Materials and Methods

The computerized records of the diagnostic laboratory at the CSU Veterinary Teaching Hospital and of PDS, University of Saskatchewan, were queried for all reports of equine cutaneous biopsy specimens submitted between July 1, 2000, and July 1, 2010. To ensure that the search was complete, additional terms and wild cards that were queried included the following: basal, carcinoma, fibro-, giant, hemangio-, leio-, lipo-, lymphoma, lymphosarcoma, mast, melano-, myxo-, papilloma, peripheral, sarcoid, skin, and tricho-. A database was created with the submission records and diagnostic reports from these biopsy specimens; the age, sex, and breed of each patient, if known, were entered into the database. Data regarding the date of histologic diagnosis and geographic location (town, state or province, and zip or postal code) were also collected. Because the address of the patient's location was not always provided, the geographic location of the submitting veterinary clinic was used as a surrogate. The submission record was used to determine the location of the lesion or lesions on the patient (head, neck, trunk, thoracic limb, pelvic limb, limb unspecified, ventrum, penis, perianal region, tail, or general area). Ocular neoplasias on the third eyelid, conjunctiva, or sclera of the eye were excluded. Lesions were classified as tack associated if the clinical history made specific mention of tack (eg, alopecia in the girth area or nodules under the saddle). Seasons were broadly classified as spring (March to May), summer (June to August), fall (September to November), and winter (December to February). The presence of erosion or ulceration was also noted in the database for each biopsy specimen. Specimens that did not include neoplastic disease were excluded from detailed analysis.

Descriptive and comparative statistical analyses were performed with statistical software.^a Frequency of lesions by categorical variables, including diagnostic laboratory, year and season of submission, sex, breed, and location on the body, was compared via χ² tests and ORs with 95% CIs as applicable. Ages were compared between groups via a t test for 2 groups and ANOVA for 3 or more groups after evaluating for equality of variance (Levine test). Stratified analysis was performed to compare trends between CSU and PDS. Where trends were similar, only the summary statistic was reported.

After preliminary exploration of the data, the following corporal locations were merged for further statistical analysis: pelvic limb, thoracic limb, and limb unknown were merged into limb; neck, body, and ventrum were merged into body; and tail and perianal region were grouped into perianal. Similarly, biopsy specimens with diagnoses of sarcoma, schwannoma, neurofibrosarcoma, peripheral nerve sheath tumor, and myxosarcoma were grouped into the category soft tissue sarcoma. The benign mesenchymal lesions leiomyoma, fibroma, lipoma, myxoma, and hemangioma were grouped into the category benign soft tissue tumors; apocrine tumors and carcinomas were grouped into an epithelial tumor category; and SCC, SCC in situ, hemangiosarcoma, and hemangioma were initially evaluated individually and then subsequently grouped together as UV-associated neoplasms on the basis of their association with UV radiation in horses or in other species.^1,6–14

Results

A total of 5,141 cutaneous biopsy specimens were submitted to CSU and PDS between July 1, 2000, and July 1, 2010. There were 3,351 specimens with histologic evidence of neoplastic disease identified (65.2% of total submissions); 2,869 of these were submitted to CSU and 482 to PDS. The mean age of horses with neoplastic lesions was significantly (P < 0.001) higher (10.9 years; 95% CI, 10.67 to 11.18 years) than the mean age of horses with nonneoplastic lesions (9.6 years; 95% CI, 9.29 to 9.95 years). Males were 1.65 (95% CI, 1.42 to 1.91) times as likely to have neoplastic versus nonneoplastic lesions. The most common breeds were similar in the neoplastic and nonneoplastic populations.

There was an annual average of 338 cases submitted/y (range, 269 to 382 submissions/y). The frequency of biopsy submissions varied significantly (P < 0.001) by season, with the most (33%) cases submitted in the summer, followed by the spring (27%), fall (23%), and winter (17%). Biopsy specimens submitted for histologic examination from Colorado, Alberta, Saskatchewan, Wyoming, Montana, South Dakota, New Mexico, Texas, Manitoba, Nebraska, Minnesota, and other midwestern states (including Kansas, Oklahoma, North Dakota, Iowa, Missouri, and Wisconsin) comprised over 80% of total submissions during the time period evaluated. Geographic location was not available for 60 (< 2%) biopsy specimens.

