12-05-2025
The Promise of AI and Cardiovascular Diseases for PCPs
Physician outlooks on artificial intelligence (AI) vary, but most will tell you they're open to using it. AI as a tool to improve the administrative side of medicine is widely embraced, as is AI in medical imaging. But what is its role in helping detect cardiovascular diseases in the primary care setting? As with all areas of AI in medicine, that's an evolving picture.
In an official statement, the American Heart Association (AHA) said it supports the creation of tools and services that would 'further the science and practice of precision medicine by enabling more precise approaches to cardiovascular and stroke research, prevention and care of individuals and populations.'
Nevertheless, the AHA continued, 'several challenges exist, and few artificial intelligence tools have been shown to improve cardiovascular and stroke care sufficiently to be widely adopted.'
Still, AI isn't going anywhere, and the push for its use in cardiovascular disease prediction and treatment will only increase.
'AI is everywhere in the healthcare system,' said Hector Garcia-Garcia, MD, an interventional cardiologist with MedStar, based in Washington, D.C. 'AI tools are involved with labs, imaging, values, and more, and it's well integrated into electronic medical records today.'
Hector Garcia-Garcia, MD
The role of primary care physicians (PCPs) continues to be critical in heart disease detection. As frontline healthcare workers, PCPs are generally the first to note something amiss when it comes to patient cardiovascular health.
'PCPs are crucial to identifying cardiovascular health risks,' said Anthony Mercando, MD, a cardiologist at White Plains Hospital Physician Associates in New York City. 'They take family histories, do physical exams, and get blood panel results, all of which may show cause for investigating heart health.'
While all these traditional tools have long led PCPs to refer patients to their cardiologist peers, today AI is increasingly involved in screening. And that may be a boon, as a high percentage of heart diseases go undiagnosed.
The AI-Enhanced Tools
If there's a traditional workhorse in the world of heart disease, it's the stethoscope. And now, AI has joined the show, promising to enhance the tool's ability to pick up subtle changes. Preliminary studies have demonstrated that AI-enhanced stethoscopes may lead to higher discovery and earlier diagnosis of cardiac murmurs associated with valvular heart disease, which affects 1 in 10 adults older than 65 years.
The new stethoscopes work by making a 30- to 60-second recording of the patient's heart and lung sounds. Doctors also hear the sounds with greater clarity, thanks to amplification and noise cancellation. From here, AI steps in, instantly analyzing the sounds and ECG data to deliver a full picture of cardiac health — sometimes in as quickly as 15 seconds. The stethoscope represents a big step forward in early diagnosis, according to a Lancet-published study.
Mercando is a fan of the new tool.
'We're doing a pilot here to screen patients who are not yet symptomatic or have abnormal test results,' he said. 'It's an exciting tool to pick up disease earlier in the process, and it's a tool that any healthcare practitioner can use.'
The stethoscope is one example of how AI can enhance diagnosis of heart diseases, and represents a low-cost, frontline tool that PCPs can add to their practices. Other tools are less obvious, at least to the patient.
One of these is imaging, a key player in heart disease diagnosis. CT scans of the heart have long been stalwarts in recognizing heart disease. For a relatively low cost, a scan can reveal three-dimensional images of a heart and its surrounding structures. Cardiac CT may involve contrast dye to reveal the moving heart and blood vessels.
In the AI-enhanced version, CT can produce faster analysis of the images it takes. 'The AI is digging into hundreds of data points and is key in making the results faster,' said Garcia-Garcia. 'The results are always supervised by a cardiologist, but the AI is giving us the ability to identify red flags faster and earlier.'
Anthony Mercando, MD
Still, as with other uses of AI in medicine, it's important to recognize the clinical correctness of the results may not be perfect— thus the continued need for human intervention. According to the AHA's statement on CT imaging using AI, the applications are now 'capable of automating the computing of coronary artery calcium score' from low-dose CT or even from nuclear imaging studies.
'There's no AI tool in the United States that isn't supervised by the doctor, so the final word lies with the doctor,' said Garcia-Garcia. 'Therefore, there is no real downside.'
At Mercando's community hospital, the newly created position of AI coordinator is looking into relevant applications for AI. One of those is a solution that looks at every chest CT and creates a calcium score. 'When you have a tool that can predict risk and allow specialists to treat patients earlier, you are saving lives,' said Mercando.
Going Forward
As a PCP, you may or may not be tuned into how AI is helping your cardiac/potential cardiac patients. But as the technology becomes more widespread and the public becomes more aware, patients may bring it into conversation.
'A classic example is how AI is now reading mammograms, identifying differences from 1 year to the next,' said Mercando. 'Patients are aware of this and now asking for it. The same may carry over to heart disease — patients are supportive of the technology.'
AI in cardiac care isn't limited to stethoscopes and CT — its tentacles are beginning to reach into structural interventions like measuring stenosis, and by automatically detecting intracranial hemorrhage in stroke diagnosis.
In its statement, the AHA warns against AI limitations specific to imaging. The group points to sourcing, curating, and sharing appropriate data as one sticking point, along with questions surrounding the appropriate reporting of results using accepted statistical measures. The AHA suggests that deep learning algorithms may be enhanced by using a new medical imaging data readiness scale for improved outcomes.
As with all medical uses of AI, the AHA also cautions against bias leading to health inequities. The organization encourages the development of AI learning models that involve underrepresented races in cardiac care. The tool is far from perfected yet, but its potential will continue to intrigue doctors of all stripes, including cardiologists.
Moving forward, the fact that AI will play an increasingly larger role in cardiovascular screening isn't up for debate. What is debatable is the point when will the tool be useful to improving patient outcomes at scale. In the AHA's opinion, the answer lies in the need to mitigate bias and ensure education and access to AI technologies. Only then, said the organization, can AI realize its full potential.
From his perspective, Garcia-Garcia views AI's role in cardiovascular care as a game changer. 'AI can change how a patient is treated,' he said. 'Early detection gets patients to specialists sooner, and that saves lives.'
