The recurrence of prostate cancer after potentially curative local therapy is becoming a significant urologic problem. There are few prospective randomized trials, and the optimal diagnostic and treatment strategies for men who fail potentially curative therapy are not known. The experience to date seems to suggest the following as a reasonable approach. A detectable serum PSA level (> or = 0.4 ng/mL) after radical prostatectomy is evidence of residual or recurrent prostate cancer. Men with low- or moderate-grade cancers (Gleason score < 7), with capsular penetration, or with positive surgical margins in whom disease recurs more than 2 years after radical prostatectomy with a PSA doubling time greater than 12 months seem likely to harbor a local recurrence and are the only good candidates for salvage therapy. Unless there is a palpable recurrence, transrectal ultrasound and biopsy are generally not recommended, and CT scanning and bone scintigraphy usually do not provide helpful information. The role of monoclonal antibody scanning is currently investigational. Men with high-grade tumors (Gleason score > or = 7) or with seminal vesicle or lymph node involvement in whom disease recurs within 2 years of radical prostatectomy are most appropriately observed or treated with early hormonal therapy. Men who do not achieve a PSA nadir of 0.5 ng/mL or less within 2 years of radiotherapy are very likely to harbor residual disease. For young healthy men who are willing to accept a substantial risk of impotency, urinary incontinence, and bladder neck contractures, salvage radical prostatectomy is a reasonable option if the preradiation tumor characteristics are acceptable (PSA < 10 ng/mL, Gleason score < or = 6) and if the current PSA is less than 10 ng/mL. Salvage cryotherapy may result in substantial morbidity and should only be offered on an investigational basis. Other men failing radiation may be observed or treated with hormonal therapy. There is seldom a role for repeat biopsy. Because the optimal time to begin hormone therapy is still not known, early or delayed treatment are both reasonable options. Testicular androgen ablation by orchiectomy or LHRH agonists is considered standard therapy. Combined therapy with an antiandrogen does not seem to be beneficial for all patients and should not be routinely used. Sexually active men in whom preservation of potency is important can be offered an investigational regimen such as a 5-alpha-reductase inhibitor combined with an oral antiandrogen or intermittent LHRH agonist therapy. It is hoped that the results of ongoing randomized trials and future research will establish efficient and effective practice guidelines to evaluate and treat men who have failed potentially curative therapy for localized prostate cancer. This remains a very important and controversial topic that will challenge many practicing urologists.