Using Imaging—and AI—to Find Those at Risk of Cardiac Events

Improvements in computed tomography (CT), changes in guidelines, and the rise of artificial intelligence (AI) could lead to a dramatic shift over the next decade in using imaging to find people who lack symptoms but are at risk for cardiac events.

ASPC 2024 Congress | Image credit: American Society for Preventive Cardiology

That was the assessment of speakers at a joint session of the American Society for Preventive Cardiology (ASPC) and the Society of Cardiovascular Computed Tomography (SCCT). The session took place Friday on the opening day of ASPC’s 2024 Congress in Salt Lake City, Utah.

Maros Ferencik, MD, PhD, MCR, assistant professor at Oregon Health Sciences University, said in the past, when patients were asymptomatic the major concern was looking at risk factors such as family history. “For decades, we've been working with this paradigm, ‘Does the patient have ischemia or a stress test? Or does this patient have obstructive disease?”

Ferencik said that work by investigators including ASPC's Martha Gulati, MD, MS, of Cedars-Sinai Heart Institute led to a critical change in the 2021 guidelines from the American College of Cardiology and American Heart Association (ACC/AHA) that expanded the definition of coronary artery disease (CAD). Today, the definition recognizes both obstructive and nonobstructive coronary plaque, which is more likely to affect women. Ferencik went through a series of slides to show how this condition, which includes patients with coronary artery calcification as revealed in a CAC test, can also be found in patients who have coronary computed tomography angiography (CCTA).

The value of CCTA comes in the ability to quantify the amount of plaque, he said. Certainly, patients with obstructive disease are likely to have cardiac events. But data from the PROMISE trial¹ showed that more than half had nonobstructive disease—it’s the patients with high level of plaque driving the event rates, Ferencik said.

PROMISE compared use of less invasive CT angiography with SPECT myocardial perfusion imaging for patients with suspected CAD. Although the trial did not show any change in outcomes for use of CT scans, the SCCT hailed the results as proof that less-invasive testing was effective and could be used on more patients.

Ferencik said CCTA not only allows for early detection of CAD, but it also gives clinicians greater insight into the composition of the plaque and how the disease changes over time. With early detection and the right therapies clinicians can move beyond halting disease. “This is our ultimate goal,” he said. “We will be able to not only stop the progression of atherosclerosis but actually regress the risk process.”

Ron Blankstein, MD, FASPC, of Harvard Medical School dug further into how cardiac CT fits into current ACC/AHA guidelines—with implications for reimbursement. The fact that both CAC and CCTA are widely supported by guidelines is important when physicians must weigh in on behalf of their patients to get CCTA covered by insurers. A warning—even with coverage, patients may face significant copayments.

Where is CAC helpful? Blankstein said scoring can be useful when clinicians have patients already taking a statin and are considering whether to intensify therapy based on other factors. “If they have a very strong family history of premature events, if they have inflammatory conditions like lupus or psoriasis—I think there is a role for calcium scoring these patients,” he said.

On the other end, some patients may have 10-year risk score above 20%, which would indicate treatment per current guidelines, but the physician wants more information. “There’s a potential role of a calcium score to decide what how aggressive you want to be,” Blankstein said.

Similarly, CCTA may be used to clarify decision-making, and it’s likely being used in clinical care in ways that are not yet recognized, he said. “Perhaps there is a concern for noncalcified plaque individuals who are going to have a large burden of noncalcified plaque. For example, patients with HIV and systemic inflammatory diseases are going to be more likely to have that.

“So there may be a role for a CCTA in these asymptomatic individuals. [This is] certainly not supported by guidelines right now. But the reality is this is being done in clinical care,” Blankstein said.

CCTA may be needed just to get a patient into a clinical trial, and evidence for its use to monitor disease over time “continues to expand,” he said. “We can use CCTA to tell us which patients are responding to therapy.”

The guidelines themselves continue to evolve, Blankstein noted. He cited a guideline from SCCT called the CAD RADS 2.0 document. “Anytime you have a 2.0 you know that version 1.0 wasn't quite perfect,” he said.

He offered additional case studies in use of guidelines in clinical practice. While the guidelines focus on therapy initiation, increasingly they are used to clarify questions about intensification of therapy—including when to add additional therapies. And today both CAC and CCTA guidelines are widely supported, which was not the case a decade ago.

Leandro Slipczuk, MD, PhD, of Montefiore Medical Center, concluded the session by highlighting potential uses of AI with CCTA. There is great promise in this area, but the challenges of reimbursement mount, as CMS continues to reduce payment for imaging.

However, developing algorithms to find individuals without symptoms who are likely to have cardiac events could have huge implications for health systems. Many companies offer products, and many institutions are already using AI, he said.

CCTA with AI can produce atherosclerosis analysis quickly, and tools are emerging to allow integration of new information, Slipczuk said, such as incorporating information from the aorta.

The emerging field of “radiomics,” which uses AI in a form of precision medicine, is gathering data to predict which individuals will develop plaque. The field is moving, he said, “from focusing only on secondary prevention to a continuum of risk.”

He especially wants to see the field look at younger patients, for whom intervention can reverse disease before it creates damage that will permanently disable patients and cause high downstream costs. Today, “We are waiting until patients are too old,” Slipczuk said. “We must [act] earlier to treat CAD and induce plaque regression with more aggressive treatment.”

Reference

Dougals PS, Hoffman U, Patel MR. Outcomes of anatomical versus functional testing for coronary artery disease. N Engl J Med. 2015;372(14):1291-1300. doi: 10.1056/NEJMoa1415516.