SCCT: Assessing the Impact of AI Detection of Coronary Plaque on Chest CT Scans and Notification of Treating Physicians

Coronary calcification findings on chest computed tomography (CT) scans are “often ignored,” according to Ron Blankstein, MD.

“ … Traditionally, radiologists often ignore that information. Sometimes it's kind of buried somewhere in the body of a report that there may be calcifications in the coronary arteries, but we are now finding that this is really important information. … This is information that’s actionable and this is information that, if used correctly, has a very important potential to improve patient outcomes,” maintained Dr. Blankstein, the director of cardiac CT and associate director of the cardiovascular imaging program at Brigham and Women’s Hospital in Boston.

Accordingly, Dr. Blankstein and colleagues are examining the use of a cloud-based AI platform (Nanox AI) for detecting coronary plaque on chest CTs and whether subsequent notification of these findings to treating physicians had an impact on patient care.

While the multicenter AI INFORM study is ongoing, six- and 12-month results at the Brigham and Women’s Hospital that were presented at the Society of Cardiovascular Computed Tomography (SCCT) conference revealed approximately 33 percent and 46 percent increases, respectively, in initiation or intensification of lipid-lowering therapy. Dr. Blankstein noted there was also a corresponding mean decrease in low-density lipoprotein cholesterol (LDL-C) of 17 mg/dL in the AI INFORM cohort vs. 4 mg/dL in the usual care arm of the study.¹

“ … The fact that the LDL is decreased shows us that not only are clinicians prescribing therapies — in this case lipid-lowering therapies — but patients are actually taking them as well. So it's a nice confirmation to see that it's not just that this leads to more initiation of treatment, but we're actually seeing LDL cholesterol go down, and obviously that is an important aspect. This is how this ultimately will improve outcomes,” emphasized Dr. Blankstein, a professor of medicine and radiology at Harvard Medical School.

In a separate multicenter study presented at the SCCT conference, Dr. Blankstein and colleagues evaluated the use of the cloud-based HealthCCSng 2.0 algorithm (Nanox AI), which provides automated coronary artery calcium (CAC) scoring and categorization of plaque based on assessment of chest CT exams.²

The study, which involved 5,468 chest CT scans, revealed greater than 90 percent agreement between the CCSng 2.0 algorithm and reviewing radiologists. Specifically, researchers noted a 91.4 percent agreement for a threshold of CAC > 100.

“ … These are great algorithms, and I think they're very useful. They improve our ability to detect and quantify how much plaque there is, and ultimately for radiologists to include this information when they are reporting routine chest CTs. But someone still has to always evaluate this to make sure that what the AI algorithm picked up and traced as coronary calcifications is correct. But I think it's reassuring to see that in the majority of cases that agreement was quite high,” added Dr. Blankstein.

References

1. Huck DM, Blair CV, Besser S, et al. Artificial intelligence detection of incidental coronary artery calcium to enhance cardiovascular disease prevention: the AI INFORM trial. Presented at the Society of Cardiovascular Computed Tomography (SCCT) conference, July 9-12, 2026, San Diego. Available at: https://cattendee.abstractsonline.com/meeting/21508/meeting-info

2. Blair CV, Huck DM, Petranovic M, et al. Agreement between AI-detected coronary artery calcium and expert visual assessment: a multi-site analysis. Presented at the Society of Cardiovascular Computed Tomography (SCCT) conference, July 9-12, 2026, San Diego. Available at: https://cattendee.abstractsonline.com/meeting/21508/meeting-info