Volumetric Quantification of Myocardial Perfusion Using Analysis of Multi-Detector Computed Tomography 3D Datasets

Multi-detector computed tomography (MDCT) assess-ment of myocardial perfusion is based on visualization of 2D slices. To overcome the subjective nature of this analysis, we developed a new technique for quantification of myocardial perfusion from MDCT 3D datasets and tested it against nuclear myocardial perfusion imaging (MPI). MDCT datasets of 44 patients were analyzed to generate bull's eye display of myocardial perfusion and calculate segmental quantitative indices of extent and severity of perfusion defect. Visual interpretation of MDCT-derived bull's eyes was in agreement with rest MPI scores (□=0.79). Quantitative MDCT perfusion data were in good agreement with MPI summed scores: r=0.84; objective detection of perfusion defects using MDCT index resulted in: area under ROC curve=0.87, sensitivity 0.92, specificity 0.91, and accuracy 0.89. Our technique for volumetric analysis of MDCT data allows accurate and objective detection of perfusion defects.