Preoperative indicators of failure of en bloc resection or perforation in colorectal endoscopic submucosal dissection: implications for lesion stratification by technical difficulties during stepwise training

Background and Aims: The technical difficulties inherent in endoscopic submucosal dissection (ESD) for colorectal neoplasms may result in the failure of en bloc resection or perforation. The aim of this retrospective study was to assess the predictors of en bloc resection failure or perforation by using preoperatively available factors. Methods: Between September 2002 and March 2013, 716 colorectal ESDs in 673 consecutive patients were performed at a tertiary cancer center. Patient characteristics, tumor location, tumor type, colonoscopy-related factors, and endoscopist experience were assessed based on a prospectively recorded institutional ESD database. Logistic regression analysis was performed to identify predictors of failure of en bloc resection or perforations, with subgroup analyses of ESDs performed by endoscopists less experienced in colorectal ESD (< 40 cases) and for colonic lesions only. Results: On multivariate analysis, independent predictors of failure of en bloc resection or perforations were the presence of fold convergence (odds ratio [OR] 4.4; 95% confidence interval [95% CI], 1.9-9.9), protruding type (OR 3.6; 95% CI, 1.8-7.1), poor endoscope operability (OR 3.5; 95% CI, 1.8-6.9), right-sided colonic lesions (OR 3.0; 95% CI, 1.5-6.3 vs rectal lesions), left-sided colonic lesions (OR 3.2; 95% CI, 1.7-6.3, vs rectal lesions), the presence of an underlying semilunar fold (OR 2.1; 95% CI, 1.3-3.6), and a less-experienced endoscopist (OR 2.1; 95% CI, 1.3-3.6). Among less-experienced endoscopists, colonic lesions were independent predictors (right-sided colonic lesions 8.1; 95% CI, 2.9-25.1; left-sided colonic lesions 8.1; 95% CI, 2.5-28.3 vs rectal lesions). For colonic lesions, the presence of fold convergence (OR 3.7; 95% CI, 1.6-8.6), poor endoscope operability (OR 3.6; 95% CI, 1.8-7.2), a less-experienced endoscopist (OR 3.0; 95% CI, 1.7-1.8), and the presence of an underlying semilunar fold (OR 2.7; 95% CI, 1.5-4.7) were identified predictors. Conclusion: This study successfully identified predictors of en bloc resection failure or perforation. Understanding these indicators could help to accurately stratify lesions according to technical difficulty and to appropriately select endoscopists.