Doctors Train AI on Preventing Heart Attacks

A new tool assesses coronary inflammation, which is often overlooked in measuring the risk of a heart attack. One doctor says this test could detect problems years in advance and save lives.

An AI tool that can detect inflammation in the coronary artery could help clinicians diagnose heart disease much earlier, even decades before the patient shows any outward signs of distress.

CaRi-Heart Technology, developed by Connecticut-based Caristo Diagnostics, was recently given its own Category III CPT code by the American Medical Association’s CPT Editorial Panel, an important step in the path to adoption after FDA approval and, just as important, payer reimbursement. The technology has also shown promising results in trials conducted in 2024 at five National Health Service hospitals in the UK, where reports indicate more than half of patients analyzed by the AI tool had their treatments changed.

To clinicians, the tool could be a critical step forward in the diagnosis and treatment of heart disease, the leading cause of death in the country.

“Cardiology disease is very different than other diseases in terms of how we treat it,” says Stephen Bloom, MD, MSCCT, FASNC, FAHA, FACP, FACC, a cardiologist with Midwest Heart and Vascular Specialists in Overland Park, Kansas, part of the HCA Midwest Health System. “We do mammograms before people have breast cancer. We do colonoscopies before people have colon cancer. And then in cardiology, we wait until they have symptoms, and then we do our best to treat our patient, now with established disease. It doesn't even make sense.”

AI has the potential to analyze data from tests, such as a CT scan, more quickly and with more detail than the human eye. While traditional imaging can identify visible plaques that cause narrowing and blockages, the CaRi-Heart tool zeroes in on perivascular fat, or coronary inflammation, which is overlooked in assessing someone’s heart health.

And since acute MIs occur when non-calcified plaque ruptures, any method for detecting non-calcified plaque better and earlier would save lives.

“[AI] could actually look at each coronary [artery], each segment, if you break it down to three, and [it] can actually summarize not only how much calcified plaque you have, but also non-calcified vulnerable plaque, which  has more of a tendency to rupture and cause a heart attack,” Bloom says.

Bloom notes that inflammation can be present in many diseases and measured by a blood test (c-reactive protein (CRP)). However, this blood test is not specific for the heart and less sensitive.

But with coronary CT with AI, clinicians can drill down further than they’ve been able to in the past. Bloom says doctors often use stress testing, but these tests only become positive for heart disease when the patient has a coronary blockage greater than 70%. This can create a false sense of security. Analyzing inflammation with plaque analysis  from a CT could create a much better definition of a patient’s cardiac risks.

And it could be done long before any signs of heart disease are evident.

“We can take patients even before they have symptoms and diagnose whether or not they have early coronary disease and treat them with appropriate medication as well as a change in their diet [and] exercise,” Bloom says.

The challenge, as always with new technologies, is reimbursement. Most payers currently don’t cover tests like nuclear cardiac stress testing, MRI stress scans or even coronary CTs unless there are symptoms. Bloom is hoping that the results of the UK tests and the newly approved CPT code will prod payers to cover the CaRi-Heart test.

“The good thing is it’s simple [and] it’s painless and less expensive than nuclear and other tests like MRI, and so it could be the gatekeeper to reducing heart disease by finding it early and treating the patient in the early stages,” he says.

“The next five years will probably dramatically change the way we treat our patients with coronary disease,” Bloom adds, adding that plaque analysis from a CT has only been approved for reimbursement this past year. As more clinicians use these new tools with coronary CT and gather data on its effectiveness, they’ll build a better argument for coverage.

“We will  be able to diagnose and treat coronary disease well before symptoms occur and finally reduce heart disease as the number one cause of death today ”.