ARTIFICIAL INTELLIGENCE IN EARLY DIAGNOSIS OF CARDIOVASCULAR DISEASES: A SYSTEMATIC LITERATURE REVIEW

Abstract

Cardiovascular diseases (CVDs) remain the leading cause of global morbidity and mortality, necessitating innovative approaches for early detection and risk stratification. Artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL), has emerged as a transformative tool in cardiovascular medicine, offering enhanced diagnostic accuracy and predictive performance. This systematic review aims to synthesize contemporary evidence on the application of AI technologies in the early diagnosis and risk prediction of cardiovascular diseases, highlighting methodological trends, clinical performance, and emerging innovations. The review was conducted in accordance with PRISMA guidelines. A comprehensive literature search was performed across PubMed, Scopus, Web of Science, IEEE Xplore, and Google Scholar for studies published between 2019 and 2025. Studies evaluating AI-based diagnostic or predictive models for cardiovascular conditions—including coronary artery disease, heart failure, arrhythmias, and multimodal risk prediction—were included. Data extraction encompassed study characteristics, AI methodologies, data modalities, validation strategies, and performance metrics such as accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). Methodological quality was assessed using standardized appraisal tools. A total of 85 studies were included in the final synthesis. Deep learning approaches, particularly convolutional neural networks for imaging and electrocardiogram analysis, demonstrated superior performance across multiple applications. Reported accuracy ranged from 82% to 97%, with AUC values frequently exceeding 0.90, especially in arrhythmia detection and coronary artery disease prediction. Multimodal AI models integrating electronic health records, imaging, and biomarker data achieved enhanced predictive capability compared to traditional risk scores. External validation was reported in 61% of studies, supporting generalizability across diverse populations.AI-driven diagnostic and predictive models consistently outperform conventional methods in cardiovascular disease detection and risk stratification. While significant progress has been achieved, challenges related to interpretability, external validation, clinical integration, and ethical implementation remain critical considerations. Continued advancement in explainable AI and multimodal data fusion is essential to translate these technologies into routine cardiovascular care.