- 20 views
- 23 downloads
Magnetic Susceptibility Separation in Brain
- Author / Creator
- Hamidi Esfahani, Javad
Quantitative susceptibility mapping (QSM) has become an area of intense interest in brain imaging due to its ability to quantify tissue magnetic susceptibility. QSM deploys a gradient echo sequence to produce phase images that are sensitive to susceptibility differences, which can then be processed into a susceptibility map. QSM exhibits promising performance in deep gray matter where iron-rich regions exist. Iron is paramagnetic and appears in QSM results as a positive source. In addition, QSM showed sensitivity to myelin due to its diamagnetic susceptibility. Myelin is abundant in white matter tracts, for example. Thus, QSM has been applied to study changes in iron and myelin in the brain. Of particular interest is the demyelinating disease multiple sclerosis (MS). However, in the same voxel, changes in both iron and myelin can occur which have confounding effects on the resulting mean susceptibility. For example, a loss of myelin or a gain in iron could have similar increases on the net susceptibility.
The challenge for conventional QSM has been distinguishing iron deposition (increasing paramagnetic component) from demyelination (decreasing diamagnetic component). Recently susceptibility separation methods were introduced as a means to solve the problem. The methods incorporate the reversible component of transverse relaxation known as R2’ and assume diamagnetic and paramagnetic dephasing effects have similar contributions to R2’, while they have opposing effects on the phase used for QSM. Utilizing both R2’ and QSM it becomes possible to separate the dia- and paramagnetic components. However, the susceptibility separation methods have only recently been introduced and there is no knowledge on the reliability of such methods. Particularly for longitudinal studies, high repeatability is necessary for any biomarker.
In this work, background information to understand susceptibility separation is introduced in Chapter 1. Then in Chapter 2, a study of scan-rescan repeatability is performed on healthy subjects. Optimal parameter choices for susceptibility separation are first investigated. In particular, variation of the relaxometric coefficient (Dr) was assessed in deep gray matter and white matter. This coefficient determines the relative weights of R2’ with respect to absolute susceptibility values, and has typically been kept constant. The results show the optimal Dr value tends to be larger for dominant component and smaller for minor component e.g. in deep gray matter where strong paramagnetic sources exist Dr is larger for paramagnetic component than diamagnetic. With optimal choices, the repeatability of the method is examined using Bland-Altman plots, scan-rescan correlations and interclass correlation coefficients (ICC). The results suggest that separation methods are not as repeatable as conventional QSM in both deep gray matter and white matter. Conventional QSM had the largest average ICC scores of 0.97 and 0.88 in deep gray matter and white matter respectively. The separated paramagnetic component had ICC of 0.85 and 0.64 in deep gray matter and white matter, respectively. Diamagnetic components had the least average ICC scores of 0.45 and 0.45 in deep gray matter and white matter. However, the separation methods do offer qualitative distinction of paramagnetic and diamagnetic components. Two limitations were found. First, the R2’ map is vulnerable to noise, thus its combination use with conventional QSM may propagate the noise into the separation maps. Second, the combination of multiple scans requires exact registration which is difficult to achieve for lower resolution imaging.
In Chapter 3, the conclusions of this thesis are discussed along with limitations and directions for the future. Overall susceptibility separation is an exciting new research field that is still in its infancy. Our work has illustrated both its current flaws and its promise for future examination of myelin and iron in the human brain.
- Graduation date
- Spring 2022
- Type of Item
- Master of Science
- 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.