Prognosis Following Surgical Excision of Canine Cutaneous Mast Cell Tumors With Histopathologically Tumor-Free Versus Nontumor-Free Margins: A Retrospective Study of 31 Cases

Introduction

Mast cell tumors (MCTs) are one of the most common cutaneous tumors in the dog, representing 7% to 21% of all canine skin tumors and 11% to 27% of all malignant canine skin tumors.¹ The biological behavior of MCTs is highly variable, ranging from solitary cutaneous tumors curable by surgery to single or multiple tumors with widespread metastasis. Because its behavior varies so greatly, it is important, yet difficult, to be able to predict how a MCT is going to behave in an individual dog. Factors that have been reported to be prognostic for survival include histopathological grade,^2–4 clinical stage,⁵⁶ location,⁶ chromosome nucleolar organizer regions stained with silver (AgNORs),⁴⁷ proliferate cell nuclear antigen (PCNA) staining,⁴ growth rate,² breed,² and recurrence.² Although these may be helpful, the biological behavior of MCTs continues to be unpredictable. Prognosis for survival following surgical removal of MCT with grossly tumor-free margins in dogs that have no other clinical evidence of MCT has been reported.²³ Limited information is available on the importance of histopathologically tumor-free margins.⁸

When a MCT is removed and the histopathology report confirms tumor-free or nontumor-free margins, it is not presently known how this information affects the prognosis and what additional therapy, if any, is appropriate. This retrospective study of dogs with surgically excised, cutaneous MCTs was performed to determine if histopathologically tumor-free versus nontumor-free margins were associated with the frequency of relapse (locally or distantly), the time to relapse, or survival. The authors’ null hypothesis was that following surgical excision of single or multiple cutaneous MCTs confined to the skin where grossly tumor-free margins are obtained and metastasis to lymph nodes or other noncutaneous sites is not identified, there is no difference in the percentage of dogs that relapse or the percentage of dogs that die due to MCT between dogs with histopathologically tumor-free margins and dogs with histopathologically nontumor-free margins.

Materials and Methods

The Veterinary Medical Data Base for Purdue was used to retrospectively identify dogs that had undergone surgical removal of cutaneous MCT at the Purdue University Veterinary Teaching Hospital (PUVTH) between January 1980 and June 1995. In all cases, preoperative histopathology or cytopathology was consistent with MCT. The diagnosis of MCT was confirmed by histopathological examination of the surgically excised tissue. Criteria required for inclusion into this study were no prior history of MCT; single or multiple cutaneous MCT with no evidence of regional lymph-node or distant metastasis; tumor-free margins grossly observed following surgery at PUVTH; and histopathological evaluation of surgically excised tissue and review of surgical margins (when margins were in question, these were identified and labeled by the attending surgeon for closer examination). Dogs that received chemotherapy or radiation therapy prior to recurrence of MCT, or for which no postoperative follow-up was available, were excluded from this study.

Data was collected from case records and included signalment, body weight, results of clinical staging procedures, tumor duration, location of the tumor (defined as head and neck, trunk, or limb), tumor size, histopathological grade, results of histopathological evaluation of surgical margins, treatment, disease-free interval, time to relapse, survival time, and cause of death. Follow-up information was obtained by examination of medical records at PUVTH or telephone or written survey of veterinarians, owners, or both. All relapses were confirmed by cytopathology, histopathology, or both. Those dogs with histopathologically tumor-free margins were compared to those with nontumor-free margins for differences in age, weight, sex, tumor duration, location and histopathological grade, disease-free interval, survival, frequency of relapse, and subsequent systemic mastocytosis. In addition, data from all dogs was evaluated to determine if age, weight, sex, tumor site, tumor duration, or histopathological grade could serve as prognostic factors for frequency of relapse, disease-free interval, survival, or the development of systemic mastocytosis. Tumor duration was measured from the time the tumor was noticed until it was removed. In this study, “relapse” is defined as the postoperative recurrence of a cutaneous MCT either locally or distant to the initial tumor. In addition, “local relapse” is defined as relapse within the surgical scar or near the surgical site.

Surgical removal of MCTs was performed or supervised by a Diplomate of the American College of Veterinary Surgeons at the PUVTH. Excision with 3 cm of normal tissue cranially, caudally, laterally, and deep to the tumor was routinely attempted. If removal of 3-cm margins deep to the tumor was not possible, the excision extended at least one fascial plane deep to the lesion.

