Mast cell tumors (MCTs) are among the most

Transcription

1 J Vet Intern Med 2006;20: Recurrence Rate, Clinical Outcome, and Cellular Proliferation Indices as Prognostic Indicators after Incomplete Surgical Excision of Cutaneous Grade II Mast Cell Tumors: 28 Dogs ( ) Bernard Séguin, M. Faulkner Besancon, Jennifer L. McCallan, Loralei L. Dewe, Matthew C. Tenwolde, Emily K. Wong, and Michael S. Kent The objectives of this study were to determine local recurrence rate, clinical outcome, and prognostic value of the number of argyrophylic nucleolar organizer regions (AgNORs), presence of proliferating cell nuclear antigen (PCNA), and number of Ki-67 positive nuclei after incomplete surgical excision of canine cutaneous grade II mast cell tumors (MCTs). This retrospective study included 30 MCTs in 28 dogs. Medical records were examined and follow-up information was obtained from owners and referring veterinarians. Only cases in which excision was incomplete and no anvcillary therapy (other than prednisone) for MCT was given were included. Paraffin-embedded tumor tissues were retrieved for AgNORs, PCNA, and Ki- 67 staining. Median follow-up time was days. Seven (23.3%) tumors recurred locally. Median time to local recurrence was not reached with a mean of 1,713 days. The estimated proportions of tumors that recurred locally at 1, 2, and 5 years were 17.3, 22.1, and 33.3%, respectively. Eleven (39.3%) dogs developed MCTs at other cutaneous locations. Median progressionfree survival was 1,044 days. Median overall survival was 1,426 days. The combination of Ki-67 and PCNA scores was prognostic for local recurrence (P 5.03) and development of local recurrence was prognostic for decreased overall survival (P 5.04). Results suggest that a minority of incompletely excised MCTs recur. Therefore, ancillary local therapies may not always be necessary. However, local recurrence can negatively affect survival of the affected dogs. Cellular proliferation indices may indicate the likelihood of MCT recurrence after incomplete excision. Key words: Canine; Ki-67; MIB-1; Proliferating cell nuclear antigen. Mast cell tumors (MCTs) are among the most common neoplasms in dogs, accounting for 7 21% of all cutaneous and subcutaneous tumors. 1,2 The clinical course of the disease is variable, from benign to highly malignant. 3,4 The best overall treatment approach for dogs with MCT remains controversial. Wide surgical excision continues to be the treatment of choice for localized cutaneous MCTs. 1,4 Local adjuvant therapy does not appear to be necessary after complete excision of MCTs. The rate of local recurrence after complete resection has been reported to be 5% in 2 studies and 3 and 11% in 2 other studies. 5 8 When excision is incomplete, additional therapy in the form of a 2nd surgery, radiation therapy, or chemotherapy is recommended. 4 However, 3 studies have failed to support the assumption that most incompletely excised MCTs will recur. Reported recurrence rates of incompletely resected MCTs in which margins were evaluated and for which no adjuvant therapy was From the Department of Surgical and Radiological Sciences (Séguin, McCallan, Dewe, Kent), and the Veterinary Medical Teaching Hospital (Besancon, Tenwolde, Wong), School of Veterinary Medicine, University of California, Davis, CA. Dr Séguin is presently affiliated with the Department of Clinical Sciences, College of Veterinary Medicine, Oregon State University, Corvallis, OR. Dr Besancon is presently affiliated with the Cape Cod Veterinary Specialists, Buzzards Bay, MA. Previously presented in part at the 24th Annual Conference, Veterinary Cancer Society, Kansas City, MO, November Reprint requests: Bernard Séguin, DVM, MS, Department of Clinical Sciences, 285 Magruder Hall, College of Veterinary Medicine, Oregon State University, Corvallis, OR 97331; bernard.seguin@oregonstate.edu. Submitted May 19, 2005; Revised September 13, October 10, November 17, 2005; Accepted November 18, Copyright E 2006 by the American College of Veterinary Internal Medicine /06/ /$3.00/0 instituted were 18, 19, and 35%. 7 9 These results call into question the recommendation to provide additional treatment to all dogs with incompletely excised MCTs. These results suggest that many dogs may suffer unnecessary morbidity from additional treatment, in addition to the increased costs incurred by their owners. On the other hand, if left untreated, 1 study determined that local recurrence was associated with progression to malignancy (ie, multicentric regrowth, metastasis, or both) in 59% of dogs. 10 Therefore, being able to prognosticate on the likelihood that an incompletely excised MCT may recur would help identify which dogs require adjuvant therapy and which do not. Several factors have been used to predict the biological behavior of MCTs, including clinical stage, growth rate, systemic signs, anatomic location, histologic grade, number of argyrophylic nucleolar organizer regions (AgNORs), presence of proliferating cell nuclear antigen (PCNA), and number of Ki-67 positive nuclei. 4,10 13 Among these, histologic grade is the most reliable to predict survival. 2,8,10,14 16 However, 2 studies have come to different conclusions regarding the predictive value of histologic grade for local recurrence. One study found grade not to be prognostic for local recurrence, and another found grade to be prognostic. 8,9 Grade II MCTs appear to have the widest range of biological behavior compared to the other 2 grades. 5,10 Although grade I MCTs have an excellent prognosis and grade III MCTs often metastasize, grade II MCTs are more difficult to prognosticate. 4,5 Indicators of cellular proliferation may provide prognostic information about likelihood of MCTs to recur locally. The more proliferative are cells in the tumor, the more likely it should be for the tumor to recur. Three cellular proliferation indices are commonly used to evaluate tumors in humans and dogs: AgNOR, PCNA, and Ki-67. 4,15,17 In 1 study, AgNOR and PCNA were predictive for

