Historically, new prognostic markers for breast cancer were assessed separately and sequentially in a series of studies, often with considerable cost and substantial time to complete each subsequent study. Candidate markers were rarely compared with one another head to head since different investigators analyzed different cohorts of patients and used different technologies. The availability of public gene expression data sets that also include pathologic and clinical outcome information on each case has changed prognostic marker research considerably. These databases provide an opportunity to rapidly evaluate almost any mRNA expression-based markers for prognostic and treatment response predictive values in silico [1]. Not only single gene markers but any combinations of genes can be tested and their performance compared with already published mRNA-based outcome predictors.
展开▼
