Predictive factors for technically difficult endoscopic submucosal dissection in the colorectum

BACKGROUND AND STUDY AIMS: Because of technical difficulty, colorectal endoscopic submucosal dissection (CR-ESD) is not widely performed. We aimed to determine risk factors for such technical difficulty as defined by long procedure duration (≥  150 min), perforation, and piecemeal resection.

PATIENTS AND METHODS: Patients with consecutive colorectal tumors treated with ESD between April 2006 and December 2010 were enrolled in a prospective cohort study. For prediction of technical difficulty, three types of factor were investigated: tumor location, tumor type, and colonoscopy-related. Cases were subsequently categorized into earlier and later periods (April 2006 - August 2008, 123 lesions; September 2008 - December 2010, 124 lesions). Variables were analyzed using multiple logistic regression, with subgroup analyses for each period.

RESULTS: 247 lesions were analyzed. Flexure location was an independent risk factor for technical difficulty as measured by longer procedure duration (odds ratio [OR] 4.1, 95 % confidence interval [95 %CI] 1.1 - 14.9), piecemeal resection (4.7, 1.1 - 17.2), or perforation (8.8, 1.1 - 56.8). Tumor with scarring or locally recurrent was a risk factor for longer procedure duration (4.7, 1.7 - 13.7), and for piecemeal resection (7.8, 2.4 - 25.0). Tumor of size  ≥  50 mm or spreading across ≥  2 folds was the strongest independent risk factor for longer duration (6.3, 2.8 - 15.4), and was an independent risk factor for longer duration in both time periods (earlier, 3.3, 1.1 - 10.4; later, 27.4, 7.4 - 138.0). Flexure location was an independent risk factor for perforation (13.9, 1.5 - 129.1) and for piecemeal resection (5.1, 0.9 - 25.2) in the earlier but not the later period.

CONCLUSIONS: Factors predicting technical difficulty of CR-ESDs were clarified. Their importance was influenced by the increasing experience of the endoscopist.