2 934 Séguin et al recurrence. 15 In that study, however, recurrence was not defined and some investigators consider MCTs arising in the skin at a distant site as recurrence. 6 Also, the status of the surgical margins (ie, complete versus incomplete) was not reported. 15 Therefore, the prognostic value of these variables is still unclear. The objectives of the present study were to evaluate the rate of local recurrence and clinical outcome after incomplete surgical resection of cutaneous grade II MCTs in dogs when ancillary treatment (other than prednisone) was not instituted, and to determine if AgNOR, PCNA, and Ki-67 alone or in combination are prognostic indicators for local recurrence after incomplete excision of grade II MCTs in dogs. The hypotheses were that incomplete surgical resection would infrequently lead to local MCT recurrence and that each cellular proliferation index or a combination of them would be predictive for local recurrence in incompletely excised MCTs in the dog. Materials and Methods Criteria for Selection of Cases Medical records from all dogs presented to the University of California, Davis, Veterinary Medical Teaching Hospital (UCD- VMTH) with grade II MCTs diagnosed between September 1994 and September 2002 were reviewed. Only those dogs with grade II MCT that were treated surgically and for which histopathologically confirmed incomplete margins were obtained were included in the study. An incomplete margin was defined as tumor cells extending to the surgical edge or within 1 mm of the edge. 18 Exclusion criteria included a 2nd surgery, chemotherapy other than prednisone, or radiation therapy for the treatment of residual MCTs. Procedures Hospital records for all dogs with histologically confirmed incomplete surgical margins were reviewed. Signalment, weight, date of surgery, location, size, pre- and postoperative treatment, date of local recurrence, date of metastasis, locations of other MCTs, date of occurrence at other location, date of death, cause of death, and time to last follow-up were recorded. Lymph node, spleen, liver, and bone marrow status with respect to the presence or absence of mast cells detected by use of cytology also was recorded. Presence of mast cells in any of those locations was considered metastasis. Additional information then was obtained via telephone call to owners. A questionnaire was sent to owners who agreed to complete and return it within a 3-month period. The questionnaire attempted to gather information regarding when the mass was 1st noticed, if there was any ancillary therapy performed by the referring veterinarian after removal of the mass (eg, 2nd surgery, chemotherapy, or radiation therapy), whether there was local recurrence (and if so when), if a new MCT was noted in a separate location and when, and, if not alive, date and cause of death. After all questionnaires were returned, a follow-up phone call was made to the owner or referring veterinarian to confirm all information. Time to local recurrence was defined as the interval between the excision of the MCT and evidence of local recurrence. Local recurrence was cytologically or histologically confirmed whenever possible. When re-evaluation was only possible by telephone conversation with the owner and the owner described growth of a mass in the site of the original MCT, this observation was considered a local recurrence. Time to a different location was defined as the time interval between surgery and cytologic or histologic evidence of another MCT developing at a different cutaneous location. Time to metastasis was defined as the interval between surgery and evidence of metastasis. Cytologic confirmation of metastases was obtained. Follow-up time was defined as the interval between surgery and the last date of follow-up or date of death if the animal had died. Progression-free survival was defined as the interval between surgery and developing local recurrence, another MCT at a different cutaneous location, or metastasis, whichever came 1st. Overall survival time was defined as the interval between surgery and death. Cause of death was classified as related to MCT, unrelated, or relationship unknown. For dogs in which the cause of death was unknown, death was attributed to MCT for statistical analysis. Dogs that died of causes other than MCT were censored at their date of death. Proliferation Indices Formalin-fixed paraffin-embedded samples were available from 20 tumors, 3 with local recurrence and 17 without. More than 1 paraffin block was available from some tumors, and a total of 31 blocks was available for review. Tissues were cut with a microtome into 5-mm sections and fixed onto a positively charged microscope slide in standard fashion. Four recuts were made from each block. One slide each was stained with hematoxylin and eosin, PCNA, Ki- 67, and AgNOR. PCNA Immunostaining Sections were deparaffinized twice in xylene for 10 minutes each time, and rehydrated in a graded ethanol series to 70% ethanol. Slides were soaked 30 minutes in 0.3% hydrogen peroxide in methanol to inactivate endogenous peroxidase, followed by a triple rinse in phosphate buffered saline (PBS). Slides then were steamed in citrate buffer a for 10 minutes, allowed to cool for 20 minutes, and triple rinsed in PBS. Serum-free protein block b was added to each slide for 10 minutes in a moist