Development of a preoperative risk-scoring system for predicting poor responders to peroral endoscopic myotomy.

ABSTRACT Background and aims Peroral endoscopic myotomy (POEM) is an effective treatment for esophageal motility disorder. However, there are poor responders who will probably need retreatments such as endoscopic pneumatic dilation or re-POEM, and a scoring-system for the prediction of poor responders preoperatively remains unestablished. We aimed to develop and validate a preoperative scoring-system for predicting poor responders. Methods Overall, 244 patients who underwent POEM for esophageal motility disorders in our hospital from April 2015 to March 2019 were retrospectively included in this study. Poor responders were defined as patients with any of following: (1) Eckardt score ≥3 at 1-year follow-up, (2) endoscopic findings of food retention at 1-year follow-up and (3) retreatments within 1 year after POEM. A risk-scoring system for poor responders was developed based on multiple logistic regression analysis and its performance was internally validated using bootstrapping. Results Forty patients were diagnosed poor responders at the 1-year follow-up. In multivariate study, points of risk scores were assigned for 4 independent risk factors as follows: pretreatment Eckardt score (1-point increments), previous treatments (4 points), sigmoid-type esophagus (4 points), and esophageal dilation grade ≥Ⅱ (4 points). The scoring system could predict an estimated risk for poor responders and provided satisfactory discrimination (area under the receiver operating characteristic curve, 0.78 [95% confidence interval, 0.68-0.88]) and calibration (slope = 0.93 [95% confidence interval, 0.62-1.31]). Conclusions A validated risk-scoring system for predicting poor responders preoperatively was established; this system could be useful in selecting treatment strategies and postoperative surveillance.