Lexington, Kentucky (July 22, 2025) – University of Kentucky researchers are working with researchers at Michigan State University (MSU) to develop cutting-edge simulations of the progression of heart disease.
The National Science Foundation (NSF) awarded the UK and MSU $1.2 million over four years A joint grant entitled “Sch: Strengthen the computational model of cardiac pathology in Machine-Learning.”
The project will be led by Dr. Jonathan Wenk, Professors Jack and Linda Gill at Stanley and Karen Pigman University in the UK, Jonathan Wenk, Ph.D., and Co-PI Kenneth Campbell, Co-Vasis Medicine and Physiology, Co-Vasis Medicine and Physiology, UK School of Medicine.
Cardiovascular disease is the main cause of death in the United States, and fibrosis – the excessive accumulation of extracellular matrix fibers in organ tissues – is a key contributor to heart failure. Current predictions from the American Heart Failure Association show that about one in four people develop heart failure in their lifetime.
This NSF-funded project brings together expertise in multiple fields of engineering, computer science, applied mathematics and physiology to build a sophisticated computational model of the heart that can simulate the development of fibrosis and the resulting adverse changes in cardiac structure and function. By integrating machine learning and artificial intelligence, this task aims to build a fundamental understanding of the progression of heart disease and can assist in the assessment of potential treatment strategies.
“The idea is to create patient-specific models that are really tailored to anatomy, histology and genetics, and test the various treatments to see which treatments respond better,” Wenk said.
This study aims to create a multiscale finite element framework that includes network models (myocardial perfusion), agent-based models (myocardial fibrosis), and timescale separation schemes (myofibril growth). This hybrid modeling approach allows for simulation of the interaction between system-level biology and molecular mechanisms across scale.
The major innovation is the development of physics-based neural networks (PINNs) and other machine learning technologies, replicating multi-scale models in computationally efficient ways. This allows for faster prediction of structural and functional changes that occur in ischemic and non-ischemic heart disease.
“Dr. Lee and I both have expertise in finite element modeling, but we're leveraging his expertise in machine learning and PINN for this grant,” Wenk explained.
By combining data-driven machine learning and mechanical modeling, the team hopes to be able to accurately predict how fibrotic disease progresses in individual patients and how they will respond to potential treatments. If successful, this can pave a new pathway towards personalized diagnosis and treatment of cardiovascular medicine.
“This model hopes to tell us whether treatment will prevent adverse modifications that lead to heart failure,” Wenk said. “If patients can maintain close to normal functioning, they need to have a better quality of life.”
The goals of the project support education and future research. The team will introduce cardiac engineering examples into the computational mechanics curriculum and develop open source software tools and databases based on the modeling framework.
The grant is a continuation of collaboration between Wenk, Campbell and Lee. The team previously secured grants from two National Institutes of Health focused on multi-scale modeling of the heart. Wenk has worked with both his MSU colleague Lee and his British colleague Campbell for over a decade.
The studies reported in this publication are supported by the National Science Foundation under award number 2406028.
