Abstract
BACKGROUND: Evidence suggests etiologic heterogeneity among breast cancer subtypes. Previous studies with six-marker IHC classification of intrinsic subtypes included small numbers of black women. METHODS: Using centralized laboratory results for estrogen receptor (ER), progesterone receptor, HER2, proliferation marker, Ki-67, EGFR, and cytokeratin (CK)5/6, we estimated case-only and case-control ORs for established breast cancer risk factors among cases (n = 2,354) and controls (n = 2,932) in the African American Breast Cancer Epidemiology and Risk (AMBER) consortium. ORs were estimated by ER status and intrinsic subtype using adjusted logistic regression. RESULTS: Case-only analyses by ER status showed etiologic heterogeneity by age at menarche, parity (vs. nulliparity), and age at first birth. In case-control analyses for intrinsic subtype, increased body mass index and waist-to-hip ratio (WHR) were associated with increased risk of luminal A subtype, whereas older age at menarche and parity, regardless of breastfeeding, were associated with reduced risk. For basal-like cancers, parity without breastfeeding and increasing WHR were associated with increased risk, whereas breastfeeding and age ≥25 years at first birth were associated with reduced risk among parous women. Basal-like and ER(-)/HER2(+) subtypes had earlier age-at-incidence distribution relative to luminal subtypes. CONCLUSIONS: Breast cancer subtypes showed distinct etiologic profiles in the AMBER consortium, a study of more than 5,000 black women with centrally assessed tumor biospecimens. IMPACT: Among black women, high WHR and parity without breastfeeding are emerging as important intervention points to reduce the incidence of basal-like breast cancer.