Survival in metastatic castration resistant prostate cancer (mCRPC) trial participants.

e15136 Background: Median overall survival (mOS) in patients (pts) with metastatic prostate cancer progressing despite castrate levels of testosterone (mCRPC) was 13-16 months (m) in the pre-docetaxel era. These data, obtained from clinical trials, were used to construct currently available prognostic nomograms. We hypothesise that these models no longer reflect survival. Pts and physicians urgently require updated prognostic data on which to base management decisions. Methods: Pts with mCRPC treated on phase I-III trials at our institution were identified and data retrospectively collected. Predicted survival by Halabi and Smaletz nomograms were compared to calculated survival using Kaplan-Meier analysis. Cox model multivariate (MV) analysis used variables at referral, including performance status (PS), Gleason (GS), prostate specific antigen (PSA), lactate dehydrogenase (LDH), alkaline phosphatase (ALP), hemoglobin (Hb), visceral disease and albumin. Results: 423 pts with CRPC treated between 2003 and 2011 were included. At diagnosis median age was 62 years (y; 41.8 – 82.7); 226 (53.4%) had metastatic disease. Median interval from diagnosis to CRPC was 2.7y (0.2 – 21.7). At referral 248 pts (58.6%) were chemotherapy-naïve. Halabi and Smaletz models predicted mOS in chemo-naïve pts of 21m and 18m respectively, however the observed mOS was 32m (95%CI 28 – 38). Survival from CRPC was 43m (CI 37 – 46) and 39m (CI 34 - 44) in pre- and post-chemo pts, respectively. Conclusions: Despite aggressive disease characteristics, our pts lived significantly longer than predicted by current nomograms. MV analysis confirmed the importance of several previously identified prognostic factors. Survival data from this large cohort of CRPC pts should encourage men considering clinical trial participation. Previously developed nomograms no longer accurately predict survival.