BACKGROUND AND AIMS: We aimed to establish a predictive model and develop a simple risk-scoring system (Zhongshan POEM Score) to help clinicians to characterize high-risk patients for clinical failure after peroral endoscopic myotomy (POEM). METHODS: A total of 1538 patients with achalasia treated with POEM with available follow-up data were included in this study and were randomly classified to the training cohort (n = 769) or internal validation cohort (n = 769). A risk-scoring system was developed using multivariate Cox regression analysis in the training cohort. The system was then internally validated by survival analysis in the validation cohort. RESULTS: During a median follow-up time of 42 months, 109 patients had clinical failure. In the training stage, 3 risk factors for clinical failure were weighted with point values: previous treatment (2 points), intraprocedural mucosal injury (2 points for type I and 6 points for type II), and clinical reflux (3 points). The patients were categorized into low-risk and high-risk groups. In the validation stage, Kaplan-Meier curves differed significantly between the 2 groups. Patients in the high-risk group had a significantly higher risk of clinical failure than those in the low-risk group (hazard ratio, 3.99; 95% confidence interval, 2.31-6.91; P < .001). Satisfactory discrimination and calibration were shown. CONCLUSIONS: This risk-scoring system demonstrated good performance in predicting clinical failure in patients who underwent POEM.
BACKGROUND AND AIMS: We aimed to establish a predictive model and develop a simple risk-scoring system (Zhongshan POEM Score) to help clinicians to characterize high-risk patients for clinical failure after peroral endoscopic myotomy (POEM). METHODS: A total of 1538 patients with achalasia treated with POEM with available follow-up data were included in this study and were randomly classified to the training cohort (n = 769) or internal validation cohort (n = 769). A risk-scoring system was developed using multivariate Cox regression analysis in the training cohort. The system was then internally validated by survival analysis in the validation cohort. RESULTS: During a median follow-up time of 42 months, 109 patients had clinical failure. In the training stage, 3 risk factors for clinical failure were weighted with point values: previous treatment (2 points), intraprocedural mucosal injury (2 points for type I and 6 points for type II), and clinical reflux (3 points). The patients were categorized into low-risk and high-risk groups. In the validation stage, Kaplan-Meier curves differed significantly between the 2 groups. Patients in the high-risk group had a significantly higher risk of clinical failure than those in the low-risk group (hazard ratio, 3.99; 95% confidence interval, 2.31-6.91; P < .001). Satisfactory discrimination and calibration were shown. CONCLUSIONS: This risk-scoring system demonstrated good performance in predicting clinical failure in patients who underwent POEM.