Artificial intelligence in early diagnosis of somatic and dental diseases: current status, evidence, and integration challenges

Abstract

The link between oral health and overall health has been widely epidemiologically confirmed, and artificial intelligence (AI) technologies provide a practical bridge for the early detection of systemic diseases in routine dental settings. This paper systematically examines the integration of AI into the early diagnosis of somatic and dental diseases. Using data from approximately 320 million dental radiographic images generated annually in the United States and containing markers of systemic diseases, the diagnostic performance of AI in detecting carotid artery calcification, osteoporosis, oral manifestations of diabetes, and precancerous oral lesions is analyzed. Existing evidence shows that FDA-approved dental AI systems achieve accuracy above 90% in detecting various pathologies, deep learning models demonstrate sensitivity of approximately 90% in detecting carotid artery calcification, and accuracy in oral cancer screening exceeds 80% in most cases. However, data fragmentation, lack of external validation, and disunity between medical and dental education remain major barriers to integration. The clinical value of AI lies in augmenting, not replacing, clinical judgment; its successful implementation requires interoperability of electronic health records, clinical trust in explainable AI systems, and support for interdisciplinary educational reform.