chamber at room temperature. After draining off the protein block, the tissue sections were incubated with PCNA c at a concentration of 1 : 400 overnight at 4uC. The slides then were triple rinsed in PBS and incubated with the biotinylated secondary antibody (goat anti-mouse immunoglobulin G [IgG]) d for 10 minutes in a moist chamber at room temperature. Slides again were triple rinsed with PBS and incubated with streptavidin horseradish peroxidase (HRP) e for 10 minutes in a moist chamber at room temperature. The slides then were soaked twice in PBS for 5 minutes. Aminoethyl carbazole f was added to each slide and monitored for staining for 1 3 minutes. Slides were rinsed in tap water, then in distilled water and counterstained in Meyers hematoxylin g for 1 2 minutes. Slides were rinsed again in distilled water and coated with Cover Care aqueous mounting media h on a plate warmer for approximately 30 minutes. The positive control consisted of paraffin-embedded canine intestinal epithelium, as well as normally dividing cells on the sample slides such as basal cells of the epidermis and germ cells of the hair follicles. Negative controls were obtained by using purified mouse myeloma IgG1 i in place of the primary antibody. The number of positive cells in 500 mast cells was counted in the tumor at 1,0003 magnification. 10,15 Ki-67 Immunostaining Sections then were deparaffinized in 2 changes of xylene for 10 minutes and rehydrated before immunostaining. Antigen retrieval was performed by using a Russell Hobbs electric pressure cooker j that was time and temperature programmable to provide a standard protocol. This method was carried out for 20 minutes at high pressure in 0.1 M citrate buffer, ph 6.0. k After completion, slides were allowed to cool to room temperature in the buffer (for approximately 20 minutes). They were washed in 0.1 M phosphate

3 Incomplete Excision of Mast Cell Tumors 935 buffer (PB), ph 7.4, which was used for all subsequent washes. They then were immersed in 3% H 2 O 2 in PB for 10 minutes to quench endogenous peroxidases, and then in 10% normal goat serum with 0.1% bovine serum albumin for 1 hour to block nonspecific staining. For KI-67 immunostaining, MIB-1 antibody l was used at a dilution of 1 : 500 in a solution of 0.1 M PBS and incubated overnight at 4uC and slides were then washed for 10 minutes. The biotinylated goat anti-mouse secondary antibody m also was used at a dilution of 1 : 500, and was incubated for 1 hour at room temperature. Slides again were washed in 0.1 M PB for 10 minutes and incubated with streptavidin HRP peroxidase complex n at a concentration of 2.5 mg/ml for 30 minutes. The reaction was developed using diaminobenzidine counterstained with Gill s hematoxylin o, dehydrated, and coverslipped. Positive controls included were normal dog lymph node and thymus as well as normally dividing cells on the sample slides such as basal cells of the epidermis and germ cells of the hair follicles. Negative controls were obtained by omitting the primary antibody. The number of positive cells in 500 mast cells was counted in the tumor at 1,0003 magnification. 10 AgNOR Staining Sections were deparaffinized twice in xylene for 10 minutes each time, and rehydrated in a graded ethanol series to distilled, deionized water. The AgNOR solution was prepared by dissolving gelatin at a concentration of 2% into 1% aqueous formic acid. Gelatin and formic acid were mixed for 45 minutes with a stir bar on a warm stir plate to ensure complete dissolution of the gelatin. This solution then was mixed with 50% aqueous silver nitrate p at a ratio of 1 : 2 under darkened room conditions. The solution was added to each slide for 45 minutes in the dark at room temperature. The slides were rinsed with distilled, deionized water and coated with Cover Care aqueous mounting media h on a plate warmer for approximately 30 minutes. The number of granules in 100 nuclei was counted in the tumor at 1,0003 magnification. 15 Statistics For identification of prognostic variables, all dichotomous divisions of continuous independent variables (eg, age, weight, or size) were determined by use of the median value of the sample population. Independent variables evaluated were sex, age, weight, tumor size, tumor location (eg, head and neck, trunk, limb, genitoperineal, or intraoral), stage at time of surgery (metastasis was defined as presence of mast cells in the lymph nodes, liver, spleen, or bone marrow), character of surgical margins (mast cells at surgical edge versus mast cells within 1 mm of surgical edge), and presence or absence of preoperative treatments, postoperative treatments, and postoperative complications. For analysis of the effect that individual prognositc factors had on progression-free survival, time to local recurrence, time to new tumor locations, and the overall survival, a log rank test was used. Local recurrence and development of another cutaneous MCT also were evaluated as prognostic factors for overall survival. Cox regression analysis was used for multivariate analysis of the effects that MIB-1, PCNA, and AgNOR staining had on progression-free survival, time to local recurrence, time to new tumor locations, and the overall survival. The model was set up such that each individual variable (MIB-1, PCNA, and AgNOR staining), the product of each possible pair of variables and the product of all 3 variables together were tested in a stepwise forward fashion. The stepwise forward model was performed independent of P-values from the univariate analysis. The same model was set up for all 4 outcomes (progression-free survival, time to local recurrence, time to new tumor locations, and overall survival). Fischer