Open Access

Robert Benedikt Höh, Christoph Würnschimmel, Rocco Simone Flammia, Benedikt Horlemann, Gabriele Sorce, Francesco Chierigo, Zhe Tian, Fred Saad, Markus Graefen, Michele Gallucci, Alberto Briganti, Carlo Terrone, Shahrokh F. Shariat, Derya Tilki, Luis Kluth, Philipp Mandel, Felix Chun, Pierre I. Karakiewicz

• Introduction: Randomized clinical trials demonstrated improved overall survival in chemotherapy exposed metastatic prostate cancer patients. However, real-world data validating this effect with large scale epidemiological data sets are scarce and might not agree with trials. We tested this hypothesis. Materials and Methods: We identified de novo metastatic prostate cancer patients within the Surveillance, Epidemiology, and End Results (SEER) database (2014-2015). Kaplan-Meier plots and Cox regression models tested for overall survival differences between chemotherapy-exposed patients vs chemotherapy-naïve patients. All analyses were repeated in propensity-score matched cohorts. Additionally, landmark analyses were applied to account for potential immortal time bias. Results: Overall, 4295 de novo metastatic prostate cancer patients were identified. Of those, 905 (21.1%) patients received chemotherapy vs 3390 (78.9%) did not. Median overall survival was not reached at 30 months follow-up. Chemotherapy-exposed patients exhibited significantly better overall survival (61.6 vs 54.3%, multivariable HR:0.82, CI: 0.72-0.96, p=0.01) at 30 months compared to their chemotherapy-naïve counterparts. These findings were confirmed in propensity score matched analyses (multivariable HR: 0.77, CI:0.66-0.90, p<0.001). Results remained unchanged after landmark analyses were applied in propensity score matched population. Conclusions: In this contemporary real-world population-based cohort, chemotherapy for metastatic prostate cancer patients was associated with better overall survival. However, the magnitude of overall survival benefit was not comparable to phase 3 trials.