Refining the "double two-thirds" rule: Genotype-based breed grouping and clinical presentation help predict the diagnosis of canine splenic mass lesions in 288 dogs.

Prediction of the likely histopathological diagnosis of canine splenic masses can guide appropriate decision-making. This study explores the predictive effect of breed and clinical presentation on the diagnosis of a canine splenic mass. Records from the Royal Veterinary College, United Kingdom (2007-2017) were reviewed. Dogs with a histopathologic or cytologic diagnosis from a splenic mass, or imaging findings consistent with disseminated metastatic disease, were included. Signalment, physical examination, haematology results, imaging findings and pathology reports were recorded. Breeds were grouped according to several permutations of their phenotype and then by clustering of breeds based on single nucleotide polymorphism analysis. Binary logistic regression was performed to identify predictors of malignancy and haemangiosarcoma. Two hundred and eighty-eight dogs were identified: 27% female and 63% male, 21% entire and 79% neutered; German Shepherd was the most common breed (11%). Median age was 10 years and median bodyweight 25 kg. Thirty-eight percent of dogs presented with haemoabdomen; a splenic mass was found incidentally in 28%. Sixty percent had a malignant tumour of which haemangiosarcoma comprised 66%. On multivariable analysis, genotype-based breed group (P = .004), haemoabdomen (P < .001) and neutrophil count (P = .025) predicted malignancy, and genotype-based breed group (P < .001) and haemoabdomen (P < .001) predicted haemangiosarcoma. Genotype-based breed group and occurrence of haemoabdomen may have predictive value to diagnose malignant splenic masses and more specifically haemangiosarcoma. The effect of genotype-based breed grouping was a superior predictor of the diagnosis of a canine splenic mass lesion compared with all phenotype-based groupings tested.