The optimal management of clinically localized prostate cancer remains unresolved. Management options range from a conservative approach to definitive treatment. While evidence suggests that expectant management yields similar 10-year survival rates and quality-adjusted life years compared to definitive treatment, this approach alone will deprive some patients with potentially curable disease of the opportunity for curative therapy. Effective management of localized prostate cancer requires differentiation between patients with biologically aggressive disease, in whom curative therapy is strongly warranted, and those with indolent malignancy, in whom conservative management would be equally efficacious. A comparison of surveillance studies in the literature yields a striking similarity: every series contains a substantial subset of long-term survivors, particularly in patients with favorable clinical parameters. We describe a prospective clinical study to evaluate a novel approach in which the decision between definitive therapy and conservative management is determined by the rate of prostate-specific antigen (PSA) increase or the development of early, rapid clinical and/or histologic progression. We enrolled 250 patients followed with active surveillance with selective delayed intervention. Patients were followed with active surveillance until they met criteria defining significant PSA, clinical, or histologic progression. At a median follow-up of 42 months, 60 patients came off observation, while 140 have remained on observation. We conclude that an approach of active surveillance with selective intervention for patients with rapid biochemical or clinical progression is feasible, and that PSA doubling time appears useful in guiding intervention in patients managed initially with expectant management. A policy of close monitoring with selective intervention for the 15% who progress rapidly is appealing, and is currently being investigated in several clinical trials.