Image-based estimation of passive myocardial properties using finite element modeling
Patient-specific cardiac finite element (FE) models, which are used to estimate the stress state of the heart, are highly dependent upon the assumed material properties defined in the models. A method is developed for determining cardiac material properties from computed tomography imaging data. Custom optimization software is developed to interact with a nonlinear large-deformation FE solver in order to perform nonlinear optimization on the material parameters of a patient-specific FE left ventricular (LV) model. The optimization method is tested using synthetic diastolic images based on an FE model with known material parameters as well as in clinical images for a normal male patient. The optimization technique produced material parameter estimates that are in good agreement with the parameters of the FE/LV model that is used to produce the synthetic images. Sensitivity studies show that the optimization process is relatively insensitive to the parameter starting values but highly sensitive to the diastolic loading pressures.
