Ultra-high field (7T+) magnetic resonance imaging (MRI) improves image quality substantially compared to normal clinical 1.5T or 3T MRI. However, 7T MRI brings new challenges pertaining to B1+ inhomogeneity, B0 inhomogeneity, increased spectral bandwidth, and RF heating.

Quantitative parametric mapping produces maps of tissue T1, T2 and other parameters in standard units rather than the weighted images normally produced in MRI. This has advantages by facilitating multi-site studies, long-term investigations, and making it easier to detect diffuse disease. At 3T, the Multi-Parametric Mapping (MPM) approach of Weiskopf [Weiskopf et al, Frontiers in Neuroscience, 2013] has been used in many neuroimaging studies, but it has been difficult to adapt for 7T. Magnetic Resonance Fingerprinting (MRF) is an alternative model-based technique to quantify multiple tissue properties after a single scan. MRF was initially created for clinical (1.5T and 3T) MRI scanners, where it has been widely applied to map T1, T2 and other parameters. MRF is more challenging at 7T due to B0 and B1+ inhomogeneity, limited peak B1+ and specific absorption rate (SAR) limits.

My thesis aims to develop the methodology for 7T MRF on the first clinically-approved model of 7T MRI: the Siemens “7T Terra”. Chapter 1 and Chapter 2 introduce my thesis and summarise key background theory. Chapter 3 reviews the theory of MRF. Chapter 4 introduces an interleaved Cartesian MRF acquisition and analysis algorithm. To my knowledge, this is the first MRF implementation on the 7T Terra platform. MRF estimates of T1, T2, PD, and B1+ for brain and abdomen applications are evaluated, emphasizing the sensitivity of the MRF technique to B1+, B0, and slice profile effects. In Chapter 5 and Chapter 6, I address the challenge of generating a model that accurately reflects experimental magnetization evolution at 7T, which is more difficult than for 3T MRF. Chapter 7 compares the challenges of MRF at 3T versus 7T. In Chapter 8 and Chapter 9, I implement two strategies to accelerate the 7T MRF sequence: GRAPPA and sliding-window radial undersampling. These are steps towards a clinically acceptable scan time for 7T MRF. Chapter 10 describes my work to understand the parallel transmit framework on our new 7T Terra scanner, culminating in a successful human in vivo liver scan utilizing an 8-channel body transmit array coil.

Overall, I have implemented key steps towards a viable 7T MRF sequence for human in vivo studies. This work will pave the way to a robust quantitative parametric mapping approach at 7T.