| Achoimre: | <p>Quantitative MRI has promise to provide precise measures for assessing neurological disorders such as amyotrophic lateral sclerosis (ALS). While MRI is very sensitive to pathological changes in tissue, it lacks the specificity needed for biological interpretability. One way to address this is to compare MRI to more direct measurements, such as histopathology. In this thesis, I present methods that enable direct comparison of quantitative T2 imaging to histopathology in the same post-mortem tissue samples, with the aim of better understanding alterations to T2 in ALS. This requires that we address several challenges to acquiring T2 maps in post-mortem brains, particularly in the context of our study at 7 tesla. First, 7T MRI exhibits considerable spatial inhomogeneity of the B1 field, leading to non-exponential signal evolution. To more accurately model the signal, we utilize the extended phase graph (EPG) model and derive voxel-wise maps of B1, T2 and S0 simultaneously. We show that EPG fitting provides improved T2 mapping under inhomogeneous B1 field. A second issue for quantitative T2 mapping in post-mortem brains is the influence of formalin fixation, which can dramatically reduce the T2 value in proportion with the formalin concentration distribution. We propose a kinetic tensor (KT) model to simulate 4D formalin flux under Fick’s diffusion law. Based on the KT prediction of formalin concentration, we propose a simple linear correction of quantitative T2 maps. Combining EPG fitting with the KT correction, we demonstrate that quantitative T2 map values in motor cortex can distinguish ALS brains from the control group, and that quantitative T2 value is significantly correlated to the tissue ferritin content. This work provides evidence that quantitative T2 mapping may provide a non-invasive marker of cortical iron deposition in ALS.</p>
|
|---|