Liquid biopsy AR/enhancer alteration detection before AR-targeted therapy and correlation with survival in metastatic castrate-resistant prostate cancer patients.

171 Background: Patients with metastatic castrate-resistant prostate cancer (mCRPC) have a variety of different systemic therapy options, yet survival outcomes remain poor. To minimize the chance of rapid resistance, it is critical to be able to personalize therapy upfront. We previously developed a cell-free DNA liquid biopsy assay that tracks alterations in the androgen receptor ( AR) locus, including its upstream enhancer (EnhanceAR-Seq). This assay demonstrated the ability to stratify patients based on resistance to AR-directed agents after treatment started (Dang and Chauhan et al., JCO PO, 2020). Here we evaluated if the same approach can accurately risk-stratify patients with mCRPC prior to the initiation of first-line AR-targeted therapy. Methods: We performed EnhanceAR-Seq on plasma cell-free DNA samples from 55 mCRPC patients prior to the initiation of AR-targeted therapy. Forty-seven patients were treated with Abiraterone and eight with Enzalutamide. Five patients were excluded due to having already received AR-targeted therapy or being lost to follow-up. Kaplan-Meier analysis was performed, and progression-free survival (PFS) and overall survival (OS) were assessed in relation to AR/enhancer locus status (wild-type vs. altered). Results: Median follow-up time was 30 months. EnhanceAR-Seq noninvasively detected AR/enhancer locus alterations in 36% (18/50) of the patients in our cohort, while 64% (32/50) were determined to be wild-type. Alterations detected by the assay were AR amplification (12/18), AR nonsynonymous single nucleotide variants (4/18), AR truncation (2/18), and AR enhancer amplification (13/18). Eleven patients had both AR gene body and enhancer amplifications present. Strikingly, patients with AR/enhancer alterations detected in plasma cell-free DNA were found to have significantly worse outcomes, with median PFS of 16.3 (wild-type) vs. 10.8 months (altered) (p = 0.046; HR = 2.10), and OS trending toward significantly different with median 34.3 (wild type) vs. 24.7 months (altered) (p = 0.19; HR = 1.05). Conclusions: AR/enhancer locus alterations in patients with mCRPC, as detected noninvasively by EnhanceAR-Seq, corresponded with significantly worse PFS and trended toward significantly worse OS. Our results support the role of cell-free DNA AR/enhancer locus alterations as prognostic, and potentially predictive, biomarkers that enable more precise upfront risk stratification and treatment personalization.