894O - Association of Radiographic Progression-Free Survival (RPFS) Adapted from Prostate Cancer Working Group 2 (PCWG2) Consensus Criteria (APCWG2) with Overall Survival (OS) in Patients (PTS) with Metastatic Castration-Resistant Prostate Cancer (MCRPC): Results from COU-AA-302

Abstract

ABSTRACT Background Imaging data have not been previously rigorously evaluated as an intermediate marker in mCRPC due to difficulties in interpretation and flare phenomenon. The APCWG2 is intended to continue treatment until clinically meaningful progressive disease (PD) occurs. In the prospective phase 3 study COU AA 302, we investigated correlation of OS with a novel rPFS endpoint using APCWG2 to recognize bone flare and avoid premature discontinuation (DC). Methods 1088 pts were randomized 1:1 to abiraterone acetate (AA) + prednisone (P) or placebo (PL) + P. Scans were obtained q 8 wks during first 24 wks, q 12 wks thereafter. APCWG2-defined PD by bone scan (BS) required confirmation or persistent lesions depending on timing of first detection of PD. All scans were investigator reviewed (IR) and blinded central radiologist reviewed (BCRR). Continuing blinded study medication after radiographic PD in absence of unequivocal clinical PD was allowed. Association of rPFS to OS was evaluated by Spearman's correlation via Clayton copula. Results Positive association between rPFS and OS was observed (r = 0.71). 229 pts (AA 108, PL 121) showed progression (wk 8; ≥ 2 bone lesions [BL]) by BCRR. Of these, 166 (AA 92, PL 74) did not show PD (follow-up; +≥ 2 BL) (ie, would have discontinued prematurely by traditional rPFS assessment). Of these, 47 (AA 33, PL 14) showed ↓ BL. There was high consistency between BCRR and IR at interim analyses (AA, 79%; PL, 76%). Interim Analysis (IA) Data Cutoff AA + P (median, mos) PL + P (median, mos) HR (95% CI) P RPFS BCRR IA1 - Dec 2010 NR 8.3 0.43 (0.35, 0.52) IR IA1 - Dec 2010 13.7 8.3 0.49 (0.41, 0.60) IR IA2 - Dec 2011 16.5 8.3 0.53 (0.45, 0.62) OS* IA2 - Dec 2011 NR 27.2 0.75 (0.61, 0.93) 0.0097 NR = not reached. *Pre-specified alpha level 0.0008. Conclusions BCRR and IR rPFS assessments and sensitivity analyses were highly consistent and reliable attesting to analytical validity of rPFS as a potential biomarker. Robust association between rPFS and OS provides preliminary support for use of rPFS based on APCWG2 as an intermediate marker of OS, and a primary/co-primary endpoint in phase 3 mCRPC studies. Disclosure A. Molina: I am an employee of Janssen Research & Development and own stock in Johnson & Johnson, P. De Souza: I have served as a paid consultant for Janssen Research Development, J. Li: I am an employee of Janssen Research & Development and own stock in Johnson & Johnson, T. Kheoh: I am an employee of Janssen Research & Development, and own stock in Johnson & Johnson, J.S. de Bono: I have served as a paid consultant for Janssen Research & Development, am an employee of The Institute of Cancer Research, and my work has been supported by Cancer Research UK, S.M. Larson: Consultant to Cougar Biotechnology and Perceptive Informatics, T. Griffin: I am an employee of Janssen Research & Development, and own stock in Johnson & Johnson, H.I. Scher: I have owned stock in Johnson & Johnson, have served as an abiraterone consultant/advisory board member, and have received consulting fees/grants/travel support fees from Janssen Research & Development, All other authors have declared no conflicts of interest.