
Variable numbers of neoplastic cells had cytoplas-
mic immunoreactivity for pancytokeratin (515%)
(Figure 3E) and vimentin (340%) (Figure 3F).
Strong cytoplasmic immunoreactivity for vimentin
also was present in cells within the supporting fibrous
stroma. In all cases, rare neoplastic cells expressed
Melan A or PNL-2. Neoplastic cells did not express
Cam 5.2, CD79a, calponin, S-100, or a-smooth muscle
actin. Fine PAS-positive cytoplasmic granules were
found in a subset of neoplastic cells (510%) in the
masses of dogs 2 and 3; the majority of PAS-positive
granules were diastase sensitive, consistent with glyco-
gen. The quantity of tissue was insufficient to evaluate
PAS-staining in the mass of dog 1.
Discussion
Histologic features of the 3 masses were consistent with
canine cutaneous clear cell adnexal carcinoma.13
Characteristic features included dermal or subcutane-
ous location, lobular architecture, and epithelial cells
with clear cytoplasm. Areas of necrosis and foci of min-
eralization were present, and melanin granules were
observed infrequently. Similar to other reports,
marked cellular pleomorphism was noted multifocally,
with cells varying in shape from polygonal to spindle-
shaped and exhibiting occasional karyomegaly and
multinucleation with up to 10 nuclei.14 Although
clear cell adnexal carcinomas appear histologically
undifferentiated and exhibit marked cellular pleomor-
phism, surgical excision is curative in most cases, with
local recurrence and regional metastasis reported in
only 15% of cases.1 Expression of both cytokeratin and
vimentin by variable numbers of neoplastic cells is
consistent with previous reports.1,3,4 Melanin has been
reported occasionally in neoplastic cells, and cells
rarely express Melan A or PNL-2, suggesting differenti-
ation toward hair matrix cells1,3; however, these fea-
tures are inconsistent and are not unique to hair
matrix cells. Melan A may be marking subcellular mel-
anocytic structures phagocytosed by the epithelial
cells.1 Although expression of Melan A or PNL-2, both
Piviani et al Cytology of canine clear cell adnexal carcinoma

A B

C D

E F

Figure 3. (A) Dermal well-demarcated, multinodular proliferation of neoplastic cells (Dog 1). A focus of mineralization is present within the neoplasm
(arrow). H&E, bar = 400 lm. (B) Nests of neoplastic cells with extensive cytoplasmic clearing (Dog 1). H&E, bar = 50 lm. (C) Lobules of neoplastic epi-
thelial cells supported by fibrous stroma (Dog 3). H&E, bar = 200 lm. (D) Neoplastic cells with variable cytoplasmic clearing. Note extensive pleomor-
phism of neoplastic cells in some regions (Dog 3). H&E, bar = 100 lm. (E,F) Neoplastic cells with positive cytoplasmic immunoreactivity for (E)
pancytokeratin and (F) vimentin (Dog 1). Horseradish peroxidase/dextran polymer, bar = 50lm.

markers for melanomas,5,6 may suggest melanocytic
origin of the neoplasm, balloon cell melanomas are not
immunoreactive for cytokeratin and often exhibit
junctional activity, which was absent in the cases
reported here.1,4 Lack of cells with definitive sebaceous
differentiation and with immunoreactivity for Melan
A and PNL-2 helped exclude sebaceous carcinoma.1,7
In addition, all neoplasms reported here were negative
for CD79a expression, which has been reported to stain
canine sebocytes.7,8 Lack of epidermal contiguity
helped exclude a diagnosis of clear cell basal carci-
noma.7,9 Negative immunoreactivity for calponin and
a-smooth muscle actin, both myoepithelial markers, is
consistent with previous reports.1,3,4 Although the
majority of clear cell adnexal carcinomas recently
reported expressed S-100 protein, a subset was nega-
tive, similar to all 3 cases in this report.1 A previous
case report4 and unpublished data10 have described
positive immunohistochemical staining of clear cell
adnexal carcinomas with Cam 5.2, a cytokeratin
antibody occasionally used as a marker for sweat
gland differentiation.7 However, immunohistochemi-
cal analysis of this marker has not been performed in a
large number of cases; therefore it is difficult to
conclude if Cam 5.2 is a reliable marker for clear cell
adnexal carcinomas. Neoplastic cells in the 3 cases
described in this report were negative for Cam 5.2.
Liposarcoma was excluded based on architecture of
the neoplasms and expression of cytokeratin.1 Thus,
immunohistochemical analysis was consistent with

