RESUMO
BACKGROUND: The prognostic role of circulating tumor DNA (ctDNA)-based molecular residual disease (MRD) detection and its utility for postsurgical risk stratification has been reported in colorectal cancer. In this study, we explored the use of ctDNA-based MRD detection in patients with colorectal liver metastases (CLM), for whom the survival benefit of adjuvant chemotherapy (ACT) after surgical resection remains unclear. METHODS: Patients with CLM without extrahepatic disease from the GALAXY study (UMIN000039205) were included. The disease-free survival (DFS) benefit of ACT was evaluated in MRD-positive and -negative groups after adjusting for age, gender, number, and size of liver metastases, RAS status, and previous history of oxaliplatin for primary cancer. ctDNA was detected using a personalized, tumor-informed 16-plex polymerase chain reaction-next-generation sequencing (mPCR-NGS) assay. ctDNA-based MRD status was evaluated 2-10 weeks after curative surgery, before the start of ACT. RESULTS: Among 6061 patients registered in GALAXY, 190 surgically resected CLM patients without any preoperative chemotherapy were included with a median follow-up of 24 months (1-48 months). ctDNA positivity in the MRD window was 32.1% (61/190). ACT was administered to 25.1% (48/190) of patients. In the MRD-positive group, 24-month DFS was higher for patients treated with ACT [33.3% versus not reached, adjusted hazard ratio (HR): 0.07, P < 0.0001]; whereas no benefit of ACT was seen in the MRD-negative group (24-month DFS: 72.3% versus 62.2%, adjusted HR: 0.68, P = 0.371). Multivariate analysis showed that the size of liver metastases (HR: 3.94, P = 0.031) was prognostic of DFS in the MRD-positive group. In the MRD-negative group, however, none of the clinicopathological factors were prognostic of DFS. CONCLUSIONS: Our data suggest that ACT may offer notable clinical benefits in MRD-positive patients with CLM. MRD status-based risk stratification could be potentially incorporated in future clinical trials for CLM.