Competitive prognostic value of clinicopathologic and bioimmunologic factors in primary breast cancer.

Fourteen clinical, pathologic, and pretreatment bioimmunologic variables were evaluated for their significance in predicting the survival or the length of disease-free interval of 55 patients with primary breast cancer. The variables studied were: patient age; clinical stage of disease according to the International Union Against Cancer TNM classification; number of involved nodes; sedimentation rate; peripheral lymphocyte, leucocyte, and monocyte counts; serum levels of immunoglobulins IgG, IgA, and IgM; percentages of E-, "active" E-, and EAC-rosettes; and finally, the lymphoblastic transformation test value (PHA-LTT). A multivariate analysis using the Cox proportional hazards regression model was carried out, in a stepwise manner, to identify those variables most highly related to survival or to the length of disease-free interval. The Cox analysis showed that clinical stage, number of involved nodes, percentage of EAC-rosettes, sedimentation rate, and T-lymphocyte reactivity, (i.e., the T-lymphocyte sensitivity to PHA, expressed as the ratio between the PHA-LTT in counts per minute and the percentage of E-rosettes) were the significant prognostic factors for survival, whereas the number of involved nodes and the sedimentation rate were independent of importance in predicting the length of disease-free interval. The results obtained from this analysis proved the importance of some immunologic parameters in the estimation of prognosis. In addition, a prognostic score for summarizing multiple factors with potential use in stratification was derived from the multivariate analysis.