Abiraterone + prednisone (Abi) +/- veliparib (Vel) for patients (pts) with metastatic castration-resistant prostate cancer (CRPC): NCI 9012 updated clinical and genomics data.

5001Background: In preclinical CRPC models, PARP1 inhibition synergizes with AR targeted therapy, especially in ETS fusion-positive tumors. We hypothesized: 1. Co-targeting PARP-1 + AR is superior to AR inhibition and 2. ETS +ve predicts response. Methods: Pts had metastatic (mets) disease biopsy (bx), stratified by IHC-ETS status and randomized to Abi (Arm A) or Abi + Vel (Arm B). Primary endpoint: PSA response rate (RR > = 50% decline). Secondary endpoints: safety, objective RR (ORR), progression free survival (PFS), and molecular analysis including if DNA repair gene deficiency (DRD: BRCA 1, BRCA 2, ATM, FANCA, PALB2, RAD51B, RAD51C) predicts response. 148 pts stratified by IHC-ETS status were randomized to detect a 20% PSA RR improvement assuming a 5% 1-sided type I error and 80% power. An elastic net multivariable Cox model was used to analyze PFS. Mets bx underwent targeted exon sequencing and capture transcriptome analysis. Results: 72 pts were randomly assigned to Arm A and 76 to Arm B. PSA RR: Arm A 63.9%, Arm B 72.4% (p = 0.27). ORR: Arm A 45%, Arm B 52.2%, p = 0.51. Median PFS: Arm A 10.1 months (m), Arm B 11.3 m, p = 0.95. More Arm-B pts were on therapy for 12+ (45% vs 38%) and 18+ cycles (22% vs 17%). ETS status had no impact. Mets tissue sequencing (N = 80): 42 pts (53%) were ETS +ve, 19 (25%) had DRD, 47 (59%) had AR amplification/copy gain, 32 (40%) had PTEN mutation (mut), 33 (41%) had TP53 mut, 37 (46%) had PIK3CA activation (a) and 12 (15%) had WNT-a. Irrespective of arm pts with DRD had a higher PSA and ORR ( > = 87%) vs wild type (58%, 39%; p = 0.013, p = 0.002, respectively), higher PSA decline rate of > = 90% (74% vs 26%, p = 0.0004) and longer median PFS (95% CI): DRD 16.6 m (11 - NR) vs wild type: 8 m (5.4 ? 13.3); p = 0.02. PFS was longer in pts with normal PTEN (13.5 vs 6.2 m, p = 0.02), TP53 (13.3 vs 7.8 m, p = 0.04) and PIK3CA (10.3 vs 8.3 m, p = 0.03). Controlling for clinical factors, DRD, PTEN, TP53 and PIK3CA are associated with PFS in this order of importance. Conclusions: There was a modest trend in favor of Abi + Vel but no difference by ETS. Pts with DRD, normal PTEN,TP53 and PIK3CA had better PFS raising new hypotheses regarding the importance of integrating molecular analysis in therapeutic trials. Clinical trial information: NCT01576172.