Dissertation > Medicine, health > Clinical > Diagnostics > Diagnostic Imaging > Magnetic resonance imaging

The Application Research of Multi b-Value Diffusion-Weighted Imaging on Normal Brain

Author LeiZhengYong
Tutor FengXiaoYuan
School Fudan University
Course Medical Informatics
Keywords Magnetic resonance imaging Dispersion b values Brain Diffusion-weighted imaging Double exponential signal attenuation High b-value
CLC R445.2
Type Master's thesis
Year 2010
Downloads 51
Quotes 0
Download Dissertation

Application purpose of the first part of different b-value diffusion-weighted imaging in the normal human brain: study different b-value diffusion-weighted imaging (DWI) signal attenuation law to explore different b-value diffusion-weighted imaging signal intensity, apparent diffusion coefficient and image quality the impact. Materials and Methods: Using three different values ??of b (1000,2000,3000 s/mm2) 1.5T MR diffusion-weighted imaging MRI scans of 30 normal adults DWI data incoming the SUN ADW2.0 workstation, measuring the amount of lobe cortex, caudate nucleus head, putamen, dorsal thalamus, corpus callosum pressure portion hind legs, semi-oval center and the signal strength of the background area, internal capsule, calculated use of Functool software to get the ADC value of each pixel in different b values . Based on the above data to calculate different b value image signal-to-noise ratio and contrast to noise ratio, and analysis of the strength of b values ??of these two indicators. Results: DWI signal intensity group for each region of interest compared with a statistically significant difference (P lt; 0.01) subgroups pairwise comparisons were statistically significant difference (P lt; 0.05), along with the increase in the value of b significantly reduced; b = 1000s/mm2 to 3000s/mm2 various gray matter regions of interest signal intensity fell between 73.2% -80.0% to the head of the caudate nucleus, the most obvious decline in signal intensity in the white matter ROIs 57.6% - 63.5%, semi-oval center is the most obvious; intensity of the background noise is a significant difference between the two groups (P lt; 0.05), but no statistically significant difference between the b = 2000s/mm2, and 3000s/mm2 subgroup (P gt; 0.05). Addition to the lenticular nucleus, ADC values ??of each region of interest groups compare with a significant difference between any two of the subgroups were statistically significant difference (p lt; 0.05), were reduced with the increase in the value of b. All ROIs SNR group compared with a statistically significant difference (P lt; 0.05), in addition to pressure portion of the corpus callosum, the rest of the region of interest between subgroups pairwise comparisons were statistically significant (P lt; 0.05), With the increase of the b value lower. Frontal cortex, putamen, caudate nucleus head contrast to noise ratio decreases with the increase in the value of b (P lt; 0.05); within PLIC and the centrum semiovale contrast noise than b = 2000s/mm2 and 3000s / mm2 subgroups no statistically significant difference (P gt; 0.05); corpus callosum in b = 2000s/mm2 contrast to noise ratio minimum. Conclusion: DWI signal intensity, ADC values ??and image quality is governed by b value strength, to fully consider the strength of the b-value of each index measured in the conduct of clinical research. High b-value of the second part of the double exponential diffusion-weighted imaging analysis model and application purposes: to build a high b-value analysis model of the double exponential diffusion-weighted imaging, to explore the high-b-value diffusion-weighted imaging signal attenuation law reveal brain tissue diffusion . Materials and Methods: Using MATLAB optimization toolbox Isqcurvefit () function to create a bi-exponential analysis model, using a set of simulation data model, the credibility of the verification. 