Speaker: Sathya Vijayakumar

Title: Evaluation of four analysis methods for the detection of coronary artery disease using myocardial perfusion mri

Coronary artery disease (CAD) is a leading cause of death in the USA and worldwide and myocardial perfusion imaging is a widely used technique for the detection and assessment of CAD. Dynamic contrast-enhanced Magnetic Resonance Imaging (MRI) is now becoming increasingly popular as a technique for the detection and assessment of CAD.

Tracer kinetic models play a crucial role in the quantification of perfusion. In this work, the performance of three kinetic models was evaluated: (1) a modified 2-compartment model; (2) the Johnson Wilson (JW) model which is based on the difference in rate of change of tracer concentration in tissue regions vs. capillaries; (3) a modified JW model which accounts for the fraction of blood in the tissue region of interest and compared with the well-established method of upslopes. These models were implemented and evaluated using a robust and generalized semi-automatic processing algorithm.

Since quantification of perfusion from compartment models strongly depends on the arterial input function (AIF), it was also important to get a better understanding of the accuracy of the AIF obtained from the MR perfusion images. The factors affecting the AIF that were addressed in this work are the effect of the choice of ROI, in the left ventricle or in the aorta, and the conversion of signal intensity to concentration of contrast agent Gd-DTPA.

The studies show promising results when the modified JW model was applied, though the 2-compartment model proved to be more robust. The analysis was performed on data obtained from 17 subjects (12 patients who had undergone coronary angiography and 5 normal volunteers). Percentage sensitivity/specificity values of 100/92, 73/100, 100/67 and 92/71 were obtained for the modified 2-compartment model, the JW model and the modified JW model and the method of upslopes respectively. It was found that the modified 2-compartment model and the modified JW model perform significantly better than the more established method of upslopes. For quantitative perfusion measurements, guidelines to choose the AIF from the MR images were established.