Analysis by DNA microarrays has led to the identification of molecular subtypes of breast carcinomas that show a distinct expression profile. Several studies have demonstrated that this ‘intrinsic subtype’ classification has a strong prognostic value. In addition, gene expression profiling techniques have been used to identify gene signatures that could be associated with the outcome of breast cancer patients. Several different genomic tests have been shown to better define the prognosis of early-stage breast cancer patients as compared with conventional clinical and pathological characteristics of the tumors, and some assays are already commercially available. However, it must be emphasized that the prognostic power of these genetic classifiers has not been confirmed yet in prospective trials. Genetic signatures that might predict the activity of specific chemotherapy agents have also been developed by using gene expression profiling techniques. The same approach has been used to identify gene signatures associated with the activation of oncogenic pathways that might represent targets for molecular therapy of breast cancer. By using these approaches, gene expression techniques might significantly improve our ability to predict the risk of recurrence and to tailor the treatment for each individual breast cancer patient.