Applications of artificial intelligence in restorative dentistry: a scoping review

Abstract

Objective: Artificial intelligence (AI) applications in restorative dentistry have remarkably increased in the past 5 years. This review outlines the applications, promises, and limitations of AI in the most performed procedures in restorative dentistry.

Method and materials: An electronic search was performed in four databases: MEDLINE/PubMed, Embase, Web of Science, and Scopus. The search included articles published in English language without date restriction. Two independent reviewers assessed the eligibility of the studies and performed data extraction. Any discrepancy was resolved by the consensus of a third reviewer.

Results: A total of 33 studies were included in this review. For AI applications in restorative dentistry, the included studies were classified into three main groups: (1) diagnosis, detection, and prediction of the disease, (2) detection and prediction of the longevity of dental restorations, and (3) teeth detection and treatments. For each study, the AI model, type of dataset, sample size, and main results (accuracy, precision, sensitivity, and specificity) were reported.

Conclusions: AI systems are promising as advantageous aids for diagnosis, prediction, and treatment in dentistry, with a high degree of accuracy. Despite the AI promises, several limitations are still unresolved and must be addressed to bridge the gap between technology and clinical applications.

Keywords: artificial intelligence; dental caries; dental prostheses; dental restoration failure; machine learning.