Perioperative immunocircadian recovery and its association with circulating tumor DNA-based minimal residual disease and long-term survival after colorectal cancer surgery.

Abstract

Postoperative outcomes after curative colorectal cancer surgery vary widely, even among patients with comparable tumor characteristics and treatment pathways. Conventional clinicopathologic factors and circulating tumor DNA predominantly characterize tumor-related risk but provide limited information regarding the dynamic host response to surgical stress. We investigated whether distinct perioperative immunocircadian recovery trajectories could be identified and whether these trajectories were associated with postoperative circulating tumor DNA-based minimal residual disease and long-term oncologic outcomes. This prospective observational cohort study included 1,000 patients with stage II-III colorectal cancer who underwent curative-intent resection between June 2017 and December 2019 and were scheduled for oncologic follow-up through 60 months. Perioperative immunocircadian recovery phenotypes were derived by integrating longitudinal wearable-based sleep, rest-activity rhythm, and heart-rate measures with serial inflammatory and immune markers collected from the preoperative period through postoperative day 30. Associations with postoperative circulating tumor DNA-minimal residual disease positivity were assessed using multivariable logistic regression, whereas associations with disease-free survival and overall survival were assessed using multivariable Cox regression and a postoperative day 30 landmark analysis. Three distinct recovery phenotypes were identified: rapid recovery (382/1,000; 38.2%), intermediate recovery (401/1,000; 40.1%), and delayed immunocircadian recovery (217/1,000; 21.7%). Postoperative circulating tumor DNA-minimal residual disease positivity increased from 8.9% (34/382) in the rapid-recovery group to 17.0% (68/401) in the intermediate-recovery group and 30.0% (65/217) in the delayed-recovery group. Compared with rapid recovery, delayed recovery was associated with greater odds of postoperative circulating tumor DNA-minimal residual disease positivity (adjusted odds ratio, 3.94; 95% confidence interval, 2.45-6.33). At 60 months, estimated disease-free survival was 82.1%, 76.4%, and 63.8%, respectively, and estimated overall survival was 90.0%, 86.9%, and 77.2%, respectively. In models adjusted for clinicopathologic factors but not circulating tumor DNA-minimal residual disease, delayed recovery was associated with worse disease-free survival (adjusted hazard ratio, 2.21; 95% confidence interval, 1.64-2.98) and overall survival (adjusted hazard ratio, 2.03; 95% confidence interval, 1.45-2.85). After additional adjustment for postoperative circulating tumor DNA-minimal residual disease, the corresponding adjusted HRs were 1.90 (95% confidence interval, 1.40-2.57) and 1.80 (95% confidence interval, 1.28-2.53), respectively. Addition of the recovery phenotype to a clinicopathologic model that included circulating tumor DNA-minimal residual disease increased the optimism-corrected C-index for disease-free survival from 0.754 to 0.781. Distinct perioperative immunocircadian recovery phenotypes were associated with postoperative circulating tumor DNA-minimal residual disease positivity and long-term oncologic outcomes after colorectal cancer surgery. The recovery phenotype provided information complementary to clinicopathologic factors and circulating tumor DNA-minimal residual disease within the study cohort. However, this investigational phenotype requires independent external validation before it can be used to guide postoperative clinical management.