Bimodal breast cancer incidence patterns provide support for a dualistic model of mammary carcinogenesis
595 Background: Although breast cancers are extremely heterogeneous with respect to clinical presentation, histopathological appearance, and molecular alterations; breast cancer is often viewed as a single biologic entity with common etiology and unified pathogenesis. Accumulating data has challenged this view, demonstrating that the stratification of tumors by gene expression profiles, immunohistochemical (IHC) staining patterns, and other techniques may permit the categorization of breast cancers into two main pathways of mammary carcinogenesis. Based upon cellular origin and/or estrogen receptor (ER) expression there are 1) tumors from stem cells committed to luminal differentiation and ER expression and 2) neoplasms from stem cells programmed to display basal differentiation and lacking ER expression. If confirmed, these observations may form the basis for revised conceptual frameworks. Methods:: We applied age-at-diagnosis density plots and a two component statistical mixture model to breast cancer cases overall (n=270,124) in the SEER program. These age distributions were reevaluated after stratification by histopathologic type, race, and ER. Results: A bimodal distribution provided a better fit for the age distribution patterns than a single distribution for breast cancers overall and for all histologic types, except medullary carcinoma. However, the proportion of early-onset and late-onset tumors varied: ductal and tubular carcinomas demonstrated relatively equal proportions; lobular, papillary and mucinous tumors were associated with late-onset populations; and inflammatory and medullary tumors were more related to early-onset disease. Medullary carcinomas were exceptional in showing a single distribution irrespective of race and/or ER status. Of note, medullary breast carcinomas are rare tumors that are associated with germline mutations in BRCA1. Conclusion: Most histopathologic types of breast carcinomas demonstrated a bimodal mixture of early- and late-onset cancer populations, possibly representing two major age-related causal pathways. These population-based results seem consistent with emerging molecular data, showing two main classes of breast cancer based upon cellular origin and ER expression. No significant financial relationships to disclose.