s exact test was used to identify effects that values greater than the mean for MIB-1, PCNA, and AgNOR staining had on the presence of local recurrence or death due to disease. To estimate progression-free survival, time to local recurrence, time to new tumor locations, and overall survival, the Kaplan-Meier method was used. 19 Dogs that were lost to follow-up, alive at the end of study period, or died of causes not related to MCTs were censored at the date of last contact or date of death, respectively. All statistics were done using a commercially available statistical software package. q Significance of all analyses was established at P,.05. Results Twenty-eight dogs with 30 tumors were included in the study. Two dogs had 2 tumors at different time points (1,044 and 1,908 days after the excision of the 1st tumor for each dog), with both tumors excised with incomplete margins. Sixteen breeds were represented: Golden Retriever (n 5 5), mixed breed (n 5 5), Pug (n 5 3), Labrador Retriever (n 5 2), Boston Terrier (n 5 2), and one each for Bernese Mountain Dog, American Cocker Spaniel, Basset Hound, Dachshund, Shar Pei, Miniature Poodle, Pit Bull, American Eskimo, Whippet, Weimaraner, and Rottweiler. There were 6 sexually intact males, 5 neutered males, 1 sexually intact female, and 16 neutered females. The median age at the time of surgery was 7.8 years (range, years). The median weight was 25.9 kg (range, 9 50 kg). Tumors were located on the limb (n 5 10 [33.3%]), trunk (n 5 9 [30%]), head and neck (n 5 5 [16.7%]), genitoperineal region (n 5 4 [13.3%]), and intraoral area (n 5 2 [6.7%]). For those MCTs in which size was noted (n 5 17 [56.7%]), median tumor diameter was 3 cm (range, 1 12 cm). Twenty-six had tumor cells extending to the surgical edge and 4 had tumor cells within 1 mm of the surgical edge. None of the dogs had another surgery resulting in complete excision, chemotherapy other than prednisone, or radiation therapy for treatment of their disease. One dog received prednisone and azathioprine to treat immune-mediated polyarthritis, which had developed before removal of the MCT. Three (10.7%) dogs were treated with prednisone before surgery, 1 of which remained on prednisone after surgery (the dog with immune-mediated polyarthritis). None of these dogs had local recurrence or evidence of metastasis; however, 2 had tumors develop at other locations. Seven (25%) dogs were treated with prednisone therapy after surgery. Two of these dogs developed local recurrence, 3 had metastasis at the time of surgery (1 in the spleen and 2 in the regional lymph node), and none developed MCTs at another location. Prednisone therapy did not affect time to local recurrence (P 5.53), progression-free survival (P 5.69), overall survival (P 5.55), or time to other location (P 5.08). The median and mean times to other location for the group treated with prednisone could not be computed because all observations were censored and the median and mean times for dogs without prednisone were 1,112 and 1,225 days, respectively. Fifteen dogs were evaluated for metastasis at the initial presentation at UCD-VMTH by fine-needle aspirate of the regional lymph nodes, liver, spleen, or bone marrow.

4 936 Séguin et al Fig 1. Kaplan-Meier curve for local recurrence of 30 grade II mast cell tumors incompletely excised. A plus (+) indicates tumors that have been censored. Four dogs had their lymph node only evaluated; 1 dog had liver only; 1 dog had spleen only; 2 dogs had lymph node and spleen; 2 dogs had liver and spleen; 3 dogs had lymph node, liver, and spleen; and 2 dogs had lymph node, liver, spleen, and bone marrow evaluated. Six dogs had at least 1 site positive for presence of mast cells. Four dogs had mast cells in the regional lymph nodes, 1 dog had mast cells in the spleen, and 1 dog had mast cells in the regional lymph node and spleen. Regional lymph nodes were aspirated in 11 (39.3%) dogs. Five (45.4%) of 11 dogs had mast cells noted on cytologic examination. Liver aspirates were performed in 8 (28.6%) dogs with no mast cells noted in any of the samples. Splenic aspirates were performed in 10 (35.7%) dogs, with mast cells noted in 2 of the samples. Bone marrow aspirates were performed in 2 dogs and were negative for the presence of mast cells. Seven (23.3%) tumors recurred locally. Two tumor recurrences were confirmed by histology, 3 were confirmed by cytology, and 2 were assumed to be MCTs on the basis of a mass that recurred in the same location of the previously excised MCT. Median time to local recurrence was not reached (time to local recurrence: 2 tumors at 14 days and 1 each at 18, 31, 152, 531, and 1,426 days). Mean time to local recurrence as computed by the Kaplan-Meier method was 1,713 days (95% confidence interval [CI]: 1,353 2,072 days). None of the tumors that recurred were tumors in which tumor cells were within 1 mm of the surgical edge at initial excision. Therefore, 26.9% of MCTs with tumor cells extending to the surgical margin had local recurrence (7 of 26). No statistical difference was found between tumors with cells at the surgical edge and tumors with cells within 1 mm for local recurrence (P 5.26). For tumors with neoplastic cells at the surgical edge, the median time to local recurrence was not reached and the mean was 1,625 days. For tumors with neoplastic cells within 1 mm of the surgical edge, the median and mean times to local recurrence could not be computed because all observations were censored. The estimated proportions of tumors that recurred locally at 1, 2, and 5 years were 17.3, 22.1, and 33.3%, respectively (Fig 1). Local recurrence occurred on the limb (n 5 4 [57.1%]), genitoperineal