- Email: [email protected]
Canine Subcutaneous Mast Cell Tumour: Diagnosis and Prognosis
ARTICLE IN PRESS J. Comp. Path. 2007,Vol. 136, 231^239
www.elsevier.com/locate/jcpa
Canine Subcutaneous Mast Cell Tumour: Diagnosis and Prognosis S. J. Newman, L. Mrkonjich, K. K. Walker and B. W. Rohrbach* Department of Pathobiology, and *Department of Comparative Medicine, University ofTennessee College of Veterinary Medicine, 2407 River Dr., Knoxville,TN 37996-4542, USA
Summary The aim of this study was to characterize the pathology and clinical outcome of the subcutaneous variant of canine mast cell tumour. Fifty-three cases satisfying the inclusion criteria were selected from the pathology archive of the College of Veterinary Medicine, University of Tennessee. Referring veterinarians provided information on outcome. These dogs had a median age of 9 years (range 3^17 years). After characterizing tumours histologically, nuclear expression of proliferating cell nuclear antigen (PCNA) and Ki67 (MIB-1 clone) was determined immunohistochemically and mast cell origin was con¢rmed with c-Kit staining. Counts of argyrophilic nucleolar organizer regions (AgNOR) were determined by silver staining. Nuclear labelling was counted in 100 tumour cells. Margins were recorded as incomplete in 66% of dogs, and metastases occurred in 6% of dogs. The estimated minimum mean survival time from date of diagnosis was 1199 days, ranging from 55 to 41780 days. The median scores from immunohistochemical labelling were PCNA 0.05 and Ki67 0.03 per 100 tumour cells. The median score for AgNOR staining was 1.25 per 100 tumour cells. The patterns of c-Kit expression included membranous labelling in 20 tumours, stippled cytoplasmic labelling in 23 tumours and di¡use cytoplasmic labelling in 10 tumours. Age (r ¼ 0.61, P ¼ 0.14) and AgNOR score (r ¼ 0.58, P ¼ 0.17) had moderate, but non-signi¢cant, negative associations with survival. PCNA (r ¼ 0.32, P ¼ 0.47), Ki67 (r ¼ 0.22, P ¼ 0.64) and c-Kit immunolabelling was not associated with survival. The subcutaneous variant of canine mast cell tumour is distinct in having features of intermediate histological grade and extended mean survival times, suggesting a slightly better long-term prognosis than for higher grade dermal variants. Expression of nuclear proliferation markers is not associated with outcome. r 2007 Elsevier Ltd. All rights reserved. Keywords: dog; mast cell tumour; immunohistochemistry
Introduction Mast cell tumours (MCTs) comprise 7^21% of all canine skin tumours (Macy and MacEwan, 1989). While ¢ne needle aspiration and cytopathology can con¢rm the tumour type, histopathological examination of biopsy tissue has typically been required for tumour grading (Bostock,1973; Patnaik et al.,1984).While these grading systems have stood the test of time, there is increased interest in better characterization of cutaneous
Correspondence to: Department of Pathobiology, University of Tennessee, College of Veterinary Medicine, Room A201, 2407 River Dr., Knoxville, TN 37996-4542, USA (e-mail: [email protected]). 0021-9975/$ - see front matter
doi:10.1016/j.jcpa.2007.02.003
MCTs so as to predict more accurately their likely clinical behaviour. Although the criteria for grading are well established, pathologists may interpret these di¡erently, and this difference produces discrepancies in tumour classi¢cation (Patnaik et al.,1984). Hence, there have been attempts to use light microscopical stains (toluidine blue, Giemsa, and periodic acid Schi¡ [PAS]) to distinguish more adequately the tumour grades (Simoes and Schoning,1994). Currently being evaluated is silver staining of the argyrophilic nucleolar organizer region (AgNOR) and an array of immunohistochemical procedures that include antibodies that identify components of mast cell granules (chymase and tryptase), suspected proto-oncogene r 2007 Elsevier Ltd. All rights reserved.
ARTICLE IN PRESS 232
S.J. Newman et al.
defects (c-Kit, p53, p21 and p27) (Ginn et al., 2000; Ja¡e et al., 2000;Wu et al., 2004), enzymes (matrix metalloproteinases) (Leibman et al., 2000), and proliferation markers (proliferating cell nuclear antigen [PCNA] and Ki67 MIB-1 clone) (Abadie et al.,1999). In the process of better characterizing MCTs by more advanced morphometric and molecular techniques, it has become apparent that many factors may be involved in prognosis (Preziosi et al., 2004). The characterization and prognosis of subcutaneous variants of canine MCTs have not been fully determined. Hence, the aim of this study was to characterize a large number of these variants by histopathological examination and enumeration of nucleolar organizer regions and proliferation markers. These results were correlated to survival times in order to formulate prognostic parameters for the subcutaneous variant of canine MCT.
Materials and Methods Case Material
The 1999^2005 medical records database from the College of Veterinary Medicine, University of Tennessee was searched for diagnoses of canine cutaneous MCT. Records with corresponding pathology reports from the same time were examined. From these cases, a subset of 53 dogs with primary subcutaneous variants of MCTs was selected. The inclusion criteria included the availability of two or more histological sections demonstrating that the MCT involved only the subcutaneous adipose tissue and did not encroach on adnexal structures or the dermis (Fig. 1). In addition, one or more of the following criteria must have been present: (1) a pathology report
listing the tumour as being subcutaneous in origin, (2) a gross description of a subcutaneous mass, or (3) a history from the veterinarian describing the neoplasm as a subcutaneous mass. The breed, age, and sex of all animals in the study were extracted from medical record data and/or pathology reports. Tumour Characteristics
The surgical pathology database provided the anatomical location of the tumours included in the study. Additionally, historical information such as documentation of previous MCTs, development of recurrence, or de novo tumour was collected from both the medical record and the pathology report. During slide review, the primary investigator determined completeness of surgical margins and evidence of intravascular invasion or metastasis. The completeness of the surgical margins was described using the accepted terminology ‘‘dirty’’ (where neoplastic cells extended to the surgical margin or within 1mm of the surgical margin) or ‘‘wide’’ (where there was a 1^3 cm wide margin around the neoplasm) (Powers, 2001; Newman, 2003). Tumour Grading
Criteria used to grade the tumour were either quanti¢able or simply determined to be present or absent. Quanti¢able criteria included signi¢cant numbers of eosinophils (440% of the cellular in¢ltrate), cytoplasmic granularity (scored as 0^3 for non-existent, mild, moderate or marked), and mitotic rate (number per 40 ¢eld). Criteria in the second category included the presence of excess stroma, foci of collagenolysis, in¢ltrative pattern of growth, prominent nucleoli, binucleate mast cells, anisocytosis and anisokaryosis. AgNOR Staining and Immunohistochemistry
Fig. 1. Photomicrograph of a subcutaneous variant of canine cutaneous mast cell tumour. Note the location of the neoplasm deep in the panniculus (arrow). HE. Bar,100 mm.
All procedures were carried out using formalin-¢xed and para⁄n wax-embedded tissues. For AgNOR staining, sections were initially de-para⁄nized with xylene and rehydrated. Slides were then placed in an aluminium foil-wrapped Coplin jar containing freshly prepared silver stain comprised of two volumes of aqueous silver nitrate solution 50% w/v, combined with one volume of gelatin B 2%, dissolved in aqueous formic acid 1% (Abadie et al., 1999). Slides were immersed in stain at room temperature for 1h before being rinsed, dehydrated, and cover-slipped. A canine lymph node with lymphoma served as a positive control tissue. All cases were characterized quantitatively for AgNORs by counting the proportion of positively stained cells in a total of 100 neoplastic mast cells in a representative 100 ¢eld. Images of these ¢elds were
ARTICLE IN PRESS 233
Canine Subcutaneous Mast Cell Tumour
digitally captured and a computer program (Microsoft Windows XP Professional, Microsoft Paint version 5.1, copyright 1981^2001) was utilized for counting. Immunohistochemistry (IHC) was employed to evaluate expression of c-Kit (CD117), PCNA, and Ki-67 (clone MIB-1). Slides were initially de-para⁄nized and re-hydrated through graded alcohols to water. Antigen retrieval was performed on slides destined for c-Kit and MIB-1 labelling by ¢rst heating in citrate bu¡er pH 6.0 (Dako; Carpinteria, USA) for 25 min at 95 1C in a steamer and subsequently leaving the slides to cool for 20 min. After retrieval, all slides were processed using a computer-controlled automated stainer (Dako; Model S3400). Slides were ¢rst soaked in Tris-bu¡ered saline (pH 7.6), for 5 min and then incubated in hydrogen peroxide 3% for 5 min. A 5 min non-serum protein block (Dako) was used on MIB-1 and c-Kit slides (20 min for PCNA). The following antibodies were applied for 60 min at room temperature: c-Kit at 1 in 1000 (Dako), PCNA at 1 in 1000 (Biogenex, San Ramon, USA), and MIB-1 at 1 in 800 (Dako). A horse radish peroxidase (HRP)-labelled polymer system (Envision+ anti-rabbit; Dako) was applied to c-Kit slides and incubated for 30 min.An HRP-labelled polymer system (Envision+ anti-mouse; Dako) was applied to all other slides and incubated for 30 min. All slides were then rinsed, and 3,30-diaminobenzidine chromagen was applied to slides for 10 min.Lastly, slides were rinsed, counterstained with Harris haematoxylin, dehydrated, and coverslipped. Negative control slides were included in which primary antiserum was replaced by Universal Negative Control+ rabbit and mouse serum (Dako) where appropriate. Para⁄n wax-embedded arrays containing di¡erent canine tissues and lymph nodes with lymphoma were used as controls to ensure positive expression. The location and intensity of c-Kit expression was determined and performed as per the method of Kiupel et al. (2004). Counts of PCNA and Ki67 expressing tumour cells were made using digital image capture as above, and strongly positive nuclei per 100 tumour cells were counted in a representative 100 ¢eld.
