APPLICATION OF ARTIFICIAL INTELLIGENCE IN DENTISTRY: LITERATURE REVIEW

Background:Artificial intelligence (AI) has rapidly emerged as a transformative technology in dentistry, enabling advanced data analysis, image interpretation, and clinical decision support. The integration of machine learning and deep learning into dental practice has significantly improved diagnostic accuracy and treatment planning. However, concerns regarding data quality, methodological variability, and risk of bias continue to limit its widespread clinical adoption. 

Objective:This study aimed to systematically evaluate the applications of AI in dentistry, assess its diagnostic performance across different specialties, and critically analyze the quality and risk of bias of the available evidence using PRISMA guidelines. 

Materials and Methods:A literature review was conducted following PRISMA 2020 guidelines. Electronic databases including PubMed, Scopus, Web of Science, and IEEE Xplore were searched for studies published between January 2018 and January 2025. Eligible studies were original research articles evaluating AI applications in dentistry with reported quantitative outcomes. A total of 412 records were identified, Full-text assessment was conducted for 112 articles, of which 40 met the inclusion criteria after screening. Data extraction included study design, AI model type, dataset characteristics, and performance metrics. Risk of bias was assessed using the PROBAST tool. 

Results:The majority of studies (52%) focused on dental radiology, followed by orthodontics (18%) and implant dentistry (15%). Deep learning models, particularly convolutional neural networks, were used in 85% of studies. AI systems demonstrated high diagnostic performance, with accuracy ranging from 82% to 95%, sensitivity from 80% to 93%, and specificity from 85% to 96%. The highest accuracy was observed in caries detection, periodontal bone loss assessment, and CBCT analysis. However, approximately 78% of studies exhibited moderate to high risk of bias, primarily due to small datasets and lack of external validation. 

Conclusion:AI shows significant potential to enhance diagnostic accuracy and optimize clinical workflows in dentistry. Despite promising results, challenges related to data heterogeneity, methodological limitations, and ethical concerns must be addressedf or its successful clinical adoption. Future research should focus on large-scale validation and standardized reporting to facilitate clinical integration.