Effectiveness of Artificial Intelligence in Endodontic Diagnosis and Treatment Evaluation: A Systematic Review.

Artificial intelligence (AI) has emerged as a transformative tool in endodontics, offering potential to enhance diagnostic accuracy, treatment evaluation, and clinical decision-making. This systematic review aimed to assess the effectiveness of AI models in diagnosing endodontic conditions and evaluating treatment outcomes compared with human clinicians. Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines, a comprehensive literature search was conducted across PubMed, ScienceDirect, Web of Science, Cochrane Library, and Google Scholar up to June 2025. Ten studies met the inclusion criteria, encompassing diverse AI models such as convolutional neural networks (CNNs), U-Net architectures, YOLOv5, and ChatGPT-4. These were applied to tasks including periapical lesion detection, root canal filling evaluation, fractured instrument identification, and caries diagnosis. Across studies, AI systems achieved diagnostic accuracy ranging from 75% to 99%, with sensitivity and specificity frequently exceeding 80%. DenseNet201 achieved the highest performance for fractured instrument detection (area under the curve (AUC) = 0.900; Matthews correlation coefficient (MCC) = 0.810), while ChatGPT-4 demonstrated superior diagnostic accuracy (99%) compared with dental students (77-80%). Cone-beam computed tomography (CBCT)-based models consistently outperformed those using panoramic or periapical images. Despite high accuracy, variability in methodologies, dataset sizes, and outcome metrics limited quantitative synthesis; therefore, a meta-analysis was not feasible. Overall, AI demonstrated comparable or superior performance to clinicians, offering advantages in speed, reproducibility, and objectivity. However, limitations such as dataset heterogeneity, lack of external validation, and challenges in detecting subtle features underscore the need for further large-scale, multicenter studies. AI integration into routine endodontic practice shows strong potential but requires continued refinement and validation for reliable clinical adoption.