Impact of perioperative complications on ctDNA-based MRD detection and prognosis: Insights from the GALAXY study.

Abstract

3600 Background: The GALAXY study (UMIN000039205) demonstrated the utility of circulating tumor DNA (ctDNA)testing to detect molecular residual disease (MRD) and monitor postoperative recurrence. This analysis evaluates the influence of perioperative complications and the timing of blood sampling on MRD detection rates and predicting clinical outcomes. Methods: From the 6,032 patients enrolled in GALAXY, 2,400 were available for this analysis after excluding those enrolled in randomized trials or with insufficient follow-up. A clinically validated, tumor-informed ctDNA assay (Signatera, Natera, Inc.) was utilized to prospectively detect and quantify ctDNA. MRD was assessed within a defined postoperative "MRD window" of 2-10 weeks post-surgery. Perioperative complications were classified as Grade 2 or higher according to the Clavien-Dindo classification. Results: Perioperative complications occurred in 302 cases (12.6%), with anastomotic leakage (2.4%), ileus (2.0%), and intra-abdominal abscesses (1.3%) being the most common. Complications were more frequent in males and patients with rectal cancer. Cell-free DNA (cfDNA) concentrations measured at 2–4 weeks post-surgery were significantly higher in cases with complications compared to those without complications (9.2 vs. 6.9 ng/mL; p < 0.001). Three-year recurrence-free survival (RFS) was significantly worse in MRD-negative cases with complications (80.7%) compared to those without complications (87.0%; HR 1.63; 95% CI 1.143–2.323; p = 0.007). In MRD-positive cases, 3-year RFS was 15.6% in patients with complications versus 19.9% in those without (HR 1.16; 95% CI 0.831–1.608; p = 0.389). Among patients with complications, ctDNA testing conducted at 2–4 weeks post-surgery versus 4–10 weeks showed marked differences in 3-year RFS based on landmark analysis at 10 weeks. For MRD-negative cases, 3-year RFS was 75.32% versus 90.71% (HR 2.875; 95% CI 1.252–6.605; p = 0.013). In MRD-positive cases, 3-year RFS was 27.7% versus 6.25% (HR 0.46; 95% CI 0.230-0.929; p = 0.030). This trend was similar even when colon and rectal cancer were analyzed separately. In contrast, no timing-related differences were observed in cases without complications. Conclusions: Perioperative complications may elevate cfDNA levels, potentially confounding MRD assessment. Our findings suggest delaying ctDNA testing to at least 4 weeks postoperatively in patients experiencing Clavien-Dindo Grade ≥2 complications. These results provide essential guidance for optimizing clinical trial designs involving MRD evaluation through ctDNA analysis. Clinical trial information: UMIN000039205 .