The experience after 10 years with anatomical radical retropubic prostatectomy at The Johns Hopkins Hospital is reviewed. Between April 1982 and March 1991, 955 men with clinically localized prostate cancer (clinical stages T1 to T2) underwent staging pelvic lymphadenectomy and anatomical radical retropubic prostatectomy. Using actuarial analysis, at 10 years the likelihood of an undetectable prostate specific antigen (PSA) level was 70%, isolated elevation of PSA 23%, distant metastases 7% and local recurrence 4%. The actuarial likelihood of an elevated serum PSA increased with increasing pathological stage: the 10-year likelihood of freedom from PSA relapse was 85% for men with organ confined disease, 82% with focal capsular penetration, 54% with established capsular penetration and Gleason score 2 to 6 disease, 42% with established capsular penetration and Gleason score 7 to 10 disease, and 43% with seminal vesicle involvement. These data indicate that radical prostatectomy cures the majority of men with organ confined disease or with well to moderately well differentiated tumors that have penetrated the prostatic capsule to the extent where it is possible to obtain a clear surgical margin. Radical prostatectomy should be reserved for patients who can be cured and who will live long enough to benefit from it. These are also the patients who have the best quality of life postoperatively.