Vet Clin Pathol 0/0 (2012) 17 2012 American Society for Veterinary Clinical Pathology 5
Cytology of canine clear cell adnexal carcinoma
the diagnosis of canine cutaneous clear cell adnexal
carcinoma.14
Cytologic features of clear cell adnexal carci-
noma have not been described previously, probably
owing to the rarity of this neoplasm. Although cyto-
logic evaluation was performed in a low number of
cases, numerous histologic similarities between the 3
cases described in this report and the other 17 cases
of this neoplasm diagnosed at our institution suggest
that the cytologic appearance of the neoplasms in
these 3 dogs can be considered representative of the
cytomorphologic spectrum of this neoplasm. Defin-
ing cytologic features included the presence of large
numbers of pleomorphic cells, predominantly oval
to spindle-shaped, with cytoplasmic projections;
arrangement of cells as individualized cells or in
loose clusters with no distinct evidence of cell-to-cell
junctions; presence of a light blue to gray back-
ground resulting from streaming of cytoplasm from
ruptured cells; marked anisokaryosis with multinu-
cleation; prominent nucleoli; cytoplasmic rarefaction
in some cells; and the presence of cytoplasmic eosin-
ophilic stippling, occasionally condensed in needle-
shaped pink structures or in globular deposits. These
needle-shaped structures were not seen on histologic
sections and may represent atypical glycogen depos-
its, as neoplastic cells of clear cell adnexal carcinoma
have been shown to be glycogen-rich,1,4 and PAS-
positive diastase-sensitive granules, consistent with
glycogen, were noted on histologic sections of the
masses from 2 dogs. However, glycogen deposits are
usually colorless on Romanowsky stains. Electron
microscopic examination would have been helpful
in further elucidating the nature of this cytoplasmic
material. Necrosis and chronic hemorrhage also
were noted and correlated with histologic findings.1
Rare neoplastic cells contained fine blue granules
interpreted as melanin. The cellular morphologic
appearance, lack of defined epithelial arrangements,
and the presence of many criteria of malignancy are
more suggestive of a malignant mesenchymal tumor
or melanocytic tumor than an epithelial neoplasm.
This may lead to a possible misdiagnosis if the
cytopathologist is not familiar with the cytomorpho-
logic features of this rare type of carcinoma. Inap-
propriate case management may follow, considering
the worse prognosis and higher tendency for recur-
rence of sarcoma and melanocytic tumors compared
with the biologic behavior of clear cell adnexal
carcinoma.
Previously reported cytologic features of a cutane-
ous mass from a 3-year-old Golden Retriever were
similar to the features reported here, and included
6 Vet Clin Pathol 0/0 (2012) 17 2012 American Society for Veterinary Clinical Pathology
Piviani et al
marked cellular pleomorphism, anisokaryosis, aniso-
cytosis, multinucleation, multiple nucleoli, prominent
paranuclear clearing, and presence of magenta
intracytoplasmic material.8 The cytologic diagnosis
was malignant neoplasia with features of mesenchy-
mal origin. Histologically, the mass consisted of multi-
ple irregular lobules, subdivided by bands of fibrous
connective tissue, with markedly pleomorphic cells
containing intracytoplasmic clear vacuoles. Thus, in
conjunction with the immunohistochemical results,
which included immunoreactivity for pancytokeratin
and lack of expression of Melan A, S-100, and vimen-
tin, the histologic diagnosis was sebaceous carcinoma.
In contrast, the neoplastic cells in the cases reported
