Illustration: Getty Images
A physical exam and echocardiogram have been the traditional tools for diagnosing aortic stenosis. Yet many patients, especially older adults who receive chest CT scans for unrelated conditions, are never referred for an echocardiogram, even when aortic valve calcification (AVC), a precursor to aortic stenosis, is visible in existing imaging.
Researchers and care teams at Stamford Health’s Heart and Vascular Institute in Connecticut are seeking to change that paradigm. They are using a pair of FDA-cleared algorithms, developed by Bunkerhill Health, to detect AVC and coronary artery calcification (CAC). The algorithms run in the background of non-contrast CT scans and quantify the amount of AVC or CAC present.
“We’re trying to standardize the way we treat cardiovascular disease,” explains David Hsi, MD, chief of cardiology and co-director of Stamford Health’s Heart and Vascular Institute. “The goal is early disease detection so appropriate patients can see a cardiologist and begin a surveillance program.”
Artificial Intelligence to Detect AVC
Catching aortic stenosis early, when interventions can have the biggest benefits, is an ongoing challenge, explains Dr. Hsi’s Stamford Health colleague, interventional cardiologist Arzhang Fallahi, MD, the director of structural heart disease at Stamford Health.
“Patients above age 75 have a significant increase of moderate-to-severe valve disease, and a large proportion of that is aortic stenosis,” he says. “Yet there’s no formal recommendation for elderly patients to get echocardiograms.”
The AI-powered AVC tool could help close that gap. To test the algorithm, Stamford Health’s team screened roughly 300 patients who had non-cardiac-related chest CT scans. They presented their findings at an internal grand rounds presentation in May.
“About a third of the cohort never previously had an echocardiogram for aortic valve disease, even though they had detected aortic valve calcification,” Dr. Hsi says. “And 11 patients were found having severe aortic valve calcification and echo-confirmed stenosis during the follow-up.”
While this is early data, “it gives us important direction to pay attention to this subpopulation of patients coming through chest CT for other reasons,” Dr. Hsi says.
AI-Enabled CAC Detection
The AVC algorithm builds on Bunkerhill Health’s contrast CAC tool. Stamford Health has been using the CAC algorithm since 2024, screening thousands of patients and reporting their findings at the American Heart Association’s Scientific Sessions this past November.
“The results demonstrated that, in patients with severe aortic stenosis and CAC, there is room to improve clinical care, including starting pharmacotherapy and aspirin therapy earlier for patients who were not yet entered into that clinical pathway,” Dr. Hsi says.
When elevated scores are detected, both the AVC and CAC tools automatically notify the ordering physician of the results and recommend a cardiology referral.
“Depending on the degree [of calcification], we’ll do a patient history to determine whether a patient has concerning symptoms that would warrant additional testing,” Dr. Fallahi says. “Those with additional symptoms may be referred to a coronary CTA, a stress test, or if we’re concerned enough, jump straight to the cath lab.”
Other Artificial Intelligence Advances
Stamford’s team has integrated additional AI-enabled tools into their workflows. One is Heartflow’s FFRCT(fractional flow reserve derived from CT), which analyzes coronary CTs for the presence of plaque. “Since implementing Heartflow this year, we’ve seen a 10% increase in volume,” Dr. Hsi says.
The AI-driven CardioCare Digital Health Platform by egnite, meanwhile, has proven beneficial for aortic stenosis workflows. “It allows us to monitor patients with six-month follow-ups with the structural heart valve team, follow-up echos, or imaging as needed,” Dr. Fallahi says.
As AI algorithms grow more sophisticated in the coming months and years, Drs. Hsi and Fallahi can envision a future in which AI findings shape clinical guidelines. “Best of all, these tools will allow us to achieve the ultimate goal of catching disease early so patients don’t have to come into the ER crashing and burning,” Dr. Fallahi says.
May 20, 2026 
