SCH: Transfer Regression to Enable Cross-Domain Cardiovascular Event Prediction

NIH RePORTER · NIH · R01 · $299,604 · view on reporter.nih.gov ↗

Abstract

SCH: Transfer Regression to Enable Cross-Domain Cardiovascular Event Prediction This project proposes fundamental novelties (e.g., transfer volume regression) in computer science, data science, and biomedical engineering to address the critical health challenge of lacking a standardized clinical risk prediction for major cardiovascular events (CVe)—defined as stroke, heart attack, heart failure (HF), and death. Our proposal includes a pioneering transfer regression learning method, combined with several novel machine learning and data science techniques, to develop the first smart, standardized, fairness-aware, and user-friendly CVe prediction model. This model leverages commonly used, low-cost (sometimes free), safe, and quick screening programs (featuring low radiation and no need for contrast agents), aiming to significantly enhance clinical outcomes. The efficacy and robustness of this model are to be validated using four large and diverse datasets, exemplifying a significant advancement in Smart Health and Biomedical Research in the era of Artificial Intelligence and Advanced Data Science (SCH). RELEVANCE (See instructions): We will develop and validate transfer volume regression to enable quantitative clinical risk predictions for major cardiovascular events (defined as stroke, heart attack, heart failure, and death in this proposal) to guide preventive therapy.

Key facts

NIH application ID
11062600
Project number
1R01HL177813-01
Recipient
CASE WESTERN RESERVE UNIVERSITY
Principal Investigator
SHUO LI
Activity code
R01
Funding institute
NIH
Fiscal year
2024
Award amount
$299,604
Award type
1
Project period
2024-09-01 → 2028-08-31