PURPOSE: The present study aimed to distinguish between glioblastomas and primary central nervous system lymphomas (PCNSLs) using (1)H-magnetic resonance spectroscopy (MRS), especially glutamate (Glu)/creatine (Cr) and Glu/Glu+glutamine (Gln) ratios. MATERIALS AND METHODS: A total of 46 patients (31 cases diagnosed with glioblastoma, 15 with PCNSL) were examined by in vivo single-voxel proton (1)H-MRS with a 3-T MR imaging system. Differences in absolute concentration of Cr, choline/Cr, lipid (1.3 ppm)/Cr, Glu+Gln/Cr, Glu/Cr, and Glu/Glu+Gln ratios among groups were evaluated with Mann-Whitney U tests. RESULTS: PCNSLs [3.408 ± 1.194 (standard deviation)] showed significantly higher Glu/Cr ratios as compared to glioblastomas (2.220 ± 0.942; P = 0.003) [Glu/Cr cutoff ratio of 2.509 showed a sensitivity of 88 % (7/8) and a specificity of 92 % (22/24)], while glioblastomas (0.539 ± 0.098) showed significantly lower Glu/Glu+Gln ratios as compared to PCNSLs (0.728 ± 0.147; P < 0.001) [Glu/Glu+Gln cutoff ratio of 0.558 showed a sensitivity of 69 % (18/26) and a specificity of 100 % (13/13)]. And PCNSLs (1.101 ± 0.387) showed significantly higher Cho/Cr ratios as compared to glioblastomas (0.850 ± 0.465; P = 0.026). CONCLUSION: Glu/Cr, Glu/Glu+Gln, and Cho/Cr ratios may be useful in distinguishing between glioblastomas and PCNSLs.
PURPOSE: The present study aimed to distinguish between glioblastomas and primary central nervous system lymphomas (PCNSLs) using (1)H-magnetic resonance spectroscopy (MRS), especially glutamate (Glu)/creatine (Cr) and Glu/Glu+glutamine (Gln) ratios. MATERIALS AND METHODS: A total of 46 patients (31 cases diagnosed with glioblastoma, 15 with PCNSL) were examined by in vivo single-voxel proton (1)H-MRS with a 3-T MR imaging system. Differences in absolute concentration of Cr, choline/Cr, lipid (1.3 ppm)/Cr, Glu+Gln/Cr, Glu/Cr, and Glu/Glu+Gln ratios among groups were evaluated with Mann-Whitney U tests. RESULTS: PCNSLs [3.408 ± 1.194 (standard deviation)] showed significantly higher Glu/Cr ratios as compared to glioblastomas (2.220 ± 0.942; P = 0.003) [Glu/Cr cutoff ratio of 2.509 showed a sensitivity of 88 % (7/8) and a specificity of 92 % (22/24)], while glioblastomas (0.539 ± 0.098) showed significantly lower Glu/Glu+Gln ratios as compared to PCNSLs (0.728 ± 0.147; P < 0.001) [Glu/Glu+Gln cutoff ratio of 0.558 showed a sensitivity of 69 % (18/26) and a specificity of 100 % (13/13)]. And PCNSLs (1.101 ± 0.387) showed significantly higher Cho/Cr ratios as compared to glioblastomas (0.850 ± 0.465; P = 0.026). CONCLUSION:Glu/Cr, Glu/Glu+Gln, and Cho/Cr ratios may be useful in distinguishing between glioblastomas and PCNSLs.
Authors: X Lin; M Lee; O Buck; K M Woo; Z Zhang; V Hatzoglou; A Omuro; J Arevalo-Perez; A A Thomas; J Huse; K Peck; A I Holodny; R J Young Journal: AJNR Am J Neuroradiol Date: 2016-12-08 Impact factor: 3.825