Can Artificial Intelligence Help Detect Heart Disease? - Cone Health

Skip to Content

Published on February 01, 2018

Can Artificial Intelligence Help Detect Heart Disease?

Can artificial intelligence help detect heart disease

No treadmills. No radiation. No hours-long visits. Just hook up a few leads, do a quick scan and get a verdict on your heart health.

The heart care team at Cone Health is taking part in clinical trials for what could be a game-changing technology, using artificial intelligence to help detect coronary artery disease.

The LeBauer-Brodie Center for Cardiovascular Research and Education at Cone Health's Heart and Vascular Center is among about a dozen healthcare facilities around the country participating in the Coronary Artery Disease Learning and Algorithm Development (CADLAD) study.

The study involves the use of a cardiac Phase Space Tomography Analysis (cPSTA) System, applying a machine-learned algorithm in an effort to greatly simplify how coronary heart disease is diagnosed.

The technology is being developed by Analytics 4 Life, a digital health company based out of Toronto.

Currently, if you have some chest pain and need to get a test, you might have to miss a day of work to do it, and it can lead to another test, which can lead to another test. On the other hand, if you can use this new technology as an alternative, it takes three minutes in the office.

Coronary artery disease is caused by the accumulation of plaque in the arteries, restricting blood flow - a condition that can lead to a heart attack. According to the Centers for Disease Control and Prevention, the disease kills about 370,000 Americans each year.

Those employing the cPSTA system use a device that looks like a tablet computer, but about an inch thick with a handle at one end.

The machine connects to the body with sensors and scans electrical signals emitted by the heart, detecting voltage differences between the sensors. Millions of data points are created, which are sent to a cloud-based repository and analyzed with machine-learned algorithms. An image is then rendered.

If there is no coronary disease, the image will be solid, appearing uniform in nature. If there's coronary disease, it will contain all these arcs and angulation in space. It'll be sort of disrupted, like a piece of Swiss cheese almost. These images are basically chest tomographs of all these electrical data points.

At present, doctors trying to diagnose coronary artery disease might have a patient do an exercise electrocardiogram (ECG) test, walking on a treadmill. They can also conduct an exercise SPECT scan, for which they inject a radioactive material into a patient's bloodstream. A special camera examines the material as it makes its way through the body while the person is exercising.

Both kinds of tests can be time consuming and expensive. The ECG test manages to detect coronary artery disease in only about half of those who have it. The SPECT test is more sensitive, finding the disease in about three-quarters of those who have it, but also has a greater tendency to produce false positives.

That can be a problem sometimes with people who have large breasts or are overweight. It can create an imaging artifact which leads you to think they've got a problem, but then you do a heart catheter, and they're completely clean.

Early data from the first phase of the study has been compared with angiograms. Thus far that data indicates the cSPTA has detected coronary artery disease in 92 percent of those who have it.

This might serve as a really good screening test in the office once the technology gets fully developed. With this test there is no risk, no radiation, no safety issues. You just lay on a table, and we collect the signals.

Phase one of the CADLAD study should be completed in 2018, followed by a second year-long trial to validate initial findings.

Since its founding in 1991, the LeBauer-Brodie Center for Cardiovascular Research and Education has completed more than 300 research trials, providing residents in and around central North Carolina with access to clinical trials and the latest treatments, while advancing cardiovascular medicine around the globe. Learn more about the center here.

About the Author

Thomas Stuckey, MDDr. Thomas Stuckey is a co-founder and Medical Director of the Lebauer-Brodie Center for Cardiovascular Research and Education, and Medical Director of Quality at Moses Cone Hospital.​