Data Analysis
Continuous data were summarized as a mean71 standard deviation or median and range, depending upon the distribution of data. Survival interval was de¢ned as the number of days from the date of diagnosis until death, and survival analysis was performed by the method of Kaplan^Meier. Patients that died from causes unrelated to MCTand those lost to follow-up or alive at the time the study was terminated were censored on the last day of contact and included in the analysis of survival time. The association of survival time with the presence or absence of granularity, excess stroma, clean surgical margins, and multiple MCTs for dogs that died from tumour-related disease was evaluated using a non-parametricWilcoxon test. Correlation of AgNOR, Ki67, PCNA and age with survival time was performed by the method of Pearson (P ¼ 0.05 signi¢cance for all tests).
Results Case Presentation
Five hundred and sixty-one dogs with MCTs were identi¢ed in the retrospective search of medical records. Fifty-three dogs met the primary inclusion criteria of the study in that two tissue sections were available that con¢rmed a nodular subcutaneous MCT. Of the 53 dogs, 22 (41.5%) had an historical notation of a subcutaneous mass, 25 (47.2%) had a gross pathology description of a subcutaneous mass, and 52 (98.1%) had con¢rmation of a subcutaneous location documented in the surgical biopsy report. The following breeds were a¡ected with subcutaneous MCT: golden retriever (8), Labrador retriever (7), boxer (5), mixed breed (5), cocker spaniel (4), beagle (3), weimaraner (2), pharaoh hound (2), and miniature schnauzer (2). Other purebreds (21) were represented by a single animal each. Neutered female animals were apparently over-represented (27 [52%]), followed by neutered male (13 [25%]), intact female (7 [13.5%]), and intact male (5 [9.6%)]). The sex of one animal was not recorded. The dogs ranged in age from 3 to 17 years, with a median age of 9 years (n ¼ 51). The age of two animals was not recorded.
Clinical Outcome
Medical records were searched for follow-up information such as tumour recurrence, development of additional MCTs, use of chemotherapy or radiation therapy, current status (alive or dead), and date and cause of death. In cases where the information in the medical record was incomplete, a questionnaire and/ or follow-up telephone call was directed to the referral veterinarian.
Tumour Characteristics
The most commonly a¡ected anatomical locations for tumour development within the subcutaneous tissue included the hindlimb (14 [26.9%]), forelimb (12 [23.1%]), abdomen (11 [21.2%]), thorax (6 [11.5%]), head/neck (6 [11.5%]), back (2 [3.8%]), and tail (1 [1.9%]). The location of the tumour mass in two animals was not recorded.
ARTICLE IN PRESS 234
S.J. Newman et al.
Sixty-six percent of the tumours (35/53) were incompletely excised, and this incompleteness was most consistent at the deep tissue margin. Twenty-eight of the tumours had neoplastic cells present at the surgical margin, and seven had neoplastic cells within 1mm of the surgical margin (both of which were criteria for ‘‘dirty’’ margins). Post-surgical metastases occurred in 3/53 (6%) of dogs. The survival times for the two patients treated with radiation therapy were 112 days and 476 days (mean ¼ 294 days); for the two patients treated with chemotherapy, survival times were 227 days and 1049 days (mean ¼ 638 days). The one animal that survived 227 days also had metastatic disease. These survival times, while based on small numbers of treated patients, are less than the mean for all study patients considered together. Observations on 12 histological parameters thought to be pertinent to this set of neoplasms are included in Table1.While none of these parameters showed a statis-
Table 1 Morphological features of canine subcutaneous mast cell tumour Variable Metastasis
Finding
Absent Present Not recorded Excess stroma Absent Present Absent Incomplete margins Present Multiple MCTsy Absent Present Not recorded Cytoplasmic granularity Absent Present Mitotic rate 0 per 40 ¢eld 0-1 per 40 ¢eld 1 per 40 ¢eld 1-2 per 40 ¢eld Anisocytosis Absent Present Anisokaryosis Absent Present Multinucleation Absent Present In¢ltrative growth Absent Present Collagenolysis Absent Present Abundant eosinophils Absent Present
Number out of 53 cases (percent) 43 (81) 3 (6) 7 (13) 21 (40) 32 (60) 18 (34) 35 (66) 36 (68) 11 (21) 6 (11) 18 (34) 35 (66) 34 (64) 17 (32) 1 (2) 1 (2) 2 (4) 51 (96) 0 (0) 53 (10 0) 45 (85) 8 (15) 7 (13) 46 (87) 47 (89) 6 (11) 10 (19) 43 (81)
Incomplete margins include those cases with neoplastic cells extending
to the margin or within 1mm of the margin. The animal has been a¡ected with mast cell tumours prior to removal of the current neoplasm.
y
tically signi¢cant association with prognosis, notable variation existed between tumours. The median scores for proliferation markers and c-Kit for all patients (n ¼ 53) are shown in Tables 2 and 3, respectively. Several of these markers demonstrated a negative association with survival, however none attained statistical signi¢cance. Age and AgNOR score had a moderate but non-signi¢cant, negative association with survival after diagnosis (r ¼ 00.61, P ¼ 0.14, and r ¼ 0.58, P ¼ 0.17, respectively). PCNA and Ki67 were not associated with survival (r ¼ 0.32, P ¼ 0.47 and r ¼ 0.22, P ¼ 0.64, respectively). There was no association between c-kit expression and survival. An example of strong cytoplasmic c-Kit labelling from an intermediate grade neoplasm is demonstrated in Fig. 2. Similarly, an example of nuclei containing multiple AgNORs from an intermediate grade tumour is shown in Fig. 3. Scattered cells within the neoplastic population exhibited strong nuclear PCNA expression (Fig. 4) and Ki67 expression (Fig. 5). The estimated minimum mean survival time from date of diagnosis was 1199 7 565 days (Fig. 6). The mean survival time (and standard error) was underestimated because dogs that were still alive when the study ended were censored (given credit for the time that they were observed to be alive).The longest period of observation was for a dog that was still alive 1780 days after diagnosis and subsequently was censored.
Table 2 Comparison of proliferation marker expression between dogs of this series (n ¼ 53) and published case series Variable
Dogs of this series Mean7standard deviation (median)
Data from published case series (mean or mean7standard deviation)
AgNOR
1.3270.27 (1.26)
Bostock et al. (1989) Low grade mean ¼ 1.4 Medium grade mean ¼ 3.2 High grade mean ¼ 6.3
Ki67
0.04170.043 (0.03)
Abadie et al. (1999) Grade I ¼ 0.03170.016 Grade II ¼ 0.09370.067 Grade III ¼ 0.20470.095
PCNA
0.11470.138 (0.05)
Abadie et al. (1999) Dogs that died from MCT 0.61870.169 Dogs that survived MCT 0.42670.235 Grade 1 ¼ 0.42770.238 Grade II ¼ 0.50770.237 Grade III ¼ 0.59070.148
There is no reason to believe that subcutaneous variants were excluded from data derived from the literature.
ARTICLE IN PRESS Canine Subcutaneous Mast Cell Tumour
235
Table 3 Comparison of c-Kit expression between dogs of this series (n ¼ 53) and published case series Pattern of Labelling
Number of dogs in this series positive (percent positive)
Data from published literature Percentage of tumours positive Kiupel et al. (2004)
Membranous
20 (38)
Tumours with local recurrence ¼ 2.4 Tumours with distant recurrence ¼ 14.3 Tumours with mortality ¼ 2.4
Stippled cytoplasmic with lesser membrane
23 (43)
Tumours with local recurrence ¼ 14 Tumours with distant recurrence ¼ 13 Tumours with mortality ¼ 25.6
Di¡use cytoplasmic
10 (19)
Tumours with local recurrence ¼ 23.1 Tumours with distant recurrence ¼ 38.5 Tumours with mortality ¼ 38.5
Fig. 3. Photomicrograph of AgNOR staining of a subcutaneous mast cell tumour. Nucleolar regions range from 1 (arrowhead) to multiple (arrow). IHC. Bar,13 mm
There is no reason to believe that subcutaneous variants were excluded from data derived from the literature.
Fig. 4. Photomicrograph of nuclear PCNA expression (arrow) from a subcutaneous mast cell tumour. IHC. Bar,13 mm.
Fig. 2. Photomicrograph of cytoplasmic expression of c-Kit from a subcutaneous mast cell tumour. IHC. Bar,13 mm.
Survival time ranged from 55 to 41780 days.The median observation time was 352 days. The median followup time was 459 days and ranged from 55 to 1780 days. The animals in this case series had a 61% survival rate at 1206 days, 30% survival at 1392 days, and 0% survival at 1780 days.
Discussion A subcutaneous variant of canine cutaneous MCT has been mentioned brie£y in the literature, where it was described as a soft, £eshy mass that could be clinically consistent with a benign lipoma (Thamm and Vail,
Fig. 5. Photomicrograph of nuclear Ki67 expression (arrow) from a subcutaneous mast cell tumour. IHC. Bar,13 mm.
ARTICLE IN PRESS 236
S.J. Newman et al.
Fig. 6. Kaplan^Meier curve depicting survival time for dogs with subcutaneous mast cell tumour.
