Assessment of Semi-automated Computed Tomographic Measures of Segmental Perfusion Defects in a Swine Model (Sus scrofa) of Intermediate Coronary Lesions

Abstract

Computed tomographic myocardial perfusion (CTP) imaging is a tool that shows promise in emergent settings for defining the hemodynamic significance of coronary artery disease. In this study, we examined the accuracy with which the transmural perfusion ratio (TPR) derived through semiautomated CTP analysis reflected segmental perfusion defects associated with intermediate coronary artery lesions in swine. Lesions (diameter stenosis, 65% ± 11%) of the left anterior descending coronary artery (LAD) were created in 10 anesthetized female swine (weight, 47.5 ± 1.9 kg) by using a pneumatic occlusion device implanted on the LAD. Occluder inflation pressures were adjusted to maintain fractional flow reserve (FFR, 74.3 ± 1.7) during adenosine infusion (140ug/kg/min). Static CTP imaging using a stress-rest protocol and segmental TPR derived from semiautomated CT perfusion software was compared with microsphere-derived TPR (mTPR) by using a 16-segment model and polar mapping. Intermediate LAD stenosis was verified through multiplanar coronary CT angiography. Receiver operating characteristic analysis identified an optimal threshold for segmental perfusion defects for intermediate lesions (TPR threshold, ≤0.80); however, the area under the receiver operating characteristic curve was 0.58, and the overall accuracy was 63%. At this threshold, the sensitivity and specificity were 65% and 61%, and the positive and negative predictive values were 61% and 65%, respectively. Although CTP–TPR illustrated segmental perfusion defects with intermediate lesions, the disparity between CTP–TPR and mTPR measures of segmental perfusion suggests that further advances in analysis software may be necessary to improve the localization of segmental defects for intermediated lesions.