Comparative Assessment of CT and MR Perfusion with a Myocardial Perfusion Phantom

Abstract

Introduction: Myocardial CT perfusion (CTP) is a relatively new technique which enables the assessment of perfusion deficits through first pass imaging in combination with pharmacological stress. Comparison of CT and MR perfusion (MRP) has been limited and in vivo assessment is affected by multiple factors. We utilised a myocardial perfusion phantom to precisely compare high resolution k–t SENSE MR perfusion at 3 T with single phase CT perfusion under identical imaging conditions. Methods:TheMRandCTcompatiblephantom includes venous input, cardiac chambers, pulmonary and aortic output, coronary arteries and two diffusion chambers to represent themyocardium.Coronary flow andmyocardial perfusion may be precisely controlled. Noise estimates were based on published clinical data. Results: CTP contrast is dependent on photon energy however the contrast to noise (CNR) of low kV acquisipreliminarydata fromacohort of 37patientswasanalysed. TwoCCTA reporters, assessed calcified plaques onCCTA, followed by one assessing ICA lesions (gold standard). Calcified plaques were assessed by three methods; residual cross-luminal area (method 1), residual mean luminal diameter (method 2) and luminal flow estimation (method 3). Lesions were classified as non-obstructive ( 70% stenosis). Results:Mean agewas 58, and 67%weremale. Fromour preliminary data of 37 patients, 152 calcified plaques were detected; 39 (25%)were classified as obstructive inmethod 1, 34 (22%) in method 2 and 16 (10%) in method 3. Kappa values for agreement between eachmethod and ICAwere as follows: method 1: 0.47, method 2: 0.38, method 3: 0.22. The greatest discordance was lesions of 50–70% category, where the best agreement was 15% (method 1). Conclusion:Agreement for quantifying calcifiedplaque luminal stenosis between ICA and CCTA is fair to moderate in this study. Residual cross-luminal area was the most accurate method. Furthermore, calcium blooming and beamhardeningmake it unfeasible to accurately subcategorise significant plaques.