Cutaneous MCTs: Associations with Spay/Neuter Status, Breed, Body Size, and Phylogenetic Cluster

Introduction

Mast cell tumor (MCT) is the most common cutaneous tumor in the dog, accounting for 16–21% of all cutaneous tumors.^1–3 Grade 2 and grade 3 (using the Patnaik system) MCTs are clinically important as they are commonly associated with substantial morbidity and mortality.⁴ Over the past 20 yr, significant advances have been made in the treatment of canine grade 2 and grade 3 MCTs with surgery, chemotherapy and radiation therapy.^4–6 These treatments have resulted in improved survival times for dogs suffering from the disease; however, treatment may place a substantial financial burden on the dogs' owners and does not result in a cure for every dog treated. Understanding the risk factors involved in the development of grade 2 and grade 3 MCTs may allow veterinarians to identify high-risk dogs and institute early detection strategies.

Associations between specific dog breeds and grade 2 and grade 3 MCT risk have been reported; however, no consistent spay/neuter or body size associations have been noted in the currently available studies of MCTs.^7–12 The primary purpose of the current study was to compare dogs with histologically confirmed grade 2 and grade 3 cutaneous MCTs to a control group of dogs without MCTs in regard to spay/neuter status, breed, phylogeny-based breed group, and breed standard weight group.¹³

Materials and Methods

In this case-control study, characteristics of dogs with MCT seen at the Animal Medical Center, New York, NY with MCT were compared with characteristics of a sample of dogs seen at the Animal Medical Center without MCT. Cases were dogs with a histopathological diagnosis of grade 2 or grade 3 cutaneous MCT (based on a surgical biopsy) diagnosed between Jan 1, 1997 and Dec 31, 1998 and between Jan 1, 2005 and Dec 31, 2006. These dates were chosen because they represented the earliest and most recent time periods when age, spay/neuter status, and breed could be obtained from the hospital's searchable computerized database. Dogs diagnosed with grade 1 MCT were excluded from this study because grade 1 tumors exhibit different biologic behavior and carry an excellent prognosis with surgical excision as the sole therapy.⁴ Cases were identified by searching the surgical biopsy database of the hospital for a diagnosis of grade 2 or 3 cutaneous MCT.

The control population was selected with a goal of identifying approximately five controls for each case. Each patient seen at the hospital had been assigned (consecutively) a six-digit medical record number at the time of the initial visit. To choose approximately 2,000 control dogs (500/yr of this 4 yr study) from the approximately 15,000 canine patients/yr visiting the hospital, the number six was randomly chosen. Every third patient number ending with the digit six was selected to obtain approximately one thirtieth of the canine hospital population during the study period as controls. Duplicate cases and controls were then eliminated. Due to the tertiary care role of the hospital, <2% of the control population presented for routine spays and neuters.

Spay/neuter status, sex, breed, and age for both the cases and controls were obtained from the medical records database of the hospital. Age at time of spay/neuter was available for neither cases nor controls and discharge diagnoses were not available for the control dogs. The cases and controls were also classified by phylogenetic clusters, as established by Parker et al. (2007), in an attempt to see if certain genetically related breeds were at an increased risk for developing MCT, as has previously been proposed by Peters (1969).^8,13 Purebreed dogs not included in the analysis were classified as “others combined.”

Body weights of individual dogs were not included in the database used in the present study, so to evaluate the impact of size on MCT development, breed standard weight groups were assigned to each breed. Purebreed dogs were classified into four size groups as described by Hayes et al. (1994): small (<9 kg), medium (9≤18 kg), large (18≤36 kg), and giant (≥36 kg).¹⁴ American Kennel Club (AKC) breed standards were reviewed for the expected body weight, then dogs were classified by size based on breed standard weights. If a dog was not recognized by the AKC, the breed specialty website was used as the source for a breed's expected weight. Dogs classified as AKC Toy breeds were included in the “small” group. Mixed-breed dogs could not be classified by this method and were not included in the expected breed weight analysis. Any patient for which the complete signalment was not available was excluded.

Statistical Analysis

Using logistic regression, odds ratios (OR, odds of disease in one group divided by odds of disease in a referent group) with 95% confidence intervals (CI) were calculated to estimate relative risks. All ORs were adjusted for age and some were adjusted for other variables as well. The criterion for inclusion of a covariate in multivariate models was that the covariate changed the OR for the variable of interest by ≥15%. Age could not be evaluated as a risk factor because of the hospital-based case-control design of the study, but age was adjusted for when other risk factors were examined. In this study, the ratio of cases to controls was lowest at the youngest ages (< 5 yr of age), highest in middle and early old age (5–11 yr of age), and intermediate at the oldest ages (>11 yr of age). When computing adjusted ORs for other variables, age was adjusted for as a quantitative variable using a quadratic term (age + age²). A sensitivity analysis was also performed (not shown), adjusting for age as a three category variable. The adjusted ORs were similar to those when adjusting for age as a quantitative variable.

