Complexity-driven anesthesia selection, procedural time, and R0 resection failure in colorectal ESD: an exploratory single-center study of a candidate 75-min threshold

Background The oncologic outcome of colorectal endoscopic submucosal dissection (ESD) critically depends on R0 resection, yet a minority of failures arise from unforeseen intraoperative complexity. Existing comparisons of R0 rates between general anesthesia (GA) and monitored anesthesia care (MAC) have not accounted for difficulty-driven anesthesia assignment. Methods In this single-center exploratory retrospective study, 370 patients undergoing colorectal ESD were enrolled. Oaxaca–Blinder decomposition characterized temporal changes in anesthesia allocation. Marginal structural models with stabilized inverse probability weighting controlled for confounding. Because only 29 R0 failures occurred (6 in the MAC group), all R0-related analyses—including mediation, nested prediction models, threshold identification, and decision curve analysis (DCA)—were reported as exploratory secondary analyses. Five complementary methods were used to derive a descriptive uncertainty range; DCA selected the primary actionable cut-point. Results The overall R0 resection rate was 92.2%, and baseline complexity was higher in the GA group. Measured lesion characteristics accounted for 43.0% of the GA–MAC procedural time gap in the early-to-middle period and 81.3% in the middle-to-recent period, consistent with a shift toward lesion difficulty–driven anesthesia selection. After covariate balancing, GA was associated with longer procedural time (geometric mean ratio 1.64, 95% CI 1.42–1.90). Procedural time was associated with part of the statistical association between GA and R0 failure, although this estimate reflects inconsistent mediation rather than a causal pathway. DCA identified 75 min as the primary actionable cut-point for increased R0 failure risk; the five complementary methods yielded a descriptive range of 71.5–95 min. Conclusions In the context of difficulty-driven anesthesia selection, a 75-min procedural-time threshold was associated with increased R0 failure risk beyond preoperative assessment in this exploratory single-center cohort. Procedural time may serve as a real-time marker of dissection complexity, but the threshold requires prospective multicenter validation before clinical use.