Discussion with Mayo Clinic’s Dr. Demilade Adedinsewo: Exploring the Future of Artificial Intelligence in Cardiovascular Medicine

Exploring the Future of Artificial Intelligence in Cardiovascular Medicine: A Discussion with Mayo Clinic’s Dr. Demilade Adedinsewo

February 2, 2022 – By Amrika Ramjewan, Principal Business Strategist – Mayo Clinic Innovation Exchange

Artificial intelligence (AI) is transforming the practice of medicine — with cardiology at the forefront of this paradigm shift. Recent studies have shown that AI-enhanced electrocardiograms (ECGs) can detect a variety of heart conditions. These include heart rhythm disorders such as atrial fibrillation, long QT syndrome, heart muscle disease (cardiomyopathy), and heart failure. By helping to identify novel risk factors and relationships in data, AI algorithms have tremendous potential to save lives and transform approaches to diagnosis, prevention, and treatment of cardiovascular diseases.

Demilade A. Adedinsewo, M.B., Ch.B., M.P.H., is a non-invasive cardiologist, assistant professor of medicine, Women’s Health Scholar, and director of research for the Women’s Heart Clinic and Cardiovascular Disease Fellowship program at Mayo Clinic in Florida. With deep interest in the application of digital tools in cardiovascular disease management, Dr. Adedinsewo is advancing the integration of AI-enabled technologies into the clinical practice at the intersection of women’s heart health, cardiovascular disease prevention, and cardiovascular health disparities.

Dr. Adedinsewo earned her medical degree from Obafemi Awolowo University before completing a research fellowship in Epidemiology at the Centers for Disease Control and Prevention, sponsored by the Oak Ridge Institute for Science and Education. She completed her residency at Morehouse School of Medicine, and a clinical fellowship in Cardiology at Mayo Clinic — where she also served as chief fellow in the Department of Cardiovascular Medicine. She has completed professional certificate programs in Machine Learning and Artificial Intelligence at the Massachusetts Institute of Technology, and at the Harvard University Extension School. Dr. Adedinsewo is a Fellow of the American College of Cardiology (FACC).

Q: Can you share an overview of state-of-the-art applications of AI in cardiovascular medicine?

DA: In the last few years, there has been exponential growth in this field. Many applications are still limited to research studies and academic publications. AI models can identify disease patterns using labeled examples, referred to as supervised learning, or sift through data to identify patterns linked to disease without providing any labeled examples, in what is termed unsupervised learning.

Applications of AI in cardiology include disease detection and prognostication spanning left ventricular function, valvular heart disease, arrhythmia, coronary artery disease, hypertension, congenital heart disease, stroke, and mortality. These AI models have leveraged existing clinical data including ECG, computed tomography (CT), ultrasound, x-ray, magnetic resonance imaging (MRI), laboratory results, and data from electronic health records.

Several new biometric monitoring devices, sensors, and algorithms have received FDA approval for cardiovascular indications. These include but are not limited to the AliveCor heart monitor, Apple Watch, Eko Stethoscope software, Fitbit ECG App, Caption Health’s AI-guided echocardiogram software, and Ultromics’ software for automated echocardiographic measurements and interpretation.

Q: How are innovative applications of artificial intelligence in cardiovascular medicine being advanced at Mayo Clinic?

DA: At Mayo Clinic, a group of physicians and researchers developed deep learning (a form of artificial intelligence) models using data from a standard 12-lead ECG (an affordable and readily available test) to predict low ejection fraction, atrial fibrillation, hypertrophic cardiomyopathy, cardiac amyloidosis, and aortic stenosis.

The team recently completed a prospective clinical trial showing that the AI-enhanced ECG improved the diagnosis of low ejection fraction without increasing echocardiography utilization. This study is one of the first to evaluate an AI-based tool in a randomized clinical trial and paves the way for further implementation research and incorporation into routine clinical practice to improve patient care.

I worked with this team to evaluate the low ejection fraction model as a screening tool in a group of pregnant and postpartum women, given that heart failure is a leading cause of death in this population, which we found to be effective. We are currently working on validation studies to evaluate the effectiveness of this model in a diverse group of pregnant and postpartum women at external institutions. We are collaborating with physicians at hospitals in Florida, Georgia, Oklahoma, North Dakota, Missouri, California, and in several international locations. We are also leading an ongoing prospective study in this space.

Q: AI-enabled ECG algorithms have been featured in many studies, from predicting physiological age to determining risk of heart failure. What opportunities do you see for using insights from AI-enabled ECGs to improve risk prediction and public health?

DA: I believe the potential opportunities are numerous. Multiple, additional studies are in the works. Studies I am currently working on with the Digital Innovation Lab at Mayo Clinic in Florida have shown promising preliminary results. These include the use of the AI-ECG to predict outcomes among heart transplant patients and conduction disturbances requiring pacemaker implantation among patients who receive a transcatheter aortic valve.

Mayo Clinic’s Artificial Intelligence team in Rochester, Minnesota and Mayo’s Digital Innovation Lab in Jacksonville, Florida, have been instrumental in the development of AI-based algorithms. They consist of an excellent group of data scientists, data engineers and statisticians committed to improving the care that we provide to our patients. AI-ECG algorithms can not only diagnose cardiovascular conditions but non-cardiovascular conditions as well, including COVID-19 disease, liver pathology, kidney disease, hyperkalemia and sleep apnea.

I think we have only just begun to scratch the surface when it comes to potential uses of the AI-enhanced ECG. In the future, I believe AI applications will have public health implications through the development of screening tools, analysis of disease patterns, and surveillance (and potentially guidance) of public health interventions.

Q: What breakthrough innovations in healthcare delivery or technology excite you most?

DA: I am excited about potential opportunities to leverage this technology to improve cardiovascular care, specifically among women and minority patients. Women and racial/ethnic minorities bear a high burden of cardiovascular disease — and are frequently excluded from large cardiovascular clinical trials from which guidelines are developed and therapeutic management decisions made.

We know now that the “one-size-fits-all” approach for managing cardiovascular disease is not ideal. My hope is that we can start to address some of these issues and move toward health equity using AI tools and technologies correctly and responsibly.

Call for Healthcare Innovations

Are you developing an innovative healthcare technology? Contact the Mayo Clinic Innovation Exchange to learn how membership can help bring your company or your idea closer to patients.


All other company, product, and service names of monitoring devices, sensors, and algorithms referenced in this article are for identification purposes only, and does not imply endorsement. All brand and product names are the property of their respective owners.