From identifying patients at-risk of developing heart conditions to improving diagnostic accuracy and mapping digital heart twins, cardiologists are increasingly developing innovative ways to harness the power of AI.
Becker’s asked four leaders how they integrate AI across the cardiovascular service lines at their hospitals and health systems.
Editor’s note: Responses have been lightly edited for clarity and length.
Question: Why is the field of cardiology uniquely positioned to “lead the way” with AI? How has your organization incorporated the use of AI?
Wayne Franklin, MD. Senior Vice President of the Children’s National Heart Center at Children’s National Hospital (Washington, D.C.): Cardiology is well suited for AI because of the high volume of data, including images, pharmacologic, interventional and surgical treatment options, patient outcomes and preventive medicine. Actually, cardiology was one of the first specialties to utilize an early form of AI with computer-assisted interpretation of electrocardiograms, started in the 1970s.
Fast-forward to today, and not only can AI help make an impact on the leading cause of death in the U.S. but cardiovascular medicine has such a vast amount of evidence already scientifically proven that the field is ready to take the next step by integrating cutting-edge AI technology.
Here at Children’s National, we have started to use AI in our pediatric cardiac intensive care unit, where we have utilized continuous surveillance monitoring to help predict when patients are likely to clinically deteriorate and early signs show that this prevents cardiac arrests.
With thoughtful and well-planned implementation, AI has the ability to transform not only cardiovascular disease but the entire spectrum of healthcare, leading to safer, faster and more personalized care for patients and families.
Blake Gardner, MD. Enterprise Senior Medical Director, and Kaley Graham, Executive Clinical Director of the Cardiovascular Clinical Program at Intermountain Health (Salt Lake City): Cardiology is uniquely positioned to lead the way with AI due to its data-rich environment and the critical nature of cardiovascular health.
The vast amount of data generated from imaging, diagnostics and patient monitoring provides a fertile ground for AI to enhance clinical decision-making and improve patient outcomes, two areas that Intermountain Health is well known for and is a national leader.
At Intermountain Health, we have incorporated AI into various aspects of our cardiology practice. For instance, AI-driven tools are used for administrative tasks, such as scheduling and resource allocation, as we work on developing AI tools for clinical care tools for real-time decision support, and screening processes to identify high-risk patients more efficiently.
Additionally, Intermountain is looking at AI capabilities for surgical planning and delivery of state-of-the-art care, ensuring precision and reducing complications. These advancements not only streamline operations but also elevate the quality of care we provide to our patients.
Joshua Lampert, MD. Director of Machine Learning at Mount Sinai Fuster Heart Hospital (New York City): The field of cardiology is an ideal setting to lead development and the incorporation of novel technologies into clinical practice. The diversity of cardiovascular disease presentations spans indolent disease, which may not manifest for decades, and imminently fatal scenarios that require immediate treatment to prevent death.
Furthermore, model classifications in cardiology can result in invasive procedures that carry unique risks while failure to identify relevant disease forfeits the opportunity to identify patients at risk of sudden cardiac death. This is a common challenge in medicine, where often we must extrapolate population-level risk estimates to individual patients. There is clear opportunity for improvement.
We recently published in NEJM AI the role of local model calibration to provide patient-level estimates of disease that better aligns with observed disease.
We have incorporated a variety of AI tools such as automated inbox letter response drafting, integrated patient status notification systems and an ECG algorithm toolkit. For atrial fibrillation catheter ablation procedures, we now incorporate AI techniques into the procedure to identify relevant electrical signals that could contribute to atrial fibrillation recurrence. Despite great optimism — which is warranted — even greater effort is needed to ensure responsible and ethical deployment of these novel solutions.
Venkatesh Murthy, MD, PhD. Associate Chief of Cardiology for Translational Research and Innovation at the University of Michigan Health Frankel Cardiovascular Center (Ann Arbor, Mich.): At University of Michigan, we are aggressively developing and implementing artificial intelligence into cardiovascular research and clinical programs.
U-M has been a leader in cardiovascular imaging. In this arena, we have pioneered the use of cardiac PET for diagnosis and management of known and suspected coronary artery disease. PET can precisely measure the blood flow to heart muscles without invasive testing. This enables measurement of the function of not only large coronary arteries, which can be seen with angiography, but also the microvasculature.
Microvascular disease is the earliest stage of CAD and particularly affects women and patients with diabetes, among other conditions. However, PET remains costly and only available in specialized centers, limiting diagnosis of this condition. We have developed a highly innovative transformer (the “T” in GPT) based foundation model for electrocardiograms, which has enabled us to build an accurate predictor for abnormal blood flow to the myocardium from either a resting or stress ECG.
We have also developed AI tools which enable us to precisely quantify inflammation in the hearts of patients with myocarditis. This enables us to track inflammation more accurately and to adjust doses of immune modulating drugs in patients with cardiac inflammation.
Together with our colleagues in radiology, we have also been using commercially developed AI tools to ensure lung cancers, which may be incidentally missed cardiac imaging studies, are not missed. These have been in use for more than four years and have had a major impact on our clinical programs.
We are continuing to explore additional applications for AI both in research and clinical practice and are committed to leading the world in this arena.
