Application of artificial intelligence in oral potentially malignant disorders: current opinions and future barriers.

Abstract

Oral potentially malignant disorders (OPMDs) refer to oral mucosal disorders with an increased risk of malignancy, primarily oral squamous cell carcinoma (OSCC), especially in South and Southeast Asia. Since not all patients with OPMDs develop oral cancer, accurate early detection and diagnosis of malignant transformation are critically important for clinicians to determine the optimal therapeutic approach. Therefore, distinguishing OPMDs from early-stage OSCC is an increasing challenge in the clinic. Artificial intelligence (AI) technology has recently been shown to quickly identify high-risk conditions/lesions for screening oral cancer early. Moreover, the AI algorithm can also be used to determine the prognosis of OPMDs. In this review, we systematically summarize the medical records, oral images, pathological examinations, biomarkers, omics data and other aspects of the main outcomes of AI applied to address OPMDs-related issues. Furthermore, we discuss automated diagnostic systems and risk prediction tools for malignant transformation with pleasant outcomes and the potential to ultimately assist clinicians. Finally, we introduce the current challenges and barriers to AI in OPMDs on the premise that more advanced AI models and larger datasets will lead to the use of AI models in OPMDs.