Practical use of meta-analyses in predicting disease risk, outcome, and therapy response in breast cancer

Breast cancer is globally the most frequent malignant disease in women with increasing incidence. Meta-analyses using data from a large set of patients combining genetic and standard clinicopathological features provide valuable models in predicting disease risk, outcome and therapy response. With the advent of molecular technologies, the amount of available data generated for each tumor and each patient is growing exponentially. The increased data availability allows the development of new innovative systems enabling to discover more effective prognostic and predictive biomarkers. The goal of this review is to summarize meta-analyses that utilize data from countless patients to provide breast cancer risk prediction (Gail-model, Claus-model, BRCApro, IBIS, BOADICEA), and prediction of prognosis and expected therapy response (PREDICT, Magee). In the last part of the review we introduce online analytical tools (KMplot, ROCplot) developed to examine, rank and validate new prognostic and predictive biomarker candidates.