Other factors needed to be taken into account in specific analyses. After finding an interaction between sex and spay/neuter status (P=0.001), the analysis of the association of spay/neuter status with MCT was stratified by sex. The analysis of the association of sex with MCT by spay/neuter status was also stratified. For analyses concerned with specific breeds, the referent group was mixed-breed dogs, chosen because of their large number (n=358) and because their proportion of MCT cases/patient visit at the hospital was equal to the average proportion for the hospital canine population as a whole (0.004). The ORs relative to mixed-breed were calculated for all breeds represented by at least five dogs in the case-control study. Analyses concerned with phylogenetic clusters and breed standard weight groups included only purebreed dogs because only purebreed dogs could be categorized in both of these ways. The small dog group was used as the referent group for breed standard weight groups. The “others combined” group was used as the referent for the phylogenetic clusters.

It was decided before the analyses that if any of the breeds had no MCTs, it would be possible to assign a probability of disease score of 0.01 (instead of 0) to one individual in that group to be able to include the breed or breed group in the logistic regression (because otherwise, groups with 0 MCTs would be rejected by the statistical procedure). The P values for each factor in multivariate models were computed with likelihood ratio tests. No adjustments were made for multiple comparisons.

Results

Distribution of Cases

A total of 321 dogs were diagnosed (via surgical biopsy) with cutaneous MCT at the hospital between 1997 and 1998 and between 2005 and 2006. Of these, 69 were excluded because of incomplete medical records or because they were diagnosed with grade 1 MCTs. In total, 252 dogs with grade 2 or grade 3 cutaneous MCT, along with 1,608 controls, were included in this case-control study (Table 1).

Table 1 Canine Patient Population at the AMC and Distribution of All Cases and All Controls by Year

*

Includes only the first visit of the year for each patient.

AMC, Animal Medical Center; MCT, mast cell tumors

View Fullscreen

Sex and Spay/Neuter Status

Spay/neuter status was associated with MCT, but the magnitude of the association was different in females and males (P for interaction=0.001). Table 2 showed that, after adjusting for age, spayed females were more than four times more likely to develop MCT than intact females (adjusted OR, 4.11; 95% CI, 2.19–7.69), whereas neutered males were possibly more likely (adjusted OR, 1.37; 95% CI, 0.90–2.09) than intact males to develop MCT. Among intact dogs, females were found to have lower odds of MCT than males (OR, 0.50; 95% CI, 0.25–0.98), whereas among altered dogs, females were more likely to develop MCT than males (adjusted OR, 1.48; 95% CI, 1.07–2.03). Similar patterns were found in each breed standard weight group (results not shown).

Table 2 Age-Adjusted ORs for MCTs by Sex and Spay/Neuter Status for All Dogs

*

Referent category

CI, confidence interval; OR, odds ratio

View Fullscreen

Breed

Of the 114 pure breeds in both the case and control populations, 56 breeds were represented by five or more dogs. Eleven of these breeds had elevated odds of MCT relative to mixed-breeds at an α probability level ≤0.10 in either unadjusted or adjusted logistic regressions. The top part of Table 3 showed these ORs within size classes. The Rhodesian ridgeback had the highest OR (adjusted OR, 34.44; 95% CI, 5.17–229.23), although the number of dogs of this breed was small (n=7) and the confidence interval therefore wide. Among breeds with at least 25 individuals in the study, the boxer (adjusted OR, 6.09; 95% CI, 2.91–12.75), Labrador retriever (adjusted OR, 3.95; 95% CI 2.34–6.66), pug (adjusted OR, 3.17; 95% CI, 1.47–6.82), and golden retriever (adjusted OR, 2.12; 95% CI, 1.13–3.98) had elevated odds for MCT. Among breeds with at least 25 individuals in the study, there were six breeds with notably lower odds of MCT compared with mixed-breeds (bottom part of Table 3). These were: shih tzu (adjusted OR, 0.44; 95% CI, 0.13–1.51), maltese (adjusted OR, 0.42; 95% CI, 0.12–1.44), Yorkshire terrier (adjusted OR, 0.34; 95% CI, 0.10–1.14), chihuahua (adjusted OR, 0.31; 95% CI, 0.07–1.35), dachshund (adjusted OR, 0.18; 95% CI, 0.02–1.38), and miniature poodle (adjusted OR, 0.16; 95% CI, 0.02–1.25).

