Molecular residual disease detection using whole-genome sequencing-based ctDNA assay for resectable colorectal cancer in the MONSTAR-SCREEN-3 project.

187 Background: Circulating tumor DNA (ctDNA) has been established as a strong predictive biomarker for post-surgical molecular residual disease (MRD) detection. For colorectal cancer (CRC), several promising ctDNA-based MRD assays are under global development, with validation studies underway. The MONSTAR-SCREEN-3 study aims to evaluate an ultra-sensitive whole-genome sequencing (WGS)-based MRD assay across cancer types, including a dedicated CRC cohort. Methods: MONSTAR-SCREEN-3 (N=3,200) is a prospective multicenter study, including the “Definitive cohort” enrolling 1,100 patients with solid tumors undergoing definitive therapy. The study utilizes Precise MRD (Myriad Genetics) to detect ctDNA, incorporating up to 1,000 tumor-specific alterations identified through WGS of tumor tissue. Longitudinal plasma samples were collected before and after definitive therapy with a planned 2-year follow-up. For CRC, patients with cStage III–IV disease undergoing curative surgery were eligible. Results: As of the end of August, 2025, 184 CRC patients were enrolled, of whom 95 underwent WGS analysis for tissue samples, and 86 had MRD data 1-month post-surgery. Primary tumors were most frequently located in the sigmoid colon (32%) and ascending colon (17%). Pathologic stage I/II/III/IV disease accounted for 9%/42%/47%/1%, respectively; 60% received adjuvant chemotherapy. Panel design was successful in 99% (94/95) of patients with tissue WGS data available, with one case failing due to insufficient panel size. Customized panels had at least 954 variants (range: 954-1,000, 72% of patients had the maximum of 1,000). ctDNA positivity was 100% (94/94) at baseline and 20% (17/86) one month after surgery. Importantly, ultra-sensitive detection (tumor fraction <100 parts per million) was observed in 6% (6/94) at baseline and 41% (7/17) of MRD-positive patients postoperatively. Six patients developed recurrence, all of whom were MRD-positive at 1-month post-surgery, with a median lead time of 3.2 months (range: 0.0-6.0), highlighting the potential clinical impact of earlier detection. Meanwhile, no recurrence occurred among MRD-negative patients. Conclusions: A WGS-based MRD assay successfully identified future recurrence and shows potential for recurrence prediction. These findings support the potential clinical utility of this assay, though longer follow-up is needed to fully establish its role in colorectal cancer management.