AI-powered opportunistic screening could help find patients with undetected aortic stenosis
The use of artificial intelligence (AI) to identify incidental cardiac findings is a growing area of research, and now the identification of aortic stenosis (AS) is being tested on noncardiac chest CT scans. Cedars-Sinai Medical Center researchers are conducting a real-world study of an opportunistic algorithm that looks at all chest scans for the hallmark signs of AS.
Some imaging experts have said this specific use of AI could serve as an early screening system to catch patients with heart disease years before they would become symptomatic, adding diagnostic value beyond why these scans were originally ordered. AS is one of the most common and serious heart valve disease problems, but most patients are unaware they have it at first and go undiagnosed until they become symptomatic. The American Heart Association (AHA), for example, believes there is still a sizable population of patients who remain undiagnosed or untreated.
Since the old treatment for AS was surgery, many patients and doctors shed away from referring patients. However, transcatheter aortic valve replacement (TAVR) has resulted in many more patients coming forward for treatment than previous was believed to make up this population. Even so, about 50% of patients who are diagnoses still go without treatment, the AHA said.
The Cedars-Sinai study includes a review of 440 routine non-contrast, nongated chest CT scans from adults age 50 or older. They used an investigational device algorithm developed by the AI company Nanox that sits on top of the PACS to perform authomated AI reads for AS on all chest CTs, regardless of why they were ordered. The AI identified aortic valve calcium in 178 patients. Of these patients, 151 findings (84.8%) were not mentioned in the radiology report, researchers said.
Only six of 440 patients (1.4%) met the study's high/severe calcium threshold, using the Multi-Ethnic Study of Atherosclerosis (MESA) long-term risk scoring for severe AS using aortic valve calcification (AVC) scores measured in Agatston units (AU) via computed tomography. These are the same units used to calculate CT coronary calcium scoring exams.
Under the lower MESA community-risk categories, 99 of 440 (22.5%) had at least 100 AU, which is considered moderately high risk of AS. An additional 93 patients below the severe-disease cutoff also had elevated long-term risk scores.
The researchers said the omitted finding of AS on the reports shows many patients with measurable aortic valve calcium at a major imaging center are being left outside the regular valve-care pathway.
"Aortic valve sclerosis and calcification are the earliest detectable stage of the AS pathway," explained study investigator Ofir Koren, MD, assistant professor of interventional and structural cardiology with the Smidt Heart Institute at Cedars-Sinai Medical Center, in a presentation at the SCCT2026 meeting in July. "Recognizing and reporting it creates a critical opportunity for risk stratification, surveillance and timely intervention."
Examples of non-contrast, non-gated CT scans performed for non-cardiac reasons that clearly show incidental aortic stenosis. This type of opportunistic CT AI imaging could change the way patients are screened for AS to enable earlier interventions. Images from the SCCT 2026 presentation by Ofir Koren, MD, Cedars-Sinai.
He said AS is not seen as a management and treatment target so it is often overlooked on CT scans. With increasing interest in treating AS earlier, this will require a change in thinking that converts what is seen as missed incidental findings into missed opportunities.
Aortic valve calcium (AVC) can accumulate for years before a patient develops symptoms or severe aortic stenosis. Traditionally patients have only been treated when they are in advanced stages of the disease and they are often high risk. However, recent TAVR trials have shown earlier intervention before a patient becomes symptomatic can improve outcomes, recovery times, and avoid a decline in quality of life.
Koren said opportunistic AI for AS is coming at a convergence of two parallel revolutions, with earlier detection of disease thanks to CT and and expanding interventional therapies to treat AS.
AS calcifications could be detected by the AI at increasing rates the older patients were. The Cedars-Sinai study found 17% of patients ages 50-59 has detectable AS, about 50% ages 70-79, and 100% after age 90.
This study included a mix of scans originally intended to be used for lung nodule/cancer screening, lung disease, cardiac symptom evaluation, trauma and falls and connective tissue disease.
"The opportunity is not only to detect AVC, but to close the care gap after detection. What is an incidental finding today, could be a critical finding tomorrow. Al can turn an incidental finding into a structured clinical pathway," Koren said.
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: [email protected]