Artificial Intelligence–Assisted Detection Of Caries And Pulpal Pathology To Support Restorative And Endodontic Treatment Planning In Children
Pediatric dentistry; restorative dentistry; endodontic diagnosis; artificial intelligence; panoramic radiography; treatment planning
AuthorsAbstractEarly identification of caries and pulpal pathology is critical for effective restorative and endodontic treatment planning in pediatric patients. This study evaluated a pediatric-focused artificial intelligence system designed to assist in the detection and localization of common dental pathologies on panoramic radiographs. A curated dataset of seventy-two de-identified pediatric radiographs containing expert-annotated lesions across six diagnostic categories was analyzed. Images were standardized and partitioned at the patient level to ensure unbiased evaluation. The system automatically identified clinically relevant pathologies and generated interpretable outputs to support diagnostic decision-making. The dataset demonstrated substantial class imbalance, with caries representing the majority of lesions, and lesion size varied considerably across categories. The model achieved high diagnostic specificity across all lesion types and strong discriminatory performance for caries, developmental anomalies, and inflammatory pathology. Visualization analysis confirmed that predictions were derived from anatomically meaningful regions. These findings indicate that artificial intelligence may serve as a supportive tool for improving diagnostic consistency and enhancing restorative and endodontic treatment planning in pediatric dentistry. Larger multi-centre clinical studies are required before routine implementation.
INTRODUCTIONPanoramic radiography is a critical diagnostic instrument in the field of pediatric dentistry as it provides the dentist with a complete picture of the dentition, the supporting structures, and the developmental pattern using a single image. Restoratively and endodontically, panoramic radiographs are highly important in evaluation of the caries level, pulpal level, periapical level, and developmental anomalies with the direct impact on children treatment planning. Nonetheless, the situation with panoramic radiography of children is distinctly difficult to interpret because of transitional dentition, overlapping structures of the organs, and rapid development that can make the process of recognizing and classifying dental pathology more complicated. This diagnostic uncertainty can have a negative impact on restorative decisions and when endodontic care is administered. Deep learning (artificial intelligence, AI) has become one of the most promising instruments to overcome these issues, improving the consistency of diagnosis and facilitating clinical decisionmaking in pediatrics restorative care. Recent research has shown that AI-based systems have the potential to enhance the process of caries, developmental defects, and inflammatory lesion detection, thus enhancing preventive and therapeutic measures in young patients¹. The development of AI in dental imaging has followed the same pattern as the development of medical imaging in general. •••••••••••••••••••••••••••••••• ejprd.org - Published by Dennis Barber Journals. Barber Ltd. All rights reserved
EJPRD
Copyright ©2025 by Dennis