Predictors of incomplete resection and perforation associated with endoscopic submucosal dissection for colorectal tumors.

Background and Objective: Colorectal endoscopic submucosal dissection (ESD) is technically challenging. Our aim was to identify predictors of incomplete resection and perforation in colorectal ESD. Design: Retrospective study. Setting: Academic Japanese endoscopy unit. Patients and Main Outcome Measurements: A total of 267 consecutive cases of colorectal tumors treated by ESD from May 2010 to February 2013 were analyzed. Predictors of incomplete resection and perforation, including lesion size, growth type, pathological diagnosis, use of hemostatic forceps, degree of fibrosis, history of biopsy, history of local endoscopic treatment, and endoscopic operability. Results: The incomplete resection rate was 4.1%. The perforation rate was 5.6%. Univariate analysis identified severe fibrosis (P = .032), submucosal (SM) deep (>1000 μm) invasion (P = .033) and poor endoscopic operability (P = .030) as predictors of incomplete resection, and severe fibrosis (P = .038), postendoscopic treatment (P = .016), and poor endoscopic operability (P = .012) as predictors of perforation. Multivariate analysis identified poor endoscopic operability and SM deep invasion as independent predictors of incomplete resection, and poor endoscopic operability and severe fibrosis as independent predictors of perforation. There was no adjustment of P values for multiple testing. Limitation: A single-center study by a single colonoscopist. All statistical results should be taken as descriptive only. Conclusions: Poor endoscopic operability and SM deep invasion were significant independent predictors of incomplete resections. Poor endoscopic operability and severe fibrosis were significant independent predictors of perforation. These features may provide helpful information when planning colorectal ESD. [Copyright &y& Elsevier]