region (n 5 2 [28.6%]), and head and neck (n 5 1 [14.3%]). Location was not prognostic for local recurrence (P 5.20). All dogs with local recurrence died during the study period. Four of the 7 dogs died for reasons related to MCT disease and 2 died of unknown causes. Median number of nuclei positive for PCNA was (range, 2 248) for nonrecurring tumors and (range, ) for recurring tumors. Median number of nuclei positive for Ki-67 was (range, 0 178) for nonrecurring tumors and (range, ) for recurring tumors. Median number of granules positive for AgNOR in 100 nuclei was (range, ) for nonrecurring tumors and (range, ) for recurring tumors. Eleven (39.3%) dogs developed MCTs at other locations, 3 of which also had local recurrence. Nine of these 11 dogs died during the study period, 5 of which died for reasons related to MCTs and 2 of which died of unknown causes. Six dogs had metastasis. All were diagnosed at the time of the surgery and all 6 of these dogs died during the study period. Three of these dogs died for reasons related to MCT disease. The median progressionfree survival was 1,044 days (95% CI: 412 1,676 days). Presence of metastasis was not prognostic for local recurrence (P 5.30), developing a MCT at a different location (P 5.27), or overall survival (P 5.77). Twenty-four (85.7%) dogs died during the study period. Median overall survival was 1,426 days (95% CI: 691 2,161 days). Seven dogs died from causes related to MCT disease, 10 died from causes unrelated to MCT disease, and 7 died from unknown causes. Median time to follow-up was days (range, 52 2,277 days) for the 30 tumors. The shortest follow-up time for dogs still alive at the end of the study period was 247 days. This dog was 1 of the 2 dogs that had 2 incompletely excised MCTs and the follow-up time on the 1st incompletely excised MCT was 1,295 days. Of all the statistical analyses performed to identify prognostic factors, only 2 were significant. First, the combination of the scores from the Ki-67 and PCNA stains in the Cox regression analysis was prognostic for local recurrence (P 5.03; SE ; CI: ). By taking into account these 2 scores, scores above the mean were prognostic for presence of local recurrence. Secondly, developing local recurrence was a negative prognostic factor for overall survival (P 5.04) (Fig 2). Discussion The goals of this study were to determine local recurrence rate, clinical outcome, and prognostic value of AgNOR, PCNA, and Ki-67 staining after incomplete surgical excision of grade II MCT in dogs. Results of incomplete resection of MCT in dogs with respect to local recurrence have been evaluated in other studies. The results of our study are similar to those of Misdorp, 9

5 Incomplete Excision of Mast Cell Tumors 937 Fig 2. Kaplan-Meier survival curves comparing dogs with tumor local recurrence and dogs without local recurrence. A plus (+) indicates dogs that have been censored. Michels et al, 7 and Murphy et al 8 in which no adjuvant therapy was performed after incomplete excision (23% versus 35, 18, and 19%). These results contradict the assumption that most MCTs will recur locally after incomplete excision. Other studies have reported local recurrence rates for incompletely excised MCTs with adjuvant therapies. Davies et al 20 reported a rate of 7% with adjuvant vinblastine and prednisolone. Rates of local recurrence with adjuvant radiation therapy have been reported to be 5, 14, and 20% Although the rates of local recurrence of studies in which no adjuvant treatments were performed cannot be compared statistically to those in which adjuvant treatments were administered, the rates are not very different and raise the question of the efficacy of adjuvant therapies. Because chemotherapy and especially radiation therapy are associated with considerable morbidity, establishing their efficacy becomes crucial. The results of all of these studies emphasize the importance of having a control group when evaluating an adjuvant treatment for its effects on local recurrence of MCT. The cause of this low rate of recurrence is unknown, but some possible reasons have been considered. Mast cells are an important part of the body s defense mechanism, being involved in hypersensitivity reactions, acute and chronic inflammatory processes, immunocyte stimulation, and protection from parasites. 24 Consequently, many of the mast cells seen at the periphery of histologic sections of resected MCTs may in fact be part of a local inflammatory reaction, or attracted to the neoplasm itself, and this possibility highlights the problem of not being able to differentiate between normal and neoplastic mast cells in the skin. 5,7 Results in 1 dog in a previous study suggest that some normal mast cells indeed may be present close to a MCT. A cluster of well-differentiated mast cells was observed 3 cm from the tumor at histology, and was it concluded that these were not neoplastic mast cells. 25 When incomplete surgical excision is represented only by microscopic disease, residual cells may be killed by the host s immune system, which likely already has mounted a response to the neoplasm. 7,9 Finally, factors inhibiting growth of the residual cells also may be released from the surrounding soft tissues. 