Staging procedures were performed at the discretion of the attending clinician. All dogs had complete physical examinations. Additional staging procedures in the 31 dogs included thoracic radiographs (n=28), abdominal radiographs (n=26), bone-marrow cytopathology (n=22), abdominal ultrasonography (n=8), buffy-coat examination (n=5), lymph-node aspiration cytopathology (n=4), lymph-node biopsy (n=2), liver aspiration cytopathology (n=1), and bone-marrow biopsy (n=1).

One pathologist (DeNicola) reviewed all histopathological specimens. Only those cases in which the biopsy material was sufficient to clearly determine if the tumor margins were tumor-free or nontumor-free were included. For the purpose of this retrospective study, histopathological tumor-free margins were defined as the absence of tumor cells in the submitted surgical margins.

Results

Thirty-one of 132 dogs evaluated at PUVTH for MCT fulfilled the criteria for inclusion in this study [Table 1]. Of these, 20 (65%) dogs had tumor-free margins, and 11 (35%) had nontumor-free margins. The 31 dogs included seven mixed-breed dogs, five Labrador retrievers, four boxers, two golden retrievers, two poodles, and one each of Brittany spaniel, Bouvier des Flandres, cocker spaniel, collie, dachshund, English pointer, Great Dane, puli, schnauzer, vizsla, and Weimaraner. There were significantly more boxers (P<0.001) and Labrador retrievers (P<0.01) compared to the canine hospital population for the same period.

All dogs were followed until relapse, death, or for ≥30 months [Tables 2, 3]. This allowed postoperative follow-up for ≥12 months for 95% (19/20) and 55% (6/11) of the tumor-free and nontumor-free margins groups, respectively. Follow-up information until death was obtained in 50% (10/20) and 100% (11/11) of the tumor-free and nontumor-free margins groups, respectively. Necropsy performed in three dogs (case nos. 11, 15, and 27) did not reveal histopathological evidence of MCT. Nineteen percent (6/31) of the dogs are still alive (4/31 were lost to follow-up), all of which are in the tumor-free margin group. Thirty-six percent (4/11) of dogs in the nontumor-free margin group died of causes unrelated to MCT <12 months postoperatively, compared to none (0/19) in the tumor-free margin group. The length of follow-up was significantly greater (P<0.05) for the tumor-free (mean±standard deviation [SD], 47.0±25.0 months; median, 43 months) versus nontumor-free margins groups (mean±SD, 24.1±20.8 months; median, 15 months).

There were no significant differences in age, weight, sex, neuter status, tumor duration, tumor site, histopathological grade, or frequency of death due to MCT between the tumor-free and nontumor-free margins groups. With respect to tumor duration, dogs in the tumor-free margin group had a preoperative tumor duration of 2.8±3.1 months (mean±SD) compared to 17.7±23.6 months (mean±SD) for those in the nontumor-free margin group (P=0.07).

Forty-eight percent (15/31) of the dogs in this study relapsed (range, 4 to 67 months). All first relapses occurred in the skin, with no evidence of systemic metastasis. Distant relapse in the skin occurred in 80% (12/15), and local relapse in the skin occurred in 20% (3/15). There was no significant difference in the frequency of local relapse or distant relapse between the tumor-free (1/20, 5%; 8/20, 40%) and nontumor-free margins groups (2/11, 18%; 4/10, 40%), respectively.

The median time to relapse in the nontumor-free margin group was 7.5 months [Table 2], compared to 13 months for the tumor-free margin group. The frequency of relapse by 6, 12, and 24 months was 20% (2/10), 71% (5/7), and 86% (6/7), respectively, for the nontumor-free margin group compared to 20% (4/20), 21% (4/19), and 28% (5/18), respectively, for the tumor-free margin group. Thus, the frequency of relapse by 6 months was the same for both groups but was significantly greater by 12 and 24 months for the nontumor-free margin group compared to the tumor-free margin group (P<0.05).

