| Object: To investigate the clinical application value of diffusion tensor imaging(DTI) and proton magnetic resonance spectroscopy(1H-MRS) in glioma grading and invasion in peritumoral edema region.Material and Methods: Twenty pathological confirmed glioma patients (twelve for low-grade,eight for high-grade ) were examinated by head conventional MRI,DTI and 1H-MRS. Nine brain metastatic carcinoma and six meningioma patients, served as peritumoral edema control group, were also given the same examinations. The peritumoral edema can be classified into three types according to the distance from the margin of enhanced parenchyma to peritumoral edema:⑴proximal peritumoral edema: the distance less than 1cm⑵intermediated peritumoral edema: the distance from 1cm to 2cm⑶distal peritumoral edema: the distance more than 2cm, parenchyma and peritumoral edema in Non-enhanced tumor was discriminated by color spectrum metabolism map. The ratio of Cho/Cr,NAA/Cr,NAA/Cho and the average diffusion coefficient(ADC),fractional anisotropy(FA) value were measured in parenchyma, peritumoral edema region and contralateral normal brain tissue respectively. The differential analysis of Cho/Cr,NAA/Cr,NAA/Cho ratio and ADC,FA value in parenchyma and peritumoral edema was performed within and between groups, which was represented as low-grade glioma,high-grade glioma,metastatic carcinoma and meningioma groups. The critical value was defined as the reference value corresponded to the average value,median value, maximum sensitivity and maximum Youden index in ROC curve of the Cho/Cr,NAA/Cr,NAA/Cho ratio and ADC value in parenchyma of gliomas. And the histopathological result was considered as golden standard, tumors classified as high grade by critical value and confirmed by histological examination were considered true-positive findings; low-grade gliomas that were histologically confirmed as low grade were considered true-negative findings. Statistic analysis was performed by SPSS 13.0 software. All data must be checked by homogeneity test for variance before analysis, and t-test was performed in the comparison of two samples, while one-way ANOVA in multiple samples, LSD and SNK(Student-Newman-Keuls) test in each two of multiple samples. Correlation analysis was performed by Pearson and Spearman method, and matched data byχ2test. P<0.05 indicated statistical significance.Result: 1. The NAA/Cr,Cho/Cr ratio and ADC value of proximal peritumoral edema region in high grade gliomas(0.84±0.22,2.79±1.03,(1167.70±125.98)×10-6mm2/s) had significant differences with low-grade glioma(1.22±0.46,1.43±0.40,(1359.98±108.78)×10-6mm2/s),metastatic carcinoma(1.24±0.24,1.40±0.63,(1709.70±199.39)×10-6mm2/s) and meningioma(1.22±0.17,1.26±0.42,(1644.15±268.01)×10-6mm2/s)(P<0.05). and no difference was observed only in Cho/Cr ratio of high-grade glioma between parenchyma and proximal peritumoral edema region, though differences existed compared with comtralateral parenchyma and only difference was observed in ADCvalue between. proximal and intermediated peritumoral edema region.The correlation coefficient of Cho/Cr and ADC value in proximal peritumoral edema region of high-grade gliomas were r=-0.855, P<0.01; while it was r=-0.177, P>0.05 in metastatic carcinoma, which demonstrated the possibility of tumor cell invasion in proximal peritumoral edema of high-grade gliomas.2. The FA value had no statistic significance in both parenchyma and peritumoral edema within and between groups of three tumors.3. There was significant difference in NAA/Cr, Cho/Cr,NAA/Cho and ADC value between high-grade glioma and low-grade glioma parenchyma. In the correlated evaluation of tumor pathological grading with upper three parameters, obvious significance was acquired and the correlation coefficient were r=-0.690, r=0.584, r=-0.586,r=-0.665, P<0.01, which showed that these parameters can be used in the evaluation of glioma grading.4. Compared the glioma grading result based on the critical value of NAA / Cr,Cho/Cr,NAA/Cho and ADC value in parenchyma with pathology, a number of sensitivity,specificity,positive predictive value and negative predictive value were obtained. Through comparing these results and considering acquiring the greatest positive predictive value as the standard of choosing critical value, we can found that greater positive predictive value can be obtained when serving the reference value corresponded to maximum Youden index in ROC curve as critical value than conventional/enhanced MRI or other parameters, though the sensitivity and specificity may not be the maximum. And then the NAA / Cr equals to 0.77, Cho / Cr 2.77 , NAA/Cho 0.16 and ADC value 1143×10-6mm2/s, the corresponding positive predictive value were 72.7%,, 75%,100%, 87.5%, Compared the result judged by these crucial results with conventional MRI, theχ2 value wereχ2NAA/Cr=0,χ2Cho/Cr=1.5,,χ2ADC=1.5,P>0.05;χ2NAA/Cho=4,P<0.05。Conclusion: No difference was observed in Cho/Cr ratio of high-grade glioma between proximal peritumoral edema region and parenchyma while differences existed when comparing with comtralateral normal parenchyma,only difference was observed in ADCvalue between. proximal and intermediated peritumoral edema region.and ADC value was negatively correlated to Cho/Cr ratio in proximal peritumoral edema region. Through these the invasion of tumor cells in proximal peritumoral edema of high-grade glioma can be predicted and MRS/DTI was considered useful for glioma pathological grading. Moreover, choosing proper critical value of related parameter can obtain greater positive predictive values, which in turn improve the diagnostic probability of grading. And it can be used in evaluation of glioma grading. But how to improve the selection of critical value needs further investigation.
|
PDF Full Text Request