Model-based Analysis of Myocardial Microstructure for Left Ventricular Mechanics using MRI

Abstract

Heart disease is the leading cause of morbidity and mortality in the western world. The underlying mechanisms of heart failure (HF) remain largely unexplained. Increased myocardial tissue stiffness, diffuse fibrosis, and tissue remodelling have been identified as contributors to HF. However, their biomechanical effects are not well understood. In this thesis, the link between intrinsic mechanical properties of the myocardium, estimated using an inverse finite element approach, and indices of fibrosis quantified from T1 magnetic resonance imaging (MRI) was investigated using patient data. No significant correlation was found between intrinsic myocardial stiffness and native or post-contrast T1 times. Another important determinant of left ventricular (LV) function is the microstructural tissue organisation. A statistical analysis of microstructural orientations and myofibre disarray derived from diffusion-weighted magnetic resonance imaging (DWI) of healthy rats and end-stage spontaneously hypertensive rats was conducted. No evidence of remodelling of orientations or increased global myofibre disarray was found in the diseased rats, but hypertension did affect regional disarray. Hypertension had no significant effect on the transmural distributions of primary or secondary eigenvectors of the diffusion tensors. Model-based parameterisations were developed to fit microstructural orientations directly to diffusion tensors, or to raw diffusion signals. These methods avoid the need to perform eigenanalysis on the diffusion tensor, or, when fitting to raw diffusion signals, eliminate the need to compute diffusion tensors using least squares fitting. Cardiac mechanics simulations are sensitive to the representation of microstructural orientations, as well as other factors such as loading conditions and kinematic constraints. To investigate the effects that variations in the microstructural orientation fields have on LV mechanics, a region of indifference was constructed around the helix angle field parameters that best represent a high resolution DWI dataset of a canine heart. This indifference region defines the field parameters to which the goodness-of-fit is insensitive to. A set of perturbed helix angle fields was determined from the decomposition of the indifference region, and the perturbations were used to describe the microstructural orientations in passive and contractile mechanics simulations. Passive mechanics predictions were found to be insensitive to variation in helix angle parameters, but end-systolic predictions of ventricular geometry were substantially affected. This thesis demonstrated the use of computational model-based analyses of myocardial microstructure, assessed through MRI, to provide tissue-specific evaluation of LV mechanical function and microstructural remodelling. A particular emphasis is placed on the role of myocardial fibrosis, intrinsic myocardial stiffness, and microstructural orientations.