Survival benefit of radical prostatectomy in lymph node-positive patients with prostate cancer.

BACKGROUND Positive lymph node (LN) status is considered a systemic disease state. In prostate cancer, LN-positive diagnosis during pelvic LN dissection (PLND) potentially leads to the abandonment of radical prostatectomy (RP). OBJECTIVE To compare the overall survival (OS) and relative survival (RS; as an estimate for cancer-specific survival) in LN-positive patients with or without RP. DESIGN, SETTING, AND PARTICIPANTS Between 1988 and 2007, a total of 35 629 men with prostate cancer were identified at the Munich Cancer Registry; of those, 1413 patients had positive LNs. INTERVENTION Of these 1413 LN-positive patients, prostatectomy was abandoned in 456 LN-positive patients, whereas 957 underwent RP despite the LN-positive finding. MEASUREMENTS Crucial analyses are based on 938 LN-positive patients (688 with RP and 250 without RP) with complete data regarding age, grade, and prostate-specific antigen (PSA). OS (Kaplan-Meier estimates) and RS are presented, and Cox regression analysis was used to show the influence of predictors such as clinical stage, age at surgery, number of positive LNs, PSA level, grade, and extent of surgery. RESULTS Median follow-up was 5.6 yr. OS of patients at 5 yr and 10 yr was 84% and 64%, respectively, with RP and was 60% and 28%, respectively, with aborted RP. The RS of patients at 5 yr and 10 yr was 95% and 86%, respectively, with RP and was 70% and 40%, respectively, with abandoned surgery. There was an imbalance, however, in the number of positive LNs: 17.2% with RP had four or more positive nodes versus 28% in the patient group without RP. In the multivariate model, RP was a strong independent predictor of survival (hazard ratio: 2.04 [95% confidence interval, 1.59-2.63; p<0.0001]). CONCLUSION LN-positive patients with complete RP had improved survival compared to patients with abandoned RP. These results suggest that RP may have a survival benefit and the abandonment of RP in node-positive cases may not be justified.