FDA Clears AI Software for Opportunistic Multisystem Computed Tomography Screening

The Food and Drug Administration has granted 510(k) clearance for the AI-CVD software, which may offer a variety of automated assessments for preventive multisystem screening based off of a single computed tomography (CT) scan.

Combining 10 FDA-cleared modules in one artificial intelligence (AI)-enabled software, the AI-CVD platform offers quantitative measurements for cardiovascular assessments — including epicardial fat volumetry, aortic valve calcium and coronary artery calcium (CAC) scoring — along with automated analysis of lung and liver attenuation, according to HeartLung Corpooration, the developer of the AI-CVD software.

Noting that more than 19 million chest CT scans are obtained annually in the United States, Nathan D. Wong, Ph.D., MPH, FACC, FAHA, FNLA, MASPC, said the AI-CVD offers significant promise for preventative assessment and early detection.

“This opens up unprecedented potential for clinicians to identify a wide range of cardiovascular and other cardiometabolic conditions earlier in their subclinical state, enabling earlier and personalized therapeutic interventions to prevent the onset of clinical cardiovascular events and other conditions such as diabetes,” emphasized Dr. Wong, a professor medicine, director of the Heart Disease Prevention Program and co-director of the Center for Global Cardiometabolic Health and Nutrition at the University of California-Irvine.

Robert Kloner, M.D., Ph.D., noted that the inclusion of 10 FDA-cleared modules within the AI-CVD software for opportunistic assessment can improve the awareness of clinicians who may not have expertise in certain disease types.

“This is clearly an example of where AI will enhance the physician’s ability to implement a better health care plan and all from one scan,” posited Dr. Kloner, the director of the Cardiovascular Research Institute at Huntington Medical Research Institutes and a professor of medicine at the Keck School of Medicine at the University of Southern California (USC). “It is important to point out that many of these findings will likely be picked up before any symptoms are manifest by the patient. So, these silent killers may be uncovered and dealt with before they can cause morbidity and mortality to the patient.”

Noting the connection between different organ systems within the spectrum of cardiometabolic disease, Dr. Wong praised the inclusion of lung, liver and muscle fat assessments in the AI-CVD software.

“This is very timely given the great attention in the medical community from the American Heart Association’s recently introduced Cardiovascular Kidney Metabolic (CKM) Health Initiative, which stages disease based on a comprehensive set of cardiometabolic measures ranging from obesity to hypertension, dyslipidemia, kidney disease and subclinical CVD (cardiovascular disease),” pointed out Dr. Wong, a past president of the American Society for Preventive Cardiology.

Could the use of the AI-CVD software lead to increased documentation needed in radiologist reports? Dr. Kloner noted that the automated quantitative analyses with the AI-CVD software should save time for radiologists. However, Dr. Kloner also acknowledged the need for increased radiology reporting would depend on the initial findings with AI-CVD.

“For example, if a high calcium score was observed in the coronary arteries of a patient, additional studies including CT angiography, stress test nuclear imaging or coronary angiography associated with percutaneous coronary intervention (stenting, angioplasty) might be required,” said Dr. Kloner. “However, if the a calcium score of the coronary arteries was zero and the patient was deemed very low risk, then no additional radiologic studies related to the patient’s coronary arteries would be needed in most cases.”

Overall, Drs. Kloner and Wong concur that multisystem detection with AI-CVD assessment of CT scans can facilitate improved patient outcomes and lower costs.

“The AI-CVD platform offers earlier detection of a wide range of cardiometabolic conditions in an opportunistic fashion on CT scans that have already been performed for other reasons, thus not requiring new scans to be done, and identification of earlier disease would hopefully lead to earlier therapeutic interventions to prevent costly clinical events and hospitalizations,” added Dr. Wong. “All of this in the long run we expect will reduce costly CVD events and procedures and their associated costs that contribute to the U.S. having the highest health-care costs in the world.”