As accumulating post-mortem evidence indicates the involvement of iron dyshomeostasis in neurodegeneration, the need to measure iron levels in the living human brain is more urgent than ever before. Magnetic resonance imaging (MRI) provides a non-invasive tool to detect brain iron in clinical and preclinical settings, by exploiting the magnetic properties of tissue iron using different pulse sequences. However, the question remains open, as to whether existing quantitative MRI techniques can be used as a robust technique to measure iron in a complex biological system like brains. Therefore, this PhD thesis seeks to extend this understanding by assessing existing techniques and developing original quantitative methods for improved iron measurement in clinical and preclinical MRI. After briefly revising the efforts that have been made in the literature to quantify brain iron using quantitative MRI, the original contribution of the PhD thesis is composed of three experiments:
The first project was to investigate whether the measurement derived from the 3.0 T dual-contrast fast-spin-echo (FSE) MRI, termed ‘effective R2’ in this work, may be used as a quantitative MRI method, an alternative to the time-consuming conventional transverse relaxation rate (R2) to evaluate the iron level in healthy (HC) and Parkinson’s disease (PD) brains. Retrospective 3.0 T FSE in vivo MRI data from case-control HC and PD subjects were selected from the Parkinson’s Progression Markers Initiative (PPMI) database, and the effective R2 was calculated from exponential fitting of these data. Linear regression analysis was then performed between the effective R2 and independently-estimated brain iron concentration (derived from the subject’s age using the empirical age-dependent formulae reported by Hallgren and Sourander). To further investigate its potential clinical use, the effective R2 was compared between groups, and linear correlation analysis was performed between the effective R2 and the functional dopamine transporter (DaT) results. The findings of the project suggested a strong linear correlation between the effective R2 and the estimated brain iron concentration, as well as a strong correlation between the putaminal effective R2 and the DaT dysfunction in PD. Therefore, it can be concluded the effective R2 may be used as a fast, quantitative XI MRI method to aid the evaluation of iron and DaT functions with 3.0 T clinical MRI, as an alternative to conventional R2.

To exceed the constraints of R2 and effective R2, in this subsequent research, the state-of-art quantitative susceptibility mapping (QSM) MRI technique was assessed for measuring iron concentration, at the local NHS Trust hospital. This project aimed to answer if a QSM method using nonlinear morphology enabled dipole inversion with L1-regularisation (nMEDI-L1) can quantify iron in ferric iron (Fe3+) phantoms and human brains accurately, using a local clinical 3.0 T MRI scanner. Using a 3D gradient-echo (GRE) sequence, the images of phantoms and healthy volunteers were obtained, and processed using the nMEDI-L1 QSM method. The resulting susceptibility images of human brains were compared with routine clinical anatomical scans, and with the (effective) transverse relaxation rates (R2, R2*, effective R2) mapping. Linear regression analysis was performed between the susceptibility measured in the phantom and the Fe3+ concentration, and between the brain susceptibility and the brain iron concentration estimated using Hallgren and Sourander’s formulae. A very strong linear correlation was found between the susceptibility and the Fe3+ concentration in the phantom, with the Fe3+-associated susceptibility increase that excellently matched the theoretical prediction of 1.30 ppb/μg·g-1. As the results showed enhanced iron contrast in the QSM images, compared with the routine anatomical scan, a significant, strong, positive linear correlation between the susceptibility and the estimated iron concentration were also observed, with a smaller iron-associated susceptibility increase, compared to the phantom. The result is consistent with the literature, supporting the assertion that QSM may be used to measure iron concentration.

Lastly, to investigate the preclinical value of ultra-high-resolution QSM, the scope of QSM for measuring ferritin-bound iron and amyloid-β (Aβ), a misfolded protein strongly associated with Alzheimer’s disease (AD) pathology, is studied in vitro, with 9.4 T preclinical MRI. Assessment of a selection of phase-processing techniques was performed, and automatic parameter optimisation techniques were tested for the nMEDI-L1, to optimise QSM for 9.4 T small-animal MRI system for the first time. R2* XII and QSM of ferritin, Aβ, and Aβ+ferritin aggregates were measured in vitro, and linear regression analysis was performed between the ferritin (Aβ) level and the R2* and susceptibility. Linear regression analysis was also performed between pixel-wise R2* and susceptibility. An optimal QSM pipeline was devised after assessments. The results show ferritin-associated R2 and susceptibility increase, which is particularly significant in Aβ+ferritin aggregates, showing in detail how QSM offers scope as a sensitive tool to detect iron-laden Aβ aggregates at ultra-high-field MRI.