Accuracy of Dynamic Computed Tomography Adenosine Stress Myocardial Perfusion Imaging in Estimating Myocardial Blood Flow at Various Degrees of Coronary Artery Stenosis Using a Porcine Animal Model

Objective: To determine the accuracy of computed tomography (CT) dynamic stress myocardial perfusion imaging to estimate myocardial blood flow (MBF) in a porcine animal model with variable degrees of induced coronary artery stenosis in comparison with microsphere-derived MBF. Methods and Materials: Seven domestic pigs (36 +/- 4 kg) received stents (confirmed 3.0 mm diameter) in the left anterior descending coronary artery distal to first diagonal branch. A balloon catheter was placed within the stent and inflated to various degrees to obtain a defined luminal narrowing (50% and 75% diameter stenosis) as confirmed by intra-arterial flow wire measurement. All models underwent adenosine-mediated (140 mu g/kg/min) dynamic stress and rest myocardial perfusion CT imaging using a dual-source CT scanner (shuttle-mode with 100 kV/300 mAs, 20 mL iopromide) with prospective acquisitions every second heartbeat for 30 seconds. CT-estimated MBF (MBFCT) was calculated using a model-based parametric deconvolution method and correlated to that of fluorescent microspheres (MBFmic) injected at each perfusion state. Results: All study procedures were performed without complications, and all animals completed the study protocol. Among 448 myocardial segments, 31 (7%) were considered nonevaluable because of motion artifacts. With stress, MBFCT increased significantly (1.10 +/- 0.25 vs. 0.80 +/- 0.28 mL/g/min, P < 0.001; at stress and rest, respectively) in all myocardial segments and correlated with MBFmic (r = 0.67, P < 0.001). MBFCT overestimated MBFmic, independently of adenosine-stress and degree of coronary stenosis (beta = 2.3, 95% confidence interval: 1.81-2.79 mL/g/min, P < 0.001). Although there were no differences in MBFCT between 50% and 75% coronary stenosis at rest (0.01 +/- 0.08 mL/g/min, P = 0.86), MBFCT was significantly lower at 75% than at 50% under stress conditions (0.53 +/- 0.19 vs. 0.71 +/- 0.24 mL/g/min, P = 0.002). Conclusions: CT-derived MBF measurements at rest and stress with varying degrees of coronary stenosis show a valid difference but an underestimated correlation with microsphere-derived MBF in a porcine animal model.