ARTIFICIAL INTELLIGENCE IN CLINICAL DIAGNOSTICS FOR EARLY DETECTION OF CHRONIC DISEASES: A SYSTEMATIC REVIEW

Early detection of chronic diseases is critical for reducing morbidity and alleviating the overall healthcare burden. Artificial intelligence (AI) has emerged as a promising tool for enhancing diagnostic accuracy, risk prediction, and clinical decision support. This review synthesizes recent evidence on AI-driven diagnostic systems across diverse chronic diseases.

A systematic review was conducted following the PRISMA guidelines. Peer-reviewed English-language studies published between January 2020 and November 2025 were retrieved from PubMed/MEDLINE, Scopus, Web of Science, IEEE Xplore, and Embase.

Thirty-two studies from 13 countries were included, with most originating from China, India, and Saudi Arabia. The studies examined metabolic/cardiometabolic conditions (20 studies), musculoskeletal disorders (3), pulmonary diseases (3), cancer/hematological conditions (3), neurodegenerative diseases (1), and ophthalmologic/dental conditions (2). Hybrid AI models were the most commonly used overall (56%), especially in metabolic diseases, followed by machine learning (25%) and deep learning (19%). Validation approaches included k-fold cross-validation, 80/20 train-test splits, electronic health record (EHR)-based validation, and external validation. Across subgroups, predictive performance was high, with AUC ranging from 0.7467 to 1.0, accuracy from 77.08% to 99.97%, sensitivity from 77% to 100%, and specificity from 59.2% to 100%.

AI models, particularly hybrid approaches, demonstrate potential for early detection of chronic diseases by integrating laboratory, clinical, and imaging-based multimodal data. However, heterogeneity in datasets, retrospective study designs, limited external validation, and inconsistent reporting constrain generalizability. These findings highlight the need for prospective multicenter trials, standardized datasets, and improved methodological transparency to support clinical implementation.