Assessment of the fibrotic myocardial tissue mechanics by image processing

We propose an image processing methodology to study the mechanical properties of fibrotic and normal tissue of the myocardium in magnetic resonance. The procedure fuses cine and late enhanced images and estimates the myocardial motion in cine images. A novel fast and robust tracking procedure is proposed in which the deformation is modeled by a Markov random field and quasi-incompressible, spatio-temporally smooth and homeomorphic transformations are promoted. The method has been applied to the analysis of the differences in the contractility patterns of hyperenhanced and normal zones in a set of 12 patients affected by hypertrophic cardiomyopathy. The absolute value of both the mean radial and circumferential components of the peak Green strain and strain rate tensors has been encountered significantly lower when hyper-enhancement is present. On the other hand, the peak rotation dispersion in the longitudinal-radial plane has been encountered significantly smaller in the fibrotic region.