In total, samples were submitted from 1,903 geldings, 114 stallions, 1,028 mares, and 227 horses for which the sex was unspecified. The top 10 most common breed categories in which tumors were identified were Quarter Horses (39.4%); American Paint Horses (12.9%); Arabians (10.4%); warmbloods (8.9%); Appaloosas (7.5%); Thoroughbreds (6.7%); burros, donkeys, or mules (3.5%); ponies (3.1%); draft breeds (2.7%); and Morgans (1.3%). Other breeds comprised the remaining 3.5% of submissions. There was significant (P < 0.001) variation in the mean age of horses with different tumor types (Table 1).

Sarcoid, SCC, and melanoma were the 3 most commonly diagnosed tumors, representing 46%, 18.9%, and 10.3% of the total neoplastic diagnoses, respectively, during this 10-year period. The frequency of tumors diagnosed in specimens received by CSU and PDS was statistically different (P < 0.001). Tumors associated with UV radiation in horses and in other species (SCC, SCC in situ, hemangiosarcoma, and hemangioma) were 2.3 (95% CI, 1.78 to 3.0) times as common in specimens received by CSU versus PDS. Appaloosa horses were 7.2 (95% CI, 5.6 to 9.2) times as likely to have tumors associated with UV radiation, compared with other breeds. American Paint Horses were 4.4 (95% CI, 3.6 to 5.4) times as likely to have tumors associated with UV radiation as were all other breeds. The mean age of horses with tumors associated with UV radiation was 14.3 years (95% CI, 13.8 to 14.8 years), which was significantly (P < 0.001) different than the mean age of 9.3 years (95% CI, 9.0 to 9.6 years) of horses with all other neoplasms.

Sarcoids comprised 53% of all tumors located on the head and body, and tumors on the head were 2.3 (95% CI, 2.0 to 2.7) times as likely to be sarcoids, compared with any other tumor type. Sarcoids were rarely multicentric (< 1% of cases). Horses with sarcoids had a mean age of 7 years (95% CI, 7.9 to 8.5 years), and there was no appreciable breed or sex predisposition.

Squamous cell carcinomas were the second most common tumor in the data set. Appaloosas were 8.0 (95% CI, 6.3 to 10.2) and American Paint Horses were 4.6 (95% CI, 3.7 to 5.7) times as likely as all other horses to have SCC. Males were 1.7 (95% CI, 1.4 to 2.1) times as likely as females to have SCC. Predisposed sites for SCC included the penis and perianal region. Squamous cell carcinomas represented 57.8% of tumors located on the penis and 44.2% of perianal tumors. Tumors located on the penis were 12.1 (95% CI, 9.8 to 15.0) times as likely to be SCC, compared with any other tumor (excluding papillomas), and tumors in the perianal region were 9.0 (95% CI, 6.5 to 12.4) times as likely to be SCC, compared with any other tumor type (excluding melanoma). The mean age of horses with SCC was 16.4 years (95% CI, 15.8 to 16.9 years).

Melanocytic tumors comprised 36% of all perianal tumors. Tumors in the perianal region were 17.4 (95% CI, 12.7 to 23.8) times as likely to be melanocytic tumors, compared with other tumor types (excluding SCC). The mean age of horses with melanocytic tumors was 10.0 years (95% CI, 9.7 to 11.2 years). Melanocytic tumors were 3.2 (95% CI, 2.4 to 4.2) times as likely to be diagnosed in Arabians, compared with any other breed. No sex predilection was identified.

Soft tissue sarcomas comprised 7.6% of all neoplastic biopsy specimens submitted during the time frame examined. Soft tissue sarcomas were distributed most commonly over the body (10.3%), limb (9.6%), and head (7.3%). Multicentric distribution was also fairly common (9.1% of soft tissue sarcomas). Horses with soft tissue sarcoma had a mean age of 11.7 years (95% CI, 10.7 to 12.6 years). Burros, donkeys, and mules were 2.5 (95% CI, 1.3 to 3.9) times as likely as other breeds to have soft tissue sarcomas. No significant sex predisposition was detected.

Papillomas were most commonly diagnosed on the head and penis, and tumors located on the head or penis were 52.2 (95% CI, 39.5 to 68.9) times as likely to be papillomas, compared with any other tumor (excluding SCC). Papillomas were occasionally multicentric (2%). The mean age of horses with papillomas was 10.5 years (95% CI, 9.3 to 11.8 years). Of papillomas, 38% occurred in horses < 5 years of age. There was no apparent breed or sex predisposition.

Benign soft tissue tumors were rare. Benign soft tissue tumors comprised 7% of all tumors located on the head and ≤ 3% of tumors in other locations. The mean age of horses with benign soft tissue tumors was 9.9 years (95% CI, 8.6 to 11.2 years). Benign soft tissue tumors were 3.1 (95% CI, 1.9 to 5.1) times as likely to be diagnosed in draft and pony breeds, compared with all other breeds. A sex predisposition was not detected.