2001). The Patnaik grading system mentions subcutaneous involvement following in¢ltration by the higher grade dermal neoplasms (Patnaik et al., 1984; London et al., 1996). The subcutaneous MCT variants were never incorporated into the Patnaik grading scheme (personal communication), but we believe pathologists have graded them as if they were typical dermal forms. Involvement of the subcutaneous tissue as part of an in¢ltrative dermal mass di¡ers from the neoplasms we refer to in this series, as these have no primary dermal involvement and present as demarcated subcutaneous masses from the outset. The results of the present study have suggested that these subcutaneous variants may carry a better long-term prognosis (mean survival time of greater than 1199 days) than higher grade dermal neoplasms.The following comparison of survival times demonstrates this more thoroughly. Patnaik determined that 83% of dogs with a grade I MCTsurvived 1500 days; 44% of dogs with a grade II MCTsurvived 1500 days, and this decreased to 3% for dogs with a grade III MCT (Patnaik et al.,1984). Similarly, Abadie et al. (1999) noted that the 730-day survival for dogs with a grade I MCT was 100%; for grade II, it was 44%, and for grade III, it was 7%. In that same study, the median follow-up time was 540 days, and there was complete surgical excision of the tumours (Abadie et al., 1999). Bostock (1973) reported that dogs with poorly di¡erentiated tumours (69%) had a mean survival time of 18 weeks (126 days), while those with intermediately di¡erentiated tumours (18%) had a mean survival of 28 weeks (196 days), and dogs with well differentiated tumours (13%) had a mean survival time of 51weeks (357 days). Dogs in the present study with subcutaneous MCT showed 61% survival at 1206 days,
30% survival at 1392 days, and 0% survival at 1780 days.The dogs in the present case series thus had better survival times than those in the previously mentioned studies, so location in the subcutaneous tissue may be an important prognostic factor despite the relatively small number of dogs having follow-up. Additional large studies are needed to address this point more clearly.While grading is paramount in the diagnosis of dermal MCTs it will be di⁄cult to create a grading system for the subcutaneous variants of canine MCT where the range of alterations is relatively small. The subcutaneous location should be de¢ned by pathologists so that an appropriate prognosis might be given. These subcutaneous tumours may behave di¡erently with respect to metastasis, recurrence at surgical sites, and response to therapy compared with the dermal forms. The majority of the cases in our series had incomplete surgical margins (35/53 [66%]). Recurrent disease occurred in only 5/53 (9%) of the dogs, and in only three of these were ‘‘dirty’’ surgical margins described previously. The recurrence rate for grade II MCTs in the published literature has been as high as 50% (Al-Sarraf et al., 1996). The lower recurrence rate seen in our study, despite substantial follow-up time (median 459 days), may be a re£ection of a less aggressive nature of the subcutaneous variants and is more favourable than that seen with complete excision of intermediate dermal MCTs. Additionally, adjunctive therapy was used in only four animals in this study, with two receiving radiation therapy and two receiving chemotherapy. Some published studies suggest that adjunctive local treatment may not be required after complete excision of grade II MCTs and that most dogs do not require systemic chemotherapy (Seguin et al., 2001). Complete excision was associated with e¡ective local control in 89% of these dogs, so adjunctive therapy may not always be required (Weisse et al., 2002). In fact, in one study, there was no recurrence or metastasis for a series of MCTs completely excised from 21 dogs which had a median survival time of 603 days and a median follow-up time of 351 days (Simpson et al., 2004). While radiation therapy may not be routinely required for completely excised dermal MCTs, megavoltage radiation treatment is an e¡ective adjunctive therapy for cases with residual disease at the surgical site (Ladue et al.,1998). Dogs (37 animals) that received radiation therapy for moderately di¡erentiated, but incompletely excised, dermal MCTs had a recurrence rate of 3% (one case) at 1 year and 7% (three cases) at 3 years (Frimberger et al., 1997). The median followup time in this case series was 724 days (range 933^1964) (Frimberger et al., 1997). Even without adjunctive therapy in most of our cases, the recurrence rates did not di¡er signi¢cantly from these values,
ARTICLE IN PRESS Canine Subcutaneous Mast Cell Tumour
suggesting a less aggressive behaviour for the subcutaneous variant of canine MCT. Metastasis was relatively infrequent in our study, with only three dogs (3/53 [5.7%]) having con¢rmed metastatic disease during the period of follow-up. All three of those dogs died of their disease, and the mean survival time was 506 days (median ¼ 227 days, range 85^1206). Metastasis has been previously reported in 55^96% of undi¡erentiated dermal MCTs (Thamm and Vail, 2001). Twenty-one percent of dogs in that study had historical evidence or concurrent development of multiple cutaneous MCTs. While predisposed breeds may have a propensity to develop multiple MCTs, this has been previously determined not to negatively a¡ect survival times (Murphy et al., 2006). Hence, based on survival times, recurrence rates, and metastasis, the subcutaneous variants of MCT had a less aggressive behaviour than higher grade dermal variants and most closely resembled those of intermediate grade. Molecular techniques are being used to characterize canine MCTs more accurately. While substantial work has been done in characterizing the dermal tumours, to the authors’ knowledge this is the ¢rst study to characterize the subcutaneous variants with molecular markers. The c-Kit gene encodes a receptor tyrosine kinase that has an extracellular domain of ¢ve immunoglobulin-like folds and an intracellular kinase domain separated by transmembrane and juxtamembrane domains (Zemke et al., 2001). Because the c-Kit receptor plays a regulatory role in cellular growth and di¡erentiation, when altered it may play a role in neoplastic transformation of mast cells (London et al., 1996). Expression of c-Kit is seen in most mast cells, but the intensity and density of cell labelling varies within neoplasms. Lower grade tumours typically have fewer labelled cells, and these label less strongly than higher grade tumours, which have more labelled cells that also label more intensely (Reguera et al., 2000). Recently, location of cellular labelling was also determined to correlate with the prognosis of dermal MCTs (Kiupel et al., 2004). The three patterns documented were membrane-associated labelling, focal to stippled cytoplasmic labelling with decreased membrane-associated labelling, and di¡use cytoplasmic labelling (Kiupel et al., 2004). While membrane localization is expected, the presence of cytoplasmic c-Kit may be due to activated internal forms or active synthesis of the receptor (Reguera et al., 2000). Increased cytoplasmic c-Kit labelling was associated with an increased rate of local recurrence and decreased survival (Kiupel et al., 2004). In the subcutaneous variants of canine MCTevaluated in this study, c-Kit was not associated with survival. However, c-Kit expression was determined to be mainly membranous
237
with lesser stippled cytoplasmic localization, indicative of a more intermediate grade. Nucleolar organizing regions are loops of ribosomal DNA involved in transcription of ribosomal RNA to form proteins that are components of the ribosomal subunits (Hung et al., 2000). These regions can be detected following modi¢ed silver staining. AgNOR counts are related to degree of proliferation and prognosis in many neoplasms, including dermal MCTs (Simoes and Schoning,1994). Bostock et al. (1989) reported AgNOR counts as 1.1^3.5 for grade I, 2.5^4.2 for grade II, and 4.1^8.1 for grade III canine MCT. Hung et al. (2000) reported an improved AgNOR staining technique that distinguished grades without overlap between categories as grade I: 3.6771.17 (2.5^4.84), grade II: 9.3672.36 (7^11.72), and grade III: 16.771.61 (15.1^18.3). The mean of AgNOR counts that distinguished grade II and III from grade I was 6.0 (Hung et al., 2000). However, the mean AgNOR counts were lower for patients in the current study (Table 2) than any lower limit previously reported for dermal variants; thus the subcutaneous variants we examined may not be as aggressive. While the relationship did not have prognostic value, there was some association between increased AgNOR counts and decreased survival.The range of values was small, making cut-o¡ determinations for separate grades within this variant form di⁄cult. PCNA is a non-histone nuclear protein that acts to enhance the function of polymerase delta and polymerase epsilon within the cell cycle (Hung et al., 2000). Although PCNA is synthesized during G1 and peaks during the S phase, there is low expression at the time of mitotic separation and in non-proliferating cells (Hung et al., 2000). Hence, PCNA expression can be used to estimate the degree of proliferating activity in a neoplastic cell population (Abadie et al.,1999). In one large study, the growth fraction of tumour cells expressing PCNA ranged from 5.3% to 65.3% (Hung et al., 2000). A cut-o¡ of 33.3% expression discriminated grade III tumours from grade I (Hung et al., 2000). In our study, PCNA counts were lower than those reported for any dermal MCT grades (Table 2) no matter which counting methodology was used. This may in part re£ect the fact that we counted the number of positive nuclei within 100, rather than in 1000, cells and because only nuclei labelled di¡usely dark brown were counted. Nevertheless, these ¢ndings support the contention that this subcutaneous MCT variant is less aggressive, based on proliferation indices, than what has been reported for the dermal variants. The range of values was small, making cut-o¡ determinations for separate grades di⁄cult. Ki67 is a useful marker for proliferative potential in canine dermal MCTs, but its speci¢c function is not understood, and it is detectable
ARTICLE IN PRESS 238
S.J. Newman et al.
