Predicting survival in node-positive prostate cancer after open, laparoscopic or robotic radical prostatectomy: A competing risk analysis of a multi-institutional database

Objectives To investigate cancer‐specific mortality and other‐cause mortality in prostate cancer patients with nodal metastases. Methods The study included 411 patients treated with radical prostatectomy and pelvic lymph node dissection for prostate cancer with lymph node metastases at 10 tertiary care centers between 1995 and 2014. Kaplan–Meier analyses were used to assess cancer‐specific mortality‐free survival rates at 8 years' follow up in the overall population, and after stratifying patients according to clinical and pathological parameters. Uni‐ and multivariable competing risk Cox regression analyses were used to assess cancer‐specific mortality and other‐cause mortality. Finally, cumulative‐incidence plots were generated for cancer‐specific mortality and other‐cause mortality after stratifying patients according to the number of positive lymph nodes and the median age at surgery, according to the competing risks method. Results Men with prostate‐specific antigen ≤40 ng/mL and those with one to three positive lymph nodes showed higher cancer‐specific mortality‐free survival estimates as compared with their counterparts with prostate‐specific antigen >40 ng/mL and >3 metastatic lymph nodes, respectively (all P < 0.001). At multivariable Cox regression analyses, preoperative prostate‐specific antigen >40 ng/mL, >3 lymph node metastases and pathological Gleason score 8–10 were all independent predictors of cancer‐specific mortality (all P‐values ≤0.001). On competing risk analysis, when patients were stratified according to the number of positive lymph nodes (namely, ≤3 vs >3), the 8‐year cancer‐specific mortality rates were 27.4% versus 44.8% for patients aged 3 positive lymph nodes, the overall mortality rate is completely related to prostate cancer in young patients.