A comprehensive prediction model of functionally significant coronary artery stenosis based on coronary computed tomography and the amount of myocardium in jeopardy assessed by fractional flow reserve

To explore a method for estimating the amount of myocardium in jeopardy based on CCTA and derive a comprehensive prediction model for coronary lesions with fractional flow reserve (FFR) value of 0.8 or less. Quantitative features and Myocardial Jeopardy Index (MJI) were measured in 104 coronary arteries of 91 CCTA exams, retrospectively. For 80 left anterior descending (LAD) arteries, the sum of vessel score, diameter stenosis, proximal lesions, apical wrap, long disease, and bifurcation disease were evaluated. The multivariate logistic regression analysis was applied to identify independent predictors for FFR ≤ 0.8. A LAD-Score was then derived to predict the hemodynamic significance of LAD lesions. MJI, diameter stenosis and minimum lesion diameter were significant independent predictors for FFR ≤ 0.8 in all 104 vessels at multivariate analysis. The integrated diagnostic accuracy of multi-predictor was increased to 83% from 72% when using stenosis of 70% alone. The area under the receiver operating characteristic curve (AUC) was increased from 0.738 to 0.839 correspondingly. For LAD lesions, the optional cutoff value of stenosis was 62.5%, with a diagnostic accuracy of 71% and AUC of 0.746. Using LAD-Score, the diagnosis accuracy was improved to 84% with AUC of 0.874. The MJI, estimated in CCTA, can help to predict the coronary lesion’s hemodynamic significance more accurately. The LAD-Score, combining LAD lesion features and the amount of myocardium in jeopardy estimated in CCTA, improves the diagnostic performance assessed using invasive FFR.