only in the active phase of the cell cycle (Abadie et al., 1999; Sakai et al., 2002). Nevertheless, Ki67 counts allow for distinction amongst di¡erent grades of MCT (Sakai et al., 2002). Additionally, Ki67 detection allows one to di¡erentiate grade II tumours into groups with signi¢cantly di¡erent survival rates (Abadie et al., 1999). Because its short half-life makes its detection a more reliable marker of cell proliferation, Ki67 may be better than PCNA as a prognostic factor for canine dermal MCTs (Abadie et al.,1999; Sakai et al., 2002). Determination of the number of nuclei positively labelled for Ki67 may be useful in predicting prognosis, particularly for dogs with grade II MCTs. In one study, there was a signi¢cantly higher mean number of Ki67 positive nuclei in the tumours of dogs that died of MCT than in those that survived, without any overlap in values between groups (Abadie et al., 1999). The number of positive nuclei was also signi¢cantly di¡erent between MCT grades (Grade 1 ¼ 31716, grade II ¼ 93767, and grade III ¼ 204795). These values had the most impact in determining the behaviour of grade II tumours. All of the dogs in our study had Ki67 counts that were within the range documented previously for grade I dermal MCTs (0.0317 0.0016) (Abadie et al., 1999), and these counts were not associated with survival. In the present study, this may in part re£ect the fact that we counted the number of positive nuclei within 100, rather than 1000, cells and because only nuclei diffusely labelled dark brown were counted. Nevertheless, these Ki67 ¢ndings support the contention that this subcutaneous variant of MCT is less aggressive than what has been reported for the dermal variants (Table 2).The range of values was small, making cut-o¡ determinations for separate grades di⁄cult. The subcutaneous variant of canine MCT is distinct, but it shares features with the intermediate histological grade of dermal MCT. Thus, less subjective criteria were needed for establishment of MCT prognosis in this newly described subcutaneous variant. This was accomplished by assessing histological parameters and prognostic indicators such as AgNOR counts and the expression of c-Kit, PCNA and Ki67. In this study, PCNA and Ki67 counts were not predictors of prognosis; additionally, AgNOR score had a moderate but non-signi¢cant, negative association with survival after diagnosis. Based on expression of proliferation markers, these tumours consistently fall within the lower range of values, suggesting a lower grade of proliferative activity than has been previously determined for dermal MCTs. Based on survival times, recurrence rates, and metastasis, these tumours have a better prognosis than higher grade dermal variants. For this reason, this subcutaneous variant of canine MCT should be recognized by pathologists. However, due to the
narrow range in both histological parameters and proliferation marker expression, it seems unnecessary to grade the neoplasms within this subcategory.
Acknowledgments The idea for this manuscript came from Dr Amiya Patnaik, a former colleague and friend, and hence is dedicated in his memory. Mrs Misty R. Bailey is acknowledged for editorial advice, and Ms Anik Vasington is acknowledged for graphical assistance. The Center of Excellence (COE) in Livestock Diseases and Human Health at the College of Veterinary Medicine, University of Tennessee is acknowledged for providing the funds for wages for Ms Kindra Walker as part of the COE summer student program.
References Abadie, J. J., Amardeilh, M. A. and Delverdier, M. E. (1999). Immunohistochemical detection of proliferating cell nuclear antigen and Ki-67 in mast cell tumors from dogs. Journal of the American Veterinary Medical Association, 215, 1629^1634. Al-Sarraf, R., Mauldin, G. N., Patnaik, A. K. and Meleo, K. A. (1996). A prospective study of radiation therapy for the treatment of grade 2 mast cell tumors in 32 dogs. Journal of Veterinary Internal Medicine, 10, 376^378. Bostock, D. E. (1973). The prognosis following surgical removal of mastocytomas in dogs. Journal of Small Animal Practice, 14, 27^40. Bostock, D. E., Crocker, J., Harris, K. and Smith, P. (1989). Nucleolar organizer regions as indicators of post-surgical prognosis in canine spontaneous mast cell tumors. British Journal of Cancer, 59, 915^918. Frimberger, A. E., Moore, A. S., LaRue, S. M., Gliatto, J. M. and Bengtson, A. E. (1997). Radiotherapy of incompletely resected, moderately di¡erentiated mast cell tumors in the dog: 37 cases (1989-1993). Journal of the American Animal Hospital Association, 33, 320^324. Ginn, P. E., Fox, L. E., Brower, J. C., Gaskin, A., Kurzman, I. D. and Kubilis, P. S. (2000). Immunohistochemical detection of p53 tumor-suppressor protein is a poor indicator of prognosis for canine cutaneous mast cell tumors.Veterinary Pathology, 37, 33^39. Hung, L. C., Pong, V. F., Cheng, C. R., Wong, F. I. and Chu, R. M. (2000). An improved system for quantifying AgNOR and PCNA in canine tumors. Anticancer Research, 20, 3273^3280. Ja¡e, M. H., Hosgood, G., Taylor, H. W., Kerwin, S. C., Hedlund, C. S., Lopez, M. K., Davidson, J. R., Miller, D. M. and Paranjpe, M. (2000). Immunohistochemical and clinical evaluation of p53 in canine cutaneous mast cell tumor.Veterinary Pathology, 37, 40^46. Kiupel, M., Webster, J. D., Kaneene, J. B., Kaneene, R. and Yuzbasiyan-Gurkan, V. (2004). The use of KITand tryptase expression patterns as prognostic tools for canine cutaneous mast cell tumors.Veterinary Pathology, 41, 371^377.
ARTICLE IN PRESS 239
Canine Subcutaneous Mast Cell Tumour
Ladue,T., Price, G. S., Dodge, R., Page, R. L. andThrall, D. E. (1998). Radiation therapy for incompletely resected canine mast cell tumors. Veterinary Radiology and Ultrasound, 39, 57^62. Leibman, N. F., Lana, S. E., Hansen, R. A., Powers, B. E., Fettman, M. J., Withrow, S. J. and Ogilvie, G. K. (2000). Identi¢cation of matrix metalloproteinases in canine cutaneous mast cell tumors. Journal ofVeterinary Internal Medicine, 14, 583^586. London, C. A., Kisseberth, W. C., Galli, S. J., Geissler, E. N. and Helfand, S. C. (1996). Expression of stem cell factor receptor (c-Kit) by the malignant mast cells from spontaneous canine mast cell tumours. Journal of Comparative Pathology, 115, 399^414. Macy, D. W. and MacEwen, E. G. (1989). Canine mast cell tumors. In: Clinical Veterinary Oncology, S.J. Withrow and E.G. MacEwen, Eds, J.B. Lippincott, Philadelphia, pp. 156^166. Murphy, S., Sparkes, A. H., Blunden, A. S., Brearley, M. J. and Smith, K. C. (2006). E¡ects of stage and number of tumours on prognosis of dogs with cutaneous mast cell tumours.Veterinary Record, 158, 287^291. Newman, S. J. (2003). Diagnostic pathology for the cancer patient. ClinicalTechniques in Small Animal Practice, 18,139^144. Patnaik, A. K., Ehler,W. J. and MacEwen, E. G. (1984). Canine cutaneous mast cell tumor: morphologic grading and survival time in 83 dogs.Veterinary Pathology, 21, 469^474. Powers, B.E. (2001). The pathology of neoplasia. In: Small Animal Clinical Oncology, 3rd Edit., S. J. Withrow and E. G. MacEwen, Eds.,W.B. Saunders, Philadelphia, pp 4^17. Preziosi, R., Morini, M. and Sarli, G. (2004). Expression of the KIT protein (CD117) in primary cutaneous mast cell tumors of the dog. Journal of Veterinary Diagnostic Investigation, 16, 554^561. Reguera, M. J., Rabanal, R. M., Puigdemont, A. and Ferrer, L. (2000). Canine mast cell tumors express stem cell factor receptor. AmericanJournal of Dermatopathology, 22, 49^54. Sakai, H., Noda, A., Shirai, N., Iidaka, T., Yanai, T. and Masegi,T. (2002). Proliferative activity of canine mast cell
tumours evaluated by bromodeoxyuridine incorporation and Ki67 expression. Journal of Comparative Pathology, 127, 233^238. Seguin, B., Leibman, N. F., Bregazzi, V. S., Ogilvie, G. K., Powers, B. E., Dernell,W. S., Fettman, M. J. andWithrow, S. J. (2001). Clinical outcome of dogs with grade-II mast cell tumors treated with surgery alone: 55 cases (19961999). Journal of the American Veterinary Medical Association, 218,1120^1123. Simoes, J. P. C. and Schoning, P. (1994). Canine mast cell tumors: a comparison of staining techniques. Journal ofVeterinary Diagnostic Investigation, 6, 458^465. Simpson, A. M., Ludwig, L. L., Newman, S. J., Bergman, P. J., Hottinger, H. A. and Patnaik, A. K. (2004). Evaluation of surgical margins required for complete excision of cutaneous mast cell tumors in dogs. Journal of the AmericanVeterinary Medical Association, 224, 236^240. Thamm, D.H. and Vail, D.M. (2001). Mast cell tumor. In: Small Animal Clinical Oncology, 3rd Edit., S. J. Withrow and E. G. MacEwen, Eds, W.B. Saunders, Philadelphia, pp. 261^282. Weisse, C., Shofer, F. S. and Sorenmo, K. (2002). Recurrence rates and sites for grade II canine cutaneous mast cell tumors following complete surgical excision. Journal of the American Animal Hospital Association, 38,71^73. Wu, H., Hayashi, T. and Inoue, M. (2004). Immunohistochemical expression of p27 and p21 in canine cutaneous mast cell tumors and histiocytomas. Veterinary Pathology, 41, 296^299. Zemke, D., Yamini, B. and Yuzbasiyan-Gurkin, V. (2001). Characterization of an undi¡erentiated malignancy as a mast cell tumor using mutation analysis in the proto-oncogene c-KIT. Journal of Veterinary Diagnostic Investigation, 13, 341^345.