Table 3 ORs (Relative to Mixed-Breed, n=358) for All Breeds* of Dogs with an Association with MCTs (P<0.10 in Either Adjusted or Unadjusted Analysis)

*

Excluding breeds with n<5.

†

Adjusted for age, sex, spay/neuter status, and sex-spay/neuter interaction.

OR, odds ratio

View Fullscreen

Breed Standard Weight Group and Phylogenetic Cluster Group Among Purebreed Dogs

Because breed standard weight group and breed phylogenetic cluster group were somewhat correlated, it was necessary to examine ORs for each of these variables adjusting for the other (Table 4). Breed standard weight group was strongly associated with odds of MCT. Adjusting for age, sex, spay/neuter status, and phylogenetic cluster, small and medium breeds were at the lowest risk of developing MCT. Large breeds had approximately twice the risk compared with small breeds (adjusted OR, 2.10; 95% CI, 1.31–3.37) and giant breeds had >5× the risk of small breeds (adjusted OR, 5.44; 95% CI, 1.93–15.33), although the number of giant dogs in the sample was small (n=32) and therefore the confidence interval was wide. Among the phylogenetic clusters identified by Parker et al. (2007), the mastiff and terrier group showed elevated odds of MCT (adjusted OR, 3.19; 95% CI, 1.95–5.20 compared with the “other” group) when adjusting for age, sex, spay/neuter status, and breed standard weight group.

Table 4 Adjusted* ORs for MCTs by Breed Standard Weight Group and Phylogenetic Cluster Group for Purebreed Dogs Only (n=1,497)

*

Adjusted for age, sex, spay/neuter status and sex-spay/neuter interaction; in addition, breed standard weight group was adjusted for phylogenetic cluster group and phylogenetic cluster group for breed standard weight group.

†

Referent category

CI, confidence interval; OR, odds ratio

View Fullscreen

Discussion

In this study, a large referral patient population was used to study the association between certain physical attributes and the development of canine grade 2 and 3 MCT. This and other studies serve as initial steps in identifying factors associated with the development of MCT.

In the current study, spayed female dogs were at a substantially increased risk for developing MCT compared with intact females, whereas castrated males had a possibly higher risk than intact males. The reasons for this are unknown. Receptors for estrogen and progesterone were found in canine cutaneous MCT in one study, but another study failed to document estrogen receptors in canine MCT.^15,16 The potential protective role of sex hormones against MCT requires further investigation in a study in which age at the time of altering can be determined.

The role of sex hormones in the development of certain types of neoplasia has been demonstrated in other canine tumors. Although it is known that spaying at an early age greatly decreases the risk for mammary cancer, neutering has been reported to be associated with an increased risk for development of prostate cancer.^17–19 Similar to the results reported here, Ware et al. (1999) showed a risk for cardiac hemangiosarcoma >5× greater in spayed females than intact females and a slightly elevated risk in castrated males compared with intact males.²⁰ Another study showed the risk for bone sarcoma was influenced by age at the time of neutering, with male and female dogs who underwent neutering before 1 yr of age being more likely to develop bone sarcomas than sexually intact dogs.²¹ An alternative interpretation of these findings that the risk for certain cancers is elevated in spayed/neutered dogs is that pet owners who have their dogs spayed or neutered are also those who pursue biopsy of tumors. A third potential explanation for this result is that spaying and neutering may be a surrogate marker for another MCT risk factor, such as obesity. For instance, spaying and neutering have been previously established as risk factors for obesity. In turn, obesity has been linked to an increased risk for certain types of cancer, including MCTs.^22–26 The well-documented health and behavioral benefits of spaying and neutering will need to be carefully weighed against the potential protective role of sex hormones in the development of MCT described here and the increased risk of developing certain other cancers described elsewhere.

