Coronary inflammation and AI-Risk scores from cardiovascular computed tomography: impact on risk prediction and clinical management in a real-world setting.

Abstract

Aims: Coronary computed tomography angiography (CCTA) is the primary investigation for stable chest pain. Despite approximately 80% of individuals undergoing CCTA not having obstructive coronary disease, this group contributes to two-thirds of major adverse cardiovascular events. Assessment of coronary inflammation using perivascular fat attenuation index (FAI) and AI-derived risk scores (AI-Risk) has demonstrated enhanced risk prediction beyond traditional clinical and CCTA parameters. We aimed to assess if FAI and AI-Risk alter risk prediction and clinical management in a real-world setting.

Methods and results: Consecutive patients undergoing CCTA with FAI calculation and AI-Risk (CaRi-Heart®) at a single centre over a 3-year period were recruited. Conventional risk scores for non-fatal and fatal myocardial infarctions (QRISK3 and SCORE, respectively) were compared with AI-Risk. Clinical management decisions based on risk factors and CCTA results were recorded. FAI and AI-Risk scores were then provided and the resultant clinical management decision recorded. One hundred and sixty-four patients were included in the study (n = 164, male 78%, 56 years). Forty-eight per cent of the patients had no evidence of coronary artery disease (CAD) on CCTA, with 41% having non-obstructive CAD and 10% with potentially obstructive CAD. AI-Risk reclassified risk in 58% and 43% of patients compared with QRISK3 and SCORE, respectively. Clinical management was changed in 33% of patients following AI-Risk analysis.

Conclusion: FAI and AI-Risk scores in a real-world setting changed risk prediction in around half of individuals and changed clinical management in around a third.