7,9 We found that local recurrence was prognostic for overall survival of dogs with incompletely excised grade II MCTs (P 5.04). Dogs developing local recurrence had shorter survival times than dogs without local recurrence (Fig 2). In another study, local recurrence was associated with progression to malignancy (multicentric regrowth, metastasis, or both) in 59% of dogs. 10 Between 25 and 50% of dogs (7 of 28 died of causes related to MCTs and 7 died of unknown causes) died of causes related to MCTs in our study. Considering these results, we must therefore still recommend adjuvant therapy for incompletely excised MCTs (because it is currently the standard of care) 4 until further information becomes available regarding the impact of omitting additional treatment on the overall prognosis for affected dogs. The results of the proliferation indices in our study were similar to those reported in other studies. The mean (6 standard deviation [SD]) for PCNA-positive cells in 500 cells was in tumors that did not recur and in tumors that did recur in one study 15 and the mean number of PCNA-positive nuclei/1,000 cells was for grade II MCTs (therefore, approximately per 500 cells) in another sudy. 10 In our study, the median was for nonrecurring tumors and for recurring tumors counted in 500 cells. The mean 6 SD number of AgNOR granules per nucleus counted in 100 nuclei was in nonrecurring tumors and in recurring tumors in a previous study. 15 Our results were a median of counted in 100 nuclei ( per nucleus) for nonrecurring tumors and counted in 100 nuclei ( per nucleus) for recurring tumors. The mean number of Ki-67 positive nuclei/1,000 cells was (therefore, approximately in 500 cells) for grade II MCTs in another previous study. 10 Our results were a median of for nonrecurring tumors and for recurring tumors counted in 500 cells. Although statistical comparison is not possible and different studies may have different definitions for an event (ie, it is not clear what was considered recurrence in the Simoes et al 15 study, meaning it is not clear if a recurrence was only at the local level or also included tumors at a different cutaneous site), the fact that our results are not logarithmically different than those previously reported suggests that the tumors examined in this study are similar in their proliferative capacity to other MCTs that have been studied. There were several important limitations in this retrospective study. The small numbers of samples for which follow-up information was available makes drawing statistical conclusions difficult. To that effect, paraffin-embedded tissue only was available from 3 tumors that had recurred as compared to 17 tumors that did not recur. Furthermore, not all information was available for each MCT, such as tumor size, staging

6 938 Séguin et al results, and specific cause of death. Differences in these factors could have skewed the results. No postmortem examinations were performed on any of the dogs in this study. Therefore, the rate of metastasis and local recurrence might be higher than reported here. Type I error may have occurred in which the null hypothesis was rejected regarding the prognostic significance of the combination of PCNA and Ki-67 scores. As mentioned previously, very few tumor samples in each group (recurring versus nonrecurring) were available to test the different staining methods. The reasons for this result are, because of the current recommendation to perform adjuvant therapy when a MCT is incompletely excised, most dogs received adjuvant therapy; the rate of local recurrence was relatively low (23%); and the paraffin blocks of some tumors were unavailable for testing. It is possible that with more samples, especially in the local recurrence group, different results and conclusions would have been obtained. Type II error also may have occurred such as the prognostic significance of developing a postoperative complication and the presence of local recurrence. This study found a P-value of With a larger number of cases in the study, it is possible that statistical significance would have been reached (P,.05). If it were true that postoperative complications do increase the likelihood of a local recurrence, a plausible explanation is that the complication leads to an enhanced inflammatory response, which in turn could stimulate and enhance the ability of the remaining mast cells to divide. Alternatively, it is possible that the complication is the effect rather than the cause of local recurrence, that is tumors that will recur obviously have cells that are proliferating and this may lead to complications. Given that these are mast cells containing a multitude of molecules such as histamine and proteolytic enzymes, as the tumor cells proliferate some may degranulate thereby increasing the likelihood of a complication. We limited our study to grade II MCTs because of the prognosis for the other 2 grades and the clinical implications. Dogs with a grade I MCT (well differentiated) have an excellent prognosis for survival and local control after local excision. 2,4,8 Therefore, additional therapy rarely is recommended. Dogs with a grade III MCT (undifferentiated) have a poor prognosis for survival and a high rate of metastasis. 2,4 Some oncologists do not routinely recommend aggressive local treatment of these tumors because of the poor prognosis for survival. 2 Additionally, chemotherapy usually is recommended for these dogs and could have an impact on the local tumor. 2 However, grade II MCTs have a wide range of biological behavior and predictions of clinical outcome are difficult for these tumors. 5,10 Therefore, we wanted to concentrate on grade II MCTs to characterize their behavior. Furthermore, by limiting the study to grade II MCTs, we have eliminated the confounding effect that grade may have on local recurrence, as evidenced by the conflicting results of the studies by Misdorp 9 and Murphy et al. 8 A larger study including grade I and III MCTs is needed to determine the rate of local recurrence after incomplete excision for all grades of MCTs and to determine if grade can be a prognostic factor for local recurrence. Treatment with prednisone as adjuvant therapy was not a reason to eliminate cases in this study. In a previous study, 20% of dogs responded to prednisone therapy for their MCT and 80% of the dogs responding underwent partial remission, with the responses generally being of short duration. 