Artificial intelligence for early diagnosis and risk prediction of periodontal-systemic interactions: Clinical utility and future directions.
Authors
Affiliations (3)
Affiliations (3)
- Department of Periodontology, Sri Sai College of Dental Surgery, Vikarabad 501102, Telangana, India. [email protected].
- Project Manager, Data Quality, Prime Healthcare Management Inc., Ontario, CA 91764, United States.
- Department of Periodontics, Kamineni Institute of Dental Sciences, Narketpally 508254, Telangana, India.
Abstract
Artificial intelligence (AI) is transforming healthcare by improving diagnostic accuracy and predictive analytics. Periodontal diseases are recognized as risk factors for systemic conditions, including type 2 diabetes mellitus, cardiovascular disease, Alzheimer's disease, polycystic ovary syndrome, thyroid dysfunction, and post-coronavirus disease 2019 complications. These conditions exhibit complex bidirectional interactions, underscoring the importance of early detection and risk stratification. Current diagnostic tools often fail to capture these interactions at an early stage, limiting timely intervention. This study hypothesizes that AI-driven approaches can significantly improve early diagnosis and risk prediction of periodontal-systemic interactions, enhancing clinical outcomes. To evaluate AI's role in diagnosing and predicting periodontal-systemic interactions in studies from 2010 to 2024. This systematic review followed PRISMA guidelines (2009) and included peer-reviewed articles from PubMed, Scopus, and Embase. Studies with large sample sizes (≥ 500 participants) were selected, focusing on AI models integrating multi-omics data and advanced imaging techniques such as cone beam computed tomography and magnetic resonance imaging. Machine learning models processed structured clinical data, deep learning models combined imaging and clinical data, and natural language processing models extracted insights from clinical notes. AI applications significantly enhanced diagnostic and predictive accuracy, reducing diagnostic time by 40% and improving predictive accuracy by 25% in periodontal patients with type 2 diabetes mellitus. Studies with sample sizes of 1000-1500 participants reported diagnostic accuracy improvements up to 92%, with specificity and sensitivity rates of 94% and 90%, respectively. Increasing sample sizes over the years reflected advancements in AI, data collection, and model training, reinforcing model reliability. AI's integration of multi-omics and imaging data has transformed early diagnosis and risk prediction in periodontal-systemic interactions, improving clinical outcomes and decision-making.