BACKGROUND AND PURPOSE: An accurate evaluation of the World Health Organization grade is critical in pediatric intracranial tumors. Our aim was to explore the correlations between parameters derived from conventional DWI, intravoxel incoherent motion, and diffusional kurtosis imaging with histopathologic features to evaluate the accuracy of diffusion parameters for grading of pediatric intracranial tumors.
MATERIALS AND METHODS:Fifty-four pediatric patients with histologically proved intracranial tumors who underwent conventional DWI, intravoxel incoherent motion, and diffusional kurtosis imaging were recruited. The conventional DWI (ADC), intravoxel incoherent motion (pure diffusion coefficient [D], pseudodiffusion coefficient [D*], perfusion fraction [f], diffusional kurtosis imaging [K], and diffusion coefficient [Dk]) parameters in the solid component of tumors were measured. The cellularity, Ki-67, and microvessel density were measured. These parameters were compared between the low-and high-grade pediatric intracranial tumors using the Mann-Whitney U test. Spearman correlations and receiver operating characteristic analysis were performed.
RESULTS:The ADC, D, and Dk values were lower, whereas the K value was higher in high-grade pediatric intracranial tumors than in low-grade tumors (all, P , .001). The K value showed positive correlations (r ¼ 0.674-0.802; all, P , .05), while ADC, D, and Dk showed negative correlations with cellularity and Ki-67 (r ¼ À0.548 to À0.740; all, P , .05). The areas under the curve of ADC VOI , D VOI , Dk VOI , and K VOI were 0.901, 0.894, 0.863, and 0.885, respectively, for differentiating high-from low-grade pediatric intracranial tumors. The area under the curve difference in grading pediatric intracranial tumors was not significant (all, P . .05).CONCLUSIONS: Intravoxel incoherent motion-and diffusional kurtosis imaging-derived parameters have similar performance compared with conventional DWI in predicting pediatric intracranial tumor grade. The diffusion metrics may potentially reflect tumor cellularity and Ki-67 in pediatric intracranial tumors. ABBREVIATIONS: AUC ¼ area under the curve; D ¼ true diffusion coefficient; D* ¼ pseudodiffusion coefficient; Dk ¼ the corrected ADC without Gaussian bias; DKI ¼ diffusional kurtosis imaging; f ¼ perfusion fraction; IVIM ¼ intravoxel incoherent motion; K ¼ diffusional kurtosis; MVD ¼ microvessel density; PIT ¼ pediatric intracranial tumor; WHO ¼ World Health Organization