Prognostic factors in canine mammary tumors: a multivariate study of 202 consecutive cases.

The prognostic variables of 223 consecutively sampled spontaneous mammary tumors from female dogs were studied. These variables included flow cytometric DNA analysis and cell proliferation measured as cells in S-phase rate evaluated from DNA histograms. The dogs were surgically treated, in most cases with unilateral mastectomy (all mammary glands), and 202 of the 223 dogs were studied temporally following surgery. Univariate analysis with correction for age indicated that the variables of lymph node metastasis, elevated S-phase rate, presence of a sarcoma, DNA aneuploidy, and ulceration and infiltrative growth into underlying tissue had a statistically significant negative influence on the survival rates of dogs with a diagnosed malignant tumor. Similar results were obtained from tests on all dogs, but tumor size and its relative hazard increased with increasing size of the tumors, regardless of whether total or disease-specific mortality was considered. Using multivariate-analysis conducted Cox's proportional hazards model, elevated S-phase rate, increased age, and presence of a sarcoma remained statistically significant risk factors. The prognostic value of DNA ploidy and lymph node status varied depending on choice of end point. The study of tumor growth pattern and tumor size provided no prognostic information in the multivariate analysis. Flow cytometric cell analysis, including S-phase rate and DNA ploidy, is of value in predicting the prognosis of canine mammary tumors and can be used as a new prognostic tool to improve the preoperative diagnostics of canine mammary tumors.