ARTIFICIAL INTELLIGENCE IN PEDIATRIC DIAGNOSTIC IMAGING: A SYSTEMATIC REVIEW OF MACHINE LEARNING APPLICATIONS FOR EARLY DISEASE DETECTION

Abstract

Background: Artificial intelligence (AI) and machine learning (ML) technologies are rapidly transforming pediatric diagnostic imaging. Objective: This study systematically reviews the performance of AI-based diagnostic tools in pediatric radiology across multiple organ systems. Methods: A systematic search of PubMed, Scopus, and Web of Science databases was conducted for studies published between 2020 and 2025. Studies comparing AI diagnostic performance against pediatric radiologists were included. Results: Forty-two studies met inclusion criteria. AI models achieved mean diagnostic accuracy of 91.3% (95% CI: 88.7–93.9%) for pediatric chest pathologies, 89.6% for musculoskeletal abnormalities, and 87.4% for neurological disorders. Deep learning models consistently outperformed traditional ML approaches. Conclusion: AI demonstrates substantial promise in augmenting pediatric diagnostic imaging workflows, though standardization of datasets and prospective clinical validation remain critical priorities.