Background: Breast cancer (BC), the most common cancer diagnosed in women in the United States, is a heterogeneous disease in which age-specific incidence rates (ASIRs) differ by race and mortality rates are higher in blacks than whites. Goals: (i) understand the reasons for the black-to-white ethnic crossover in the ASIRs; (ii) formulate a plausible explanation for higher BC mortality in blacks; and (iii) develop statistical models and computing tools for estimating BC survival. Methods: We collected incidence, survival, demographic, clinical and treatment data of BC in black and white women from NCI's SEER database. Incidence rates were stratified by the status of estrogen and progesterone receptor (ER, PR). Time period and birth cohort effects on incidence rates were estimated by age-period-cohort analysis. Multivariate Cox regression analysis was utilized to develop statistical models for prediction of BC survival. The accuracy of the models was assessed by c-indexes and calibration plots. Results: Analysis of ASIRs of BC phenotypes stratified by the ER and PR status showed that: (i) the black-to-white crossover in the corresponding incidence rates was absent; (ii) ASIRs for a given BC phenotype in black and white women are proportional in each age interval; (iii) blacks have lower risk for the [ER+, PR+] BC phenotype, higher risk for [ER-, PR-] and almost equal risk for [ER+, PR-] compared to whites; and (iv) mortality rates differ by race only for the [ER-, PR-] phenotype. This suggests that: (i) carcinogenesis in aging, leading to the development of each of the considered BC phenotypes, is similar in whites and blacks; (ii) the crossover is a result of not taking into account hormone receptor status; and (iii) the difference in race-specific mortality is due to larger death rates for [ER-, PR-] blacks. Six Cox Proportional Hazards models (stratified by age group and lymph node status) were developed. The models fit well the observed data, are well calibrated, and have good discrimination power (c-indexes) between cases. Based on these models, a web-based tool was developed to instantaneously estimate BC survival probability when given a patient's personalized demographic and clinical data. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com.bibliotheek.ehb.be/en-US/products/dissertations/individuals.shtml.]