Abstract Background First-generation panel-based ctDNA sequencing is widely used to identify actionable genomic alterations, but limited coverage and sensitivity hinder broader clinical utility, especially in cancers driven by DDR LOF. Further, tissue-free, variant-based approaches for response monitoring have several challenges, including the limited number and low variant allele frequencies (VAF) of monitorable variants and contamination from non-tumor-derived sources (eg, clonal hematopoiesis [CH]). Newer panels may address these challenges with broader coverage, optimized bioinformatics, and the integration of epigenetic analysis for molecular response (MR). Methods Blood samples were collected from 2 phase 1 studies (TRESR; NCT04497116 and ATTACC; NCT04972110) of the ataxia telangiectasia and Rad3-related (ATR) inhibitor camonsertib in combination with gemcitabine or PARPi (talazoparib, niraparib, or olaparib). Cell-free DNA and PBMCs were sequenced using the combined genomic (>800 genes) and methylation (15Mb) GuardantINFINITY platform. Results The study included 117 patients (pts) with ovarian (n=29), breast (n=18), pancreatic (n=18), prostate (n=16) and other (n=36) cancers, enrolled with pathogenic alterations in BRCA2 (n=41), ATM (n=32), BRCA1 (n=23), PALB2 (n=8) and other genes (n=13). Enrollment alterations identified by local NGS were confirmed in 83% (94/114) of pts excluding 3 enrolled by protein loss. 50% (5/10) of complex alterations (large deletions or rearrangements) with sufficient circulating tumor fraction (cTF) were confirmed. Among pts with breast, ovarian, prostate, or pancreatic cancers previously treated with PARPi and/or platinum and harboring BRCA1/2 or PALB2 alterations, reversions were detected in 31% (17/54), with 6 pts reverted by 1 or more large intragenic deletions. A CRC pt with a clonal biallelic sATM alteration had 2 missense reversions in cis, an event not previously reported to occur in ATM. In pts with matched PBMCs, confirmed CH alterations were detected in 79% (44/56), including 59% (33/56) with at least 1 CH variant over 1% VAF. 106 pts had evaluable paired samples at baseline and on-treatment (taken between 2-12 weeks on-treatment [median, 3 weeks]). Using mVAF of the Guardant360 74-gene subset or methylation-based cTF, 72% or 86% were monitorable, respectively. In the subset of evaluable pts with matched PBMC sequencing (n=54), after PBMC-informed CH filtering, monitorability was reduced from 69% to 61% between G360 mVAF and G360 mVAF-CH, respectively. PBMC-informed CH filtering improved the correlation between variant-based and methylation-based monitoring (G360 mVAF, R=0.89, p=1.1e-11 vs G360 mVAF-CH, R=0.95, p=3e-15; Pearson correlation). Correlation of ctDNA monitoring metrics with clinical outcomes is ongoing. Conclusions Broad panel coverage enabled accurate diagnosis of complex DDR LOF and detection of reversion alterations. Methylation-based cTF allows monitoring of most pts in this cohort and consistent with previous observations may avoid signal contamination from CH for robust MR analysis. Citation Format: Ezra Rosen, Joseph D Schonhoft, Ian M Silverman, Arielle Yablonovitch, Sunantha Sethuraman, Parham Nejad, Danielle Ulanet, Julia Yang, Insil Kim, Kezhen Fei, Yi Xu, Errin Lagow, Shile Zhang, Mingyang Cai, Maria Koehler, Benedito A Carneiro, Stephanie Lheureux, Michael Cecchini, Benjamin Herzberg, Jorge S Reis-Filho, Victoria Rimkunas, Timothy A Yap. Circulating tumor DNA (ctDNA) genomic and epigenomic profiling (GuardantINFINITY) for diagnosis of DNA damage repair (DDR) loss of function (LOF) and response monitoring in the TRESR and ATTACC trials [abstract]. In: Proceedings of the AACR-NCI-EORTC Virtual International Conference on Molecular Targets and Cancer Therapeutics; 2023 Oct 11-15; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2023;22(12 Suppl):Abstract nr A123.
Read full abstract