| PART 1:Application of Different b-Value in Normal Human Brain Diffusion-Weighted ImagingObjective:To research the signal attenuation of diffusion-weighted imaging at different b value, To study the effect of different b-value on signal intensity (SI), ADC and the quality of image.Materials and Methods:Three different weightings (b=1000,2000,3000 s/mm2) were applied in thirty normal human brains in diffusion-weighted imaging on 1.5T MR machine. DWI data were imported into SUN ADW 2,0 workstation and then detected the signal intensity on frontal gray matter, caudate head, lentiform nucleus, dorsal thalamus, splenium of corpus callosum, posterior limb of Internal capsule, centrum semiovale and backgound area. ADC of each pixel was computed in different conditions using the Functool. And then SNR, CNR at different b value were calculated according to those data. The relationship between these parameters and b-value was analysed.Results:The Sls of all ROIs were significantly decreased with the increase of b-value (P<0.01). From b=1000s/mm2 to 3000s/mm2, the drops of ROIs in gray matter were between 73.2%-80.0%, it was most obvious in caudate head, and the drops of ROIs in white matter were from 57.6% to 63.5%, it was most obvious in centrum semiovale.There were significantly abnormality between groups of background area, but in which, there was no significant abnormality between the subgroups comparsion of 2000s/mm2 and 3000s/mm2. Except lentiform nucleus, the ADC of ROIs decreased with the increase of b-value, which was statistically different compared between subgroups (p<0.05). SNR of all ROIs between groups were statistically different (P<0.05), which decreased with the increase of b-value. CNR of frontal white matter, lentiform nucleus and caudate head decrease with the increase of b-value, but there was no significant difference on CNR between internal capsule and centrum semiovale from b=2000s/mm2 to 3000s/mm2, while CNR of splenium of corpus callosum was lowest at b=2000s/mm2.Conclusion:The strength of b-value influenced measurement of SI, ADC and the quality of image, which should be considered on clinical research.PART 2:The Construction and Application of Biexponential Model on High b-value Diffusion-Weighted ImagingObjective:Using the biexponential model, to research the decaying regulation of DWI signal intensity and study the human brain diffusion information.Materials and Methods:We used the function lsqcurvefit() in MATLAB optimization toolbox to construct a biexponential model, and used emulated data to validate the reliability of the new model. Eight different diffusion factors were applied to diffusion-weighted imaging on 6 normal adult at the direct of x, y, z. Eight different diffusion-sensitive factor were 0,500,1000,1500,2000, 2500,3000,3500s/mm2. We used the biexponential model to analyse the signal attenuation, based on each pixel of lentiform nucleus, posterior limb of internal capsule, dorsal thalamus, frontal gray matter, and got the fitting residual and three new parameters, Df, Ds, As. The bixeponential fitting result was compared with that of the monoexponential model.Results:The biexponential model was reliable, which was confirmed by the emulated data. The results of the biexponential fitting for four ROIs were (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 fitting residuals of monoexponential model were 0.232,0.431,0.635,0.323. From the data fitting, the biexponential fitting model was better than the monoexponential one.Conclusion:Three new parameters were obtained by the biexponential model, which not only descripted the diffusion speed, but also provided the ratio of two component.The biexponential model was more fit to the signal intensity at high-b-value.PART 3:The comprehensive application research of Multi b-Value Diffusion-weighted ImagingObjective:Using multi b-value diffusion-weighted imaging data, a new diffusion parameter (Rd) was constructed. Quantitative analysis of Rd in different brain structures was conducted. According to comparing SNR and CNR of Rd map, ADC1000 map and ADC3000 map, the preliminary evaluation of Rd for the clinical feasibility and superiority was conduct.Materials and methods:Eight different diffusion factors were applied to diffusion-weighted imaging on 25 normal adult at the direct of x, y, z. Eight different diffusion-sensitive factor (b-value) were 0,500,1000,1500,2000, 2500,3000,3500s/mm2.22 images were obtained at each slice after examine. After image coregistration,8 different b-values and their signal intensity were applied to calculate and achieve a new diffusion parameter based on each pixel of the preliminary images. And a new parameter map (Rd map) was obtained. Rd values of caudate head, lentiform nucleus, dorsal thalamus, splenium of corpus callosum, posterior limb of internal capsule and posterior horn of lateral ventricle were measured on Rd map. The database were statistically analyzed and compared with b1000 and b3000 ADC map on SNR and CR.Results:The maximum value of Rd was at posterior horn of lateral ventricle, reaching 23.19±2.45. The comparisons between caudate nucleus, putamen, and dorsal thalamus were no significant abnormality, P>0.05. ROIs'Rd value of frontal white matter was lower than that of ROIs in cauhate head and lentiform nucleus, P<0.05. The change trend of Rd, ADC1000 and ADC3000 was similar in different ROI. Each ROI's SNR of Rd map was higher than that of ADC1000 and ADC3000 map, P<0.05. There was no significant abnormality for posterior limb of internal capsule on three parameter maps, P>0.05. CR of each ROI was lower at Rd map compared with ADC1000 and ADC3000 maps, except the posterior horn of lateral ventricle.Conclusion:The new diffusion parameter (Rd) can improve the SNR apparently, distinguish the different brain construct effectively, compared to the apparent diffusion coefficient. The new method can improve the sensitive to diffusion. |
PDF Full Text Request