Quantitative Susceptibility Mapping Using Magnetic Resonance Imaging

Abstract

Magnetic susceptibility is an intrinsic tissue property that reflects underlying concentration of iron, calcification or contrast agents, which are useful for the investigation of a wide range of physiological or pathological conditions. Due to this promising outlook, there has been a long-standing interest in quantifying magnetic susceptibility. Although methods to quantify susceptibility of certain material samples have been proposed in the past, a practical means to measure an arbitrary susceptibility distribution in a living organism was lacking. Consequently, many of the potential applications were still in speculation. This thesis reports a framework that allows quantitative mapping of magnetic susceptibility in human brain using magnetic resonance imaging (MRI). Two major building blocks were proposed to overcome the technical hurdles. First, a background field removal method was developed to obtain the magnetic field of interest free of contamination from background sources. Second, two independent methods were proposed to solve a classical ill-posed inverse problem of determining susceptibility sources from measured magnetic field. With these technical developments, quantitative susceptibility mapping was realized. Its utility was demonstrated in a molecular MRI application, where identification and quantification of iron-based contrast agents are now feasible, and in cerebral MRI, where susceptibility provides a more objective measurement of hemorrhage, allowing cross-center comparisons and longitudinal studies.