Significance of criteria for an artificial intelligence–based platform in postoperative rehabilitation of children with anorectal malformations

Abstract

Background. Anorectal malformations (ARM) represent one of the most complex congenital anomalies in pediatric surgery and require prolonged, structured rehabilitation even after technically successful surgical correction. Conventional postoperative rehabilitation strategies are often based on subjective clinical assessment and lack reliable tools for individualized outcome prediction. Aim. To develop and substantiate criteria for an artificial intelligence (AI)–based rehabilitation platform designed to individualize postoperative rehabilitation in children with anorectal malformations. Objectives. To analyze clinical and instrumental parameters influencing postoperative functional recovery, identify prognostically significant features suitable for AI modeling, and evaluate the effectiveness of predictive tools for rehabilitation outcomes. Results. The proposed AI-based platform enabled the formation of individualized functional profiles for each patient by integrating clinical, instrumental, and dynamic follow-up data. Machine learning models demonstrated a predictive accuracy ranging from 82% to 91% for rehabilitation outcomes. Early identification of functional deterioration and postoperative complications was achieved, allowing timely modification of rehabilitation strategies and resulting in a reduction in the overall rehabilitation duration. Conclusion. Integration of artificial intelligence into postoperative rehabilitation management significantly enhances functional recovery, optimizes clinical decision-making, and improves long-term outcomes in children with anorectal malformations.