Mast cell tumors were rare, comprising approximately 2% of all tumors in the database. There was no apparent site of predilection for mast cell tumors. Horses with mast cell tumors had a mean age of 12.7 years (95% CI, 11.6 to 13.9 years). No sex or breed predisposition was apparent.

Cases of cutaneous lymphoma comprised 1.7% of the biopsy specimens. Cutaneous lymphoma was 3.5 (95% CI, 2.0 to 6.2) times as likely to be located on the body, compared with any other location. Lymphoma was occasionally reported to be multicentric (7% of lymphomas). Horses with lymphoma had a mean age of 15.4 years (95% CI, 12.5 to 16.3 years). Thoroughbreds were 2.5 (95% CI, 1.2 to 5.3) times as likely to have lymphoma, compared with all other breeds combined. Males and females were equally affected.

Giant cell tumors of soft parts (also referred to as malignant fibrous histiocytoma) were rare tumors in the database (1% of all tumors in the database). Giant cell tumor of soft parts was diagnosed 10 times as frequently in specimens received by CSU as by PDS. Tumors on the limb were 6.0 (95% CI, 3.2 to 11.2) times as likely to be GCTSP as other tumors. No breed predisposition was evident. Horses with GCTSP had a mean age of 12.9 years (95% CI, 11.1 to 14.8 years). No sex predisposition was detected.

Basal cell tumors were also rare (1.2% of diagnoses). Horses with basal cell tumors had a mean age of 14.7 years (95% CI, 12.7 to 16.7 years). Morgans and ponies were 4.6 (95% CI, 2.0 to 10.4) times as likely as other breeds to have basal cell tumors. No sex predisposition was detected. Basal cell tumors were evenly distributed over the body, including the head (18%), limb (15%), trunk (13%), ventrum (10%), tail (7.7%), and perianal area (7.6%).

Hemangiosarcomas were rare, comprising 0.6% of all tumors. Hemangiosarcomas were 5.6 (95% CI, 2.2 to 14.3) times as likely to be located on the body as opposed to other locations. No significant breed or sex predisposition was identified. Horses with cutaneous hemangiosarcomas were a mean age of 12.8 years (95% CI, 9.4 to 16.3 years).

Ulceration was most commonly noted in sarcoids and SCCs (69.5% and 21.2% of ulcerated tumors, respectively). No other tumor type comprised > 5% of tumors in which ulceration was noted. Only 31 biopsy specimens (< 0.6%) were described as tack associated. No significant associations were detected between tack association and tumor type.

Discussion

In the present study, evidence of cutaneous neoplasia was present in 65% of cutaneous biopsy specimens from equids submitted to diagnostic laboratories in Colorado and Saskatoon over the 10-year study period. Whereas general trends may be similar to those reported for horses in other regions,^1,3,15–17 there was variation in the frequency of lesion by breed, sex, anatomic location, and location of residence. These results may benefit equine practitioners as they evaluate patients with dermatologic conditions, formulate differential diagnoses, and initiate treatment.

Biopsy submissions were more common in summer than in other seasons in the present study. Rather than suggesting a relation of season with the incidence of neoplasia, this finding may be explained by increased use of horses during the summer months, allowing for closer and more consistent observation of the horse and potential skin lesions. A horse's winter coat and the propensity for blanketing during the cooler seasons may hinder the detection of cutaneous lesions in the fall and winter.

Similar to previous reports,^1,15,17 sarcoid was the most common neoplastic entity of equine skin, accounting for 46% of samples in this study. Squamous cell carcinoma and melanoma were the next most common cutaneous equine neoplasias, representing 18.9%, and 10.3% of the total neoplastic diagnoses, respectively, during this 10-year period. Interestingly, sarcoids appear more common in the central than in the northeastern United States. The basis for this difference is unknown but may involve differential exposure to bovine papillomavirus (ie, through proximity to cattle), virus serotypes, or viral vectors in the area. Consistent with prior data, sarcoids were more common on the head and body than other locations.¹⁷ There were no clear breed or sex predisposition for sarcoids, although the median age and broad range suggest this lesion is common in younger horses.

Squamous cell carcinoma, including SCC in situ, was the second most common cutaneous neoplasm, accounting for almost 20% of cutaneous neoplasms in the study population. Appaloosas and American Paint Horses were predisposed to the diagnosis of these tumors, presumably because of the frequency of white hair and pink skin in these breeds. Predisposed sites for SCC were the penis and perianal area. Males were predisposed to the development of SCC, possibly because of the predilection for this tumor to be on the penis. Overall, SCCs appeared in older horses.

