Researchers at the University of Kentucky are teaming up with researchers at Michigan State University (MSU) to develop cutting-edge simulations of heart disease progression.
The National Science Foundation (NSF) awarded UK and MSU a four-year, $1.2 million collaborative grant titled "SCH: Machine‑Learning Enhanced Computational Models of Cardiac Pathophysiology."
The project is led by principal investigator (PI) Jonathan Wenk, Ph.D., Jack and Linda Gill Professor in the UK Stanley and Karen Pigman College of Engineering, and co‑PI Kenneth Campbell, Ph.D., professor of cardiovascular medicine and physiology in the UK College of Medicine, alongside Lik Chuan Lee, Ph.D., professor of mechanical engineering at MSU College of Engineering.
Cardiovascular disease is the leading cause of death in the United States, and fibrosis - excessive buildup of extracellular matrix fibers in organ tissue - is a key contributor to heart failure. Current projections from the Heart Failure Society of America indicate approximately one in four people will develop heart failure in their lifetime.
This NSF-funded project brings together expertise from multiple fields - engineering, computer science, applied mathematics and physiology - to build advanced computational models of the heart capable of simulating the development of fibrosis, and the adverse changes to cardiac structure and function that result. By integrating machine learning and artificial intelligence, the work aims to build fundamental understanding of heart disease progression and could aid in the evaluation of potential treatment strategies.
"The idea is to create a patient-specific model, which is truly tailored to their anatomy, tissue structure and genetics, to test different therapeutics to see which one they will respond to better," Wenk said.
At its core, the research aims to create a multiscale finite‑element framework that includes network models (myocardial perfusion), agent‑based models (myocardial fibrosis) and timescale separation schemes (myofiber growth). This hybrid modeling approach will allow simulation of interactions between system‑level biology and molecular mechanisms across scales.
A major innovation will be the development of physics‑informed neural networks (PINNs) and other machine learning techniques to replicate the multiscale model in computationally efficient form. This will enable faster predictions of structural and functional changes arising in ischemic and nonischemic heart disease.
"Dr. Lee and I both have expertise in finite element modeling, but on this grant, we are leveraging his expertise in machine learning and PINNs," Wenk explained.
By combining data‑driven machine learning with mechanistic modeling, the team hopes to enable accurate prediction of how fibrotic disease progresses in individual patients and how they may respond to potential treatments. If successful, this could open new avenues toward personalized diagnostics and therapy in cardiovascular medicine.
"The model will hopefully tell us whether the treatment will prevent adverse remodeling that leads to heart failure," Wenk said. "If we can maintain close to normal function, the patient should have a better quality of life."
The goals of the project support education and future research. The team plans to introduce cardiac engineering examples into computational mechanics curricula and develop open-source software tools and databases based on their modeling framework.
This grant is a continuation of collaborative work between Wenk, Campbell and Lee. The team had previously secured two National Institutes of Health grants focused on multiscale modeling of the heart. Wenk has collaborated with both his MSU colleague, Lee, and his UK colleague, Campbell, for more than a decade.
Research reported in this publication was supported by the U.S. National Science Foundation under Award No. 2406028. The opinions, findings, and conclusions or recommendations expressed are those of the author(s) and do not necessarily reflect the views of the U.S. National Science Foundation.
