Predicting Cardiovascular Disease Using Machine Learning: A Systematic Review

Max Suasnabar Perez J.; Junior Sulca Gomez K.; Mary Lozada Flores R.; Bernardo Herrera K. · Proceedings of the LACCEI international Multi-conference for Engineering, Education and Technology (2025) · DOI: 10.18687/laccei2025.1.1.2322

Resumen

Artificial intelligence (AI) is currently used in various techniques for disease prediction, including cardiovascular disease (CVD). Therefore, this review aims to demonstrate the most effective AI technology methods for the prediction of CVD risk factors and diagnoses during early detection in humans, compared to traditional methods such as electrocardiogram (ECG) signal reporters, intensive care unit (ICU) review, and echocardiographic examinations. PICO and PRISMA methodologies were used for the search and selection of relevant documents. Along these lines, 397 documents were identified, including articles and systematic reviews in the following databases: Scopus and IEEE. According to the inclusion and exclusion criteria, 26 open access articles were selected. Where, the use of AI technological methods allowed analyzing and capturing predictive values on people with or without CVD problems in an effective way, achieving a range greater than 0.99. It should be noted that the representations were made with spreadsheet tools and the data manipulation, analysis and graph generation with Python libraries. Finally, it is concluded that the most effective AI technological methods for CVD prediction are based on machine learning (ML) techniques for predicting measurement values such as sensitivity, specificity, precision, F1 score and area under the curve (AUROC), with measurement indices between 0.97 and 0.99, compared to deep learning (DL) techniques, whose indices are between 0.88 and 0.90.

Instituciones: Universidad Tecnológica del Perú