Coronary inflammation quantified on contrast and non-contrast cardiac CT identifies high risk individuals independently from coronary calcification: implications for those with zero calcium score
K Chan, M C Gissler, E Wahome, E Nicol, D Adlam, A Kardos, J Rodrigues, S E Petersen, J P Greenwood, P Tomlins, R Blankstein, S Neubauer, K M Channon, J Deanfield, C AntoniadesAbstract
Introduction
Coronary calcification is a hallmark of atherosclerosis, where individuals with zero/low coronary artery calcium (CAC) scores are often reassured. However, inflammation is a critical process that could drive the rupture of vulnerable non-calcified plaques, even in the absence of coronary calcification. Among patients undergoing coronary Computed Tomography Angiography (CCTA), coronary inflammation can be quantified using the Fat-Attenuation Index (FAI) Score, which maps perivascular adipose tissue changes in the inflamed artery. However, the feasibility and value of the combined assessment of coronary inflammation and calcification on non-contrast cardiac CT is unknown.
Purpose
To assess whether measurement of coronary artery inflammation on CCTA or non-contrast cardiac CT can be used to identify individuals at high cardiovascular (CV) risk, particularly among individuals with low or no CAC burden.
Methods
Patients who underwent paired CCTA and non-contrast CT (for Agatston CAC Score) were included from the Oxford Risk Factor and Non-invasive Imaging (ORFAN) study. FAI Score of each coronary artery was measured on CCTA (FAI ScoreCCTA) with CaRi-Research. We also developed a fully automated deep-learning pipeline for measuring FAI Score on non-contrast CT (FAI ScoreNC), and its performance is tested now for the first time. Mean FAI Scores were calculated from the average of the three coronary arteries and classified into clinically relevant centiles. The total CAC Score was obtained for the whole coronary tree. Patients were prospectively followed for CV mortality.
Results
A total of 17,280 patients (mean age 56±12 years, 51% male) were followed for a median of 6.6yr (IQR 4.6-10.0). In the whole cohort, FAI Score stratified the risk of CV mortality within each CAC Score category, with the strongest associations among individuals with CAC Score 0 using either FAI ScoreCCTA (A) or FAI ScoreNC (B). The addition of FAI ScoreCCTA or FAI ScoreNC to CAC Score significantly improved the risk discrimination of CV mortality (C). By incorporating inflammatory risk, FAI ScoreNC significantly improved risk reclassification for CV mortality over CAC Score (D). CV death occurred in 75 (0.9%) patients among those with no coronary calcification (CAC Score 0, n=7,947). In this patient group, individuals with FAI Score >90th centile showed significantly higher risk of CV mortality compared to those <25th centile, with HR 9.39 (95%CI 4.48-19.7, p<0.001) for FAI ScoreCCTA and HR 8.17 (95%CI 4.07-16.4, p<0.001) for FAI ScoreNC respectively.
Conclusion
Quantification of coronary inflammation on CCTA or non-contrast CT using FAI Score improves prediction of cardiovascular outcomes beyond CAC Score. Assessment of coronary inflammation enables identification of patients with substantially increased cardiovascular risk, particularly among individuals with a CAC score of zero, thereby allowing timely initiation of preventive therapy.Figure