The relations of Wolfe mammographic patterns, quantitative mammographic densities, and mammographically estimated breast size to breast cancer risk were investigated prospectively in a case-control study nested in the New York University Women's Health Study, a cohort of 14,291 women in New York City, NY (United States). The archived mammograms of 197 breast cancer cases who were identified during the first 5.5 years of the study and of 521 individually matched controls from the same cohort were retrieved and classified according to Wolfe parenchymal patterns and mammographic densities by two expert radiologists. Breast size and volume were estimated on the mammogram's cranio-caudal projection. In both premenopausal and postmenopausal subjects, the risk of breast cancer increased progressively with increasing density and percent density area. A significantly increased risk was found also for Wolfe pattern DY in premenopausal women and P2 pattern in postmenopausal subjects. In premenopausal women, mammographically determined breast volume and breast height also were associated positively with breast cancer risk. Although the results of the present study confirmed that mammographic parenchymal patterns and densities were important predictors of breast cancer risk, their practical use in screening seems limited due to the high prevalence of high risk patterns.