Compared with similar surveys overseas, our cats were older and male cats were over-represented. There was a notable subgroup of young cats with mediastinal involvement. Siamese/Oriental cats were over-represented in this subgroup as well as in the larger population of cats with lymphosarcoma. Compared with overseas surveys, renal involvement, mixed cases and atypical cases (including nasal lymphosarcoma) were more common. A new subcategory of nodal lymphosarcoma, with involvement restricted to node(s) of head and neck, was identified.

Objective To determine and analyse the immunophenotype and histological appearance of naturally occurring cases of lymphosarcoma in Australian cats. Design A prospective multi‐institutional study of naturally occurring feline lymphosarcoma. Methods One hundred and eighteen cats were referred for diagnosis and/or management of suspected lymphosarcoma. Tissue samples for histopathological analysis and immunophenotyping were collected as biopsies or at necropsy from 109 cases. Histological classification of the neoplasms followed the Working Formulation Classification System. Four multi‐species cross‐reactive antibodies were used to classify tumours as having a B or T cell phenotype. Results Seventy‐six (70%) cases were B cell tumours and 28 (26%) were T cell tumours. The remaining 5 (4%) specimens failed to stain with the four antibodies. Histologically, 11 (10%) cases were classified as low‐grade, 72 (66%) were medium‐grade and 26 (24%) were high‐grade tumours. There were no significant associations between age and either histological grade or immunophenotype. Mediastinal and leukaemic cases were significantly more likely to be T cell tumours (P < 0.001 and P < 0.001, respectively). Conclusions In contrast to previously documented studies in the cat, the majority of cases of lymphosarcoma were of B cell phenotype and intermediate histological grade. Based on our data, the histological phenotype of lymphosarcoma is unlikely to predict immunotype, nor are cases of certain immunotypes likely to be of specific histological subtype. Considered in relation to previous reports, the findings suggest that epidemiological factors operating in these cats to produce lymphosarcoma may be different to those operating in North American and UK cat populations.

The prevalence of FeLV antigen or provirus was considerably lower in our cohort of cats compared with studies of lymphosarcoma conducted in the Northern hemisphere. This suggests that factors other than FeLV are important in the development of lymphosarcoma in many Australian cats. No firm conclusions could be drawn concerning whether FeLV provirus contributed to the development of lymphosarcoma in PCR-positive/ELISA-negative cats.

It was possible to cure approximately one quarter of cats with lymphosarcoma using sequential multi-agent chemotherapy and/or surgery. FeLV-negative cats younger than 4 years (typically with mediastinal lymphosarcoma) had a particularly favourable prognosis. The decision to embark on chemotherapy should be based on the results of induction chemotherapy with l-asparaginase, vincristine and prednisolone, as the response to this was a good predictor of long-term survival. Cats surviving the first 16 weeks of chemotherapy generally enjoyed robust remissions (in excess of 1 year) or were cured of their malignancy.

The prevalence of FIV infection was considerably higher in our cohort of cats compared with series of lymphosarcoma cases from the Northern hemisphere. A positive FIV status was strongly associated with lymphosarcoma in Australian cats and it is possible that this infection may predispose to the development of lymphoid neoplasia. The presence of FIV infection would have been underestimated if commercial kits alone had been used for serology.