Previous studies have reported associations between the development of MCT and specific dog breeds, including the boxer, pug, golden retriever, Boston terrier, bull terrier, Labrador retriever, and shar-pei.^7–12 The current study confirmed the previously reported predisposition of each of these breeds (with the exception of the bull terrier, with only four dogs in this sample) and suggested that several other breeds (e.g., Rhodesian ridgeback, weimaraner, vizsla, boxer, great dane, and Bernese mountain dog) may also be at high-risk (Table 3). Parker et al. (2007) used a clustering algorithm based on allele frequencies to divide 132 dog breeds into 5 primary breed groups in an attempt to identify breeds that likely share traits inherited from the same ancestral source.¹³ The current study found that the mastiff and terrier group of Parker et al. (2007) had increased odds of developing MCT. This breed group contains boxers, Boston terriers, and Labrador retrievers, which were all shown to have an increased risk of MCT development in this breed analysis and in other studies. If these breeds have a common ancestry, as Parker et al. (2007) indicated, they may share a common trait leading to an increased risk of developing MCT. To date, the genetic alterations that predispose a particular animal to developing a MCT are not entirely understood.

With only a few exceptions (e.g., pug), this study found that larger breed dogs were at high-risk and smaller breed dogs at low-risk for MCT. In particular, the dachshund, miniature poodle, shih tzu, maltese, chihuahua, and Yorkshire terrier each had a relatively low-risk for MCT development (adjusted ORs <0.45). The authors are not aware of other studies documenting this phenomenon. Comparison of genetic material from breeds at a low-risk of MCT to those with a high-risk of MCT may allow for the identification of genetic factors that contribute to MCT development. One such factor could be insulin-like growth factor 1 (IGF1). IGF1 is produced in the liver, binds its receptor (the IGF1 receptor), and promotes cell proliferation and survival. The IGF1 receptor is a tyrosine kinase signal transducer. Interaction of IFG1 with its receptor has been shown to play a role the development and proliferation of malignant cells.²⁷ Already, one tyrosine kinase (c-kit) has been shown to be important in the pathogenesis of canine MCT.²⁸ Small breed dogs share a common haplotype of the IGF1 gene, which is a major influence in determination of body size.²⁹ This single IGF1 nucleotide polymorphism haplotype is absent in nearly all giant dog breeds. Lower levels of IGF1 found in small breed dogs could confer a protective effect against the development of MCT. The hypothesis implicating IGF1 in pathophysiological processes has already been shown in other conditions. Several studies have correlated longevity with small size.^30,31 The positive correlation between small size and longevity appears to be controlled by the common haplotype of the IGF1 gene found in small breed dogs.³² This information may be useful in ongoing investigations of genetic alterations in canine MCT.

This study was intended to provide preliminary data for further study of risk factors for MCT. As such, this work has several limitations, predominantly related to the retrospective nature of this study and the databases used for selection of both the cases and controls. Some dogs with MCT may have been diagnosed based on cytology samples and were therefore not included in this study. Further, dogs with MCT may not have been included in the surgical biopsy database if the biopsy was performed at another hospital. Standard procedure at the authors' hospital was to obtain histology samples and internally review all MCT biopsies as part of the pretreatment evaluation, but not all specimens were available for review. Additionally, a single pathologist did not review all biopsies for tumor grade prior to study inclusion. This may have resulted in some misclassification between grade I and grade 2 tumors.

Controls were selected randomly and were not subject to selection based on diagnosis, as the searchable database did not contain this information. It is likely that the control group contained a mixture of critically ill animals, animals with other malignancies, and a small number of healthy animals seen for routine care. Furthermore, because the hospital was a tertiary referral facility, the case and control populations did not represent a random sample of the canine population at large.

Because of the retrospective nature of this study, the signalment of the cases and controls could not be confirmed and could be subject to owner error or data entry error. The authors assumed these errors occurred with similar frequency in both the controls and cases and would have biased the data toward finding smaller associations than actually existed. Further, the authors did not know the actual weights of the dogs. Instead, breed standard weights were used as a surrogate indicator of body size. Future studies should consider the actual weight (and other body characteristics) of dogs within a breed. Finally, as mentioned above, the medical record database did not allow the authors to ascertain the age at time of spaying/neutering and thus, the length of exposure to altered sex hormones. Ultimately, this information will be critical in understanding the impact of current spay/neuter recommendations on MCT development. A prospective study will help overcome many of these limitations. A larger study could also help confirm the elevated risk in certain uncommon breeds, such as the Rhodesian ridgeback.

Conclusion

These results suggest that the etiology of MCT is multifactorial with genetic, hormonal, and other factors all playing a role in tumorigenesis. In this study, breed, sex, spay/neuter status, body size, and phylogenetic cluster each contributed to the risk for MCT. These factors alone did not fully explain risk of MCT development, suggesting that additional factors are probably involved. Other studies, particularly if prospective, will undoubtedly add to this list of likely risk factors as knowledge of the etiology of MCT gradually evolves.