To differentiate pyogenic brain abscess from other brain diseases such as necrotic glioblastomas is very important for clinic treatment. Cytosolic animo acids, lactate, alanine, succinate and acetate have been recognized as potential abscess markers. LCModel is a well-known tool to analyze the MRS data, as it provides opportunity of quantitative of metabolite concentration. Using MRS with LCModel to identify and quantitate these metabolites would benefit more precisely noninvasive diagnosis and treatment of pyogenic brain abscess. However, to differentiate the MR spectra of strongly overlapping metabolites are not easy. In this study, we validate the accuracy of LCModel on detecting these overlapping metabolites. We use some GAVA-simulated resonance spectra as our input signals and figure out the performance of LCModel analysis in different conditions. Our goal is to find an optimal analysis method to help the clinic diagnosis of abscess patients. Our result shows that the determination of basis sets is very important since the analyzed result might be different due to the improper selection of basis sets.

Phased array coils are composed of several surface coils receiving individual element signals simultaneously. Each individual surface coil provides the equivalent of the coil diameter range, and higher SNR. Therefore, combining these non-interactive phased array coils, can achieve a wide range of scan areas, uniform sensitivity and better SNR. Therefore our experiment was performed with two different coils of quadrature coil and phased array coil. Phased array MRS data were compared using various combination approaches. Data acquired by quadrature coil was regarded as a standard to verify the reliability and accuracy of metabolite concentration. The aim of our study is to do eddy current correction before and after the combination of each element coil data with LCModel analysis for quantitative comparison of metabolite concentrations. Our result shows that doing eddy current correction for each phased array coil before signal combination can achieve higher reliability and accuracy of SNR and quantitative concentrations of MR spectra in vivo.

Using Graphic Processor Unit (GPU) to process the parallel operation via Compute Unified Device Architecture (CUDA) is a new technology in recent years. In the past, the GPU has been used in parallel operation but it was not easy for programming so that it couldn¡¦t be widely used in applications. CUDA is the newly-developed environment based on C language mainly for improving the complexity in programming with CUDA. The applications of GPU with CUDA has been expending to various fields gradually due to support of IEEE floating point as well as its lower cost in hardware while comparing to the super computers. Magnetic Resonance Spectroscopy (MRS) has the feature of non-invasive to probe the concentration distributed of metabolites in vivo. It can assist doctor in clinical diagnosis. The Magnetic Resonance Spectroscopy Imaging (MRSI) is imaging by many Signal Voxel Spectroscopy (SVS) to become multi-dimension MRS image. In MRSI, it can offer more information than SVS. CUDA are applied to MR image widely such as accelerating the image reconstruction and promoting the image quality, but in MRS it is seldom for the related application. In this paper, we using the CUDA to applied in MRS, the MRSI data pre-processing, to accelerate the spatial location in MRSI. In this work, we firstly use random data with different dimensions: 1D (one-dimension), 2D and 3D to evaluate the performance of Fourier transformation by using CUDA. We also finally apply some GE 2D/3D MRSI data to see how the acceleration of using CUDA works. Our results show that the acceleration rate of Fastest Fourier Transform (FFT) with CUDA in 1D, 2D and 3D random data largely increases as the data size increases. In the experiment of 2D/3D MRSI data, we find that using CUDA for accelerating the MRSI RAW-file generating procedure would avoid the data moving times, and it is not good for CUDA 1D FFT with parallel architecture while too small data amount processing in kernel. Therefore, how to solve the relationship between MRSI data format with CUDA FFT library and how to decrease the data moving time will discuss in the study.

Conventional magnetic resonance (MR) imaging has been widely applied to clinical analysis studies due to its non-invasive property. Proton MR spectroscopy complements conventional MR imaging by enabling better lesion characterization. Thus, proton MR spectroscopy is used to assist in the differential diagnosis of intracranial pathologies. LCModel is a reliable and user-friendly post-processing tool which is used to analyse absolute concentrations in our thesis. Our phantom are solution of alanine (Ala), cytosolic amino acids (AAs), lactate (Lac), and n-acetyl aspartate (NAA) in a spherical flasks of glass. We used three basis sets with difference echo time (TE) to experiment. We also performed a retrospective study of subjects with brain abscesses referred during a span of 10 years. All subjects underwent conventional MR imaging and in vivo proton MR spectroscopy, and subjects are classified four groups according to the spectrum characteristics described in the literatures. In this thesis, phantom experiments as well as GAVA simulation are included for the basis sets comparison. Then, abscesses subjects are analyzed by LCModel using these basis sets and compared with clinical diagnosis. Our result shows that using GAVA simulation as the basis sets may provide better consistency among all metabolites and thus achieve more reliable quantification of magnetic resonance spectroscopy.

Magnetic resonance spectroscopy has been widely used in medical applications, rendering precise evaluation and diagnosis in clinics. As the development of various tools for automatic spectra analysis, providing objective quantification of metabolites, absolute concentrations has been playing an important role in clinical studies and applications as well. In this study, we investigate the reliability and accuracy of absolute concentration quantified by LCModel. Ten healthy subjects were included. We compared the resultant concentrations calculated by internal water scaling and phantom calibration, both of which are provided by LCModel. Partial volume effect was also taken into account to improve the accuracy of absolute concentrations. Automatic segmentation was applied to volume of interest in order to separate gray matter and white matter, which will facilitate the further partial volume correction and thus better accuracy of absolute quantification.

Magnetic Resonance Spectroscopy (MRS) can be applied to probe noninvasively the concentrations and distribution of metabolites of human tissue in vivo. As the improving of hardware and localization techniques, MRS becomes more and more important in clinical applications. Furthermore, some post-processing software, like LCModel, provide a graphical user interface for efficient and convenient analysis of MR spectroscopic imaging and thus increase the value of MRS applications. Although LCModel provides an efficient analysis and produces stable results, it can not provide metabolite map to observe the distribution of metabolite concentrations. For this reason our study processes the output data of LCModel and Digital Imaging and Communications in Medicine (DICOM) format MR images for 2D/3D metabolite map displaying. Users can use this software to observe the metabolic distribution in AP, SI and RL slice of brain tissue. In the meanwhile, as the absolute quantification of MRS has played more and more important role in clinical applications, this study also provides the LCModel end users an easy way for interpretation.

Magnetic resonance spectroscopy is now in widespread use, which with various tools of spectra analysis can provide concentrations of metabolites. The influence of metabolites on human physiology is greatly. Due to the tiny variation of the concentration in various metabolites, the analytic method used in the quantitative determination of the absolute concentrations of metabolites plays an important role in this research area. In this thesis we present an analysis tool for segmentation of white matter, gray matte and cerebrospinal fluid using region growing with spatial space, and provide manual interaction for exception handling in this subject. Then we use this tool to analyze different percentages of white matter and gray matter with the default parameter by LCModel and correct partial volume effect. The results show that the proposed tool can improve significantly the accuracy in absolute quantitative analysis of concentration.

Thesis (Ph. D.)--University of Hong Kong, 2010. / Includes bibliographical references (leaves 338-340). Also available in print.

Thesis (Ph. D.)--University of Hong Kong, 2010. / Includes bibliographical references (leaves 338-340). Also available online.

Thesis (Ph. D.)--University of Texas at Austin, 2000. / Vita. Includes bibliographical references (leaves 126-139). Available also in a digital version from Dissertation Abstracts.