Analysis of myocardial contractility with magnetic resonance

Abstract

Heart failure has considerable morbidity and poor prognosis. An understanding of the underlying mechanics governing myocardial contraction is a prerequisite for interpreting and predicting changes induced by heart disease. Gross changes in contractile behaviour of the myocardium are readily detected with existing techniques. For more subtle changes during early stages of cardiac dysfunction, however, it requires a sensitive method for measuring, as well as a precise criterion for quantifying, normal and impaired myocardial function. Cardiovascular Magnetic Resonance (CMR) imaging is emerging as an important clinical tool because of its safety, versatility, and the high quality images it produces that allow accurate and reproducible quantification of cardiac structure and function. Traditional CMR approaches for measuring contractility rely on tagging of the myocardium with fiducial markers and require a lengthy and often subjective dependant post-processing procedure. The aim of this research is to develop a new technique, which uses velocity as a marker for the visualisation and assessment of myocardial contractility. Two parallel approaches have been investigated for the assessment of myocardial velocity. The first of these is haimonic phase (HARP) imaging. HARP imaging allows direct derivation of myocardial velocity and strain without the need of further user interaction. We investigated the effect of respiration on the accuracy of the derived contractility, and assessed the clinical applicability and potential pitfalls of the technique by analysing results from a group of patients with hypertrophic cardiomyopathy. The second technique we have investigated is the direct measurement of myocardial velocity with phase contrast myocardial velocity mapping. The imaging sequence used employs effective blood saturation for reducing flow induced phase errors within the myocardium. View sharing was used to improve the temporal resolution, which permitted acquisition of 3D velocity information throughout the cardiac cycle in a single breath-hold, enabling a comprehensive assessment of strain rate of the left ventricle. One key factor that affects the derivation of myocardial contractility based on myocardial velocity is the practical inconsistency of the velocity data. A novel iterative optimisation scheme by incorporating the incompressibility constraint was developed for the restoration of myocardial velocity data. The method allowed accurate assessment of both in-plane and through-plan strain rates, as demonstrated with both synthetic and in vivo data acquired from normal subjects and ischaemic patients. To further enhance the clinical potential of the technique and facilitate the visual assessment of contractile abnormality with myocardial velocity mapping, a complementary analysis framework, named Virtual Tagging, has been developed. The method used velocity data in all directions combined with a finite element mesh incorporating geometrical and physical constraints. The Virtual Tagging framewoik allowed velocity measurements to be used for calculating strain distribution within the 3D volume. It also permitted easy visualisation of the displacement of the tissue, akin to traditional CMR tagging. Detailed validation of the technique is provided, which involves both numerical simulation and in vitro phantom experiments. The main contribution of this thesis is in the improvement of the effectiveness and quality of quantitative myocardial contractility analysis from both sequence design and medical image computing perspectives. It is aimed at providing a sensitive means of detecting subtle as well as gross changes in contractile behaviour of the myocardium. The study is expected to provide a clinically viable platform for functional correlation with other functional measures such as myocardial perfusion and diffusion, and to serve as an aid for further understanding of the links between intrinsic