This is a randomized, double-blind study, with two arms and 100 women in each arm. Premenopausal women will be recruited and randomly allocated to take one of the two different dietary supplements in pills daily for 2 years. The two supplements are soy isoflavones (treatment) and placebo (carbohydrates). Both treatment and placebo pills will contain multi-vitamins and minerals. Multiple blood, urine, and breast fluid samples will be obtained before and during the dietary supplement periods and analyzed for biomarkers of breast cancer risk. At baseline and after the intervention period, breast density and bone density will be assessed by radiologic techniques. The efficacy of the dietary intervention will be determined by comparing mean changes of serum markers for breast cancer risk,dense breast tissue, and bone mineral density over the two year dietary intervention period in the two supplement groups with adjustment for baseline values and individual patient characteristics of interest. We predict that 2 years of soy isoflavone supplement will reduce breast density, which may be explained by individual changes in serum markers of breast cancer risk.