The median survival time for dogs in the nontumor-free margin group was 15 months (range, 5 to 60 months). Two dogs (case nos. 25, 26; Tables 2, 3) had local relapse in the skin and subsequently died of systemic MCT at 11 and 15 months, postoperatively. A third dog (case no. 31) had distant relapse in the skin at two sites at 7 months. The dog was treated with prednisone for several months because one of the tumors was not amenable to surgery. The tumor responded transiently, and the owner discontinued the prednisone. The dog died of unrelated causes at 60 months with multiple cutaneous MCTs of approximately 2 years’ duration. There was no evidence that MCT was the cause of death in the remaining dogs in the nontumor-free margin group, resulting in an 18% (2/11) frequency of death due to MCT.

The median survival time for dogs in the tumor-free margin group, to date, is 53.5 months. However, the median survival time has not been reached, because six dogs are still living. At 24 months postoperatively, 10% (2/19) of dogs in the tumor-free margin group were dead, compared to 64% (7/11) of those in the nontumor-free margin group. Of those dogs that have died, the postoperative survival time was longer for the tumor-free margin group (mean±SD, 50.6±26.8 months) versus the nontumor-free margin group (mean±SD, 21.6±19.9 months) (P<0.05). One dog (case no. 9) in the tumor-free margin group was treated surgically for local relapse (6 inches from the primary tumor) at 30 months, but was lost to follow-up at 36 months. A second dog (case no. 13) was treated surgically by another veterinarian for distant relapse in the skin at 34 and 49 months. This dog had local relapse in the skin, with metastasis to a regional lymph node following the second relapse, and died from systemic mastocytosis 54 months after the initial surgery. In the remaining nine dogs for which follow-up was available until death, there was no evidence that MCT was the cause of death. Thus, the frequency of death due to MCT in the tumor-free margin group was 10% (1/10). There was no significant difference in frequency of death due to MCT between the tumor-free and nontumor-free margins groups.

In addition to the comparison between the two surgical margins groups, data from all dogs was analyzed to determine if factors such as age, weight, gender, neuter status, tumor duration, site, or histopathological grade were prognostic indicators for frequency of relapse, time to relapse, survival, or frequency of death due to MCT. In this study, age was the only prognostic indicator identified, with older dogs having shorter postoperative survival times compared to younger dogs (P<0.01). In addition, older dogs appeared to relapse more quickly than younger dogs, although the difference was not significant (P=0.07). The median age for all dogs in this study was 8 years, with a range of 2 to 14 years.

Discussion

The purpose of this study was to determine if the presence of histopathologically tumor-free versus nontumor-free margins was prognostic for relapse or tumor-related death in dogs with single or multiple dermal MCT. A lack of distant metastasis, lymph-node involvement, or previous history of MCT was required for inclusion in the study. In addition, the two groups were compared to determine if there were differences in tumor duration, location, histopathological tumor grade, time to relapse, and survival. Data from all dogs was evaluated to determine if prognostic factors could be identified for frequency of relapse, time to relapse, survival, or tumor-related death.

There were significantly more boxers and Labrador retrievers in this study compared to the hospital population for the same period. A breed predisposition has been suggested to occur in boxers, Labrador retrievers, and several other breeds that were not over-represented in this study, including Boston terriers, bull terriers, schnauzers, and beagles.³⁹ To the authors’ knowledge, no published study has compared a patient population with a reference population to determine if cutaneous MCTs are over-represented in certain breeds.

Surprisingly, failure to obtain tumor-free margins often did not result in local relapse. Only two of 11 dogs in the nontumor-free margin group relapsed in the surgical field. The reasons for this lack of recurrence are not known but could be theorized to include immunological response eliminating microscopic tumor; alteration in a subpopulation of tumor cells in the tumor, not allowing sustainment of growth (such as lack of key growth factor removed by one subpopulation of cells); tumor dormancy; or the observed mast cells in the tumor margins being a component of the inflammatory response versus tumor. It is of interest to note, however, that all three tumor-related deaths (case nos. 13, 25, and 26) occurred following local relapse (one following local recurrence at a metastatic site [case no. 13] and two following local recurrence at the primary site [case nos. 25, 26]). These results suggest that obtaining histopathological tumor-free margins may decrease tumor-related deaths. The lack of statistical support for such a statement in this study may be a result of small sample size.

The percentage of dogs that relapsed (48%; 15/31) is similar to that reported by Patnaik, et al.³ and Bostock.² However, recurrence was associated with a poor prognosis in these studies where 100% and 86% of relapses, respectively, resulted in death. In contrast, only 20% (3/15) of relapses led to tumor-related deaths in this study.

