The effectiveness of diffusion kurtosis imaging metrics for distinguishing between brainstem glioma and cerebellar medulloblastoma.

Abstract

In some clinical situations, distinguishing between cerebellar medulloblastoma and brainstem glioma is important. We assessed whether diffusion kurtosis imaging (DKI) metrics could be used to distinguish cerebellar medulloblastomas from brainstem gliomas in children. This prospective study was approved by the institutional review board. Seventy patients were separated into two groups according to eventual diagnosis: brainstem glioma (n = 30) and cerebellar medulloblastoma (n = 40). Both groups underwent brain magnetic resonance imaging (MRI), including DKI. The Kurtosis value for the tumor region and the ratio between Kurtosis values between the tumor and the normal parenchyma (rKurtosis) were compared between groups using the Mann-Whitney U test. Receiver operating characteristic curve analysis and the Youden's Index were applied to identify a cutoff value for distinguishing between the two tumor types, and the area under the curve (AUC), sensitivity, and specificity for the selected cutoff value were calculated. Compared with brainstem gliomas, cerebellar medulloblastomas had significantly higher Kurtosis and rKurtosis values (p < 0.05). Medulloblastoma could be differentiated from brainstem gliomas using a Kurtosis value of 0.91 or an rKurtosis value of 0.90, both of which achieved 100% sensitivity, 96.7% specificity, and AUC values of 0.990. DKI measurements can contribute to distinguishing between cerebellar medulloblastoma and brainstem glioma in children.