ARTIFICIAL INTELLIGENCE IN DENTAL CARIES DIAGNOSIS: A NARRATIVE REVIEW

Georgiana-Andreea Frumuzache, Antonia -Theodora Vrabie, Sorin Andrian, Gianina Iovan, Irina Nica, Alice-Teodora Rotaru-Costin, Simona Stoleriu

ABSTRACT

Background: Dental caries is one of the most prevalent chronic disease worldwide. It is essential to use fast and accurate methods for caries detection. In this context, increased interest was recorded in using artificial intelligence (AI) deep learning techniques for dental images analysis in order to detect and evaluate dental caries. Objective: The aim of this narrative review is to evaluate the applications of artificial intelligence in the diagnosis of dental caries and to highlight the limitations of these technologies. Methods: Analysis of the scientific literature was conducted by consulting the PubMed/MEDLINE, Scopus, Web of Science and ScienceDirect databases. Articles investigating the use of AI in the detection and classification of dental caries published from 2021 to 2026 were included. Results: Studies analysis showed that artificial intelligence–based systems, especially based on convolutional neural networks (CNN), can establish caries diagnosis very simmilar to experienced clinicians. In the studies the main investigated data source was bitewing radiography. However, factors such as the annotation method, image quality, and the size and diversity of datasets have an impact on model performance. Regarding external validation, most studies rely on retrospective datasets. Conclusions: Artificial intelligence is a promising tool for the diagnosis of dental caries, improving the efficiency and accuracy of this process. However, the validation in clinical practice requires prospective studies, standardization of methodologies and it addressing ethical and regulatory aspects.

DOI : 10.62610/RJOR.2025.1.18.10