WOFAPS 2025 8th World Congress of Pediatric Surgery

Mohammed J Aboud ¹, Manal M Kadhim ², Mustafa J Radif ³

¹ Ministry of Health
² College of Medicine University of Al Qadisiya
³ College of Computer Science and Information Technology University of Al Qadisiya

Background: Anorectal malformations (ARMs) present significant challenges in predicting long-term continence and psychosocial outcomes. Existing classification systems offer limited prognostic value. We developed an AI-integrated decision support model that incorporates validated scoring systems to forecast functional severity and quality of life in ARM patients.

Objective: To validate a novel artificial intelligence (AI) tool integrating the Sacral Anomaly Incontinence Severity (SAIS) and Anorectal Function Quality of Life (AFQoL) scores, enabling early and individualized outcome prediction in pediatric ARM management.

Methods: A retrospective cohort of 157 surgically managed ARM patients was analyzed. Sacral morphology was classified via convolutional neural networks (CNNs) with 92% accuracy. Clinical inputs, ARM subtype, sacral vertebral count, continence history, and sex, were processed through machine learning models to predict Kelly continence scores and AFQoL burden. A natural language processing (NLP) module extracted continence descriptors from clinical notes. Output scores guided surgical risk stratification and follow-up.

Results: AI-predicted continence outcomes achieved an AUC of 0.91, and AFQoL burden prediction reached an AUC of 0.90. SAIS score correlated strongly with Kelly continence scores (r = -0.79), and Kelly scores with AFQoL (r = -0.84). The AI tool stratified patients into high, moderate, and low-risk tiers and enabled an algorithm-driven clinical workflow for decision-making, postoperative follow-up, and telemonitoring.

Conclusion: This is one of the first studies to integrate deep learning, validated clinical scores, and NLP into a cohesive ARM decision support system. The model facilitates real-time risk prediction, supports surgical planning, and enhances longitudinal care, especially in low-resource or high-burden settings. Its predictive accuracy and scalability position it as a transformative tool in the era of precision pediatric surgery.