Squamous cell carcinomas and sarcoids were more likely to be ulcerated than other tumors; thus, these tumor types should be considered as important differential diagnoses when evaluating an ulcerated mass lesion. The median age for horses with SCC was almost twice that of horses with sarcoids, suggesting host factors may aid in clinical differentiation of the 2 lesions. In conjunction with anatomic site, breed, and age of the horse, the presence of ulceration may help narrow the differential diagnosis list prior to surgical biopsy.

Melanoma was the third most common cutaneous neoplasm, consistent with findings from a study¹⁵ of horses of the Pacific northwest. Similar to the Pacific northwest, Arabians were predisposed to melanocytic tumors, possibly because of genetic predisposition or the common occurrence of gray coloration in this breed, and these tumors were common in the perianal area.¹⁵ Various histologic and clinical subtypes of melanocytic tumors have been identified in gray horses and specific breeds.^18–20 As the clinical history did not usually include the color of the horse and the diagnostic reports did not always subclassify tumors according to standardized nomenclature, the melanocytic tumors in this study were not subcategorized.

Interestingly, tumors associated with UV radiation were twice as likely for CSU patients, compared with those from PDS. Factors such as living at increased altitude or in locations with decreased ozone permit increased levels of UV radiation to reach the earth's surface and result in increased incidence of skin cancers in human beings.¹³ Colorado has a mean elevation of 2,073 m (range, 1,011 to 4,401 m), whereas the mean elevation of Saskatchewan is 457 m (range, 213 m to 1,468 m).^21,22 Thus, animals in the Colorado area may experience increased UV exposure because they live at higher elevations with decreased ozone protection. Interestingly, the CDC lists many central states as having the highest melanoma-associated human death rates (Colorado, Wyoming, Idaho, Kansas, Nevada, Oklahoma, and South Dakota) and the second highest melanoma incidence interval (Colorado, Montana, North Dakota, and Kansas) in the United States.²³ In humans, melanoma is strongly associated with UV exposure, and increased prevalence of human melanoma has been noted with increasing altitude in Spain.^13,24,25 A certain subset of melanomas in horses is thought to represent a genetic predisposition (such as gray horse melanoma), and so these tumors were not included in the analysis of UV-associated neoplasia. However, risk factors for many equine cutaneous neoplasms have not been extensively evaluated and may be underestimated. For example, although the etiologies of basal cell tumors in horses are unknown, these tumors are strongly correlated with UV exposure in humans and are commonly associated with concurrent actinic lesions or SCC in cats and dogs.^1,26 Given the importance of UV radiation in the development of neoplasia in other species, the role of UV radiation in equine cutaneous neoplasia should be further investigated.

The present study found that Appaloosas and American Paint Horses were predisposed to UV-associated neoplasms. Appaloosas and American Paint Horses are more likely than other breeds to have extensive patches of white hair, which allows for increased cutaneous UV exposure. Similar trends have been identified in white-faced Hereford cattle with facial and corneal SCCs, lightly pigmented dogs that develop dermal hemangiomas and hemangiosarcomas, and human beings with malignant melanoma.^9,12,13

Breed predispositions for cutaneous lymphoma, basal cell tumors, and benign soft tissue tumors have not been reported previously. The biologic reasons for the increased prevalences of lymphoma in Thoroughbreds, basal cell tumors in Morgans and ponies, benign soft tissue tumors in draft and pony breeds, and soft tissue sarcomas in burros, donkeys, and mules are unknown. It is important to note that many soft tissue sarcomas may have actually been sarcoids; the presence of acanthotic cells overlying the epithelium is helpful in differentiating sarcoids from soft tissue sarcomas, and specimens that did not include epithelium may have been misdiagnosed in the broader category of soft tissue sarcoma. Although cutaneous lymphoma has been reported to be more common in males,²⁷ the diagnosis in the present study was evenly divided between males and females.

Location on the body was important for several types of tumors. Papillomas were more common on the head and penis, and similar to previous case series,^28,29 GCTSP tended to be distributed on the limbs and the body. Although basal cell tumors have been recently reported as having a predisposition for the extremities of horses,³⁰ this finding had not been documented in a previous case series.³¹ The present study had considerably more cases than either of these case series and demonstrates that these tumors were fairly evenly distributed over the body, with no significant difference between the limbs and other locations, suggesting that a true anatomic predilection may not exist.

The results of the present study not only support numerous epidemiological trends from previous studies but also identify new risk factors for specific tumor types in equids. Further research is necessary to identify the biological basis for many trends identified.