This study assesses the ability of an experienced surgical pathologist to predict from microscopic examination of primary mammary carcinomas whether axillary metastases existed in the radical mastectomy specimens. Primary mammary carcinomas treated by radical mastectomy from 1956 to 1960 were reexamined and classified by type of cellular growth pattern and by configuration of the neoplastic borders (pushing or infiltrating). The only information available to the pathologist was the size of the primary tumor. The predictions were compared with the microscopic findings in the axillary nodes and with the survival of the patients. The accuracy of predicting axillary nodal metastases did not exceed 70% except when Type I neoplasms existed. The incidences of axillary nodal metastases differed significantly between the less‐well‐differentiated mammary carcinomas with pushing borders and those with infiltrating borders. However, the incidence of nodal metastases in the former group was high enough to prevent accurate assessment of the status of the axillary nodes from study of the borders of the primary tumors. Survival rates after radical mastectomy correlated with the cellular growth patterns of the primary tumors but not with the character of the tumor borders. The survival rates were adversely affected by the presence of axillary metastases only among those patients treated for the less‐well‐differentiated Type III and Type IV tumors. Radical mastectomy with axillary dissection but without sacrifice of the pectoral muscles would have removed the tumor‐bearing lymph nodes in at least 85 of the 89 Type I and Type II tumors. These tumors constituted over 20% of the mammary caracinomas treated by radical mastectomy in a five‐year period. On the basis of this study, a more conservative approach to the Type I and Type II tumors appears to be justified.
