The number of treatment options for patients with prostate cancer is expanding, as mechanism-based therapies enter clinical testing. Such strategies are based on manipulating tumoral genetics, cell signaling, apoptosis, angiogenesis, the immune system, and others. To properly develop these agents, traditional trial designs and outcome measures require reconsideration. This review describes these novel drugs and their targets using a clinical states model. Investigators use a clinical states model as a framework for matching specific agents with the needs of the treatment population and the biology of the tumor at any given point in the natural history of the disease.