Canine Cutaneous Neoplasms: Prevalence and Influence of Age, Sex and Site on the Presence and Potential Malignancy of Cutaneous Neoplasms in Dogs from Zimbabwe

a INTRODUCTION The skin is the largest organ system of the body and the organ most commonly affected by neoplastic conditions in dogs 1,5. Dorn et al. 6 showed the frequency of skin neoplasms in the United States to be about 450 new cases for 100 000 dogs yearly and this represents 1 in 3 of all dog tumours. Surveys on the prevalence, predilection sites, and effects of age, sex and breed of dog on the occurrence of these neoplasms have been published in These surveys have shown regional variations in the occurrence and type of tumours that appear in dogs. Part of this variation may be due to environmental carcinogenic exposure of different dog popula


INTRODUCTION
The skin is the largest organ system of the body and the organ most commonly affected by neoplastic conditions in dogs 1,5 .Dorn et al. 6 showed the frequency of skin neoplasms in the United States to be about 450 new cases for 100 000 dogs yearly and this represents 1 in 3 of all dog tumours.Surveys on the prevalence, predilection sites, and effects of age, sex and breed of dog on the occurrence of these neoplasms have been published in Australia 7,8,14,17 , the United Kingdom 2,3 , the USA 4 , Greece 12 and South Africa 1 .These surveys have shown regional variations in the occurrence and type of tumours that appear in dogs.Part of this variation may be due to environmental carcinogenic exposure of different dog popula-tions and differences in breed popularity reflecting differences in genetic susceptibility.There is no information on the prevalence of various histopathological types of neoplasms and the influence of sex, age or breed on the prevalence of these neoplasms or on predilection sites in dogs found in Zimbabwe.The objectives of the present study were 1) to assess the prevalence, sex, age, breed and site distribution of various histopathological types of canine skin neoplasms in Zimbabwe, 2) to investigate the effect of age, sex, site distribution on the likelihood of developing cutaneous neoplasm, and 3) to investigate the effect of sex, age, type of tumour, and site distribution on potential malignancy

Case materials
Cutaneous biopsies from 900 dogs submitted for histopathology at the Section of Diagnostic Pathology, Department of Paraclinical Veterinary Studies, University of Zimbabwe, from 1996 to 2000 were studied.Cases came from the University of Zimbabwe Teaching Hospital and from various private veterinary practitioners in Zimbabwe.All cases with a histopathological diagnosis of skin neoplasia were reviewed, excluding neoplasms of the mammary or ceruminous glands.All tissues had been fixed in 10 % neutral buffered formalin, embedded in paraffin wax, and sectioned and stained with haematoxylin and eosin.Special stains, such as toluidine blue, periodic acid-Schiff and alkaline Congo red, were used at the discretion of the case pathologist.These neoplasms were classified according to published criteria 9,16 .

Statistical analysis
Initially, the histological type, prevalence, sex, age and site distribution of the neoplasms were described.The neoplasms were grouped into 4 broad classes, namely epithelial, mesenchymal, melanocytic and lymphohistiocytic.The location of the neoplasm on the body was grouped into 4 groups, namely trunk, limbs, multiple, and head and neck.The 10 most common types of neoplasms were determined using the PROC FREQ procedure of the Statistical Analysis System (SAS) 18 .The odds of developing a cutaneous neoplasm in a particular sex and breed were estimated using the PROC LOGISTIC procedure of SAS 18 .Logistic regression was performed in SAS 18 to evaluate the usefulness of age and sex of dogs in predicting the likelihood of any one of the 4 types of tumours encountered in this study, given that a dog had already been diagnosed with a tumour and the location of that tumour had been established.The association between tumour site and the tumour 's potential for malignancy was determined using the chi-square test for association.

RESULTS
A total of 900 canine cutaneous biopsies were submitted during the study period.Of these cutaneous biopsies 540 (60 %) were diagnosed as neoplasms and 460 (40 %) as cases of non-neoplastic inflammatory or degenerative diseases.The prevalence, sex and age distribution of the various histopathological types of cutaneous neoplasms are presented in Table 1.

