Preliminary studies suggest that the presence of atrophy in benign breast biopsies is an important risk factor for subsequent breast cancer. The breast cancer risk in women with atrophic lesions may be as low as one-third that of the general population, or as high as five times this general risk, depending on other prognostic factor such as age at biopsy or a family history of breast cancer. This wide variation in cancer risk may result from different types of atrophy that are similar in morphologic appearance, but have different etiologies and are associated with different cancer risks. In this case, the wide variation in cancer risk associated with atrophy in the presence of other risk factors may result from a positive correlation between the type of atrophy associated with high cancer risk and these other factors. Papillary apocrine change is a distinctive lesion seen in 27% of benign breast biopsies that has never been rigorously evaluated as a breast cancer risk factor. We will conduct a retrospective cohort study of approximately 8100 consecutive women who underwent benign breast biopsy at one of three Nashville hospitals between 1952 and 1968 (Vanderbilt or Baptist hospitals) or 1950 and 1968 (St. Thomas Hospital). All biopsies will be reviewed and reclassified for the presence and degree of atrophic or papillary apocrine changes. The degree of atrophy will be assessed through objective criteria applied by our pathologists. Biopsies that have not been analyzed in our previous studies will receive complete histologic diagnosis. Information on subsequent breast cancer outcome and epidemiologic risk factors will be obtained from interviews with patients or their next of kin. This information has already been obtained on 60% of our study subjects who participated in our prior studies. We expect to obtain follow-up on 90% of our study subjects. Variation in cancer risk among women who present for benign biopsy will be assessed by means of intra-study control groups of biopsied patients. In addition, we will use external control groups from the Third National Cancer Survey in Atlanta and the Cancer in Connecticut data base. After follow-up is completed, our data base will be analyzed using hazard regression methods and other statistical techniques that can elucidate the cancer risks associated with various constellations of risk factors from a longitudinal data base. This study will substantially increase the precision with which we can predict breast cancer risk on the basis of histologic and non-anatomic findings.