Construction of a preoperative scoring system to predict the difficulty level of colorectal endoscopic submucosal dissection

BACKGROUND: We attempted to examine the factors contributing to the difficulty in performance of colorectal ESD, with the aim of constructing a scoring system that could help in prediction of the difficulty level of the procedure. METHODS AND MATERIALS: The data were analyzed from two viewpoints: to determine the factors contributing to 1) non-en bloc resection and the factors contributing to 2) a slow resection speed. Factors falling under these two categories contributing to difficulty in performance of ESD were extracted and used to construct a scoring system. The validity of this scoring system was evaluated by calculating the correlation between the score and the resection speed in a different dataset. RESULTS: Based on the results of our analysis, we assigned scores for various factors as follows: 4 points for EMR of a scarred lesion, 1 point for tumors with a diameter of ≥ 30 mm, 2 points for lesions located in the liver/splenic flexure, 1 point for lesions located in the transverse colon, 3 points for LST-NG-PD/depressed lesions, 1 point for protruded lesions and LST-NG-F lesions (range 0–10). In the validation study, the rank correlation coefficient between the score according to the scoring system and the resection speed was -0.130, representing a weak and negative correlation (P = 0.03). We defined the difficulty level depending on the sum of the scores: 0–2, low difficulty level; 3–5, intermediate difficulty level; ≥ 6, high difficulty level. The average resection speed was 12.6 mm(2)/min in the group with scores of 0–2, 8.1 mm(2)/min in the group with scores of 3–5, and 5.5 mm(2)/min in the group with scores of ≥ 6 (11.2 mm(2)/min in all lesions). CONCLUSION: Our colorectal ESD scoring system would be useful for selection of operators with the appropriate skill level in the procedure for colorectal ESD cases.