| Part I The Application Research of Three-dimensional Arterial Spin Labeling MR Perfusion Imaging in the Evaluation of Brain TumorsObjective To evaluate the diagnostic performance of three-dimensional arterial spin labeling(3D ASL) MR perfusion imaging technology in different brain tumor quantitative analysis and imaging characterization.Materials and Methods Use GE Discovery7503.OT superconductive MRI scanner, collection57cases of brain tumor patients from January2013to January2015for MRI scans and three-dimensional arterial spin labeling (3D ASL) MR perfusion imaging,28patients with glioma sweep plus dynamic susceptibility contrast imaging (Dynamic Susceptibility Contrast, DSC) MRI perfusion imaging. All primary brain tumor cases were confirmed by surgical pathology, Part metastases were diagnosed according to the diagnosis of primary tumor, including16Gliomas(grade Ⅲ-Ⅳ),12Gliomas(grade I-II),12Metastases,3Choroid plexus papilloma,3Hemangioblastoma,3Cholesteatoma,1Medulloblastoma,1Lymphoma. By analyzing the different brain tumor imaging and3D ASL sequence parameter map feature. In different brain tumor perfusion parameters resulting figure, a substantial portion of the tumor were measured maximum tumor blood flow (the maximal tumor blood flow, TBFmax) and contralateral white matter, gray matter and blood flow to the cerebral hemispheres (cerebral blood flow, CBF, using paired samples t test, taking a=0.05significance level for statistical analysis to P<0.05was considered statistically significant,3D ASL and DSC for both technologies in glioma surgery before income average maximum relative cerebral blood flow (rCBFmax) the ratio of the line of linear regression analysis, to understand its relevance. Draw the receiver operating characteristic curve (receiver operating characteristic curve, ROC curve), low-grade gliomas to determine the optimal classification threshold and calculating the sensitivity, specificity and accuracy. Finally, paired comparison test X2conventional MRI sequences and3D ASL perfusion imaging techniques combined diagnosis of brain tumors and grading consistent rate.Results (1) Perfusion index a low-grade glioma tumor solid areas were lower than the other groups, and the difference was statistically significant (P<0.05), than among the groups, the difference was not significant, including high levels of glial tumors, meningiomas, brain metastases group, the average TBFmax/contra lateral hemisphere, TBFmax/contra lateral gray matter, each ratio TBFmax/contra lateral white matter respectively:(3.30±0.86,3.01±1.2,3.81±1.26;3.55±0.47,2.73±0.44,3.85±0.95;3.29±0.64,2.71±0.34,3.35±0.85).(2)28cases of glioma patients,3D ASL and DSC two technologies resulting in low perfusion index, were statistically significant (P<0.05), but the comparison between the two techniques of high-grade gliomas, the difference was not statistically significant (P>0.05), the two techniques of solid tumor perfusion index area TBFmax/contra lateral hemisphere, TBFmax/contra lateral gray matter, TBFmax/contra lateral white matter ratio for each line of linear regression analysis showed a significant positive correlation, the correlation coefficient respectively:r=0.953, r=0.892, r=0.781.(3)3D ASL perfusion imaging perfusion index TBFmax/contralateral hemisphere in solid tumor area, TBFmax/contra lateral gray matter, each ratio TBFmax/contra lateral white matter in distinguishing high glioma, the area under the ROC curve between low levels (AUC), respectively to0.969,0.956,0.896; which when in TBFmax/contralateral hemisphere the largest ratio (0.969), with TBFmax/contralateral hemisphere ratio>2.11for diagnostic threshold, low high-grade gliomas diagnostic sensitivity, specificity, positive predictive value, and negative predictive values were100%,91.7%,94.1%and100%.(4) DSC perfusion imaging in tumor perfusion index solid areas TBFmax/contra lateral hemisphere, TBFmax/contra lateral gray matter, each ratio TBFmax/contra lateral white matter in distinguishing high glioma, the area under the ROC curve between low levels (AUC), respectively0.948,0.943,0.904; which when in TBFmax/contralateral hemisphere the largest area ratio (0.958), with TBFmax/contralateral hemisphere ratio>1.91for the diagnostic threshold, low high-grade gliomas diagnostic sensitivity, specificity, positive predictive value, and negative predictive value were100%,83.3%,88.9%and100%.(5) Conventional MRI sequences,3D ASL perfusion imaging sequences for low and high grade glioma compared with histopathological grading results, the difference was not statistically significant (P>0.05), the combination of the two scanning methods, the sensitivity and the correct index significantly improved, respectively:93.8%and77.1%.Conclusion3D ASL and DSC measurement glioma solid areas rCBF values were well correlated,3D ASL perfusion imaging has high repeatability, completely non-invasive, etc., where the solid tumor area TBFmax/contralateral hemisphere ratio possible to distinguish high and low-grade glioma is the most sensitive parameter.3D ASL can quantitatively react tumor microcirculation perfusion with conventional MRI sequences in combination, can be used as an important supplement for brain tumor diagnosis and before surgery grade gliomas have important reference value. Part Ⅱ The Application Research of Intravoxel Incoherent Motion Diffusion-weighted MR Imaging in the Evaluation of Brain TumorsObjective To evaluate the diagnostic performance of mono-exponential model and the bi-exponential modeling of intravoxel incoherent motion(IVIM)MR imaging technology in different brain tumor quantitative analysis and imaging characterization.Materials and Methods Use GE Discovery7503.0T superconductive MRI scanner, collection60cases of brain tumor patients from January2013to January2015for MRI scans and enhanced-diffusion-weighted imaging (eDWI). All primary brain tumor cases were confirmed by surgical pathology, Part metastases were diagnosed according to the diagnosis of primary tumor. including16Gliomas(gradeⅢ-Ⅳ),12Gliomas(grade Ⅰ-Ⅱ),12Metastases,6Choroid plexus papilloma,3Hemangioblastoma,3Cholesteatoma,1Medulloblastoma,1Lymphoma. By analyzing the different brain tumor imaging and eDWI sequence parameter map feature. To measure tumor parenchyma and normal white matter of the mono-exponential model and the bi-exponential modeling ADCsmndard value、 D value f value、D value, Paired sample t test, taking α=0.05significance level for statistical analysis, with P<0.05was considered statistically significant Methods and differential diagnosis of benign and malignant brain tumors and the best low-grade glioma grading parameters by receiver operating curve (receiver operating characteristic curve, ROC),The last line of diagnostic test evaluation (sensitivity and specificity).Results:(1) Mean ADCstandard、D、f、D*had significant difference between tumors and Contralateral white matter regions,(0.81±0.14)×10-3,(0.67±0.05)×10-3,(t=73.26,P<0.001),(0.80±0.19)×l0-3,(0.64±0.13)×10-3,(t=-5.54,P<0.001),(8.47±2.30)×10-3,(5.59±1.85)×10-3,(t =-7.56,P<0.001),(28.91±11.75)×10-3,(8.20±3.57)×10-3,(t=-13.06,P<0.001), respectively.(2) Benign brain tumor diffusion coefficient D value (0.94±0.15)×10-3and ADCstandard value (0.92±0.14)×10-3were significantly higher than malignant brain tumor group (0.71±0.16,0.74±0.09)×10-3, the difference there was statistically significant (t=5.77,5.86, P <0.001), and the diffusion coefficient D values biggest difference between benign and malignant brain tumors; benign brain tumor perfusion coefficient D*is greater than the malignant brain tumor perfusion fraction f value is less than the vicious, suggesting benign microcirculation scores less than cancer, but the difference was not statistically significant (t=1.75,-0.96, P>0.05). Correlations were compared between benign and malignant brain tumor group ADCstandard, D, f, D*parameter values found in both groups, D and D*value, D*values ADCStandard value differences were statistically significant (benign:t=-13.26,13.28, P<0.01; malignancy:t=-13.62,-22.15, between P<0.01), D values ADCStandard value was no significant difference (benign:t=-1.61, P>0.05; malignancy:t=1.56, P>0.05), description D values closer ADCStandard value. Low-grade gliomas Slightly lower than normal contralateral white matter, but no significant difference (P>0.05) between the two low-grade gliomas ADCstandard value and D-value, high-grade gliomas D value, f value, D*values above contralateral normal white matter, between the two was significant difference (P<0.05).(3) Conventional MRI sequences, eDWI sequence of benign diagnosis and histopathological results of malignant brain tumors, the difference was not statistically significant (P>0.05), the combination of the two scanning methods, the sensitivity and the correct index were significantly improved, respectively as follows:91.7%and93.3%.(4) Low-grade glioma group, ADCstandard value slightly lower than the contra lateral normal white matter areas, high-grade glioma group ADCstandard value slightly higher than normal contra lateral white matter, but the difference was not statistically significant (P>0.05), the D value, f value, D*values were normal contra lateral white matter differences (P<0.05).(5)Low and high grade gliomas D value, f value, D*values were significantly different (P<0.05), ADCstandard value of the difference was not statistically significant (P>0.05).(6) ADCstandard value, D value, f value and the D*value in distinguishing high glioma, the area under the ROC curve between low levels (AUC) were0.698,0.719,0.719,0.969; D*values of which the largest (0.969) to D*value>18.7threshold for the diagnosis, the diagnosis of high-grade gliomas sensitivity, specificity, positive predictive value, and negative predictive values were93.8%,83.3%,88.2%,90.9%.(7) Conventional MRI sequences, eDWI sequence of low and high grade glioma compared with the histopathologic grading results, the difference was not statistically significant (P>0.05), the combination of the two scanning methods, the sensitivity and the correct index were significantly increased, respectively:93.7%and89.3%.Conclusion Diffusion-weighted imaging sequences eDWI brain tumor diagnosis and effective complementary technologies, combined with conventional MRI scans, eDWI can improve the accuracy of diagnosis of brain tumors. and the bi-exponential modeling than the mono-exponential model to more accurately assess the level of preoperative glioma, D*value may be in distinguishing high and low-grade glioma is the most sensitive parameter. |
PDF Full Text Request