Estimation of myocardial strain from MRI image sequence using a deformable mesh model

Analysis of myocardial strain plays an important role in cardiac MRI, since fluctuations in the strain pattern of the heart may be a sign of cardiac abnormalities. In order to extract strain information, cardiac deformation must be estimated from a set of cardiac-gated MR images. In previous works, we developed a deformable mesh model (DMM) to describe 3D myocardial motion from MRI. In DMM, left ventricle motion was estimated by deforming a 3D mesh to match with inter-frame pixel-intensity variations of a combined gated tagged and untagged MRI image sequence of the heart. In this work, from an estimated DMM we compute myocardial 3D strain maps by applying a prolate spheroidal B-spline model (PSB), and we compare it to a reference model which uses an active-contour (i.e. snakes) tracking algorithm. Furthermore, we show a comparison of DMM-based strain measurements (torsion) obtained from healthy and abnormal MRI datasets, showing the diagnostic potential of the DMM technique.