Artificial Intelligence in Endodontics: A Systematic review of Diagnostic Applications and Clinical Performance

Background: The integration of Artificial Intelligence (AI) in endodontics has gained substantial attention in recent years, offering innovative solutions for diagnosis, treatment planning, and clinical decision-making. Given the increasing complexity of endodontic procedures and the limitations of human analysis in imaging, AI systems present a promising alternative to support dental professionals. Objective: This systematic literature review aimed to synthesize current scientific evidence regarding the accuracy, performance, and applicability of AI models in endodontic diagnosis and treatment, with a focus on periapical lesion detection, root morphology analysis, and decision-making support. Methods: A comprehensive search was conducted in databases such as PubMed, Scopus, Web of Science, Google Scholar, and ScienceDirect for articles published between 2018 and 2024. The PRISMA methodology was followed for study selection, and 45 relevant articles were included based on predefined inclusion and exclusion criteria. Results: The findings reveal that AI algorithms, especially those based on deep learning and convolutional neural networks, demonstrate high accuracy in detecting periapical lesions,identifying root canal morphology, and predicting treatment outcomes. Most studies reported performance metrics comparable or superior to those of human examiners. Despite these advances, challenges remain in data standardization, external validation, and ethical regulation. Conclusion: AI has significant potential to enhance diagnostic precision and clinical outcomes in endodontics. However, further multicenter trials, real-world validations, and ethical frameworks are needed to support its routine clinical implementation