Diagnostic accuracy of artificial intelligence for detecting oral cancer and potentially malignant disorders on clinical photographs

Introduction: Artificial intelligence applied to clinical photographs could enable early detection of oral cancer and potentially malignant disorders, but its diagnostic accuracy has not been consolidated. Methods: A systematic review was conducted following PRISMA 2020. Scopus, PubMed and Crossref were searched up to 2026 for diagnostic accuracy studies of artificial intelligence detecting oral cancer or oral potentially malignant disorders from clinical photographs against a histologic or expert reference standard. Screening was assisted by artificial intelligence and verified by the authors. Results: Fourteen primary studies published in 2025 and 2026 were included. Reported accuracy was high under internal validation but varied by model, image source and lesion type, with convolutional and object-detection models reaching near-ceiling accuracy on curated datasets and general-purpose vision-language models performing less consistently. External validation was reported in a minority of studies. Conclusions: Artificial intelligence detects oral cancer and potentially malignant disorders from clinical photographs with high accuracy under controlled conditions, but the evidence is young, dominated by internal validation and uneven across models. Prospective external validation against biopsy-confirmed diagnoses is required before these tools can support screening.