26 Because of the low number of responders, we elected not to disqualify dogs that received prednisone. As a matter of fact, 25% of dogs that received prednisone after surgery developed local recurrence. If dogs that received postoperative prednisone had been eliminated, the rate of local recurrence would have been 22.7%. In our study, treatment with prednisone did not affect time to local recurrence (P 5.53), time to other location (P 5.08), progression-free survival (P 5.69), or overall survival (P 5.55). Because our study was retrospective, dogs that received prednisone were not randomized. As a consequence of this bias and small sample size, we may have found no effect of prednisone while committing a type II statistical error. One dog also was on prednisone and azathioprine. This dog was included in the study despite having received azathioprine. The prednisone and azathioprine were administered to treat immune-mediated polyarthritis, which had developed before the onset of the MCT. Although azathioprine is immunosuppressive and may have an effect on the biological behavior of mast cells, azathioprine is not known to affect MCTs. In our study, 4 MCTs had tumor cells within 1 mm of the surgical edge but not extending to the surgical margin. Tumor cells within 1 mm of the surgical edge has been defined as incomplete excision. 18 Because of the very close proximity of the tumor cells to the surgical edge when they are within 1 mm, it may be argued that these tumors should be considered incomplete resections. For example, it could be argued that if the paraffin-embedded tissue had been cut only a few millimeters deeper, tumor cells may have been identified in the surgical edge. However, in 2 other studies, tumor cells within 1 mm of the surgical edge were considered complete but close. 5,25 Combining the data of these 2 studies and the present study, 8 tumors were excised with tumor cells within 1 mm of the surgical edge and without adjuvant therapy (except 2 tumors in our study in which the dogs were administered prednisone postoperatively). Of those 8, only 1 developed local recurrence. In our study, having cells within 1 mm of the surgical edge did not decrease the chance of local recurrence compared to cells extending to the edge (P 5.26). However, none of the tumors in our study with cells within 1 mm of the margins developed local recurrence. It is possible that our study did not have sufficient power to detect a statistical difference. However, another study did not find a difference in outcome for local recurrence between complete and incomplete margins. 7 It is unknown if MCTs developing at a different cutaneous location represents metastatic spread of a neoplastic cell from the original cutaneous MCT or

7 Incomplete Excision of Mast Cell Tumors 939 if these neoplastic cells arise de novo. Consequently, in our study MCTs developing at a different cutaneous location and metastasis were considered different events. Although the staging system of the World Health Organization (WHO) considers multiple dermal tumors as stage III and therefore having a worse prognosis than a single dermal tumor (stage I), 2,27 this classification does not appear to be relevant clinically. 21,28,r Because of a lack of difference in prognosis, it seems inappropriate to consider development of other cutaneous locations necessarily as metastasis. The proportion of dogs developing MCTs at a different location in our study (39%) may be higher than that reported in other studies (11 33%). 5,25 The higher proportion in our study may be due in part to the longer follow-up time. Our study had a median follow-up time of days compared to 540 days 5 and 351 days 25 in other studies. Another controversy regarding the staging protocol lies in what constitutes metastasis, judged on the basis of cytologic examination. Clinically healthy dogs have mast cells in their lymph nodes, liver, spleen, and bone marrow. 29,s Because WHO guidelines suggest that cytologic specimens need only be evaluated for the presence or absence of mast cells, presence of mast cells in regional lymph node, spleen, liver, or bone marrow was considered as metastasis in our study regardless of numbers. 27,29 Some of the dogs in our study that were considered to have metastasis may not truly have had MCT metastasis. Furthermore, in our study, presence of metastasis was not prognostic for local recurrence (P 5.30), developing MCTs at a different location (P 5.27), or overall survival (P ). Therefore, the criteria of presence or absence of mast cells may not be adequate to determine the status of metastasis. In our study, only 23% of incompletely excised grade II MCTs recurred locally with no additional therapy. Therefore, ancillary local therapies may not always be necessary. The combination of the scores from PCNA and Ki-67 staining may help determine which tumors are more likely to recur. Developing local recurrence was a negative prognostic factor for overall survival. Therefore, additional prospective studies to determine which incompletely excised grade II MCTs require additional therapy are needed before the standard of care is changed. Footnotes a Citrate buffer, Dako, Carpinteria, CA b Serum-free protein block, Dako, Carpinteria, CA c Proliferating cell nuclear antigen, Dako, Carpinteria, CA d Biotinylated goat anti-mouse immunoglobulin G, BioCare, Walnut