Diffusion Kurtosis Imaging Combined With Arterial Spin Labeling In The Evaluation Of Grading Of Brain Glioma

Objective: To use diffusion kurtosis imaging(DKI) and arterial spin labeling perfusion imaging(ASL) to explore the manifestation of the brain gliomas and assess the value of these two technologies in grading of brain gliomas before operation Method: 37 cases diagnosed with brain gliomas(all the cases were pathologically confirmed) in the Affiliated Hospital of Zunyi Medical College during January 2015 to January 2016 were collected, with an average age of 41±15 years old. 23 cases were high grade gliomas and 14 cases were low grade gliomas. Conventional MRI examination, DKI and ASL were performed in all patients by using a GE Signa HDxt 3.0T scanner before operation. DKI and ASL were performed before the enhancement. The DKI and ASL original data were processed to get the image of relevant parameters respectively. DKI and ASL parameters of regions of interest(ROI) in parenchymal tumor, peritumoral edema and contralateral normal white matter were measured respectively, including mean kurtosis(MK), axial kurtosis KA), radial kurtosis(Kr), fractional anisotropy(FA), mean diffusivity(MD) and cerebral blood flow(CBF). Using contralateral normal white matter as reference, normalized MK(n MK), normalized KA(n KA), normalized KR(n KR), normalized FA(n FA), normalized MD(n MD) and normalized CBF(n CBF) values were got. A Mann-Whitney test was used to compare the parameters and normalized parameters between high and low-grade glioma. Receiver operating characteristic curve(ROC) with parameters which has significant difference between high and low-grade glioma was drawn to assess the diagnostic value of the parameters in grading of gliomas through the under ROC curve area(AUC). Binary logic regression to obtain the prediction value of DKI combining ASL parameters and ROC curve was drawn to assess the parameter with the highest diagnostic value through AUC. All statistical analyses were performed by SPSS 17.0 statistical software package. p<0.05 was considered statistically different. Result: MK, KA, KR and MD were significantly different between high and low grade gliomas. MK, KA, and KR of high-grade gliomas were higher than those of low-grade gliomas(p<0.05). When the cut-off was 0.56, 0.57, 0.59 and AUC were 0.898, 0.859, 0.843, the sensitivity and specificity were 91.3% and 92.9%, 87% and 85.7%, 71.4% and 87%, respectively. The coincidence were 91.9%, 86.5% and 81%, respectively. The MD of high-grade gliomas was lower than low-grade gliomas(p=0.013). When the cut-off was 1.51 and AUC was 0.745, the sensitivity, specificity and coincidence were 100%, 47.8% and 67.6%, respectively. There was no significance significant between CBF and FA between high and low-grade gliomas(p=0.481, 0.052). n MK, n KA, n KR, n FA, n MD and n CBF in the two groups were significantly different(p<0.05). n MK had the largest AUC among all the normalized DKI parameters(0.879) and the cut-off was 0.63. The sensitivity and specificity were 78.3% and 100%, and the coincidence rate was 89.2%. AUC of n CBF was 0.860, and the cut off was 1.27. The sensitivity, specificity and coincidence rate were 82.6%, 92.9% and 86.5%, respectively. A total of 26 cases had peritumoral edema(20 cases of high grade, 6 cases of low grade). MK and n CBF were significantly different between high and low-grade glioma(p=0.041, 0.048, respectively). The sensitivity, specificity and coincidence of MK were 75%, 83.3% and 76.9%, respectively. The sensitivity, specificity and coincidence of n CBF were 85%, 66.7% and 80.8%, respectively. The coincidence rate was significantly increased when DKI combined with ASL. MK combined n CBF had the largest AUC(0.960). The sensitivity, specificity and coincidence were 91.3%, 92.9% and 94.6%, respectively. It can be used as an index of non-invasive grading of glioma before operation. Conclusion: DKI and ASL techniques can be used for preoperative grading of gliomas and the coincidence in grading of brain glioma is superior to conventional MRI. DKI is superior to DTI in the grading of gliomas, and the MK value is can be used as the classification index when DKI is used alone. When applying ASL to the grading of gliomas, the n CBF should be used as the classification index. DKI combining ASL can significantly improve the diagnostic efficacy. The coincidence of MK combined n CBF was 94.6%, which is suitable for the classification of gliomas.