This is the first article of a 3-part series. Click here for part 2 and part 3.

When Michael Nanna, MD, reviews his patient load and demographics, the shifts are glaring. Compared to 5-10 years ago, more of his patients are older than 65 years and are becoming more challenging to treat.

“We’re absolutely seeing more older adults than ever before, and they’re clinically more complex, with multiple chronic conditions,” said Nanna, an interventional cardiologist at Yale School of Medicine in New Haven, Connecticut.

Cardiologists across the US are reporting similar changes, driven by historic levels of Baby Boomer aging. Adults aged 65 years or older now represent 18% of the US population, up from 12.4% in 2004, according to a Medscape Medical News analysis of data from the government and medical societies. This population grew 13% from 2020 to 2024. By 2030, people older than 65 years will outnumber people younger than 18 years for the first time in US history.

Expand

While longevity points to widespread improvements in medicine and public health, it also brings a glut of older patients in a field with a shrinking workforce and rising costs. A Presidential Advisory from the American Heart Association projects that by 2050 more than 61% of Americans — over 184 million people — are likely to have cardiovascular disease, potentially tripling related costs to the healthcare system. Heart failure, the leading cause of hospitalization in Medicare patients, is projected to affect 45 million US adults by 2050, Medscape Medical News found.

New Challenges for Clinicians

The coming decades are likely to test cardiologists’ endurance and demand greater sensitivity to the nuanced priorities of their older patients. The changes have elevated the importance of minimally invasive procedures and AI tools that can help identify disease and candidates for intervention.

“The biggest shift is that decisions are less about a single lesion or procedure and more about how treatment aligns with quality of life, independence, and what matters most to the patient,” Nanna said. “A technically ‘successful’ intervention may not be the right choice if it risks prolonged recovery or loss of function.”

Some catheter-based procedures such as percutaneous coronary intervention (PCI) and transcatheter aortic valve replacement are already central to the care of older adults, he said. Looking ahead, Nanna expects continued growth in minimally invasive approaches, including improved transcatheter valve therapies that can be used in frailer patients.

For now, Nanna urges careful selection of older patients for transcatheter tricuspid valve interventions, including edge-to-edge repair and replacement. In clinical trials for both procedures, positive results were mainly patient-reported quality-of-life gains, rather than improvements in all-cause death or heart failure hospitalizations. Both trials lacked sham-control arms, raising questions about possible placebo effects.

Cardiologists are still learning the unique risks of invasive procedures in older adults, too. Nanna’s 2023 research published in the Journal of the American Heart Association found adults aged 75 years or older receiving complex PCI typically had a higher risk for death within a year. But they were less likely to need additional heart procedures compared to those who received noncomplex PCI.

Concerns About Imaging, Interventions

Minimally invasive imaging tests are also increasingly aligned with older patients’ needs, said Krishna Patel, MD, an assistant professor of medicine at the Icahn School of Medicine at Mount Sinai and director of cardiac PET service at Mount Sinai Morningside Hospital in New York City.

Patel’s research centers on how to use imaging more effectively for older adults.

Photo-counting CT, for instance, can see through the excess coronary artery calcium that many older adults have due to atherosclerosis, buildup that otherwise reduces accuracy when clinicians look at blockages.

Many older patients cannot tolerate a cardiac MRI, leading to incomplete results, but recent innovations such as shorter scans and free breathing protocols are making the test more accessible for this population, Patel said. And compared with traditional stress tests, cardiac PET scans can more precisely diagnose issues in older adults, who often face a high risk for coronary artery disease with multivessel involvement and small-vessel disease.

Although imaging cannot improve outcomes on its own, imaging findings could improve clinical management of older patients — if there were stronger evidence, Patel said.

“All of the imaging prognostic data we have for any modality focuses on heart attacks or mortality as outcomes, and we know that older adults don’t care about that as much as they do things like quality of life or their independence or functional status,” Patel said. “So we need more information on how imaging-guided decision making can improve outcomes that are meaningful for older adults.”

