Differentiation of primary central nervous system lymphoma from high-grade glioma and brain metastasis using arterial spin labeling and dynamic contrast-enhanced magnetic resonance imaging

• Published in: European journal of radiology Vol. 112; pp. 59 - 64
• Main Authors: Xi, Yi-bin, Kang, Xiao-wei, Wang, Ning, Liu, Ting-ting, Zhu, Yuan-qiang, Cheng, Guang, Wang, Kai, Li, Chen, Guo, Fan, Yin, Hong
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier B.V 01.03.2019
• Subjects: Arterial spin labeling; Dynamic contrast-enhanced perfusion; High-grade glioma; Metastasis; Primary central nervous system lymphoma; Primary central nervous system lymphoma; Dynamic contrast-enhanced perfusion; High-grade glioma; Metastasis; Arterial spin labeling
• ISSN: 0720-048X; 1872-7727; 1872-7727
• DOI: 10.1016/j.ejrad.2019.01.008

•PCNSL demonstrated lower rCBF, higher Ktrans and Ve compared with HGG and metastasis.•Both Ktrans and rCBF had good diagnostic performance for discriminating PCNSL.•The combination of rCBF and Ktrans has the best diagnostic ability for PCNSL. Conventional magnetic resonance imaging (MRI) is sometimes difficult to distinguish primary central nervous system lymphoma (PCNSL) from other malignant brain tumors effectively. The study aimed to evaluate the diagnostic performance of arterial spin labeling (ASL) and dynamic contrast-enhanced (DCE)-derived permeability parameters to differentiate PCNSL from high-grade glioma (HGG) and brain metastasis. Eight patients with PCNSL, twenty one patients with HGG and six brain metastasis underwent preoperative 3.0-T MR imaging including conventional, ASL and DCE. Quantitative parameters including relative cerebral blood flow (rCBF), extravascular extracellular volume fraction (Ve) and the volume transfer constant (Ktrans) among PCNSL, HGG and metastasis were compared with a one-way analysis of variance. In addition, the area under the receiver-operating characteristic (ROC) curve (AUC) was constructed to evaluate the differentiation diagnostic performance of each parameter and the combination. The PCNSL demonstrated significantly lower rCBF, higher Ktrans and Ve compared with HGG and metastasis. For the ROC analyses, both Ktrans and rCBF had good diagnostic performance for discriminating PCNSL from HGG and metastasis, with the AUC of 0.880 and 0.889. With the combination of rCBF and Ktrans, the diagnostic ability for PCNSL was improved with AUC of 0.986. rCBF and Ktrans are useful parameters for differentiating PCNSL from HGG and brain metastasis. The combination of rCBF and Ktrans further helps to improve the diagnostic performance of PCNSL.