With the help of artificial intelligence, a new tool is allowing doctors at Methodist Healthcare to analyze plaque in greater detail during standard coronary CT angiograms.
Soft, non-calcified plaque in the arteries is harder to detect through standard scans and more likely to rupture — triggering a heart attack or stroke — than hardened, calcified plaque.
Dr. Michael Lane, the medical director of Advanced Cardiovascular Imaging at Methodist Healthcare, helped the company behind the new tool validate the technology, which is now in use across 11 Methodist hospitals and four outpatient clinics.
“A year from now, this will probably be one of the most common exams we get in the emergency department,” Lane said, “because it’s just too valuable. It’s just too valuable to know how much plaque you have.”
Cardiologists care about plaque composition because it can be a better predictor of heart attack risk than artery narrowing alone. The new AI tool is trained on mountains of heart disease data, helping differentiate between soft and hardened plaque more quickly.
Cardiovascular disease is the leading cause of death worldwide and in the United States. Bexar County’s heart disease death rate is higher than both the national and state averages, according to 2019–2023 data from the National Institutes of Health.
Heart disease and plaque
In his office, Lane is analyzing a 360-view of the arteries of a patient who arrived at the hospital with chest pain. The calcified plaques are easy to spot: they stand out as bright spots on the greyscale images.
But the softer plaque is nearly invisible to the naked eye.
With one click, the AI-powered tool highlights the soft plaque, and several sections of the arteries light up in bright yellow. Lane doesn’t take the data as gospel, but the predictions are impressively accurate, he said.
“This patient is not going to go home, they’re going to [cath lab], they’re going to get a stent right there and save that heart tissue, because that’s what that patient needs,” Lane said.
Over years to decades, plaque — made up of fat, cholesterol, and other substances — builds up in the walls of arteries, which carry oxygen-rich blood from the heart to the rest of the body.
This process, known medically as atherosclerosis, can go undetected for years to decades until plaque ruptures, leading to a complete or partial blockage of the artery, resulting in a heart attack or ischemic stroke.
“When I talk to my patients, the one thing I want them to always remember is that for that person who has a heart attack at age of 60, when did it actually start?” Lane said. “At age 40.”
Lipid-rich soft plaque, what doctors call low-attenuation plaque, is more likely to rupture compared to hardened, calcified plaque. The soft plaque is the stuff that kills you, as Lane put it.
Many common heart disease screening tools don’t take plaque composition into account, Lane explained, missing a portion of patients who have high-risk soft plaque but little-to-no narrowing of their arteries.
Common screenings includes calcium score testing, for example, a non-invasive scan of hardened plaque to gauge short-term cardiovascular risk and artery narrowing.
Several studies have found that a meaningful share of patients with a calcium score of zero still have non-calcified plaque that isn’t visible on standard calcium scans.
Having a clearer picture of plaque with the aid of the new AI-powered tech might mean some patients who would have otherwise been cleared get treatment, and that physicians are able to more quickly analyze plaque in more patients, Lane said.
How the tech works
The technology is part of a growing wave of artificial intelligence tools developed by companies such as California-based HeartFlow that analyze coronary CT scans to identify and measure plaque in the arteries. Health care systems across the U.S. are incorporating artificial intelligence to supplement a range of clinical and bureaucratic functions.
The plaque technology utilizes AI to analyze and compare vast amounts of training data from coronary CT scans with intravascular ultrasounds (IVUS), a minimally invasive procedure that involves a specialized catheter with ultrasound technology to produce a 3D image of plaque buildup from inside the arteries.
“They found out that [the AI models?] were not only really good at picking up blockages in the arteries that were causing chest pain, but they were picking up all kinds of other things too,” Lane said. “We were able to start picking up soft plaque on those same CT scans [while] looking for blockages.”
Lane and his team at Methodist’s flagship campus in the South Texas Medical Center campus, which sees a significant volume of cardiac patients, were a part of early testing of the tech. Methodist now maintains 16 of these machines across 11 hospitals and four outpatient clinics capable of integrating the AI piece into its scans.
Patients can request one of these scans if their hospital has access. They’re best suited for patients who are presented with atypical chest pain, meaning it doesn’t fit with the standard signs of a heart attack, or if they have known coronary artery disease, Lane said.
Lane said the new tool could be especially helpful in diagnosing cardiovascular disease in women. Although women tend to have less total plaque than men on average, they are more likely to have non-calcified plaque, which is harder to detect with traditional imaging.
“I really see it going to be very valuable in that age range of that 40-year-old to 60-year-old female, because we can see plaque that we didn’t know existed before,” Lane said.
