A retrospective cohort study was performed with a ten-year follow-up of 70 consecutive patients who were diagnosed as having invasive, operable ductal breast cancer of no special histologic type and without lymph node metastases. All tumors were evaluated by using established histopathologic criteria that reflected either the growth potential of the tumor or the host defense reaction in form of a cellular immunologic response. The evaluation was undertaken to study the interrelationship of the various histopathologic factors and to assess the risk of subsequent breast cancer mortality associated with histopathologic factors and to assess the risk of subsequent breast cancer mortality associated with different constellations of these factors. Multivariate analysis of the data by Cox's proportional hazard regression techniques confirmed the importance of the histologic risk factors reflecting the tumor growth potential and demonstrated that lack of tubule formation was the most important single risk factor. Patients with absent tubule formation were at 6.7 times the risk of cancer mortality than women with tubule formation (P = 0.002). Patients with frequent mitotic figures were at 4.2 times the risk of patients with infrequent mitoses (P = 0.014). Interactions between these two risk factors and nuclear grade were striking. Women who either lacked tubule formations or had tumors with anaplastic nuclei were at 20 times the risk of women with neither of these variables (P = 0.004). The pathologic parameters reflecting the immunologic host defense response were of much less importance. The multivariate classification of all examined histopathologic parameters was helpful in predicting the prognosis of the individual breast cancer patient.