Received, June 5th, 2006 Accepted, February 5th, 2007
www.elsevier.com/locate/jcpa
Canine Subcutaneous Mast Cell Tumour: Diagnosis and Prognosis S. J. Newman, L. Mrkonjich, K. K. Walker and B. W. Rohrbach* Department of Pathobiology, and *Department of Comparative Medicine, University ofTennessee College of Veterinary Medicine, 2407 River Dr., Knoxville,TN 37996-4542, USA
Summary The aim of this study was to characterize the pathology and clinical outcome of the subcutaneous variant of canine mast cell tumour. Fifty-three cases satisfying the inclusion criteria were selected from the pathology archive of the College of Veterinary Medicine, University of Tennessee. Referring veterinarians provided information on outcome. These dogs had a median age of 9 years (range 3^17 years). After characterizing tumours histologically, nuclear expression of proliferating cell nuclear antigen (PCNA) and Ki67 (MIB-1 clone) was determined immunohistochemically and mast cell origin was con¢rmed with c-Kit staining. Counts of argyrophilic nucleolar organizer regions (AgNOR) were determined by silver staining. Nuclear labelling was counted in 100 tumour cells. Margins were recorded as incomplete in 66% of dogs, and metastases occurred in 6% of dogs. The estimated minimum mean survival time from date of diagnosis was 1199 days, ranging from 55 to 41780 days. The median scores from immunohistochemical labelling were PCNA 0.05 and Ki67 0.03 per 100 tumour cells. The median score for AgNOR staining was 1.25 per 100 tumour cells. The patterns of c-Kit expression included membranous labelling in 20 tumours, stippled cytoplasmic labelling in 23 tumours and di¡use cytoplasmic labelling in 10 tumours. Age (r ¼ 0.61, P ¼ 0.14) and AgNOR score (r ¼ 0.58, P ¼ 0.17) had moderate, but non-signi¢cant, negative associations with survival. PCNA (r ¼ 0.32, P ¼ 0.47), Ki67 (r ¼ 0.22, P ¼ 0.64) and c-Kit immunolabelling was not associated with survival. The subcutaneous variant of canine mast cell tumour is distinct in having features of intermediate histological grade and extended mean survival times, suggesting a slightly better long-term prognosis than for higher grade dermal variants. Expression of nuclear proliferation markers is not associated with outcome. r 2007 Elsevier Ltd. All rights reserved. Keywords: dog; mast cell tumour; immunohistochemistry
Introduction Mast cell tumours (MCTs) comprise 7^21% of all canine skin tumours (Macy and MacEwan, 1989). While ¢ne needle aspiration and cytopathology can con¢rm the tumour type, histopathological examination of biopsy tissue has typically been required for tumour grading (Bostock,1973; Patnaik et al.,1984).While these grading systems have stood the test of time, there is increased interest in better characterization of cutaneous
Correspondence to: Department of Pathobiology, University of Tennessee, College of Veterinary Medicine, Room A201, 2407 River Dr., Knoxville, TN 37996-4542, USA (e-mail: [email protected]). 0021-9975/$ - see front matter
doi:10.1016/j.jcpa.2007.02.003
MCTs so as to predict more accurately their likely clinical behaviour. Although the criteria for grading are well established, pathologists may interpret these di¡erently, and this difference produces discrepancies in tumour classi¢cation (Patnaik et al.,1984). Hence, there have been attempts to use light microscopical stains (toluidine blue, Giemsa, and periodic acid Schi¡ [PAS]) to distinguish more adequately the tumour grades (Simoes and Schoning,1994). Currently being evaluated is silver staining of the argyrophilic nucleolar organizer region (AgNOR) and an array of immunohistochemical procedures that include antibodies that identify components of mast cell granules (chymase and tryptase), suspected proto-oncogene r 2007 Elsevier Ltd. All rights reserved.
ARTICLE IN PRESS 232
S.J. Newman et al.
defects (c-Kit, p53, p21 and p27) (Ginn et al., 2000; Ja¡e et al., 2000;Wu et al., 2004), enzymes (matrix metalloproteinases) (Leibman et al., 2000), and proliferation markers (proliferating cell nuclear antigen [PCNA] and Ki67 MIB-1 clone) (Abadie et al.,1999). In the process of better characterizing MCTs by more advanced morphometric and molecular techniques, it has become apparent that many factors may be involved in prognosis (Preziosi et al., 2004). The characterization and prognosis of subcutaneous variants of canine MCTs have not been fully determined. Hence, the aim of this study was to characterize a large number of these variants by histopathological examination and enumeration of nucleolar organizer regions and proliferation markers. These results were correlated to survival times in order to formulate prognostic parameters for the subcutaneous variant of canine MCT.
Materials and Methods Case Material
The 1999^2005 medical records database from the College of Veterinary Medicine, University of Tennessee was searched for diagnoses of canine cutaneous MCT. Records with corresponding pathology reports from the same time were examined. From these cases, a subset of 53 dogs with primary subcutaneous variants of MCTs was selected. The inclusion criteria included the availability of two or more histological sections demonstrating that the MCT involved only the subcutaneous adipose tissue and did not encroach on adnexal structures or the dermis (Fig. 1). In addition, one or more of the following criteria must have been present: (1) a pathology report
listing the tumour as being subcutaneous in origin, (2) a gross description of a subcutaneous mass, or (3) a history from the veterinarian describing the neoplasm as a subcutaneous mass. The breed, age, and sex of all animals in the study were extracted from medical record data and/or pathology reports. Tumour Characteristics
The surgical pathology database provided the anatomical location of the tumours included in the study. Additionally, historical information such as documentation of previous MCTs, development of recurrence, or de novo tumour was collected from both the medical record and the pathology report. During slide review, the primary investigator determined completeness of surgical margins and evidence of intravascular invasion or metastasis. The completeness of the surgical margins was described using the accepted terminology ‘‘dirty’’ (where neoplastic cells extended to the surgical margin or within 1mm of the surgical margin) or ‘‘wide’’ (where there was a 1^3 cm wide margin around the neoplasm) (Powers, 2001; Newman, 2003). Tumour Grading
Criteria used to grade the tumour were either quanti¢able or simply determined to be present or absent. Quanti¢able criteria included signi¢cant numbers of eosinophils (440% of the cellular in¢ltrate), cytoplasmic granularity (scored as 0^3 for non-existent, mild, moderate or marked), and mitotic rate (number per 40 ¢eld). Criteria in the second category included the presence of excess stroma, foci of collagenolysis, in¢ltrative pattern of growth, prominent nucleoli, binucleate mast cells, anisocytosis and anisokaryosis. AgNOR Staining and Immunohistochemistry
Fig. 1. Photomicrograph of a subcutaneous variant of canine cutaneous mast cell tumour. Note the location of the neoplasm deep in the panniculus (arrow). HE. Bar,100 mm.
All procedures were carried out using formalin-¢xed and para⁄n wax-embedded tissues. For AgNOR staining, sections were initially de-para⁄nized with xylene and rehydrated. Slides were then placed in an aluminium foil-wrapped Coplin jar containing freshly prepared silver stain comprised of two volumes of aqueous silver nitrate solution 50% w/v, combined with one volume of gelatin B 2%, dissolved in aqueous formic acid 1% (Abadie et al., 1999). Slides were immersed in stain at room temperature for 1h before being rinsed, dehydrated, and cover-slipped. A canine lymph node with lymphoma served as a positive control tissue. All cases were characterized quantitatively for AgNORs by counting the proportion of positively stained cells in a total of 100 neoplastic mast cells in a representative 100 ¢eld. Images of these ¢elds were
ARTICLE IN PRESS 233
Canine Subcutaneous Mast Cell Tumour
digitally captured and a computer program (Microsoft Windows XP Professional, Microsoft Paint version 5.1, copyright 1981^2001) was utilized for counting. Immunohistochemistry (IHC) was employed to evaluate expression of c-Kit (CD117), PCNA, and Ki-67 (clone MIB-1). Slides were initially de-para⁄nized and re-hydrated through graded alcohols to water. Antigen retrieval was performed on slides destined for c-Kit and MIB-1 labelling by ¢rst heating in citrate bu¡er pH 6.0 (Dako; Carpinteria, USA) for 25 min at 95 1C in a steamer and subsequently leaving the slides to cool for 20 min. After retrieval, all slides were processed using a computer-controlled automated stainer (Dako; Model S3400). Slides were ¢rst soaked in Tris-bu¡ered saline (pH 7.6), for 5 min and then incubated in hydrogen peroxide 3% for 5 min. A 5 min non-serum protein block (Dako) was used on MIB-1 and c-Kit slides (20 min for PCNA). The following antibodies were applied for 60 min at room temperature: c-Kit at 1 in 1000 (Dako), PCNA at 1 in 1000 (Biogenex, San Ramon, USA), and MIB-1 at 1 in 800 (Dako). A horse radish peroxidase (HRP)-labelled polymer system (Envision+ anti-rabbit; Dako) was applied to c-Kit slides and incubated for 30 min.An HRP-labelled polymer system (Envision+ anti-mouse; Dako) was applied to all other slides and incubated for 30 min. All slides were then rinsed, and 3,30-diaminobenzidine chromagen was applied to slides for 10 min.Lastly, slides were rinsed, counterstained with Harris haematoxylin, dehydrated, and coverslipped. Negative control slides were included in which primary antiserum was replaced by Universal Negative Control+ rabbit and mouse serum (Dako) where appropriate. Para⁄n wax-embedded arrays containing di¡erent canine tissues and lymph nodes with lymphoma were used as controls to ensure positive expression. The location and intensity of c-Kit expression was determined and performed as per the method of Kiupel et al. (2004). Counts of PCNA and Ki67 expressing tumour cells were made using digital image capture as above, and strongly positive nuclei per 100 tumour cells were counted in a representative 100 ¢eld.
