Use of circulating tumor DNA (ctDNA) to complement tumor tissue homologous recombination repair (HRR) gene alteration testing in TALAPRO-2, a phase 3 study of talazoparib (TALA) + enzalutamide (ENZA) versus placebo (PBO) + ENZA as first-line (1L) treatment in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC).

5056 Background: In TALAPRO-2 (NCT03395197), pts unselected for tumor genetic alterations in DNA damage repair pathways, directly or indirectly involved with HRR, received TALA + ENZA (n=402) or PBO + ENZA (n=403) as 1L treatment for mCRPC. HRR status was informed by tumor tissue for 804 pts (99.9%), by tumor tissue and ctDNA for 114 (14.2%), and by ctDNA only for 1 (0.1%). TALA + ENZA demonstrated statistically significant and clinically meaningful improvements in imaging-based progression-free survival over standard of care PBO + ENZA regardless of HRR gene alteration status (presented by Agarwal N at ASCO-GU 2023). Here we assessed the overall agreement of solid tissue vs ctDNA-based testing in assessing tumor HRR gene alteration status. Methods: HRR status for pts randomized in TALAPRO-2 was determined prospectively by testing for the presence of HRR gene alterations using FoundationOne CDx and/or FoundationOne Liquid CDx. Pts were considered HRR-deficient if they had ≥1 alteration in ≥1 of 12 HRR genes by either test. Plasma samples collected at screening were retrospectively tested using FoundationOne Liquid CDx to examine retrospective test performance and to support exploratory resolution of pts with unknown prospective HRR status. Results were reported as HRR-deficient, non-HRR-deficient, or unknown. Results: In prospective testing, high concordance in HRR status was observed between ctDNA and tumor tissue sample results: 95% agreement (90/95) between tests was observed in pts evaluable as either HRR-deficient or non-HRR-deficient by both tests. Retrospective ctDNA test results were reported for 739 pts (207 [28%] with a status of HRR-deficient, 480 [65%] non-HRR-deficient, and 52 [7%] unknown; 66 pts had samples not collected or reported). There was high concordance in pt-level alteration status between prospective molecular profiling (mainly based on solid tumor tissue per above) and retrospective ctDNA testing: the agreement was 84% (438/519) between tests for pts evaluable as HRR-deficient or non-HRR-deficient both prospectively and retrospectively (421/501 [84%] for prospective tumor tissue only vs retrospective ctDNA). Conclusions: Results show 95% agreement between prospective tissue and ctDNA-based HRR status using FoundationOne, consistent with tumor vs ctDNA results recently reported (Tukachinsky. Clin Cancer Res. 2021;27:3094–3105). The high agreement between prospective test results based primarily on tumor tissue and retrospectively assessed ctDNA collected at screening is supportive of exploratory retrospective testing of ctDNA to inform the HRR status of pts whose alteration status is unknown based on prospective testing. Clinical trial information: NCT03395197 .