Computer aided diagnosis of the myocardial ischemia based on a spatio-temporal deformation features analysis

The non-invasive measurement of the spatio-temporal deformation of the myocardium and its analysis are very useful for improving the clinical diagnosis of cardiac ischemic pathologies. Magnetic resonance tissue tagging is a privileged technique able to provide a spatio-temporal quantification of the myocardial contraction. Based on this technique, existing methods to estimate the deformation generally provide a great number of 3D spatio-temporal parameters that are often quite difficult to interpret. In this paper, we propose software tools to help clinicians to analyze these deformation parameters. We also propose a strategy for the automatic detection of dysfunctional regions using approximate reasoning. The results of this simple approach, illustrated through 5 examinations, are reasonable for large and moderate infarct sizes.