Whole-genome sequencing–based ultra-sensitive ctDNA molecular residual disease assessment in resectable gastric cancer: Results from MONSTAR-SCREEN-3.

Abstract

4081 Background: Circulating tumor DNA (ctDNA) has emerged as a promising biomarker for post-surgical molecular residual disease (MRD) detection; however, its clinical utility in gastric cancer remains poorly defined. The MONSTAR-SCREEN-3 evaluates the clinical performance of a whole-genome sequencing (WGS)–based MRD assay in a pan-cancer cohort, including patients with gastric cancer. Methods: MONSTAR-SCREEN-3 is a prospective multicenter study targeting 1,100 patients with solid tumors curative-intent treatment, including gastric cancer. Personalized panels were constructed using Precise MRD (Myriad Genetics), incorporating up to 1,000 tumor-specific alterations identified through whole-genome sequencing (WGS) of matched tumor tissue. Serial plasma samples were collected at baseline, post-neoadjuvant treatment (NAT) when applicable, 1-month post-surgery, every 3 months in year 1, and every 6 months thereafter up to 2 years. Assay performance was evaluated for ctDNA detection and recurrence monitoring. Results: As of November 2025, 88 patients with resectable gastric cancer were enrolled. Median age was 69 years (range: 33-87), with female predominance (52.2%). Treatment strategies included upfront surgery (56%, n = 49) and NAT (44%, n = 39). MRD results were available for 220 samples from 57 patients. Clinical staging included Stage I (6%), Stage II (39%), Stage III (51%), and Stage IV (5%). Personalized panel creation succeeded in 100% of patients (57/57), yielding panels containing 588-1,000 alterations. Baseline ctDNA detection was achieved in 96.4% (54/56), with 31.5% at ultra-sensitive levels (tumor fraction < 100 parts per million [ppm]). Post-operative MRD positivity was 7.8% (4/51) at 1 month and 19.5% (8/41) at 3 months, with 25.0% and 50.0% detected at ultra-sensitive levels, respectively. Among 24 patients receiving NAT with pathological assessment, post-NAT MRD status demonstrated 70.0% sensitivity and 100% specificity for pathological complete response (pCR) (MRD positivity, 70% in non-pCR vs. 0% in pCR; P < 0.01); 57.1% of MRD-positive cases were detected at ultra-sensitive levels. Two MRD-positive patients developed radiological recurrence, with MRD detection preceding imaging by 2.0 and 4.3 months. Conclusions: This WGS-based personalized ctDNA assay demonstrated high technical feasibility and baseline detection sensitivity in gastric cancer, with notable ultra-sensitive detection capability. Post-NAT MRD status demonstrated potential as a predictor of pathological complete response. Extended follow-up and comprehensive longitudinal ctDNA dynamics will be presented.