Li K, et al. - Using data from two large cohorts, researchers examined the influence of modeling the heterogeneity of predictor variables between estrogen-receptor positive (ER+) and negative (ER-) on prediction made using breast cancer (BC) risk prediction models. For ER+ (ModelER+) and ER- tumors (ModelER-), two models, respectively, were built in 281,330 women (51% postmenopausal at recruitment) from the European Prospective Investigation into Cancer and Nutrition cohort. Outcomes indicate the possibility of little improvement in BC risk prediction with modeling heterogeneous epidemiological risk factors. Nonetheless, more applicability is suggested for model specifically predictive of ER+ tumor risk than an omnibus model in risk assessment for chemoprevention.