In order to analyze which features determine a poorer prognosis we undertook a study of 91 consecutive patients with pathological Stage I breast cancer operated on at Beth Israel Medical Center (1968-71). Tumor tissue slides were reviewed and features such as: tumor size, histologic type, nuclear grade, lymphocytic and perivenular lymphocytic infiltration, as well as sinus histiocytosis in the lymph nodes removed. Records were reviewed and classified according to age and ethnic background. Survival and recurrence data were recorded up to 14 years post-mastectomy. Also determined was the presence of an antigen, previously detected in certain human breast cancer tumor tissues, which has been found to cross-react immunologically with the 52,000 dalton major envelope protein of the mouse mammary tumor virus (MMTV--gp52). Ten-year cumulative disease-free survival was 0.65. Univariate analysis of survival within various factors revealed that the only statistically significant influencing factor was the presence or absence of the antigen. Factors such as perivenous lymphocytic infiltration, diffuse lymphoid infiltration and sinus histiocytosis in regional lymph nodes also showed trends in favor of improved survival but the sample may be too small for statistical significance. The presence or absence of the antigen was independent of the host immunological reaction. There was no relationship between localization of tumor, whether medial or lateral and survival, nor with presence or absence of antigen.