Yamashita Koji, Hiwatashi Akio, Togao Osamu, Kikuchi Kazufumi, Kitamura Yoshiyuki, Mizoguchi Masahiro, Yoshimoto Koji, Kuga Daisuke, Suzuki Satoshi O, Baba Shingo, Isoda Takuro, Iwaki Toru, Iihara Koji, Honda Hiroshi
Department of Clinical Radiology, Graduate School of Medical Sciences, Kyushu University, Maidashi, Higashi-ku, Fukuoka, Japan.
Department of Neurosurgery, Graduate School of Medical Sciences, Kyushu University, Maidashi, Higashi-ku, Fukuoka, Japan.
J Magn Reson Imaging. 2016 Nov;44(5):1256-1261. doi: 10.1002/jmri.25261. Epub 2016 Apr 19.
To evaluate the diagnostic performance of intravoxel incoherent motion (IVIM) MR imaging and F-fluorodeoxyglucose positron emission tomography (FDG-PET) in differentiating primary central nervous system lymphoma (PCNSL) from glioblastoma multiforme (GBM).
Fifty patients, 17 with PCNSL and 33 with GBM, were retrospectively studied. From the 3 Tesla IVIM data, the perfusion fraction (f) and diffusion coefficient (D) were obtained. In addition, the maximum standard uptake value (SUV ) was obtained from the FDG-PET data. Each of the three parameters was compared between PCNSL and GBM using Mann-Whitney U-test. The performance in discriminating between PCNSL and GBM was evaluated using receiver-operating characteristics analysis and area-under-the-curve (AUC) values for the three parameters.
The f and D values were significantly higher in GBM than in PCNSL (P < 0.01 and P < 0.0001, respectively). In addition, the SUV value was significantly lower in GBM than in PCNSL (P < 0.0005). The AUC values for f , D , and SUV were 0.756, 0.905, and 0.857, respectively. The combination of the f and D increased the diagnostic performance (AUC = 0.936) of f (P < 0.05), but this value was not significantly different from the values for D (P = 0.30).
IVIM-MR imaging noninvasively provides useful quantitative information in distinguishing between PCNSL and GBM. J. Magn. Reson. Imaging 2016;44:1256-1261.
评估体素内不相干运动(IVIM)磁共振成像(MRI)和氟脱氧葡萄糖正电子发射断层扫描(FDG-PET)在鉴别原发性中枢神经系统淋巴瘤(PCNSL)和多形性胶质母细胞瘤(GBM)方面的诊断性能。
回顾性研究了50例患者,其中17例为PCNSL,33例为GBM。从3特斯拉IVIM数据中获取灌注分数(f)和扩散系数(D)。此外,从FDG-PET数据中获取最大标准摄取值(SUV)。使用Mann-Whitney U检验比较PCNSL和GBM之间的这三个参数。使用受试者操作特征分析和这三个参数的曲线下面积(AUC)值评估区分PCNSL和GBM的性能。
GBM中的f和D值显著高于PCNSL(分别为P < 0.01和P < 0.0001)。此外,GBM中的SUV值显著低于PCNSL(P < 0.0005)。f、D和SUV的AUC值分别为0.756、0.905和0.857。f和D的联合提高了f的诊断性能(AUC = 0.936)(P < 0.05),但该值与D的值无显著差异(P = 0.30)。
IVIM-MRI成像在区分PCNSL和GBM方面无创地提供了有用的定量信息。《磁共振成像杂志》2016年;44:1256 - 1261。