End points of clinical trials in metastatic castration-resistant prostate cancer: A systematic review.

Abstract

To review the definition and performance of the commonly used end points in trials of systemic therapies in metastatic castration-resistant prostate cancer patients. A literature search was undertaken on PubMed database to identify studies meeting established criteria, with the aim of selecting randomized clinical trials and study definition and performance of their end points. The end points were grouped into three categories: overall survival (OS), time-to-event end points, and response end points. A special analysis was performed for secondary end points of the studies which documented a benefit in OS in the experimental arm. Finally, publishes analyses for surrogacy of the included end points were also reported. OS, time-to-event and response end points in 31 selected trials were analyzed. OS was the primary end point in 14 trials, and the secondary end point in 17. A time-to-event end point was the primary end point in 8 studies, and the secondary end point in 22; the most reported time-to-event end points were composite end points, and the events changed among trials. A response end point was the primary end point in 9 studies, in 3 it was prostate-specific antigen (PSA)-related, in 3 pain-related and in 3 mixed. A response end point was the secondary end point in 19 studies: PSA response and radiologic response were the most frequently used secondary end points in 19 and 11 trials, respectively, while pain response was used in 5 studies. A homogeneous definition of progression in future trials is mandatory. Among response end points, pain-response and PSA-response appear to be the most reliable.