ACCURACY AND APPLICABILITY OF ARTIFICIAL INTELLIGENCE AS AN AUXILIARY TOOL IN DENTAL DIAGNOSIS

Abstract

The incorporation of Artificial Intelligence in Dentistry has reshaped image-based diagnosis, particularly in radiology. Deep learning models have been applied to the detection of caries, periapical lesions, periodontal bone loss, cephalometric alterations, and oral neoplasms, aiming to improve diagnostic accuracy and reduce clinical variability. This study consists of a scoping review guided by the PICo strategy, with a search conducted in the MDPI database between 2010 and 2025, using MeSH and DeCS descriptors related to Artificial Intelligence and dental diagnosis. After screening according to inclusion and exclusion criteria and applying the PRISMA protocol, 18 studies were included in the final analysis. The findings demonstrate sensitivity and specificity often comparable to those of specialists, particularly in interproximal, panoramic, and CBCT radiographs. However, methodological limitations, lack of external validation, and variability among training datasets remain concerns. It is concluded that Artificial Intelligence acts as a complementary tool, enhancing diagnostic accuracy and optimizing clinical workflow without replacing the dentist’s professional judgment.