Preventing Unnecessary Procedures

A combination of improved data collection and less invasive procedures could help cardiologists avoid overtesting in older adults.

“In the future, and as we get more data, we will see more use of more advanced imaging modalities that can give us more accurate answers in a single study vs just getting echocardiogram and stress tests in all older adults,” Patel said.

AI tools may also help cardiologists rule out unnecessary procedures or prompt them to order tests they might not have considered for older patients. Barriers to widespread adoption of AI tools by cardiologists persist, however, including a lack of education on how the tools work — which contributes to mistrust and avoidance — and financial constraints for smaller and rural hospitals.

But AI is an emerging tool for helping identify and prevent heart disease in older patients with complex needs and histories, said Faraz S. Ahmad, MD, MS, a heart failure specialist and assistant professor of cardiology at Northwestern Medicine in Chicago.

“We’ve shown that Medicare Advantage patients with heart failure have worse quality of life than those with different types of cancers, including lung cancer,” Ahmad said. “So having earlier diagnosis and the right diagnosis is really important.”

AI-based detection tools for heart failure and amyloidosis, for instance, may help cardiologists identify more cases of heart failure with preserved ejection fraction (HFpEF) and its causes faster and with greater accuracy, Ahmad said. Heart failure is the leading cause of hospitalization in Medicare patients, affecting about 5%-7% of 60- to 69-year-olds and 8%-12% of adults in their seventies.

Ahmad sees many patients suspected of having HFpEF, but sometimes additional tests — such as exercise stress echocardiogram or exercise right heart catheterization — are needed to confirm the diagnosis, he said. In a subset of patients, the AI tools developed by Ultromics, which Northwestern adopted in 2024, can increase his confidence in the diagnosis or help him avoid extra tests. Other times, he might have a low suspicion that a patient’s symptoms are heart-related, but AI picks something up that points to a high probability, pushing him to do additional testing.

“We know that HFpEF is often delayed in diagnosis or missed, and so we think this is a real opportunity to find patients earlier and get them on treatments that change outcomes,” said Ahmad, who is also the associate director of Bluhm Cardiovascular Institute Center for Artificial Intelligence at Northwestern Medicine.

The ‘Bottleneck’

Older patients’ diagnostic results are piling up, prompting some hospitals to adopt AI models that can help process images captured by minimally invasive tests.

“There’s rarely a patient that I see in clinic who doesn’t have an echo. This is where I see the greatest impact for AI, particularly for older adults, because we’re really working in a resource-constrained system. As there’s more older adults, there’s more need for imaging, and there’s more need for more precise imaging,” said David Ouyang, MD, a cardiologist at Kaiser Permanente Santa Clara Medical Center, Santa Clara, and a research scientist at the Kaiser Permanente Northern California Division of Research, Pleasanton, both in California.

As echocardiography devices become more affordable to health systems, clinicians are seeing their adoption in the emergency department, primary care settings, and ICU, Ouyang said. “Now the bottleneck is actually accurate assessment of the images.”

Sonographers often write a preliminary report of findings after scanning patients, but this “intermediate phase” of echocardiogram assessment can take about half an hour, Ouyang said. He developed EchoPrime, an AI model that generates a preliminary report for the cardiologist. A randomized clinical trial comparing EchoPrime- and sonographer-generated preliminary reports for accuracy and reliability is expected to conclude by the end of 2026.

“Our north star is what the cardiologist ultimately writes,” Ouyang said. “It’s the comparison of that preliminary to final assessment that gives a sense of how accurate the AI and the sonographers are.”

Separate clinical trials led by Ouyang’s team are using AI models to investigate the extent to which cardiologists may be overlooking amyloidosis and liverdisease in echocardiogram images. Amyloidosis is underdiagnosed and can lead to heart failure, while fatty liver disease and cardiovascular disease also have similar risk factors, including obesity, diabetes, hypertension, and metabolic syndrome.

