Background: This study sought to evaluate the association between coronary plaque characteristics, changes in the fractional flow reserve (FFR) derived from computed tomography across the lesion (ΔFFRCT), and lesion-specific ischemia using the FFR in patients with suspected or known coronary artery disease.
Methods: The study assessed coronary computed tomography (CT) angiography stenosis, plaque characteristics, ΔFFRCT, and FFR in 164 vessels of 144 patients. Obstructive stenosis was defined as stenosis ≥50%. An area under the receiver -operating characteristics curve (AUC) analysis was conducted to define the optimal thresholds for ΔFFRCT and the plaque variables. Ischemia was defined as a FFR of ≤0.80.
Results: The optimal cut-off value of ΔFFRCT was 0.14. Low-attenuation plaque (LAP) ≥76.23 mm3 and a percentage aggregate plaque volume (%APV) ≥28.91% can be used to predict ischemia independent of other plaque characteristics. The addition of LAP ≥76.23 mm3 and %APV ≥28.91% improved the discrimination (AUC, 0.742 vs. 0.649, P=0.001) and reclassification abilities [category-free net reclassification index (NRI), 0.339, P=0.027; relative integrated discrimination improvement (IDI) index, 0.093, P<0.001] of the assessments compared to the stenosis evaluation alone, and the addition of information about ΔFFRCT ≥0.14 further increased the discrimination (AUC, 0.828 vs. 0.742, P=0.004) and reclassification abilities (NRI, 1.029, P<0.001; relative IDI, 0.140, P<0.001) of the assessments.
Conclusions: The addition of the plaque assessment and ΔFFRCT to the stenosis assessments improved the identification of ischemia compared to the stenosis assessment alone.
Keywords: Plaque; coronary computed tomography angiography (CCTA); fractional flow reserve (FFR); machine learning.
2023 Quantitative Imaging in Medicine and Surgery. All rights reserved.