Impact of the peri coronary artery inflammation and Ai based risk predictors using cardiac CTA on the stratification of patients with suspected CAD Insight from the NHS England Chest Pain Pilot

Abstract Background The AI-based CariHeart Risk calculator has been developed using cardiac CT angiography (CCTA) assessment of the peri-coronary artery inflammation (FAI) and conventional cardiovascular risk factors to predict 8 years fatal and non-fatal cardiovascular events. As one of the 5 chosen Hospitals in the United Kingdom we have performed and interim analysis to assess the impact of the proposed new chest pain pathway on the risk stratification of patients referred for suspected angina to the Rapid Access Chest Pain Clinic (RACPC). Objective We set out to investigate the interaction (risk stratification predictive power) between the degree of coronary artery disease (CAD), the FAI risk score and the AI-based CariHeart Risk score. Methods 135 consecutive patients referred to the RACPC are part of this analysis. 2 patients had no contrast CCTA images at patient request and were excluded. Risk stratification was defined as no or mild CAD (50% stenosis), moderate, (50-69% stenosis), severe (70% stenosis) based on the degree or coronary artery luminal stenosis, low FAI score defined less then 50 percentile for any coronary artery, intermediate FAI score defined 50-75 percentile for LAD/RCA or 50-90 percentile for LCX, and high FAI score defined 75 percentile for LAD/RCA and 90 percentile for LCX. Low AI- CariHeart risk score was defined 1% 8 years CV mortality, intermediate risk 1-5% CV mortality, and high risk if 5% CV mortality. Results We found that patients with No or mild atherosclerotic CAD had a low FAI score in 4/114 (3.5%), intermediate score in 16/114 (14%), and high score in 94/114 (82%), (p0.00001). Low CariHeart risk score was found in 4 (3.6%), intermediate risk score in 47 (41%), and high-risk score in 63 (55%) (p0.0001). In No-or mild CAD patients the FAI score identified larger number of patients then CariHeart risk score as high risk 82% vs 55%, p0.0001. However, merging the intermediate and high FAI score patients and compare it with the merged intermediate and high CariHeart risk score patients the two scores identified the same proportion of patients 97%. The 13 patients with moderate CAD on CCTA, low FAI score was found in none, intermediate score in 38%, and high score in 62%, (p=0.025). Low CariHeart risk score was found in none, intermediate risk score in 8%, and high-risk score in 92%, (p0.007). High FAI score was found in all 3 cases whereas high CariHeart score in 2 cases and intermediate in 1 with severe CAD on CCTA. Conclusion The AI-based new chest pain pathway utilising CCTA images has identified higher proportion of at-risk patients particularly those with non-obstructive CAD. FAI score and CariHeart risk score seems to perform similarly in the aggregate intermediate and high-risk groups. To assess the value of the NP on real world outcome and the effect of new and conventional cardiovascular preventative therapy is warranted.