“Particularly as people get older and there’s more comorbidities, there might be confusion of what disease a patient might actually have,” Ouyang said.

A typical echocardiogram captures one or two images of the liver out of roughly 100 images overall, but cardiologists “might not recognize that the liver is abnormal,” Ouyang said. The EchoNet-Liver platform, which Ouyang developed and is being evaluated in a clinical trial uses AI to flag possible problems, such as fatty liver disease or cirrhosis. Cirrhosis and HFpEF can be confused with each other, Ouyang said, and AI can help “distinguish between cardiac and extra cardiac causes of swelling.”

Ahmad is participating in the amyloidosis trial using Ouyang’s EchoNET-LVH model. Ahmad and his physician colleagues have already reached out to patients, including older adults, whose echocardiograms were flagged by the AI and invited them into the clinic for further testing.

In 2025, there were 277 FDA-approved AI algorithms with cardiology applications, mostly for imaging and diagnostics. Cardiologists themselves have given AI mixed reviews, according to recent research. In 2024, specialty cardiologists rated AI-generated responses on diagnosis, triage, and management equal to or better than those of general cardiologists.

But cardiologists were less aligned with an AI-enhanced echocardiogram tool’s assessments of multimorbid patients compared with less complex patients, according to a 2026 study. An accompanying survey of 60 cardiologists suggested that they feel comfortable using AI to double-check — but not replace — their own judgment.

“I think of AI as a safety net,” Ouyang said. “The physician is definitely in control and oversees everything, but AI helps identify areas or information that might be missed on casual review.”

As time-strapped cardiologists see many more aging patients in the coming years, AI that quickly analyzes diagnostic test results — potentially finding problems earlier or identifying good candidates for intervention — could become essential.

“We feel like we need as much help as we can get,” Ouyang said.

Resources for Cardiologists

Below are several online resources for cardiologists practicing in an aging America.

ACC’s Geriatric Cardiology Section

ESC’s Geriatric-Cardiology Resources

Heart Failure Society of America Learning Page

Overview of CVD Conditions for Older Adults

AHA Scientific Statement on Coronary Artery Revascularization in Older Adults

Geriatrics Healthcare Professionals Site

GeriKit Assessment App (Apple)

GeriKit Assessment App (Google)

Gerontological Society of America Learning Center

Michigan Surgical Quality Collaborative’s Frailty Resources

American Geriatrics Society Updated Beers Criteria® for Potentially Inappropriate Medication Use in Older Adults

The 10-TaGA Geriatric Screening Instrument

Disclosures

Nanna reported current research support from the American College of Cardiology Foundation, the Patient-Centered Outcomes Research Institute (PCORI), the Yale Claude D. Pepper Older Americans Independence Center (P30AG021342), the National Institute on Aging (K76AG088428), and Merck, Inc.; personal fees from HeartFlow, Inc, Merck, Novartis, and Novo Nordisk. 

Ahmad has received research support from Atman Health, Tempus, AstraZeneca, IDoven, Ultromics, Abiomed/Johnson & Johnson, and Anumana; has received consulting fees and honoraria from AstraZeneca, Alnylam Pharmaceuticals, and Omada Health. 

Patel reported funding from NIH/NIA and PCORI. 

Ouyang reported the following disclosures: NIH R00 HL176421, R01 HL173526, R01 HL173487, U01 RD008704, AHA AI Grant Consultant (Anthem, Pfizer, AstraZeneca, Tempus, Dandelion Health, JNJ, Abbott, EchoIQ, Ultromics) Sponsored Research (Alexion, Apple), CoFounder (InVision).

Sarah Amandolare is a freelance journalist and independent science and health reporter who regularly contributes to Medscape Medical News. Her work spans a variety of topics at the intersection of medicine, technology, and longevity.