Circulating Tumor DNA Status and Adjuvant Chemotherapy in Resected Colorectal Liver Metastases

Importance: The benefit of adjuvant chemotherapy (ACT) following curative resection of colorectal liver metastases (CRLM) remains uncertain, with no consistent overall survival (OS) benefit and optimal patient selection for postsurgical treatment not established. Objective: To evaluate the association between circulating tumor DNA (ctDNA)-defined molecular residual disease (MRD) and survival outcomes and whether MRD status is associated with differential benefit from ACT in resected CRLM. Design, Setting, and Participants: This prospective analysis included patients with resected CRLM enrolled in the CIRCULATE-Japan GALAXY study between May 2020 and July 2024 with available ctDNA results 2 to 10 weeks after surgery. Analyses were landmarked at 70 days and conducted separately for patients receiving neoadjuvant chemotherapy (NAC) and those undergoing upfront surgery. Data were analyzed from January to April 2026. Exposures: Postsurgical ACT vs observation. ctDNA was assessed using a personalized, tumor-informed assay. Main Outcomes and Measures: Disease-free survival (DFS) and OS, evaluated using Cox proportional hazards models. Results: Of 298 included patients, the median (range) age was 67 (33-85) years, and 192 (64.4%) were male. The median (range) follow-up was 43.2 (1-68) months. In the upfront surgery cohort (n = 191), postsurgical MRD positivity was associated with worse DFS (hazard ratio [HR], 4.14; 95% CI, 2.69-6.36; P < .001) and OS (HR, 9.13; 95% CI, 4.44-18.78; P < .001) compared with MRD negativity. Among patients with postsurgical MRD positivity, ACT was associated with improved DFS (HR, 0.07; 95% CI, 0.02-0.24; P < .001; 48-month DFS: 37.5% [95% CI, 15.4-59.8] vs not reached) and OS (HR, 0.27; 95% CI, 0.08-0.88; P = .03; 48-month OS: 65.3% [95% CI, 30.8-85.7] vs 32.9% [95% CI, 17.4-49.3]) compared with observation. In contrast, among patients with MRD negativity, ACT was not associated with improved DFS (HR, 0.69; 95% CI, 0.32-1.50; P = .36) or OS (HR, 0.54; 95% CI, 0.10-2.88; P = .47) benefit. In the NAC cohort (n = 107), postsurgical MRD positivity remained prognostic, with inferior DFS (HR, 4.82; 95% CI, 2.92-7.97; P < .001) and OS (HR, 9.43; 95% CI, 2.78-31.96; P < .001). ACT was not associated with improved DFS or OS regardless of MRD status. ctDNA positivity at the post-NAC presurgical time point was associated with worse OS (HR, 11.42; 95% CI, 1.58-1452.50; P = .009). Conclusions and Relevance: In this prognostic study, postsurgical ctDNA-defined MRD identified patients with resected CRLM who derived OS benefit from ACT. In contrast, patients with MRD negativity experienced favorable long-term OS without clear benefit from ACT. These findings support prospective validation of ctDNA-guided adjuvant strategies, particularly in patients with CRLM undergoing upfront surgery.