Publication

Abstract : In this work, an attempt is made to analyze cardiac magnetic resonance (CMR) images to study the pathophysiology of Myocardial Ischemia (MI). It is a cardiac disorder that causes irreversible damage to heart muscles. The images considered for this study are obtained from Medical Image Computing & Computer Assisted Intervention (MICCAI) database. Adaptive fuzzy based multiphase level set method is used to extract endocardium and epicardium of left ventricle from short-axis view of CMR images. The segmentation results are validated with similarity measures such as Dice coefficient and Jaccard index. Further, five indices are derived from the segmentation results. From the results, it is observed that the average Dice coefficient for endocardium and epicardium are found to be 0.867 and 0.918 respectively. The mean Jaccard index for endocardium and epicardium are 0.766 and 0.855 respectively. It is observed that the algorithm is able to segment the left ventricle more precisely from CMR images. The ischemic subjects show a reduced mean ejection fraction (32.52) compared to the normal subjects (59.04). The average stroke volume is found to be 70.16ml and 64.05ml for healthy subjects and ischemic subjects respectively. Reduction in stroke volume and ejection fraction for ischemic subjects indicates lower quantity of blood drained by heart. It is also observed that, there is an increase in myocardial mass for ischemic subjects (182.11g) compared to healthy subjects (127.47g). The thickened heart muscle contributes to the increased myocardial mass in abnormal subjects. Further, ischemic subjects show an increase in endocardium volume at end-diastolic and end-systolic phase when compared to normal subjects. Thus, the clinical indices evaluated from adaptive fuzzy based multiphase level set method could differentiate the normal and ischemic subjects. Hence this study can be a useful supplement in diagnosis of myocardial ischemic disorder.