AI model targets inflammation, helping cardiologists find 'invisible' heart patients
Cardiologists have been focused on treating coronary plaques as a way to prevent heart attacks for decades now. Coronary inflammation, however, has been largely ignored. That changed in 2023, when the U.S. Food and Drug Administration (FDA) cleared colchicine as the first drug to treat coronary inflammation for the prevention of heart attacks. Now there is a cardiac CT imaging artificial intelligence (AI) tool that will enable clear imaging of coronary inflammation in routine scans.
At the American Heart Association (AHA) 2023 meeting, a groundbreaking study presented by the University of Oxford revealed how Caristo's CaRi-Heart AI assessment software can visualize inflammation in the perivascular fat surrounding arteries and its ability to assess cardiovascular risk and prevent heart attacks. About 40,000 coronary CT angiography (CCTA) scans were reviewed by the AI model from the U.K.'s National Health Service (NHS) records. It was able to link those CT scans from 10 years previously with health outcomes over the following 10 years.
Almost half of patients without coronary narrowings were reclassified when coronary inflammation, rather than only plaque alone, was taken into account.
Cardiovascular Business spoke with interventional cardiologist Keith Channon, MD, MBChB, FRCP, the British Heart Foundation Professor of Cardiovascular Medicine at the University of Oxford, a co-founder of Caristo, whose team conducted the study.
"What it found is that, although the presence of significant coronary artery narrowings was associated with about a twofold increased risk of a heart attack or death over the following 10 years, more than 80% of people who underwent a CT coronary scan for clinical reasons didn't have significantly narrowed arteries. And although the risk was lower in percentage terms, the absolute number of people dying or having a heart attack after their heart CT scan was substantially higher in the group who didn't have major narrowings," Channon explained.
He explained this technology may offer a paradigm shift in how cardiac patients are screened, risk stratified and treated in the future.
"This has major implications for how we use routine coronary CT scans, the advice we give patients, how we identify those at highest risk, and ultimately what treatments we then recommend for these higher risk patients. They are currently invisible and they're not treated effectively. And if we could identify this very substantial number who are at risk, we could target treatments so much more effectively and improve outcomes in the long-term," Channon said.
As cardiac CT scans become more prevalent in the U.K. and the U.S., the study's findings may have significant implications for healthcare practice. With around 350,000 coronary CT scans conducted annually in the U.K. alone and utilization rising in the U.S., the AI-driven risk assessment could be a game-changer, identifying the hidden threats and optimizing treatment prevention strategies for a large portion of patients who might be overlooked with traditional assessments.
Inflammation is the silent risk factor in myocardial infarction
Traditional risk assessments often focus on visible signs of coronary artery narrowings from plaque. Current risk assessments classify patients without coronary stenosis or only minor coronary narrowings as low risk. These patients are often given lifestyle advice and may be given low statins, but Channon said they often do not receive the follow-up care they need to ensure nothing progresses.
"Drug compliance is not driven very hard, and ultimately many of those patients don't get the focus that a patient would have if they were at much higher risk. Conversely, if a patient is known to be at high risk of a heart attack or death, even though they don't have major coronary artery narrowings right now, they will be selected and prioritized for high-intensity statin treatment, intensive follow-up lifestyle management programs, and now new treatments that specifically target coronary inflammation such as colchicine," Channon explained.
Implications of AI assessing coronary inflammation
Cardiology is focused on plaques and the impact these stenoses have on coronary blood flow, but Channon said it is just the tip of the iceberg when looking at coronary disease. AI assessments like this have the potential to detect disease earlier by visualizing the currently inflammation that has not been assessed on a large scale before, he added.
"We've been focused on physical stenosis," Channon explained. "We've then thought ourselves that we're refining the message by focusing on the physiological consequences of that stenosis, but ultimately those features are the late stages of the disease. We've had no way of detecting or quantifying the underlying biological disease process, and ultimately that's inflammation in the wall of the coronary artery driven by cholesterol, smoking, diabetes, etc."
CT calcium scoring has become a big risk assessment tool, but Channon said it only shows calcified lesions, which only appear in old age or at the end stage of the coronary disease process. The Caristo AI technology enables assessment of soft plaques and inflammation in coronary walls years before patients develop calcium or develop symptomatic disease.
Regulatory progress and future prospects for Caristo
Caristo's CaRi-Heart software has European CE mark approval and the company is currently in discussions with the U.S. FDA. Channon said the company expects to file a de-novo application for regulatory review in early 2024.
Dave Fornell has covered healthcare for more than 17 years, with a focus in cardiology and radiology. Fornell is a 5-time winner of a Jesse H. Neal Award, the most prestigious editorial honors in the field of specialized journalism. The wins included best technical content, best use of social media and best COVID-19 coverage. Fornell was also a three-time Neal finalist for best range of work by a single author. He produces more than 100 editorial videos each year, most of them interviews with key opinion leaders in medicine. He also writes technical articles, covers key trends, conducts video hospital site visits, and is very involved with social media. E-mail: dfornell@innovatehealthcare.com