Predicting outcomes in colorectal endoscopic submucosal dissection: a United States experience

Objective Endoscopic submucosal dissection (ESD) allows for en bloc resection of superficial gastrointestinal neoplasms; however, US experience has been limited. We aimed to evaluate our clinical outcomes in colorectal ESD. Design This prospective study included consecutive patients undergoing colorectal ESD at a major US center. Demographics, lesion and technical characteristics, outcomes, adverse events, and pathological diagnoses were recorded. Factors affecting resection outcomes and procedure time were evaluated. Results 77 patients who underwent colorectal ESD were analyzed. Mean colorectal lesion diameter was 49.4 mm. Mean procedure time was 104.7 min, and 97.4% of patients were discharged home on the same day. En bloc, complete, and curative resection was achieved in 97.4%, 97.4%, and 93.5% of colorectal ESD cases. Microperforation and delayed bleeding rates were 1.3% and 3.9%. On univariable analysis, the presence of tattoo adversely affected en bloc resection (p=0.002), complete resection (p=0.002), and curative resection (p=0.008). Prior EMR attempts adversely affected en bloc resection (p=0.028), complete resection (p=0.028), and procedure time (p=0.008). On multivariable analysis, the presence of tattoo predicted failure to achieve curative resection (OR 0.13; 95% CI 0.02-0.98; p=0.048). Lesion size>50 mm (OR 3.89; 95% CI 1.13-13.41; p=0.031), presence of tattoo (OR 9.38; 95% CI 1.05-83.83; p=0.045), and prior EMR attempts (OR 7.13; 95% CI 1.76-28.90; p=0.006) predicted procedure time >= 90 min. A scoring system was created to predict prolonged ESD procedure time and was externally validated, with AUC 0.78 (95% CI 0.73-0.83). Conclusion This study demonstrates the effects of multiple risk factors on resection outcomes and procedure time in colorectal ESD. Tattoo placement and attempted EMR should be avoided for lesions being considered for ESD.