SHM Converge Scientific Abstract ePoster Gallery-Accuracy of Artificial Intelligence and Machine Learning in Cardiovascular Diseases: A Systematic Review

Pdf Summary

This systematic review evaluated how accurately artificial intelligence (AI) and machine learning (ML) diagnose cardiovascular diseases (CVDs), a major global health burden affecting an estimated 523 million people in 2019. Using PRISMA and MOOSE guidelines, the authors searched PubMed for studies from the past 10 years and included observational studies and clinical trials that reported diagnostic accuracy measures such as sensitivity and specificity. Twenty-one studies met the criteria.<br /><br />Overall, AI and ML showed strong potential in cardiovascular diagnosis. The models performed especially well in cardiac imaging, including cardiac MRI, where they achieved accuracy comparable to human experts in tasks such as heart segmentation and functional assessment. Deep learning systems also automated heart function analysis and improved efficiency in clinical workflows. In some applications, AI improved 3D imaging quality and reduced scan times. Diagnostic models for hypertrophic cardiomyopathy and coronary artery disease demonstrated high sensitivity and specificity, suggesting that AI may reduce reliance on invasive procedures or contrast agents.<br /><br />However, the review also identified important limitations. Many studies had small sample sizes, were conducted at single centers, and used heterogeneous datasets and imaging protocols. These issues limit generalizability and make widespread adoption difficult. The authors emphasize that standardization, larger multi-center studies, and improved algorithm validation are needed before AI can be fully integrated into routine cardiovascular care.<br /><br />In conclusion, AI and ML are promising tools for improving the accuracy, speed, and efficiency of CVD diagnosis, particularly in cardiac imaging, but further work is needed to address data and methodological challenges.

Asset Subtitle

Urvish Patel

Meta Tag

Author List Arjun Gandhi, Bhavya Raja Gopal Yarlagadda, Keerthi Sripathi, Mahmoud Aboukhreis, Parth Shah, Rutvij Patel, Sharanya Rajgopal, Sudha Dirisanala, Tarun Parvataneni, Urvish Patel, Venkata Harshabhargav

Category Research

Concept Cardiovascular Disease

Concept Artificial Intelligence

Concept Machine Learning

Concept Sensitivity

Concept Specificity

Distinguished Non-Finalist

Presenter Organization Department of Public Health, Icahn School of Medicine at Mount Sinai, New York, NY

Presenting Author Urvish Patel

Track Technology in Hospital Medicine

Keywords

artificial intelligence

machine learning

cardiovascular disease

diagnostic accuracy

cardiac imaging

cardiac MRI

deep learning

hypertrophic cardiomyopathy

coronary artery disease

systematic review

Cardiovascular Disease

Artificial Intelligence

Machine Learning

Sensitivity

Specificity