6 patients without symptoms of nervous system and other systems with a history of normal adults randomly selected as the object of study, diffusion-weighted imaging diffusion gradient is applied in the x, y, z directions, respectively, eight different b values ??were 0,500,1000 1500,2000,2500,3000,3500 s/mm2. Denoising, registration of image pre-processing of the lenticular nucleus, dorsal thalamus, frontal white matter, internal capsule each pixel of the hind legs and other areas of interest, the use of a double exponential analysis model to analyze and get fit the residual χbi2 and three new dispersion parameters, namely the fast diffusion coefficient (Df), slow diffusion coefficient (Ds) and slow diffusion component share the proportion (As), and then fitting the double exponential and single exponential analysis model residuals for statistical analysis. The results: double exponential model simulation data to verify the double exponential analysis model confidence level is estimated to be 0.8715, 0.0460 risk factor model credible. Double exponential analysis model of the lenticular nucleus, dorsal thalamus, frontal white matter, PLIC analysis of the four regions of interest (Df, D, As, χbi2) (1.53,0.56,0.58 0.002) (1.28,0.24,0.35,0.001), (1.24,0.22,0.30,0.005), and (1.54,0.53,0.65,0.002). The residuals were four regions of interest were analyzed by single exponential analysis model 0.232,0.431,0.635,0.323. The four ROIs fitting residuals statistical analysis of the results shows that the residuals of the two models have a statistically significant difference (p lt; 0.05). Conclusion: The three double exponential analysis model parameter Df, Ds and As, not only quantitatively describe the fast and slow two Dispersion ingredients of diffusion speed and quantitative description of the respective proportion of both diffuse component. The the residual statistical analysis of the results of the two model fitting fit double exponential analysis model of diffusion-weighted imaging data analysis than the single exponential model more fitting residual smaller third part of the multi-b-value diffusion-weighted imaging. Comprehensive Application Research Objective: multi-b-value diffusion-weighted imaging data to build a new dispersion parameters Rd, quantitative analysis of different brain structure the Rd value distribution and quantitative image quality evaluation Rd, this preliminary evaluation Rd for clinical feasibility and superiority. Materials and methods: the use of eight different values ??of b on the 25 normal adult head in the x, y, z directions are respectively applied to the diffusion gradient diffusion weighted imaging, the eight different values ??of b, respectively 0,500,1000,1500 , 2000,2500,3000,3500 s/mm2, the scan after the end of each level, a total of 22 images. , Based on each of the pixels of the image after the image registration, with eight different b value of signal intensity values ??is calculated and to come to a new dispersion parameter map (Rd charts), and then were measured on the RD FIG caudate nucleus head, putamen, dorsal thalamus, corpus callosum, frontal white matter, lateral ventricle angle Rd value of the area of ??interest, and statistically analyzed. Rd graph quality evaluation using a signal-to-noise ratio and contrast, and were statistically analyzed with the two-point method the calculated ADC1ooo and ADC3000 Figure. Results: In 25 normal brain different ROI Rd of value statistics the lateral ventricle angle Rd highest, reaching 23.19 ± 2.45; Rd between the head of caudate nucleus, lenticular nucleus and dorsal thalamus pairwise comparisons without obvious statistical difference (P gt; 0.05); frontal white matter Rd value than the head of caudate nucleus and lentiform nuclear (P lt; 0.05); corpus callosum body pressure the Ministry of the Rd value than the head of caudate nucleus and lentiform nuclear low (P lt; 0.05). Rd, ADC1000 and ADC3000 trends between different ROI. Different parameters of image SNR statistics Rd Figure ROI of the signal-to-noise ratio is significantly higher than the ADC1000 and ADC3000 (P lt; 0.05); statistical analysis of the CR, the lateral ventricle posterior horn of CR in the three diffusion parameters figure was not statistically difference (P gt; 0.05), the contrast of the remaining ROI in RD diagram contrast (P LT; 0.05) lower than that in the diagram ADC1ooo and ADC3000. Conclusion: Rd and apparent diffusion coefficient, effective response of the dispersion of the different parts of the brain Rd diagram than ADC map has a better signal-to-noise ratio of the image to improve the sensitivity of the movement of water molecules diffuse.

Related Dissertations
More Dissertations