Circulating Tumor DNA in Resected Colorectal Liver Metastases: From Minimal Residual Disease Detection to Treatment Guidance—A Systematic Review

Background: Circulating tumor DNA (ctDNA) has emerged as a potential marker of minimal residual disease (MRD) after resection of colorectal liver metastases (CRLM), but its prognostic value and possible role in guiding postoperative management remain uncertain. This systematic review evaluated the available evidence on ctDNA for predicting recurrence and survival after CRLM resection, with particular attention to perioperative and serial ctDNA measurements and their potential value in guiding adjuvant treatment. Methods: PubMed/MEDLINE, Scopus, and the Cochrane Library were searched from inception to 21 July 2026 for studies evaluating ctDNA in patients undergoing resection of CRLM. Prospective and retrospective cohort studies and relevant randomized trial sub-studies were eligible. The primary outcome was the association between ctDNA status and recurrence, while secondary outcomes included recurrence-free survival (RFS), disease-free survival (DFS), overall survival (OS), serial ctDNA dynamics, and the potential predictive value of ctDNA for adjuvant chemotherapy. Risk of bias was assessed using the QUIPS tool. Findings were synthesized narratively, with an additional exploratory random-effects meta-analysis of early postoperative tissue-informed ctDNA positivity and RFS/DFS. Unadjusted and adjusted HRs were pooled separately. Results: Twenty-eight studies published between 2018 and 2026 were included, representing 2079 unique patients after accounting for overlapping cohorts. In six studies, postoperative ctDNA positivity was associated with poorer RFS/DFS (pooled unadjusted HR 5.15, 95% CI 3.65–7.27; I2 = 0%). Five studies provided adjusted estimates (pooled HR 4.95, 95% CI 2.45–10.01; I2 = 60.5%). Serial monitoring provided additional prognostic information: persistent ctDNA positivity or conversion from negative to positive during follow-up was associated with poor outcomes, whereas ctDNA clearance was generally associated with a lower risk of recurrence. ctDNA could precede radiological recurrence by approximately 2.5 to >20 months, although detection was less reliable in patients with small-volume pulmonary recurrence. In contrast, the prognostic value of preoperative ctDNA was less consistent and appeared to depend on assay methodology, quantitative ctDNA burden, and previous systemic treatment. Observational studies reported associations between postoperative MRD status and outcomes following adjuvant chemotherapy, but non-randomized treatment allocation and potential time-related bias limit conclusions about treatment benefit. Considerable variation in assay platforms and postoperative sampling schedules remained across studies. Conclusions: Postoperative ctDNA positivity is associated with poorer recurrence-related outcomes after CRLM resection. However, the moderate-to-high risk of bias and differences between studies limit confidence in the evidence. The predictive value of ctDNA for adjuvant chemotherapy benefit remains unproven. Prospective randomized trials are needed to determine whether ctDNA-guided management improves patient outcomes.