CTC and AR-V7 analysis in the ECOG-ACRIN 8153 (CHAARTED2) randomized trial: Abiraterone (Abi) with or without cabazitaxel (Abi-Cabazi) in extensive mCRPC following docetaxel.

Abstract

181 Background: The CHAARTED2 trial showed improved PFS with Abi-Cabazi over Abi-alone in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC) who previously received docetaxel for metastatic castration-sensitive disease. Here, we evaluated the prognostic and predictive impact of CTC and AR-V7 circulating biomarkers. Methods: The modified AdnaTest (Qiagen) was used to define CTC detection (+ vs –) and AR-V7 detection (+ vs –) at baseline, C5D1, as well as at the time of progression. We first assessed the prognostic value of baseline and C5D1 CTC and AR-V7 status in the two arms. We also assessed the predictive potential of baseline and C5D1 CTC and AR-V7 status to explore if these biomarkers may aid in treatment selection. Logrank test and Cox proportional hazards models were used to assess for associations between clinical outcomes and baseline biomarker status. Landmark analysis was performed to evaluate associations at C5D1. Results: A total of 223 pts were randomized between the two arms. CTC data were evaluable for 94/112 and 94/111 pts in the Abi-alone and Abi-Cabazi arms, respectively. Baseline CTC detection was prognostic for worse PFS in both the Abi-alone (7.1 vs 31.7 months, HR 3.64, P <0.001) and the Abi-Cabazi (9.6 vs 36.9 months, HR 4.13, P <0.001) arms. Likewise, baseline AR-V7 detection portended worse PFS in the Abi-alone (4.1 vs 12.8 months, HR 3.34, P <0.001) and the Abi-Cabazi (6.9 vs 15.1 months, HR 3.17, P <0.001) arms. At C5D1, there was a higher rate of CTC(-) conversion (60 vs 46%) and AR-V7(-) conversion (69 vs 58%) in the Abi-Cabazi arm. CTC(-) and AR-V7(-) status at C5D1 was prognostic of improved PFS in both arms (all P ≤0.02). In terms of predictive impact, PFS was numerically but not statistically longer in the Abi-Cabazi arm in all biomarker groups; the greatest relative PFS benefit to Abi-Cabazi was in the baseline AR-V7(+) group (HR 0.68, 95%CI 0.36–1.3, P =0.2). Similarly, at C5D1, the greatest relative PFS benefit to Abi-Cabazi was in pts remaining AR-V7(+) (HR 0.57, 95%CI 0.15–2.16, P =0.4). Conclusions: Baseline CTC and AR-V7 detection were both prognostic for inferior PFS in both study arms. PFS outcomes were broadly longer in the Abi-Cabazi vs Abi-alone arm in all biomarker groups. The presence of AR-V7 at baseline and at C5D1 was the strongest relative predictor of Abi-Cabazi benefit, but did not meet statistical significance. Clinical trial information: NCT03419234 .