Histogenetical types and prevalence
Thirty different cutaneous neoplasms were recognised.Of these, 213 (39.4 %) were of epithelial origin (13 types), 240 (44.4 %) of mesenchymal origin (13 types), 40 (7.4 %) of lymphohistiocytic origin (2 types) and 47 (8.7 %) of melanocytic origin (2 types) (Table 2).Thirty-seven per cent of the tumours were malignant.Only the lymphohistiocytic tumours had a higher number of malignant tumours than benign tumours (Table 2).The 10 most common tumours, comprising 73.7 % of all cutaneous neoplasms diagnosed during the study, are shown in Table 3 and are compared with results from studies carried out in Greece, Australia, the UK and the USA.Mast cell tumours and squamous cell carcinoma were the 2 most common tumours, comprising about a third of all cutaneous neoplasms diagnosed during this period.
The mean age of dogs with neoplasia, according to the histopathological type, ranged from 0.5 to 17 years with an overall mean of 12.2 years (Table 1).Tumours of epithelial and melanocytic origin had a higher mean age of occurrence (8.8 and 8.6 years, respectively) than those of mesenchymal (6.1 years) and lymphohistiocytic (4.1 years) origin.The breed distribution of the most common breeds with cutaneous tumours is shown in Table 4.There was wide variation in the number of epithelial, mesenchymal, lymphohistiocytic and melanocytic tumours between breeds.For example, the terriers contributed 16.9 % of the total number of epithelial tumours and 12.5 % of the mesenchymal tumours, while Rottweilers contributed 0.5 % of the epithelial tumours and 5.0 % of the mesenchymal tumours.
Twenty-two per cent (120/540) of the neoplasms were located on the head and neck, 46.1 % (249/540) on the trunk and 24.3 % (131/540) on the limbs.Only 7.4 % (40/540) of the neoplasms were found at multiple sites (Table 5).Most of the lymphohistiocytic tumours (40 %) were found at multiple sites whereas none of the melanomas were found at multiple sites.

Effect of histogenetical origin of tumour and sex of dog on the tumour's potential for malignancy, and effect of sex of dog and location of neoplasm on the likelihood of developing epithelial tumours or non-epithelial tumours
Mesenchymal tumours were significantly (P < 0.05) less likely to be malignant compared with melanocytic tumours.The odds of malignancy did not significantly differ between epithelial or lymphohistiocytic tumours and melanocytic tumours (Table 6).The sex of dog had no effect on the likelihood of occurrence of both epithelial and nonepithelial tumours and neither did it have a significant effect on the likelihood of any tumour type becoming malignant.Development of both epithelial and non-epithelial tumours did not show site specificity.

Association between tumour site and the tumour's potential for malignancy
Tumours on the trunk and limbs were, respectively, 3 and 4 times more likely to be malignant compared with those occurring at multiple sites (odds ratios, 3.19 and 4.68, respectively) (Table 7).Tumours on the head and neck were 2.93 times more likely to be malignant compared with those on multiple locations.This effect was, however, not statistically significant.

DISCUSSION
Epidemiological studies of canine cutaneous neoplasms have been per- formed in various geographical regions including the USA 4 , Australia 7,8,14,17 , the UK 2,3 , Greece 12 and South Africa 1 .The prevalence of various histological types of tumours varies among the reports published in these regions.In the present study, two-thirds of cutaneous biopsies submitted during the study period were neoplasms and of these 37.4 % were malignant.Similar results were obtained in the UK where 37.5 % of the tumours were malignant 3 .Therefore, of all the skin tumours, less than 40 % require chemotherapy, radiotherapy or immunotherapy after initial surgical excision.Forty-four per cent of the canine skin tumours were of mesenchymal origin, 39.4 % of epithelial origin, 8.7 % of melanocytic origin, and 7.4 % of lymphohistiocytic origin.Prevalence of skin epithelial versus non-epithelial tumours in the dog varied in different studies 12,17 .In 1 such study, epithelial tumours constituted 48 % of all skin tumours and mesenchymal tumours constituted 40 % 12 .The observed differences may be a reflection of canine breed population and environmental influences.
Thirty histological types of tumour were diagnosed, but 10 types accounted for 69.6 % of all canine skin tumours.These 10 tumours, in order of prevalence in our study, were mast cell tumour, squamous cell carcinoma, perianal gland adenoma, lymphoma, benign melanoma, haemangiosarcoma, sebaceous gland adenoma, fibrosarcoma and lipoma.The prevalences of these tumours are compared in Table 2