Related Experiment Video
Updated: May 9, 2026
From a 2DE-Gel Spot to Protein Function: Lesson Learned From HS1 in Chronic Lymphocytic Leukemia
Published on: October 19, 2014
Utility of proton MR spectroscopy for differentiating typical and atypical primary central nervous system lymphomas
1From the Department of Radiology and Research Institute of Radiology (S.-S.L. S.J.K., H.S.K., C.G.C., D.Y.K., S.H.C.).
Background And Purpose:
It may be challenging to differentiate primary CNS lymphomas, especially primary CNS lymphomas with atypical MR features, from tumefactive demyelinating lesions by the use of conventional MR. This study aimed to investigate the usefulness of (1)H-MR spectroscopy for making this discrimination.
Materials And Methods:
Forty-four patients with primary CNS lymphomas and 21 with tumefactive demyelinating lesions were enrolled. Single-voxel (TE = 144 ms) (1)H-MR spectroscopy scans with the use of the point-resolved spectroscopy sequence were retrospectively analyzed. The Cho/Cr and Cho/NAA area ratios were calculated. The lipid and/or lactate peak was visually categorized into 5 grades on the basis of comparison with the height of the Cr peak. The (1)H-MR spectroscopy findings were compared in all of the primary CNS lymphomas and the tumefactive demyelinating lesions and in the subgroup of atypical primary CNS lymphomas and tumefactive demyelinating lesions. The thresholds and added value of (1)H-MR spectroscopy to conventional MR were calculated by use of receiver operating characteristic curves.
Results:
Discrepancies between all of the primary CNS lymphomas and tumefactive demyelinating lesions were found in the Cho/Cr ratio (P = .000), Cho/NAA ratio (P = .000), and the lipid and/or lactate peak grade (P = .000). Lymphoma rather than tumefactive demyelinating lesions was suggested when the Cho/Cr ratio was >2.58, the Cho/NAA ratio was >1.73, and a high lipid and/or lactate peak grade (grade >3) was seen. Higher Cho/Cr ratios, Cho/NAA ratios, and lipid and/or lactate peak grades were found in atypical primary CNS lymphomas when compared with those of tumefactive demyelinating lesions. The area under the receiver operating characteristic curve of conventional MR was improved from 0.827 to 0.870 when Cho/NAA ratio was added in the uncertain cases.
Conclusions:
(1)H-MR spectroscopy may be useful for differentiating primary CNS lymphomas from tumefactive demyelinating lesions. Cho/NAA ratio could provide added value to conventional MR imaging.
