Survival benefit of definitive therapy in patients with clinically advanced prostate cancer: estimations of the number needed to treat based on competing-risks analysis.
Academic Article
Overview
abstract
OBJECTIVE: To describe the survival benefit associated with radical prostatectomy (RP), as compared with initial observation, in patients with locally advanced prostate cancer (PCa). PATIENTS AND METHODS: Overall, 1382 patients with locally advanced PCa treated with RP or initial observation between 1995 and 2009 were identified from the Surveillance, Epidemiology and End Results Medicare insurance programme-linked database. Patients were matched using propensity-score methodology, then 10-year cancer-specific mortality (CSM) rates were estimated and the number needed to treat (NNT) was calculated. Competing-risks regression analyses tested the relationship between treatment type and CSM. RESULTS: Overall, the 10-year CSM rates were 11.8 and 19.3% for patients treated with RP and initial observation, respectively (P < 0.001). The corresponding 10-year NNT was 13. The 10-year CSM rates for the same treatment groups were 8.9 vs 13.9%, respectively, for Gleason score ≤7, 16.8 vs 27.8%, respectively, for Gleason score 8-10, 10.1 vs 15.8%, respectively, for clinical stage T3a, and 17.0 vs 29.3%, respectively, for T3b/T4, respectively (all P ≤ 0.04). The corresponding NNTs were 20, 9, 17 and 8, respectively. In multivariable analyses, RP was an independent predictor of more favourable CSM rates in all categories (all P ≤ 0.04). In separate sensitivity analyses, no differences were recorded when patients treated with radiotherapy were compared with those receiving RP (P = 0.4). Conversely, patients undergoing initial observation had a higher risk of CSM compared with those treated with radiotherapy (P = 0.03). CONCLUSIONS: RP leads to a significant survival advantage compared with observation in patients with locally advanced disease. The highest benefit was observed in patients with T3b/T4 and Gleason score 8-10 disease.