Creek, CA e Steptavidin horseradish peroxidase, BioCare, Walnut Creek, CA f Aminoethyl carbazole, Zymed, South San Francisco, CA g Meyers hematoxylin, Sigma, St Louis, MO h Cover Care aqueous mounting medium, BioCare, Walnut Creek, CA i Mouse melanoma IgG1, Zymed, South San Francisco, CA j pressure cooker, Russell Hobbs, New Haven, CT k 0.1 M citrate buffer, ph 6.0, Chemicon, Temecula, CA l MIB-1 antibody, Dako, Carpinteria, CA m Biotinylated goat anti-mouse secondary antibody, Chemicon, Temecula, CA n Steptavidin horseradish peroxidase complex, Zymed, South San Francisco, CA o Gill s hematoxylin, Vector Laboratories, Burlingame, CA p 50% aqueous silver nitrate, Sigma, St Louis, MO q SPPS Version 11 for Mac OS X, SPSS Inc Headquarters, Chicago, IL r Mullins M, Dernell W, Withrow S, et al. Multiple cutaneous mast cell tumors: A retrospective analysis. Proc Vet Cancer Soc 2004;13 (abstract) s Leibman NF, Guilpin VO, Fettman MJ, et al. Cytologic comparison of mast cell numbers in liver and spleen of normal dogs and dogs with mast cell tumors. Proc Vet Cancer Soc 1999;28 (abstract) Acknowledgments This work was supported by a grant from the Center for Companion Animal Health, School of Veterinary Medicine, University of California, Davis. We thank Idexx Reference Laboratories for providing the tissue samples processed at their facilities, Diane K. Naydan for technical assistance, and Dr Philip J. Bergman and Diane Craft for their technical advice. References 1. Macy DW. Canine mast cell tumors. Vet Clin North Am Small Anim Pract 1985;15: Thamm DH, Vail DM. Mast cell tumors. In: Withrow SJ, MacEwen EG, eds. Small Animal Clinical Oncology, 3rd ed. Philadelphia, PA: WB Saunders; 2001: O Keefe DA. Canine mast cell tumors. Vet Clin North Am Small Anim Pract 1990;20: London CA, Seguin B. Mast cell tumors in the dog. Vet Clin North Am Small Anim Pract 2003;33: Seguin B, Leibman NF, Bregazzi VS, et al. Clinical outcome of grade II mast cell tumors treated with surgery alone in dogs: 55 cases ( ). J Am Vet Med Assoc 2001;218: Weisse C, Shoefer FS, Sorenmo K. Recurrence rates and sites for grade II canine mast cell tumors following complete surgical excision. J Am Anim Hosp Assoc 2002;38: Michels GM, Knapp DW, DeNicola DB, et al. Prognosis following excision of canine cutaneous mast cell tumors with histopathologically tumor-free versus nontumor-free margins: A retrospective study of 31 cases. J Am Anim Hosp Assoc 2002; 38: Murphy S, Sparkes AH, Smith KC, et al. Relationships between the histological grade of cutaneous mast cell tumours in dogs, their survival and efficacy of surgical excision. Vet Rec 2004; 154: Misdorp W. Incomplete surgery, local immunostimulation and recurrence of some tumour types in dogs and cats. Vet Q 1987;9: Abadie JJ, Amardeilh MA, Delverdier ME. Immunohistochemical detection of proliferating cell nuclear antigen and Ki-67 in mast cell tumors from dogs. J Am Vet Med Assoc 1999;215: Rogers KS. Mast cell tumors. Dilemmas of diagnosis and treatment. Vet Clin North Am Small Anim Pract 1996;26: Bostock DE, Crocker J, Harris K, et al. Nuclear organizer regions as indicators of post-surgical prognosis in canine spontaneous mast cell tumors. Br J Cancer 1989;59:

8 940 Séguin et al 13. Kravis LD, Vail DM, Kisseberth WC, et al. Frequency of argyrophilic nuclear organizer regions in fine-needle aspirates and biopsy specimens from mast cell tumors in dogs. J Am Vet Med Assoc 1996;209: Patnaik AK, Ehler WN, MacEwen EG. Canine cutaneous mast cell tumors: Morphologic grading and survival time in 83 dogs. Vet Pathol 1984;21: Simoes JPC, Schoning P, Butine M. Prognosis of canine mast cell tumors: A comparison of three methods. Vet Pathol 1994; 31: Bostock DE. The prognosis following surgical removal of mastocytomas in dogs. J Small Anim Pract 1973;14: Fenoglio-Preiser CM. Selection of appropriate cellular and molecular biologic diagnostic tests in evaluation of cancer. Cancer 1992;69: Powers BE. The pathology of neoplasia. In: Withrow SJ, MacEwen EG, eds. Small Animal Clinical Oncology, 3rd ed. Philadelphia, PA: WB Saunders; 2001: Kaplan G, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Soc 1958;53: Davies DR, Wyatt KM, Jardine JE, et al. Vinblastine and prednisolone as adjunctive therapy for canine cutaneous mast cell tumors. J Am Anim Hosp Assoc 2004;40: Turrel JM, Kitchell BE, Miller LM, et al. Prognostic factors for radiation treatment of mast cell tumor in 85 dogs. J Am Vet Med Assoc 1988;193: LaDue T, Price GS, Dodge R, et al. Radiation therapy for incompletely resected canine mast cell tumors. Vet Radiol Ultrasound 1998;39: Frimberger AE, Moore AS, LaRue SM, et al. Radiotherapy of incompletely resected, moderately differentiated mast cell tumors in the dog: 37 cases ( ). J Am Anim Hosp Assoc 1997;33: Scott MA, Stockham SL. Basophils and mast cells. In: Feldman BF, Zinkl JG, Jain NC, eds. Schalm s Veterinary Hematology, 5th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2000: Simpson AM, Ludwig LL, Newman SJ, et al. Evaluation of surgical margins required for complete excision of cutaneous mast cell tumors in dogs. J Am Vet Med Assoc 2004;224: McCaw DL, Miller MA, Ogilvie GK, et al. Response of canine mast cell tumors to treatment with oral prednisone. J Vet Intern Med 1994;8: Owen LN. TNM Classification of Tumors in Domestic Animals, 1st ed. Geneva, Switzerland: World Health Organization; 1980: Thamm DH, Mauldin EA, Vail DM. Prednisone and vinblastine chemotherapy for canine mast cell tumor: 41 cases ( ). J Vet Intern Med 1999;13: Bookbinder PF, Butt MT, Harvey HJ. Determination of the number of mast cells in lymph node, bone marrow, and buffy coat cytologic specimens from dogs. J Am Vet Med Assoc 1992;200: