Objective To validate the efficacy of the prediction model for difficulty of peroral endoscopic myotomy (POEM) through an independent cohort.

Methods A total of 617 patients with achalasia who underwent POEM at the Endoscopy Center of Zhongshan Hospital, Fudan University from January 2021 to December 2023 were included. The general data of patients were collected, and the predictive value of the prediction model for POEM difficulty in the validation cohort was estimated. The stratified analysis was undergone according to the difficulty risk scores.

Results In 617 consecutive patients, technical difficulty was observed in 90 cases (14.6%). The predictive model demonstrated moderate discriminatory capacity with an area under the receiver operating characteristic curve (AUC) of 0.711 (95%CI 0.643-0.780). Patients were stratified into three risk categories according to the difficulty risk scores: low-risk (＜0.1), medium-risk (0.1-0.25), and high-risk (≥0.25). The corresponding technical difficulty rates were 7.3%, 16.9%, and 51.6%, respectively.

Conclusion The prediction model for POEM difficulty built by our center shows good stability and discrimination, and has good clinical application value.