We now recognize that all breast cancers are not the same. Different characteristics in gene expression profiles result in differential clinical behavior. With the use of gene microarrays, different subtypes of breast cancer have been characterized. The basal subtype is characterized by high expression of keratins 5 and 17, laminin, and fatty acid-binding protein 7. The ERBB2+ subtype is characterized by high expression of genes in the ERBB2 amplicon. The luminal A subtype is characterized by the highest expression of the ER α gene. The luminal B and C subtypes have a lower expression of the ER cluster. The importance of these different subtypes lies in the fact that they differ in clinical outcome, with the basal and ERBB2+ subtypes having the worse prognosis and the luminal A group having the best prognosis. Different strategies for evaluating tumors in a clinical setting have been developed. Two such strategies are the 21-gene assay (Oncotype DX; Genomic Health, Redwood City, CA), which is currently in commercial use in the United States, and the 70-gene assay, which has been developed by a group in the Netherlands. These assays have been shown to predict clinical outcome and response to therapy. However, to date these gene assays have not been studied in a prospective manner. Over the next year, prospective clinical trials will be initiated using these predictive tools in the treatment of breast cancer. In the near future, clinical decisions will most likely be dictated by the genetic characteristics of the tumor, with the clinical characteristics becoming less important. Tailoring our treatment based on individual tumor characteristics will help us develop better therapeutic strategies and save many patients from receiving unnecessary toxic therapy.