Data Analysis
Continuous data were summarized as a mean71 standard deviation or median and range, depending upon the distribution of data. Survival interval was de¢ned as the number of days from the date of diagnosis until death, and survival analysis was performed by the method of Kaplan^Meier. Patients that died from causes unrelated to MCTand those lost to follow-up or alive at the time the study was terminated were censored on the last day of contact and included in the analysis of survival time. The association of survival time with the presence or absence of granularity, excess stroma, clean surgical margins, and multiple MCTs for dogs that died from tumour-related disease was evaluated using a non-parametricWilcoxon test. Correlation of AgNOR, Ki67, PCNA and age with survival time was performed by the method of Pearson (P ¼ 0.05 signi¢cance for all tests).
Results Case Presentation
Five hundred and sixty-one dogs with MCTs were identi¢ed in the retrospective search of medical records. Fifty-three dogs met the primary inclusion criteria of the study in that two tissue sections were available that con¢rmed a nodular subcutaneous MCT. Of the 53 dogs, 22 (41.5%) had an historical notation of a subcutaneous mass, 25 (47.2%) had a gross pathology description of a subcutaneous mass, and 52 (98.1%) had con¢rmation of a subcutaneous location documented in the surgical biopsy report. The following breeds were a¡ected with subcutaneous MCT: golden retriever (8), Labrador retriever (7), boxer (5), mixed breed (5), cocker spaniel (4), beagle (3), weimaraner (2), pharaoh hound (2), and miniature schnauzer (2). Other purebreds (21) were represented by a single animal each. Neutered female animals were apparently over-represented (27 [52%]), followed by neutered male (13 [25%]), intact female (7 [13.5%]), and intact male (5 [9.6%)]). The sex of one animal was not recorded. The dogs ranged in age from 3 to 17 years, with a median age of 9 years (n ¼ 51). The age of two animals was not recorded.
Clinical Outcome
Medical records were searched for follow-up information such as tumour recurrence, development of additional MCTs, use of chemotherapy or radiation therapy, current status (alive or dead), and date and cause of death. In cases where the information in the medical record was incomplete, a questionnaire and/ or follow-up telephone call was directed to the referral veterinarian.
Tumour Characteristics
The most commonly a¡ected anatomical locations for tumour development within the subcutaneous tissue included the hindlimb (14 [26.9%]), forelimb (12 [23.1%]), abdomen (11 [21.2%]), thorax (6 [11.5%]), head/neck (6 [11.5%]), back (2 [3.8%]), and tail (1 [1.9%]). The location of the tumour mass in two animals was not recorded.
ARTICLE IN PRESS 234
S.J. Newman et al.
Sixty-six percent of the tumours (35/53) were incompletely excised, and this incompleteness was most consistent at the deep tissue margin. Twenty-eight of the tumours had neoplastic cells present at the surgical margin, and seven had neoplastic cells within 1mm of the surgical margin (both of which were criteria for ‘‘dirty’’ margins). Post-surgical metastases occurred in 3/53 (6%) of dogs. The survival times for the two patients treated with radiation therapy were 112 days and 476 days (mean ¼ 294 days); for the two patients treated with chemotherapy, survival times were 227 days and 1049 days (mean ¼ 638 days). The one animal that survived 227 days also had metastatic disease. These survival times, while based on small numbers of treated patients, are less than the mean for all study patients considered together. Observations on 12 histological parameters thought to be pertinent to this set of neoplasms are included in Table1.While none of these parameters showed a statis-
Table 1 Morphological features of canine subcutaneous mast cell tumour Variable Metastasis
Finding
Absent Present Not recorded Excess stroma Absent Present Absent Incomplete margins Present Multiple MCTsy Absent Present Not recorded Cytoplasmic granularity Absent Present Mitotic rate 0 per 40 ¢eld 0-1 per 40 ¢eld 1 per 40 ¢eld 1-2 per 40 ¢eld Anisocytosis Absent Present Anisokaryosis Absent Present Multinucleation Absent Present In¢ltrative growth Absent Present Collagenolysis Absent Present Abundant eosinophils Absent Present
Number out of 53 cases (percent) 43 (81) 3 (6) 7 (13) 21 (40) 32 (60) 18 (34) 35 (66) 36 (68) 11 (21) 6 (11) 18 (34) 35 (66) 34 (64) 17 (32) 1 (2) 1 (2) 2 (4) 51 (96) 0 (0) 53 (10 0) 45 (85) 8 (15) 7 (13) 46 (87) 47 (89) 6 (11) 10 (19) 43 (81)
Incomplete margins include those cases with neoplastic cells extending
to the margin or within 1mm of the margin. The animal has been a¡ected with mast cell tumours prior to removal of the current neoplasm.
y
tically signi¢cant association with prognosis, notable variation existed between tumours. The median scores for proliferation markers and c-Kit for all patients (n ¼ 53) are shown in Tables 2 and 3, respectively. Several of these markers demonstrated a negative association with survival, however none attained statistical signi¢cance. Age and AgNOR score had a moderate but non-signi¢cant, negative association with survival after diagnosis (r ¼ 00.61, P ¼ 0.14, and r ¼ 0.58, P ¼ 0.17, respectively). PCNA and Ki67 were not associated with survival (r ¼ 0.32, P ¼ 0.47 and r ¼ 0.22, P ¼ 0.64, respectively). There was no association between c-kit expression and survival. An example of strong cytoplasmic c-Kit labelling from an intermediate grade neoplasm is demonstrated in Fig. 2. Similarly, an example of nuclei containing multiple AgNORs from an intermediate grade tumour is shown in Fig. 3. Scattered cells within the neoplastic population exhibited strong nuclear PCNA expression (Fig. 4) and Ki67 expression (Fig. 5). The estimated minimum mean survival time from date of diagnosis was 1199 7 565 days (Fig. 6). The mean survival time (and standard error) was underestimated because dogs that were still alive when the study ended were censored (given credit for the time that they were observed to be alive).The longest period of observation was for a dog that was still alive 1780 days after diagnosis and subsequently was censored.
Table 2 Comparison of proliferation marker expression between dogs of this series (n ¼ 53) and published case series Variable
Dogs of this series Mean7standard deviation (median)
Data from published case series (mean or mean7standard deviation)
AgNOR
1.3270.27 (1.26)
Bostock et al. (1989) Low grade mean ¼ 1.4 Medium grade mean ¼ 3.2 High grade mean ¼ 6.3
Ki67
0.04170.043 (0.03)
Abadie et al. (1999) Grade I ¼ 0.03170.016 Grade II ¼ 0.09370.067 Grade III ¼ 0.20470.095
PCNA
0.11470.138 (0.05)
Abadie et al. (1999) Dogs that died from MCT 0.61870.169 Dogs that survived MCT 0.42670.235 Grade 1 ¼ 0.42770.238 Grade II ¼ 0.50770.237 Grade III ¼ 0.59070.148
There is no reason to believe that subcutaneous variants were excluded from data derived from the literature.
ARTICLE IN PRESS Canine Subcutaneous Mast Cell Tumour
235
Table 3 Comparison of c-Kit expression between dogs of this series (n ¼ 53) and published case series Pattern of Labelling
Number of dogs in this series positive (percent positive)
Data from published literature Percentage of tumours positive Kiupel et al. (2004)
Membranous
20 (38)
Tumours with local recurrence ¼ 2.4 Tumours with distant recurrence ¼ 14.3 Tumours with mortality ¼ 2.4
Stippled cytoplasmic with lesser membrane
23 (43)
Tumours with local recurrence ¼ 14 Tumours with distant recurrence ¼ 13 Tumours with mortality ¼ 25.6
Di¡use cytoplasmic
10 (19)
Tumours with local recurrence ¼ 23.1 Tumours with distant recurrence ¼ 38.5 Tumours with mortality ¼ 38.5
Fig. 3. Photomicrograph of AgNOR staining of a subcutaneous mast cell tumour. Nucleolar regions range from 1 (arrowhead) to multiple (arrow). IHC. Bar,13 mm
There is no reason to believe that subcutaneous variants were excluded from data derived from the literature.
Fig. 4. Photomicrograph of nuclear PCNA expression (arrow) from a subcutaneous mast cell tumour. IHC. Bar,13 mm.
Fig. 2. Photomicrograph of cytoplasmic expression of c-Kit from a subcutaneous mast cell tumour. IHC. Bar,13 mm.
Survival time ranged from 55 to 41780 days.The median observation time was 352 days. The median followup time was 459 days and ranged from 55 to 1780 days. The animals in this case series had a 61% survival rate at 1206 days, 30% survival at 1392 days, and 0% survival at 1780 days.
Discussion A subcutaneous variant of canine cutaneous MCT has been mentioned brie£y in the literature, where it was described as a soft, £eshy mass that could be clinically consistent with a benign lipoma (Thamm and Vail,
Fig. 5. Photomicrograph of nuclear Ki67 expression (arrow) from a subcutaneous mast cell tumour. IHC. Bar,13 mm.
ARTICLE IN PRESS 236
S.J. Newman et al.
Fig. 6. Kaplan^Meier curve depicting survival time for dogs with subcutaneous mast cell tumour.