here lacked distinct vacuoles on H&E, stained sections
were variably immunoreactive for vimentin and
Melan A, in addition to pancytokeratin, and did not
express CD79a.
In conclusion, the cytologic appearance of clear
cell adnexal carcinoma is suggestive of a mesenchymal
or melanocytic neoplasm consisting of pleomorphic
cells, cytoplasmic streaming, prominent cytoplasmic
rarefaction/clearing, and intracytoplasmic pink nee-
dle-shaped inclusions. If these cytologic features are
evident, clear cell adnexal carcinoma, although a rare
entity, should be included in the differential diagnosis
for a cutaneous mass in dogs, even if characteristic epi-
thelial features, such as highly cohesive cell clusters,
evident cell-to-cell junctions, and distinct cytoplasmic
borders, are absent.
Acknowledgments
The authors gratefully acknowledge Drs. Ida Piperisova and
Scott Moroff, Antech Diagnostics, Lake Success, NY, for
their generous donation of cytologic samples from dog 3;
Drs. Michael H. Goldschmidt, Raquel Walton, and Amy
Durham, Department of Pathobiology, University of Penn-
sylvania, for their invaluable input; and Mrs. Jackie Ferra-
cone, New Bolton Center, University of Pennsylvania, for
assistance with immunohistochemical analysis.
Disclosure: The authors have indicated that they have
no affiliations or financial involvement with any organiza-
tion or entity with a financial interest in, or in financial
competition with, the subject matter or materials discussed
in this article.
References
1. Schulman FY, Lipscomb TP, Atkin TJ. Canine cutane-
ous clear cell adnexal carcinoma: histopathology,
immunohistochemistry, and biological behavior of 26
cases. J Vet Diagn Invest. 200517:403411.
Piviani et al
2. Jabara AG, Finnie JW. Four cases of clear-cell hidrade-
nocarcinomas in the dog. J Comp Pathol. 1978 88:525
532.
3. Mikaelian I, Wong V. Follicular stem cell carcinoma:
histologic, immunohistochemical, ultrastructural, and
clinical characterization in 30 dogs. Vet Pathol. 2003;40:
433444.
4. Yasuno K, Nishiyama S, Suetsugu F, Ogihara K, Mada-
rame H, Shirota K. Cutaneous clear cell adnexal carci-
noma in a dog: special reference to cytokeratin
expression. J Vet Med Sci. 2009;71:15131517.
5. Kurotaki T, Tomonari Y, Kanno T, Wako Y, Tsuchitani
M. A novel immunohistochemical marker of normal and
neoplastic melanocytes in formalin-fixed, paraffin-embed-
ded tissues of albino rats. Vet Pathol. 2008;45:383387.
6. Ramos-Vara JA, Miller MA. Immunohistochemical
identification of canine melanocytic neoplasms with
antibodies to melanocytic antigen PNL2 and tyrosinase:
comparison with melan A. Vet Pathol. 2011;48:443450.
Vet Clin Pathol 0/0 (2012) 17 2012 American Society for Veterinary Clinical Pathology
Cytology of canine clear cell adnexal carcinoma
7. Gross TL, Ihrke PJ, Walder EJ, Affolter VK. Skin Diseases
of the Dog and Cat. 2nd ed. Oxford, UK: Blackwell Pub-
lishing; 2005: 421833.
8. Dickinson RM, Young KM. Cutaneous mass aspirate
from a Golden Retriever: glandular guile. Vet Clin
Pathol. 2005;4:421424.
9. Goldschmidt MH, Schneider AC, Ferracone JD. Use
of monoclonal antibodies as an adjunct to the diag-
nosis of epithelial skin tumors of the dog and cat
using the EnVision Plus (R) system for the immuno-
histochemical detection of tissue antigens. In: Tho-
day KL, Foil CS, Bond R, eds. Advances in Veterinary
Dermatology. Vol 4. Oxford, UK: Blackwell Publish-
ing; 2002:142149.
10. Walder EJ. Comment on Follicular stem cell carci-
noma: histologic, immunohistochemical, ultrastructur-
al, and clinical characterization in 30 dogs. Vet Pathol.
2005;42:107108.
7