An automatic data assimilation framework for patient-specific myocardial mechanical parameter estimation

Abstract

We present an automatic workflow to extract myocardial constitutive parameters from clinical data. Our framework assimilates cine and 3D tagged Magnetic Resonance Images (MRI) together with left ventricular (LV) cavity pressure recordings to characterize the mechanics of the LV. Dynamic C 1-continuous meshes are automatically fitted using both the cine MRI and 4D displacement fields extracted from the tagged MRI. The passive filling of the LV is simulated, with patient-specific geometry, kinematic boundary and loading conditions. The mechanical parameters are identified by matching the simulated diastolic deformation to observed end-diastolic displacements. We applied our framework to two heart failure patient cases and one normal case. The results indicate that while an end-diastolic measurement does not constrain the mechanical parameters uniquely, it does provide a potentially robust indicator of myocardial stiffness. © 2011 Springer-Verlag Berlin Heidelberg.