2001). The Patnaik grading system mentions subcutaneous involvement following in¢ltration by the higher grade dermal neoplasms (Patnaik et al., 1984; London et al., 1996). The subcutaneous MCT variants were never incorporated into the Patnaik grading scheme (personal communication), but we believe pathologists have graded them as if they were typical dermal forms. Involvement of the subcutaneous tissue as part of an in¢ltrative dermal mass di¡ers from the neoplasms we refer to in this series, as these have no primary dermal involvement and present as demarcated subcutaneous masses from the outset. The results of the present study have suggested that these subcutaneous variants may carry a better long-term prognosis (mean survival time of greater than 1199 days) than higher grade dermal neoplasms.The following comparison of survival times demonstrates this more thoroughly. Patnaik determined that 83% of dogs with a grade I MCTsurvived 1500 days; 44% of dogs with a grade II MCTsurvived 1500 days, and this decreased to 3% for dogs with a grade III MCT (Patnaik et al.,1984). Similarly, Abadie et al. (1999) noted that the 730-day survival for dogs with a grade I MCT was 100%; for grade II, it was 44%, and for grade III, it was 7%. In that same study, the median follow-up time was 540 days, and there was complete surgical excision of the tumours (Abadie et al., 1999). Bostock (1973) reported that dogs with poorly di¡erentiated tumours (69%) had a mean survival time of 18 weeks (126 days), while those with intermediately di¡erentiated tumours (18%) had a mean survival of 28 weeks (196 days), and dogs with well differentiated tumours (13%) had a mean survival time of 51weeks (357 days). Dogs in the present study with subcutaneous MCT showed 61% survival at 1206 days,
30% survival at 1392 days, and 0% survival at 1780 days.The dogs in the present case series thus had better survival times than those in the previously mentioned studies, so location in the subcutaneous tissue may be an important prognostic factor despite the relatively small number of dogs having follow-up. Additional large studies are needed to address this point more clearly.While grading is paramount in the diagnosis of dermal MCTs it will be di⁄cult to create a grading system for the subcutaneous variants of canine MCT where the range of alterations is relatively small. The subcutaneous location should be de¢ned by pathologists so that an appropriate prognosis might be given. These subcutaneous tumours may behave di¡erently with respect to metastasis, recurrence at surgical sites, and response to therapy compared with the dermal forms. The majority of the cases in our series had incomplete surgical margins (35/53 [66%]). Recurrent disease occurred in only 5/53 (9%) of the dogs, and in only three of these were ‘‘dirty’’ surgical margins described previously. The recurrence rate for grade II MCTs in the published literature has been as high as 50% (Al-Sarraf et al., 1996). The lower recurrence rate seen in our study, despite substantial follow-up time (median 459 days), may be a re£ection of a less aggressive nature of the subcutaneous variants and is more favourable than that seen with complete excision of intermediate dermal MCTs. Additionally, adjunctive therapy was used in only four animals in this study, with two receiving radiation therapy and two receiving chemotherapy. Some published studies suggest that adjunctive local treatment may not be required after complete excision of grade II MCTs and that most dogs do not require systemic chemotherapy (Seguin et al., 2001). Complete excision was associated with e¡ective local control in 89% of these dogs, so adjunctive therapy may not always be required (Weisse et al., 2002). In fact, in one study, there was no recurrence or metastasis for a series of MCTs completely excised from 21 dogs which had a median survival time of 603 days and a median follow-up time of 351 days (Simpson et al., 2004). While radiation therapy may not be routinely required for completely excised dermal MCTs, megavoltage radiation treatment is an e¡ective adjunctive therapy for cases with residual disease at the surgical site (Ladue et al.,1998). Dogs (37 animals) that received radiation therapy for moderately di¡erentiated, but incompletely excised, dermal MCTs had a recurrence rate of 3% (one case) at 1 year and 7% (three cases) at 3 years (Frimberger et al., 1997). The median followup time in this case series was 724 days (range 933^1964) (Frimberger et al., 1997). Even without adjunctive therapy in most of our cases, the recurrence rates did not di¡er signi¢cantly from these values,
ARTICLE IN PRESS Canine Subcutaneous Mast Cell Tumour
suggesting a less aggressive behaviour for the subcutaneous variant of canine MCT. Metastasis was relatively infrequent in our study, with only three dogs (3/53 [5.7%]) having con¢rmed metastatic disease during the period of follow-up. All three of those dogs died of their disease, and the mean survival time was 506 days (median ¼ 227 days, range 85^1206). Metastasis has been previously reported in 55^96% of undi¡erentiated dermal MCTs (Thamm and Vail, 2001). Twenty-one percent of dogs in that study had historical evidence or concurrent development of multiple cutaneous MCTs. While predisposed breeds may have a propensity to develop multiple MCTs, this has been previously determined not to negatively a¡ect survival times (Murphy et al., 2006). Hence, based on survival times, recurrence rates, and metastasis, the subcutaneous variants of MCT had a less aggressive behaviour than higher grade dermal variants and most closely resembled those of intermediate grade. Molecular techniques are being used to characterize canine MCTs more accurately. While substantial work has been done in characterizing the dermal tumours, to the authors’ knowledge this is the ¢rst study to characterize the subcutaneous variants with molecular markers. The c-Kit gene encodes a receptor tyrosine kinase that has an extracellular domain of ¢ve immunoglobulin-like folds and an intracellular kinase domain separated by transmembrane and juxtamembrane domains (Zemke et al., 2001). Because the c-Kit receptor plays a regulatory role in cellular growth and di¡erentiation, when altered it may play a role in neoplastic transformation of mast cells (London et al., 1996). Expression of c-Kit is seen in most mast cells, but the intensity and density of cell labelling varies within neoplasms. Lower grade tumours typically have fewer labelled cells, and these label less strongly than higher grade tumours, which have more labelled cells that also label more intensely (Reguera et al., 2000). Recently, location of cellular labelling was also determined to correlate with the prognosis of dermal MCTs (Kiupel et al., 2004). The three patterns documented were membrane-associated labelling, focal to stippled cytoplasmic labelling with decreased membrane-associated labelling, and di¡use cytoplasmic labelling (Kiupel et al., 2004). While membrane localization is expected, the presence of cytoplasmic c-Kit may be due to activated internal forms or active synthesis of the receptor (Reguera et al., 2000). Increased cytoplasmic c-Kit labelling was associated with an increased rate of local recurrence and decreased survival (Kiupel et al., 2004). In the subcutaneous variants of canine MCTevaluated in this study, c-Kit was not associated with survival. However, c-Kit expression was determined to be mainly membranous
237
with lesser stippled cytoplasmic localization, indicative of a more intermediate grade. Nucleolar organizing regions are loops of ribosomal DNA involved in transcription of ribosomal RNA to form proteins that are components of the ribosomal subunits (Hung et al., 2000). These regions can be detected following modi¢ed silver staining. AgNOR counts are related to degree of proliferation and prognosis in many neoplasms, including dermal MCTs (Simoes and Schoning,1994). Bostock et al. (1989) reported AgNOR counts as 1.1^3.5 for grade I, 2.5^4.2 for grade II, and 4.1^8.1 for grade III canine MCT. Hung et al. (2000) reported an improved AgNOR staining technique that distinguished grades without overlap between categories as grade I: 3.6771.17 (2.5^4.84), grade II: 9.3672.36 (7^11.72), and grade III: 16.771.61 (15.1^18.3). The mean of AgNOR counts that distinguished grade II and III from grade I was 6.0 (Hung et al., 2000). However, the mean AgNOR counts were lower for patients in the current study (Table 2) than any lower limit previously reported for dermal variants; thus the subcutaneous variants we examined may not be as aggressive. While the relationship did not have prognostic value, there was some association between increased AgNOR counts and decreased survival.The range of values was small, making cut-o¡ determinations for separate grades within this variant form di⁄cult. PCNA is a non-histone nuclear protein that acts to enhance the function of polymerase delta and polymerase epsilon within the cell cycle (Hung et al., 2000). Although PCNA is synthesized during G1 and peaks during the S phase, there is low expression at the time of mitotic separation and in non-proliferating cells (Hung et al., 2000). Hence, PCNA expression can be used to estimate the degree of proliferating activity in a neoplastic cell population (Abadie et al.,1999). In one large study, the growth fraction of tumour cells expressing PCNA ranged from 5.3% to 65.3% (Hung et al., 2000). A cut-o¡ of 33.3% expression discriminated grade III tumours from grade I (Hung et al., 2000). In our study, PCNA counts were lower than those reported for any dermal MCT grades (Table 2) no matter which counting methodology was used. This may in part re£ect the fact that we counted the number of positive nuclei within 100, rather than in 1000, cells and because only nuclei labelled di¡usely dark brown were counted. Nevertheless, these ¢ndings support the contention that this subcutaneous MCT variant is less aggressive, based on proliferation indices, than what has been reported for the dermal variants. The range of values was small, making cut-o¡ determinations for separate grades di⁄cult. Ki67 is a useful marker for proliferative potential in canine dermal MCTs, but its speci¢c function is not understood, and it is detectable
ARTICLE IN PRESS 238
S.J. Newman et al.
