by Layla Tabea Riemann
Abstract:
In vivo magnetic resonance spectroscopy (MRS) is an important research tool to gain a deeper understanding of the biochemical processes underlying neurodegenerative or psychiatric diseases such as the Alzheimer’s disease or depression, respectively. However, it is not yet integrated into the clinical routine as a diagnostic tool since there are still some obstacles to overcome, such as 1) missing estimation of measurement uncertainties, and a lack of understanding of the influence of acquisition parameters on the precision, as well as 2) a difficult and time-consuming application which causes patient discomfort, high costs and limits the possible range of applications, e.g., for functional MRS questions. The focus of this thesis will be to address the two aforementioned challenges. To tackle the first challenge, different contributions of reproducibility and repeatability on the measurement uncertainty were evaluated by introducing a study design and statistical analysis framework to determine the minimally detectable changes (MDCs) of the investigated brain metabolites. These MDCs were then compared to the commonly used stand-in for measurement uncertainties, the Cramér-Rao lower bounds (CRLBs), which only represent a fraction of the full measurement uncertainty. As an example of a potential influence on the precision by acquisition parameters, the impact of the choice of a specific pulse within the employed sequence was investigated. Details on the determination of the measurement uncertainty and influences thereupon can be found in chapter 2. In chapter 3, the second challenge was addressed by the extension and thereby acceleration of single-voxel spectroscopy (SVS) to simultaneously obtain the spectroscopic signal of multiple spatially distinct regions by the introduction of the two spin-echo, full intensity acquired localized (2SPECIAL) sequence. To simultaneously acquire two voxels, the multi-band (MB) technique was applied, which allows the simultaneous excitation or inversion of two spatially distinct frequency bands. Here, the results from the investigation on the influence of the choice of the radio-frequency pulse from chapter 2 were utilized, and a pulse was chosen that was shown to not negatively affect the precision of the measurement but provides optimized properties regarding the MB requirements. Part of this challenge is to retrospectively decompose and reassign the acquired signals to their region of origin. To this end, a new decomposition algorithm – voxel generalized autocalibrating partial parallel acquisition (vGRAPPA) – was introduced and its performance was rigorously compared to a previously existing decomposition algorithm based on the sensitivity encoding (SENSE) technique. Both parts of this work are important pieces on the way of understanding and overcoming issues to make in vivo brain MRS faster and more precise to ultimately allow the transition of this high potential research tool towards clinical diagnostic application.
Reference:
Towards faster and more precise MR spectroscopy at 7 T (Layla Tabea Riemann), PhD thesis, Otto-von-Guericke-Universität Magdeburg, Fakultät für Elektro- und Informationstechnik, 2023.
Bibtex Entry:
@phdthesis{riemann_towards_2023,
	title = {Towards faster and more precise {MR} spectroscopy at 7 {T}},
	url = {http://dx.doi.org/10.25673/103382},
	abstract = {In vivo magnetic resonance spectroscopy (MRS) is an important research tool to gain a deeper understanding of the biochemical processes underlying neurodegenerative or psychiatric diseases such as the Alzheimer’s disease or depression, respectively. However, it is not yet integrated into the clinical routine as a diagnostic tool since there are still some obstacles to overcome, such as 1) missing estimation of measurement uncertainties, and a lack of understanding of the influence of acquisition parameters on the precision, as well as 2) a difficult and time-consuming application which causes patient discomfort, high costs and limits the possible range of applications, e.g., for functional MRS questions. The focus of this thesis will be to address the two aforementioned challenges. To tackle the first challenge, different contributions of reproducibility and repeatability on the measurement uncertainty were evaluated by introducing a study design and statistical analysis framework to determine the minimally detectable changes (MDCs) of the investigated brain metabolites. These MDCs were then compared to the commonly used stand-in for measurement uncertainties, the Cramér-Rao lower bounds (CRLBs), which only represent a fraction of the full measurement uncertainty. As an example of a potential influence on the precision by acquisition parameters, the impact of the choice of a specific pulse within the employed sequence was investigated. Details on the determination of the measurement uncertainty and influences thereupon can be found in chapter 2. In chapter 3, the second challenge was addressed by the extension and thereby acceleration of single-voxel spectroscopy (SVS) to simultaneously obtain the spectroscopic signal of multiple spatially distinct regions by the introduction of the two spin-echo, full intensity acquired localized (2SPECIAL) sequence. To simultaneously acquire two voxels, the multi-band (MB) technique was applied, which allows the simultaneous excitation or inversion of two spatially distinct frequency bands. Here, the results from the investigation on the influence of the choice of the radio-frequency pulse from chapter 2 were utilized, and a pulse was chosen that was shown to not negatively affect the precision of the measurement but provides optimized properties regarding the MB requirements. Part of this challenge is to retrospectively decompose and reassign the acquired signals to their region of origin. To this end, a new decomposition algorithm – voxel generalized autocalibrating partial parallel acquisition (vGRAPPA) – was introduced and its performance was rigorously compared to a previously existing decomposition algorithm based on the sensitivity encoding (SENSE) technique. Both parts of this work are important pieces on the way of understanding and overcoming issues to make in vivo brain MRS faster and more precise to ultimately allow the transition of this high potential research tool towards clinical diagnostic application.},
	language = {Englisch},
	school = {Otto-von-Guericke-Universität Magdeburg, Fakultät für Elektro- und Informationstechnik},
	author = {Riemann, Layla Tabea},
	year = {2023}
}