Diagnostic Accuracy of Artificial Intelligence- Based Models in Periodontitis: A Systematic Review Meta-Analysis
Open Access
- 18 February 2025
- journal article
- review article
- Published by Uphills Publishers LLC in South Eastern European Journal of Public Health
- p. 896-910
Abstract
Introduction: Periodontitis is an irreversible disease caused by host-microbe interactions leading to the destruction of tooth-supporting structures. Its complex aetiology makes early diagnosis, staging, and treatment planning challenging but crucial to prevent disease progression. Artificial intelligence (AI) models, particularly Convolutional Neural Networks (CNNs), analyze complex variables, identify patterns, and make accurate predictions. Their use in periodontitis diagnosis can enhance diagnostic accuracy, reduce human error, and provide consistent results. Objectives: This review evaluates the current landscape of AI applications in diagnosing periodontitis, with a focus on CNN-based models used directly or through proxy indicators. Methods: A systematic literature search was conducted in PubMed, Web of Science, CINAHL, Embase, Cochrane Library, and ClinicalTrials.gov up to December 2019. Included studies assessed the diagnostic accuracy of AI models for periodontitis using cross-sectional, case-control, or cohort designs. Aggressive periodontitis cases were excluded. Risk of bias was assessed using the PROBAST tool, and results are presented as a narrative synthesis. Results: AI models, particularly CNNs, demonstrated high diagnostic accuracy for periodontal bone loss using radiographic evidence, often surpassing expert performance. Models like DenseNet and U-Net excelled in segmentation and classification. Challenges included poor image quality, imbalanced datasets, and reliance on proxy indicators, highlighting the need for multivariable approaches. Discussion: AI shows promise in standardizing and scaling periodontitis diagnosis, addressing manpower shortages, and improving outcomes. However, future research should focus on integrating multivariable diagnostic approaches and refining model interpretability for clinical applicability.Keywords
This publication has 37 references indexed in Scilit:
- Exploring the role of Convolutional Neural Networks (CNN) in dental radiography segmentation: A comprehensive Systematic Literature ReviewEngineering Applications of Artificial Intelligence, 2024
- Deep learning for the early identification of periodontitis: a retrospective, multicentre studyClinical Radiology, 2023
- Automated periodontitis bone loss diagnosis in panoramic radiographs using a bespoke two-stage detectorComputers in Biology and Medicine, 2022
- Application of deep machine learning for the radiographic diagnosis of periodontitisClinical Oral Investigations, 2022
- Automating Periodontal bone loss measurement via dental landmark localisationInternational Journal of Computer Assisted Radiology and Surgery, 2021
- Local and systemic mechanisms linking periodontal disease and inflammatory comorbiditiesNature Reviews Immunology, 2021
- Convolutional neural networks in medical image understanding: a surveyEvolutionary Intelligence, 2021
- Automatic and Interpretable Model for Periodontitis Diagnosis in Panoramic Radiographs2020
- Developing a periodontal disease antibody array for the prediction of severe periodontal disease using machine learning classifiersJournal of Periodontology, 2019
- Deep Learning for the Radiographic Detection of Periodontal Bone LossScientific Reports, 2019