OBJECTIVE: Radical prostatectomy is commonly believed not to achieve the eradication of locally advanced disease. This retrospective study aimed to elucidate the role of radical prostatectomy in this condition. METHODS: A retrospective study of 158 patients surgically treated for clinical stage T3N0M0 prostate cancer was undertaken. Thirty patients had postoperative hormonal treatment, rendering prostate-specific antigen (PSA) follow-up unreliable, and were considered to be progressive at 1 month. Eighteen other patients received postoperative radiotherapy. One hundred and ten patients had radical prostatectomy only. PSA-relapse-free survival was analyzed. The mean follow-up time was 30 months. RESULTS: Seventy-nine percent of the resected specimens were pathologically T3 (pT3), and about 25% were pT3c. Thirteen percent were pT2 and 8% were pT4. Ninety-five specimens (60%) had positive surgical margins. There was poor accordance between the biopsy Gleason score and that of the specimen. A multivariate analysis showed that seminal vesicle and nodal invasion, margin status and a PSA level above 10 ng/ml were independent prognostic factors. In 47 cT3a patients with PSA <10 ng/ml, the PSA-free survival rate exceeded 70% at 24 months and the 5-year estimated PSA-free survival rate was more than 60%. CONCLUSIONS: Radical prostatectomy has a place in the treatment of clinical stage T3 prostate cancer patients with a PSA value lower than 10 ng/ml. There is a need to definitively rule out nodal or seminal vesicle invasion in order to select those patients that can benefit from surgery.