Myelin Water Imaging Using Transverse Relaxation

  • Author / Creator
    Nima Mehdizadeh
  • Tracking myelination and demyelination in the brain is a crucial part of studies on neurodegenerative diseases such as Multiple Sclerosis (MS). Classic MRI techniques are unable to provide enough information about the pathology of tracts and lesions. Multi-component analysis of multi-echo data has shown promising results in the field of myelin water imaging (MWI). Such analyses are predicated on the assumption that multi-echo signal acquired from a specific voxel in the brain can be sourced to various water compartments, which go through relaxation at different paces independently. Numerous techniques have been developed to model MRI signals to acquire the proportion of water compartments to track myelin content. This ratio is typically called the myelin water fraction (MWF) value. In this thesis, we implemented multi-component analyses for MWF from multi-echo gradient or spin echo sequences based on past literature. Once implemented, we investigate advances in analysis for each method. Using multi-echo spin-echo (MESE), constraining the transmit RF field B1+ parameter for the process of MWI analysis was investigated by supplying B1+ using an independent mapping sequence. This new approach was compared to the standard MWI method of estimating the B1+ from the decay curve. Simulations were performed using the same principles and modeling prescribed for the fitting procedure to analyze our proposition. By comparing the estimation results and reference data, we found a notable difference which resulted in significant changes in MWF maps using the proposed method. In 3D MESE data, approximated by a 3D gradient and spin echo (GRASE) sequence, MWF values were generally underestimated in white matter regions using the standard method. Similar patterns were observed in 2D MESE data, however, estimations in the in-vivo results showed major underestimations where B1 + was more than 1.05; and we noticed stronger deviation from the reference B1+ maps in 2D MESE compared to 3D data. The MWF brain maps from the standard method were in alignment with estimated maps demonstrated in previous literature. Results were partly in alignment with simulations, but the in-vivo estimations showed skewness and bias in the optimization step that is responsible for estimation of B1+; which could be due to the artifacts present in MRI data. In short, supplying the B1+ parameter via an independent flip angle map was found to offer improvements in both 2D and 3D MWF methods. The major improvement was supplying greater certainty in this key parameter, rather than attempting estimation from the MWI sequence that may be limited by SNR, artifacts or duplicate B1+ solutions. For MWI from multi-echo gradient echo (MEGRE) data, standard two and three pool analysis models were implemented and tested in human brain subjects. The effects of previously-introduced spatial saturation were examined and a non-local filter to overcome low SNR issues was proposed. The application of a saturation pulse in ME-GRE data acquisition helped with reducing the physiological noise arising from the blood inflow as previously suggested by the literature. Using the non-local filter showed improvements in image quality of MWF maps, and the correlation coefficient of ME-GRE results compared to the reference method increased when using the three-pool model. In conclusion, this thesis work has implemented spin echo and gradient echo-based analysis methods for myelin water imaging in human brain on a 3T system. Advances in MESE-type approaches introduced an independent flip angle (B1+) map to constrain solutions for both 2D and 3D sequences. MEGRE methods were briefly studied as well including applying a non-local filter.

  • Subjects / Keywords
  • Graduation date
    Fall 2021
  • Type of Item
  • Degree
    Master of Science
  • DOI
  • License
    This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.