only in the active phase of the cell cycle (Abadie et al., 1999; Sakai et al., 2002). Nevertheless, Ki67 counts allow for distinction amongst di¡erent grades of MCT (Sakai et al., 2002). Additionally, Ki67 detection allows one to di¡erentiate grade II tumours into groups with signi¢cantly di¡erent survival rates (Abadie et al., 1999). Because its short half-life makes its detection a more reliable marker of cell proliferation, Ki67 may be better than PCNA as a prognostic factor for canine dermal MCTs (Abadie et al.,1999; Sakai et al., 2002). Determination of the number of nuclei positively labelled for Ki67 may be useful in predicting prognosis, particularly for dogs with grade II MCTs. In one study, there was a signi¢cantly higher mean number of Ki67 positive nuclei in the tumours of dogs that died of MCT than in those that survived, without any overlap in values between groups (Abadie et al., 1999). The number of positive nuclei was also signi¢cantly di¡erent between MCT grades (Grade 1 ¼ 31716, grade II ¼ 93767, and grade III ¼ 204795). These values had the most impact in determining the behaviour of grade II tumours. All of the dogs in our study had Ki67 counts that were within the range documented previously for grade I dermal MCTs (0.0317 0.0016) (Abadie et al., 1999), and these counts were not associated with survival. In the present study, this may in part re£ect the fact that we counted the number of positive nuclei within 100, rather than 1000, cells and because only nuclei diffusely labelled dark brown were counted. Nevertheless, these Ki67 ¢ndings support the contention that this subcutaneous variant of MCT is less aggressive than what has been reported for the dermal variants (Table 2).The range of values was small, making cut-o¡ determinations for separate grades di⁄cult. The subcutaneous variant of canine MCT is distinct, but it shares features with the intermediate histological grade of dermal MCT. Thus, less subjective criteria were needed for establishment of MCT prognosis in this newly described subcutaneous variant. This was accomplished by assessing histological parameters and prognostic indicators such as AgNOR counts and the expression of c-Kit, PCNA and Ki67. In this study, PCNA and Ki67 counts were not predictors of prognosis; additionally, AgNOR score had a moderate but non-signi¢cant, negative association with survival after diagnosis. Based on expression of proliferation markers, these tumours consistently fall within the lower range of values, suggesting a lower grade of proliferative activity than has been previously determined for dermal MCTs. Based on survival times, recurrence rates, and metastasis, these tumours have a better prognosis than higher grade dermal variants. For this reason, this subcutaneous variant of canine MCT should be recognized by pathologists. However, due to the
narrow range in both histological parameters and proliferation marker expression, it seems unnecessary to grade the neoplasms within this subcategory.
Acknowledgments The idea for this manuscript came from Dr Amiya Patnaik, a former colleague and friend, and hence is dedicated in his memory. Mrs Misty R. Bailey is acknowledged for editorial advice, and Ms Anik Vasington is acknowledged for graphical assistance. The Center of Excellence (COE) in Livestock Diseases and Human Health at the College of Veterinary Medicine, University of Tennessee is acknowledged for providing the funds for wages for Ms Kindra Walker as part of the COE summer student program.
References Abadie, J. J., Amardeilh, M. A. and Delverdier, M. E. (1999). Immunohistochemical detection of proliferating cell nuclear antigen and Ki-67 in mast cell tumors from dogs. Journal of the American Veterinary Medical Association, 215, 1629^1634. Al-Sarraf, R., Mauldin, G. N., Patnaik, A. K. and Meleo, K. A. (1996). A prospective study of radiation therapy for the treatment of grade 2 mast cell tumors in 32 dogs. Journal of Veterinary Internal Medicine, 10, 376^378. Bostock, D. E. (1973). The prognosis following surgical removal of mastocytomas in dogs. Journal of Small Animal Practice, 14, 27^40. Bostock, D. E., Crocker, J., Harris, K. and Smith, P. (1989). Nucleolar organizer regions as indicators of post-surgical prognosis in canine spontaneous mast cell tumors. British Journal of Cancer, 59, 915^918. Frimberger, A. E., Moore, A. S., LaRue, S. M., Gliatto, J. M. and Bengtson, A. E. (1997). Radiotherapy of incompletely resected, moderately di¡erentiated mast cell tumors in the dog: 37 cases (1989-1993). Journal of the American Animal Hospital Association, 33, 320^324. Ginn, P. E., Fox, L. E., Brower, J. C., Gaskin, A., Kurzman, I. D. and Kubilis, P. S. (2000). Immunohistochemical detection of p53 tumor-suppressor protein is a poor indicator of prognosis for canine cutaneous mast cell tumors.Veterinary Pathology, 37, 33^39. Hung, L. C., Pong, V. F., Cheng, C. R., Wong, F. I. and Chu, R. M. (2000). An improved system for quantifying AgNOR and PCNA in canine tumors. Anticancer Research, 20, 3273^3280. Ja¡e, M. H., Hosgood, G., Taylor, H. W., Kerwin, S. C., Hedlund, C. S., Lopez, M. K., Davidson, J. R., Miller, D. M. and Paranjpe, M. (2000). Immunohistochemical and clinical evaluation of p53 in canine cutaneous mast cell tumor.Veterinary Pathology, 37, 40^46. Kiupel, M., Webster, J. D., Kaneene, J. B., Kaneene, R. and Yuzbasiyan-Gurkan, V. (2004). The use of KITand tryptase expression patterns as prognostic tools for canine cutaneous mast cell tumors.Veterinary Pathology, 41, 371^377.
ARTICLE IN PRESS 239
Canine Subcutaneous Mast Cell Tumour
Ladue,T., Price, G. S., Dodge, R., Page, R. L. andThrall, D. E. (1998). Radiation therapy for incompletely resected canine mast cell tumors. Veterinary Radiology and Ultrasound, 39, 57^62. Leibman, N. F., Lana, S. E., Hansen, R. A., Powers, B. E., Fettman, M. J., Withrow, S. J. and Ogilvie, G. K. (2000). Identi¢cation of matrix metalloproteinases in canine cutaneous mast cell tumors. Journal ofVeterinary Internal Medicine, 14, 583^586. London, C. A., Kisseberth, W. C., Galli, S. J., Geissler, E. N. and Helfand, S. C. (1996). Expression of stem cell factor receptor (c-Kit) by the malignant mast cells from spontaneous canine mast cell tumours. Journal of Comparative Pathology, 115, 399^414. Macy, D. W. and MacEwen, E. G. (1989). Canine mast cell tumors. In: Clinical Veterinary Oncology, S.J. Withrow and E.G. MacEwen, Eds, J.B. Lippincott, Philadelphia, pp. 156^166. Murphy, S., Sparkes, A. H., Blunden, A. S., Brearley, M. J. and Smith, K. C. (2006). E¡ects of stage and number of tumours on prognosis of dogs with cutaneous mast cell tumours.Veterinary Record, 158, 287^291. Newman, S. J. (2003). Diagnostic pathology for the cancer patient. ClinicalTechniques in Small Animal Practice, 18,139^144. Patnaik, A. K., Ehler,W. J. and MacEwen, E. G. (1984). Canine cutaneous mast cell tumor: morphologic grading and survival time in 83 dogs.Veterinary Pathology, 21, 469^474. Powers, B.E. (2001). The pathology of neoplasia. In: Small Animal Clinical Oncology, 3rd Edit., S. J. Withrow and E. G. MacEwen, Eds.,W.B. Saunders, Philadelphia, pp 4^17. Preziosi, R., Morini, M. and Sarli, G. (2004). Expression of the KIT protein (CD117) in primary cutaneous mast cell tumors of the dog. Journal of Veterinary Diagnostic Investigation, 16, 554^561. Reguera, M. J., Rabanal, R. M., Puigdemont, A. and Ferrer, L. (2000). Canine mast cell tumors express stem cell factor receptor. AmericanJournal of Dermatopathology, 22, 49^54. Sakai, H., Noda, A., Shirai, N., Iidaka, T., Yanai, T. and Masegi,T. (2002). Proliferative activity of canine mast cell
tumours evaluated by bromodeoxyuridine incorporation and Ki67 expression. Journal of Comparative Pathology, 127, 233^238. Seguin, B., Leibman, N. F., Bregazzi, V. S., Ogilvie, G. K., Powers, B. E., Dernell,W. S., Fettman, M. J. andWithrow, S. J. (2001). Clinical outcome of dogs with grade-II mast cell tumors treated with surgery alone: 55 cases (19961999). Journal of the American Veterinary Medical Association, 218,1120^1123. Simoes, J. P. C. and Schoning, P. (1994). Canine mast cell tumors: a comparison of staining techniques. Journal ofVeterinary Diagnostic Investigation, 6, 458^465. Simpson, A. M., Ludwig, L. L., Newman, S. J., Bergman, P. J., Hottinger, H. A. and Patnaik, A. K. (2004). Evaluation of surgical margins required for complete excision of cutaneous mast cell tumors in dogs. Journal of the AmericanVeterinary Medical Association, 224, 236^240. Thamm, D.H. and Vail, D.M. (2001). Mast cell tumor. In: Small Animal Clinical Oncology, 3rd Edit., S. J. Withrow and E. G. MacEwen, Eds, W.B. Saunders, Philadelphia, pp. 261^282. Weisse, C., Shofer, F. S. and Sorenmo, K. (2002). Recurrence rates and sites for grade II canine cutaneous mast cell tumors following complete surgical excision. Journal of the American Animal Hospital Association, 38,71^73. Wu, H., Hayashi, T. and Inoue, M. (2004). Immunohistochemical expression of p27 and p21 in canine cutaneous mast cell tumors and histiocytomas. Veterinary Pathology, 41, 296^299. Zemke, D., Yamini, B. and Yuzbasiyan-Gurkin, V. (2001). Characterization of an undi¡erentiated malignancy as a mast cell tumor using mutation analysis in the proto-oncogene c-KIT. Journal of Veterinary Diagnostic Investigation, 13, 341^345.
Received, June 5th, 2006 Accepted, February 5th, 2007