Study: AI Measurement of Coronary Inflammation from CT Scans Can Identify High CV Risk Patients with Low or Zero CAC Scores

Automated deep learning measurement of coronary inflammation from coronary computed tomography angiography (CCTA) scans may be a game changer for improving risk stratification in patients with low or zero coronary artery calcium (CAC) scores, according to research recently presented at the Society of Cardiovascular Computed Tomography (SCCT) conference.

For the study, the authors evaluated a research application of the AI software CaRi-Heart (Caristo Diagnostics) with a review of data from paired CCTA and non-contrast CT in 19,092 patients (mean age of 56) with a median follow-up period of 6.7 years.¹ The AI software provides automated measurement of coronary inflammation through a Fat Attenuation Index (FAI) Score with the mean FAI Score representing an average of the three coronary arteries, according to the study.

Among patients with a CAC score below 10, cardiovascular death occurred in 127 individuals (1.2%). Within this subgroup, those in the top decile (>90th centile) of FAI Score_CCTA had an approximately 13-fold higher cardiovascular (CV) mortality risk compared to those below the 25th centile, while those in the top decile of FAI Score_NC had an approximately eightfold higher risk.¹

In a recent interview with Diagnostic Imaging, lead study author Kenneth Chan, MBBC, who was honored with the Young Investigator Award at the SCCT conference, said the study findings reveal the interplay between coronary calcification and inflammation, and the need for increased risk stratification clarity beyond what CAC scoring can provide.

“We recognize that is a large proportion of patients with no calcium on CT, or very minimal (CAC). These are the patients with a (CAC) score of less than 100, but they still have a significant number of (cardiovascular) events. Inf act,(this) group of patients (account for a) third of all the events that happened in a large population study as well as in our study. … The key is to identify those are high risk within these people with no or low (CAC scores), and we found that coronary inflammation indeed can help us to identify those at risk,” explained Dr. Chan, a Clinical Research Fellow in Cardiology within the Division of Cardiovascular Medicine at the University of Oxford in Oxford, England.

Dr. Chan also emphasized that for the first time, the study findings demonstrated the capability to obtain automated assessment of coronary inflammation from non-contrast CT, which may have significant implications for facilitating timely screening in a wider population of patients.

"It's potentially applicable in the screening setting for primary prevention, because the non-contrast scan is usually lower cost with no needles and no contrast, which is better for the patient. If they're already getting the calcium score, we can also extract information to complement the risk assessment, really enhancing the power of cardiac CT for screening of cardiovascular risk,” noted Dr. Chan, who is also affiliated with the Acute Multidisciplinary Imaging and Interventional Centre (AMIIC) at the University of Oxford.

(Editor’s note: For related content, see “Can the Combination of Photon-Counting CT and AI Change Our Approach to CAC Scoring?,” “SCCT: Assessing the Impact of AI Detection of Coronary Plaque on Chest CT Scans and Notification of Treating Physicians” and “Can AI Assessment of CT Attenuation Correction Mapping Be a Viable Alternative for CAC Risk Stratification?”)

Reference

1. Chan K, Gissier MC, Sobirov I, et al. Quantifying coronary inflammation on contrast or non-contrast cardiac CT enhances CV risk stratification beyond coronary calcium score and plaque volume. 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 .