Abstract 962: Monitoring on-treatment tumor molecular dynamics using plasma whole genome sequencing (pWGS) and whole exome sequencing (pWES) of circulating tumor DNA (ctDNA) in advanced esophageal cancer: An exploratory analysis of KEYNOTE-590

Abstract Background: We evaluated on-treatment tumor molecular dynamics using ctDNA-based pWGS and pWES in patients (pts) with locally advanced or metastatic esophageal cancer that responded to pembro + chemo or chemo alone and later experienced PD in KEYNOTE-590 (NCT03189719). Methods: Pts with locally advanced, unresectable, or metastatic esophageal or Siewert type 1 gastroesophageal junction cancer were randomized 1: 1 to first-line pembro 200 mg or placebo + chemo (5-fluorouracil + cisplatin) Q3W for ≤35 cycles. Cell-free DNA (cfDNA) was isolated and sequenced from plasma samples collected at baseline (BL), during on-treatment response (cycle 5 C5), and at PD (discontinuation D) from pts with response and subsequent PD. ctDNA burden (tumor fraction TF) was estimated using low-pass pWGS by quantifying tumor aneuploidy in cfDNA. All samples with TF 1. 5% underwent high-depth pWES. ctDNA burden from pWES was estimated as maximum somatic allele frequency (MSAF). Associations of TF with MSAF were assessed using Spearman correlation coefficient r. For each subject, Jaccard index (JI) was calculated to evaluate the overlap of variants between BL and D. Variant allele frequency (VAF) and normalized VAF (nVAF) were used to determine variant clonality. Results: 108 pts (pembro + chemo, 59; chemo, 49) with initial response and subsequent PD had plasma samples available from BL, C5, and D. Samples from 94 pts had TF 1. 5% by pWGS and underwent pWES. MSAF correlated with TF at BL (r = 0. 84), C5 (0. 56), and D (0. 86). ctDNA TF was lower at C5 than at BL and D, with a mean of 4. 4% (C5) vs 20. 9% (BL) and 14. 0% (D) with pembro + chemo and 4. 0% (C5) vs 15. 8% (BL) and 15. 5% (D) with chemo. In variant-level analyses of pWES data from samples with high TF (≥2. 5%) at BL and D (n = 65), 39. 2% and 44. 2% of variants were shared between BL and D with pembro + chemo and chemo, respectively. Higher overlap was observed for variants between BL and D for the chemo arm (JI: median, 0. 54; range, 0. 04-0. 91) vs pembro + chemo (0. 37; 0. 01-0. 92). Shared variants between BL and D had higher clonality compared with variants at BL-only or D-only with pembro + chemo (BL and D mean nVAF = 0. 50 vs BL-only = 0. 39, and D-only = 0. 38) and chemo (0. 49 vs 0. 38 and 0. 38). In pooled variant analyses, TP53 was the most commonly observed mutation and was rarely lost or acquired at D. Conclusions: Comprehensive pWES was feasible with sufficient ctDNA burden (TF ≥2. 5%). ctDNA burden dynamics followed high disease burden at BL and D, consistent with radiographic determination of response/PD. Tumor variants at BL and D generally had a common clonal trunk and a higher overlap of variants was observed at D in the chemo arm. These results provide preliminary molecular evidence of immune editing in pts receiving immunotherapy that warrants further study. Citation Format: Minita Shah, E. J. Dettman, Andrew Albright, Cai Chen, Julie Kobie, Sukrut Shah, Carol Peña, Razvan Cristescu, Z Alexander Cao. Monitoring on-treatment tumor molecular dynamics using plasma whole genome sequencing (pWGS) and whole exome sequencing (pWES) of circulating tumor DNA (ctDNA) in advanced esophageal cancer: An exploratory analysis of KEYNOTE-590 abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts) ; 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84 (6Suppl): Abstract nr 962.