PT  - JOURNAL ARTICLE
AU  - Fudaba, H.
AU  - Shimomura, T.
AU  - Abe, T.
AU  - Matsuta, H.
AU  - Momii, Y.
AU  - Sugita, K.
AU  - Ooba, H.
AU  - Kamida, T.
AU  - Hikawa, T.
AU  - Fujiki, M.
TI  - Comparison of Multiple Parameters Obtained on 3T Pulsed Arterial Spin-Labeling, Diffusion Tensor Imaging, and MRS and the Ki-67 Labeling Index in Evaluating Glioma Grading
AID  - 10.3174/ajnr.A4018
DP  - 2014 Nov 01
TA  - American Journal of Neuroradiology
PG  - 2091--2098
VI  - 35
IP  - 11
4099  - http://www.ajnr.org/content/35/11/2091.short
4100  - http://www.ajnr.org/content/35/11/2091.full
SO  - Am. J. Neuroradiol.2014 Nov 01; 35
AB  - BACKGROUND AND PURPOSE: Pulsed arterial spin-labeling, DTI, and MR spectroscopy provide useful data for tumor evaluation. We evaluated multiple parameters by using these pulse sequences and the Ki-67 labeling index in newly diagnosed supratentorial gliomas. MATERIALS AND METHODS: All 32 patients, with grade II (3 each of diffuse astrocytoma, oligodendroglioma, and oligoastrocytoma), grade III (3 anaplastic astrocytomas, 4 anaplastic oligodendrogliomas, and 1 anaplastic oligoastrocytoma), and grade IV (14 glioblastomas and 1 glioblastoma with an oligodendroglioma component) cases underwent pulsed arterial spin-labeling, DTI, and MR spectroscopy studies by using 3T MR imaging. The following variables were used to compare the tumors: relative cerebral blood flow, fractional anisotropy; ADC tumor/normal ratios; and the Cho/Cr, NAA/Cho, NAA/Cr, and lactate/Cr ratios. A logistic regression and receiver operating characteristic analysis were used to assess parameters with a high sensitivity and specificity to identify the threshold values for separate grading. We compared the Ki-67 index with various MR imaging parameters in tumor specimens. RESULTS: Significant correlations were observed between the Ki-67 index and the mean, maximum, and minimum ADC, Cho/Cr, and lactate/Cr ratios. The receiver operating characteristic analysis showed that the combination of the minimum ADC and Cho/Cr ratios could differentiate low-grade and high-grade gliomas, with a sensitivity and specificity of 87.0% and 88.9%, respectively. The mean and maximum relative cerebral blood flow ratios were used to classify glioblastomas from other-grade astrocytomas, with a sensitivity and specificity of 92.9% and 83.3%, respectively. CONCLUSIONS: Our findings indicate that pulsed arterial spin-labeling, DTI, and MR spectroscopy are useful for predicting glioma grade. Additionally, the parameters obtained on DTI and MR spectroscopy closely correlated with the proliferative potential of gliomas. ASLarterial spin-labelingC1 erroraverage observed sensitivity and specificityC2 errorobserved number of instances of tumor-grade misclassificationFAfractional anisotropyLaclactateNPVnegative predictive valuePASLpulsed arterial spin-labelingPPVpositive predictive valuerCBFrelative cerebral blood flowROCreceiver operating characteristic