Survival after radical prostatectomy or radiotherapy for locally advanced (cT3) prostate cancer

PURPOSE: No prospective data examined the effect of radical prostatectomy (RP) vs. external beam radiotherapy (EBRT) in locally advanced prostate cancer (PCa). We aimed to compare survival outcomes of RP and EBRT in patients harboring cT3N0-1 PCa. METHODS: Within the SEER database (2004-2014), we identified 5500 cT3N0-1 PCa patients. Cumulative incidence plots and competing-risks regression models (CRRs) tested cancer-specific mortality (CSM) and other cause of mortality (OCM) according to treatment type. The multivariable relationship between baseline prostate-specific antigen (PSA) values and 10-year CSM after either RP or EBRT was graphically depicted using the LOESS smoothing method. Sensitivity analyses were performed in cT3N0-only patients, after OCM propensity score matching, and through landmark analyses. RESULTS: Ten-year CSM and OCM rates were significantly higher after EBRT (15.8 and 28.2%) than RP (8.1 and 10.4%) (all p < 0.0001). In multivariable CRRs, RP yielded lower CSM [hazard ratio (HR): 0.64] than EBRT. Significantly lower 10-year CSM rate was recorded after RP vs. EBRT through the entire range of baseline PSA values. The same results were recorded in cT3N0 subgroup, as well as after OCM propensity score matching. Finally, landmark analyses at 6, 12, 24, and 36 months rejected the effect of favorable survival bias after RP. CONCLUSIONS: CSM was significantly lower after RP than EBRT in cT3N0-1 PCa. A lower CSM was recorded throughout the entire range of baseline PSA and even in cT3N0 subgroup, as well as after OCM propensity score matching and landmark analyses.