The Application Value Of Diffusion Weighted Imaging,Intravoxel Incoherent Motion Imaging And Diffusion Kurtosis Imaging In Grade Differentiation Of Pediatric Brain Tumors

Objective:To investigate the diagnostic efficiency of diffusion parameters derived from conventional diffusion-weighted imaging（DWI）,intravoxel incoherent motion（IVIM）,and diffusion kurtosis imaging（DKI）for pediatric brain tumors（PBT）grading,and to explore the correlations between diffusion parameters with histopathologic features in order to improve the accuracy of preoperative diagnosis.Materials and methods:Fifty-four pediatric patients with histologically proved intracranial tumors underwent conventional magnetic resonance imaging（c MRI）,DWI,IVIM and DKI were retrospectively reviewed in this study from July 2017 to August 2020.The c MRI features,including tumor location,cyst degeneration,necrosis,hemorrhage,edema,tumor margin,and contrast enhancement were analyzed.The mean ADC_ROI、D_ROI、D^*_ROI、f_ROI、Dk_ROI and K_ROI values were obtained from the region of interest（ROI）on the maximum slice of tumor parenchyma on each diffusion maps drawn manually by Image J.Then the mean ADC_VOI、D_VOI、D^*_VOI、f_VOI、Dk_VOI and K_VOI values were obtained from the volume of interest（VOI）of the whole tumor parenchyma,which were generated from ROI drawn manually slice by slice on each diffusion maps by 3D slicer.Meanwhile,thirty-five samples were processed and histopathologic features,including cellularity,Ki-67 labeling,and microvessel density were measured.The categorical variables were analyzed by using Pearsonχ² tests or Fisher’s exact test.These parameters were compared between the low-and high-grade pediatric intracranial tumors by using the Student’s t test or Mann-Whitney U test.The correlation between the diffusion parameters and histopathological indicators was analyzed by Spearman.The receiver operating characteristic（ROC）curve was performed to evaluate the diagnostic efficacy of these diffusion parameters for differentiation between low-and high-grade brain tumors.Results:Comparing with pediatric low-grade brain tumors,the high-grade brain tumors tended to be necrotic（50%vs 19.2%,P=0.018）.Most of the diffusion metrics exhibited significant differences between low-and high-grade brain tumor groups.The ADC,D and Dk values were lower,whereas the K value was higher in high-grade pediatric brain tumors than those in low-grade pediatric brain tumors（all P<0.001）.The K value was positively correlated with the cellularity and Ki-67（r ranging from 0.674 to 0.802,both P<0.001）,while the ADC,D and Dk values were negatively correlated with the cellularity and Ki-67（r ranging from-0.548 to-0.740,all P<0.001）.Furthermore,K_VOI value had the strongest Ki-67-related correlation among all parameters（r=0.802,P<0.001）.However,the MVD was not significantly correlated with any diffusion parameters（all P>0.05）.Among all diffusion metrics,the ADC_VOI achieved the highest diagnostic accuracy for discriminating high-from low-grade tumor groups,with an area under the ROC curve value of 0.901.Nevertheless,the differences of diagnostic efficacy among quantitative diffusion parameters did not reach a significant point.（all corrected p>0.05 compared with each other）.The optimal cutoff value of the ADC_VOI in the differentiation of two groups was 1.163×10^-3mm²/s,and the corresponding sensitivity and specificity were 85.7%and 84.6%respectively.Conclusions:IVIM and DKI-derived parameters may contribute to differentiating between low-and high-grade pediatric brain tumors,and these diffusion metrics have similar performance compared with conventional DWI in predicting PBT grading.The diffusion metrics may potentially reflect tumor cellularity and Ki-67labeling in PBT.