HeartFlow’s noninvasive AI-enabled plaque technology showed 95% accuracy in determining total plaque volume and type in the coronary arteries, according to the outcomes of a study published on May 3 in the European Heart Journal Cardiovascular Imaging.

The multicenter international study, which enrolled 258 patients across 15 sites in the U.S. and Japan, compared HeartFlow’s automated deep-learning-based method for segmenting coronary atherosclerosis in coronary computed tomography angiography (CCTA) against the reference standard of intravascular ultrasound (IVUS). The researchers found:

• Total Plaque Volume vs. IVUS: 95% agreement
• Calcified Plaque Volume vs. IVUS: 95% agreement
• Non-Calcified Plaque Volume vs. IVUS: 93% agreement

“I’m thrilled to see such an overwhelming result showcasing that CT angiography, coupled with automated AI-driven Plaque Analysis is an effective way to assess plaque burden,” said Jagat Narula, M.D., Ph.D., co-principal investigator and K. Lance Gould Distinguished University Chair of Coronary Pathophysiology at UTHealth Houston. “This is a futuristic step forward for patients and clinicians to have an accurate, noninvasive test that gives insights into their total plaque volume, and is every bit as good as the invasive reference standard. It will allow clinicians to feel more informed to deliver the best treatment recommendations possible for each patient.”

The FDA-cleared HeartFlow Plaque Analysis technology is an automated method which analyzes coronary computed tomography (CT) scans and enables clinicians to visualize, characterize, and quantify plaque. The plaque technology is based on HeartFlow’s proprietary deep learning-enabled algorithms, which have been trained on more than 15 million coronary CT images.