Currently, the standard treatment for metastatic hormone-refractory prostate cancer includes a combination regimen with docetaxel and prednisone. Although such therapy has been demonstrated to improve survival compared with previous standards of treatment, the regimen is not curative. In efforts to identify alternative therapeutic strategies that would target single or multiple pathways activated in cancer, many small-molecule inhibitors and recombinant humanized mAbs have been tested in the preclinical setting, and many of these compounds have entered phase I/II clinical trials. However, no guidelines for the use of these drugs are available to date. Results from clinical trials are often difficult to interpret, primarily as a result of the lack of a molecular characterization of prostate tumors; such characterization would help in the search for predictive markers of response to specific drugs. Thus, a molecular classification of human prostate tumors is actively sought to assess predictive models for the stratification of patients with metastatic prostate cancer who are participating in targeted clinical trials.