Prognostic and predictive value of ctDNA-based MRD and actionable biomarkers in patients with resectable colorectal cancer: CIRCULATE-Japan GALAXY.

3609 Background: The role of biomarker-guided therapy for resectable colorectal cancer (CRC) remains unestablished. Circulating tumor DNA (ctDNA) has emerged as a minimally invasive biomarker for detecting molecular residual disease (MRD) and predicting recurrence. Tumor-informed ctDNA testing offers the advantage of tumor sequencing and MRD detection in a single workflow, suggesting its potential for facilitating integration of biomarker-guided therapy into perioperative treatment. Methods: GALAXY, a large-scale prospective observational study, monitored post-surgical ctDNA MRD status in patients with clinical stage II to IV or relapsed CRC (UMIN000039205). Tumor tissue and matched normal DNA were processed for whole-exome sequencing (WES) to identify and track patient-specific somatic single nucleotide variants (SNVs), for ctDNA detection in the associated patients’ plasma samples. Actionable biomarkers having OncoKB therapeutic levels 1-3 were explored and identified separately from the Signatera test (through WES data). Results: Of the 4,590 patients enrolled between May 2020, and October 2022, 2,695 who were not enrolled in randomized trials and had available WES data were included in this analysis. Identified actionable biomarkers included RAS/ BRAF wild-type in 1,369 (50.8%) patients, tumor mutational burden (TMB) high in 256 (10.3%), MSI high in 236 (9.5%), BRAF V600E in 176 (6.5%), KRAS G12C in 61 (2.3%), ERBB2 amplification in 47 (1.7%), TP53 Y220C in 33 (1.2%), NTRK1/ 2/ 3 fusion in 1 (0.04%), and RET fusion in 1 (0.04%). ctDNA positivity during the MRD window (2 to 10 weeks post-surgery) varied from 5.1% in those with TMB high and MSI high to 29.8% in those with ERBB2 amplification. ERBB2 amplification was the only actionable biomarker significantly associated with ctDNA positivity during the MRD window (odds ratio 2.20). Positive ctDNA MRD status correlated with significantly worse DFS across all actionable biomarkers, with hazard ratios ranging from 8.46 for TP53 Y220C to 5.53e+8 for BRAF V600E. In MSI high disease, recurrence was observed in only 3 of 225 (1.3%) patients with negative ctDNA compared to 5 of 12 (41.7%) with positive ctDNA during the MRD window. Adjustment by the propensity score revealed that adjuvant chemotherapy provided survival benefits in ctDNA-positive patients with pathological stage II and III disease across various biomarkers. In patients with one or more actionable mutations selected among the16-plex of the ctDNA MRD assay, those variants were detected at time of disease recurrence. Conclusions: Our study highlights the importance of ctDNA-based MRD as both a prognostic and predictive tool for CRC across different actionable biomarkers. These findings suggest the potential of MRD-driven targeted therapeutic strategies for CRC with specific actionable biomarkers. Clinical trial information: UMIN000039205.