Accuracy of artificial intelligence for dental age and sex estimation from panoramic radiographs: a systematic review

Introduction: Artificial intelligence is increasingly used to estimate chronological age and biological sex from dental radiographs for forensic and clinical purposes, but its accuracy across models and populations has not been consolidated. Methods: A systematic review was conducted following PRISMA 2020. Scopus, PubMed and Crossref were searched up to 2026 for studies reporting the accuracy of artificial intelligence models for age or sex estimation from panoramic radiographs, periapical radiographs or cone-beam computed tomography. Screening was assisted by artificial intelligence and verified by the authors. Results: Thirty primary studies published between 2024 and 2026 were included. For age estimation, the best models achieved a mean absolute error below one year, with errors ranging from 0.67 to 2.9 years across populations and datasets. For sex determination, accuracy ranged from 0.80 to 0.95, with convolutional and machine learning models performing comparably. Larger, more diverse datasets and explainable architectures were associated with the most robust estimates. Conclusions: Artificial intelligence estimates dental age with a mean error around one year and determines sex with high accuracy, but performance varies with dataset size, population and imaging modality. External, multi-centre validation remains limited, and the results should be applied with population-specific calibration in forensic and clinical settings.