The postoperative survival times for dogs in the tumor-free margin group were significantly longer (median, 53.5 months to date) compared to those in the nontumor-free margin group (median, 15 months). However, there was no difference in the number of tumor-related deaths. The dogs in the tumor-free margin group may have survived longer because they were younger when the tumors were first removed, compared to the dogs in the nontumor-free margin group. It is interesting to note that there is a similar difference between the median ages at diagnosis (4-year difference) and the median survival times (3.2-year difference to date; median survival time not reached in the tumor-free margin group) in the tumor-free and nontumor-free margins groups. In addition, none of the dogs in the tumor-free margin group died of causes unrelated to MCT ≤12 months postoperatively, compared to 36% (4/11) in the nontumor-free margin group. Therefore, it cannot be determined how many dogs in the nontumor-free margin group would have relapsed or died of MCT if they had lived longer.

Dogs in the nontumor-free margin group had a shorter preoperative tumor duration compared to those in the tumor-free margin group, although the difference was not significant. In contrast to a previous study,² there was no association between preoperative tumor duration and recurrence, tumor-related death, or survival.

Older dogs appeared to relapse more quickly following surgery than younger dogs, although the difference was not significant. One recent study of dogs treated with radiation for MCT reported a similar finding, with dogs <7.5 years of age having a longer disease-free interval than dogs >7.5 years of age.⁹ It would be necessary to compare dogs grouped by age and weight to determine if such a difference was meaningful. However, the number of animals in the present study was too small to evaluate in this manner.

Recurrence of MCTs has been associated with a poor prognosis in other studies. In a study by Patnaik, et al., where some dogs were followed as long as 1,500 days, 46% (38/83) of dogs relapsed, and 100% of these died of MCT following removal of cutaneous MCT with “wide margins.”³ Although the frequency of distant versus local skin relapse was not stated, the dogs had no other clinical evidence of MCT at the time of surgery. Bostock² reported that 43% (49/114) of dogs that had no other clinical evidence of MCT at the time of surgery relapsed following tumor excision with “grossly tumor-free margins.” Of those that had recurrence, death as a direct result of MCT was documented in 86% (42/49). The remaining 14% (7/49) were not followed to death or died of other causes.²

Histopathological tumor grade was similar between dogs with tumor-free versus nontumor-free margins and was not associated with frequency of relapse or frequency of tumor-related death. Only two of 11 dogs in the nontumor-free margin group relapsed within the surgical site. In contrast, other studies have made a correlation between tumor grade and frequency of relapse.² It is possible that histopathological tumor grade is not an important prognostic indicator in dogs where histopathological tumor-free margins are obtained.

There was a significant association between tumor-free versus nontumor-free margins and the frequency of relapse by 12 and 24 months. In the tumor-free versus nontumor-free margins groups, 21% (4/19) versus 71% (5/7) and 28% (5/18) versus 86% (6/7) relapsed by 12 and 24 months, respectively. The percentage of dogs that relapsed by 6 months was the same in both groups (20%; 4/20 and 2/10, respectively). Thus, failure to obtain histopathological tumor-free margins may be associated with a greater frequency of relapse by 12 and 24 months.

Conclusion

In this retrospective study, the authors evaluated dogs surgically treated at PUVTH for their first episode of cutaneous MCT. All dogs had no evidence of lymph-node involvement or distant metastasis at the time of surgery. In each case, the surgeon attempted to obtain 3 cm of normal tissue cranially, caudally, and laterally. Deep margins extended at least one fascial plane deep to the lesion if 3-cm margins in this plane could not be obtained. Only those cases where biopsy material was sufficient to clearly determine if the tumor margins were tumor-free or nontumor-free were included. Cases were followed until relapse or death or for at least 30 months. Results of this study showed that significantly more dogs in the nontumor-free margin group relapsed by 12 and 24 months postoperatively compared to the tumor-free margin group. However, histopathological tumor-free versus nontumor-free margins were not associated with a different frequency of tumor-related death. Only two of 11 dogs in the nontumor-free margin group relapsed within the surgical site. It is important to note that all tumor-related deaths occurred following local relapse. The lack of statistical support for an association between prognosis and histopathological tumor-free versus nontumor-free margins may be a result of small sample size. In conclusion, there were no significant differences in tumor-related death or relapse rate between dogs in the tumor-free and nontumor-free margins groups following surgical excision of cutaneous MCTs.