Evaluation of cerebral glioma grade by using normal side creatine as an internal reference in multi-voxel 1H-MR spectroscopy.

It is important to determine brain tumor grade before establishing effective therapy. Although conventional magnetic resonance imaging (MRI) is a useful method (1, 2), it is not always possible to determine brain tumor grade (3, 4). Several noninvasive neuroimaging methods, including proton magnetic resonance spectroscopy imaging (1H-MRS), diffusion-weighted MRI, and perfusion MRI, have been used to determine brain tumor grade (3, 5–7). Numerous studies have shown that 1HMRS improves preoperative diagnosis of brain tumor grading, although the accuracy of noninvasive advanced neuroimaging methods is controversial (3, 5, 8–10). Further investigations have been recommended to increase the clinical utility of these methods (5, 9). 1H-MRS provides data related to biochemical changes in the tumor and may therefore help in predicting brain tumor grade as a noninvasive diagnostic tool. Choline (Cho) is known as a radiological marker suggesting cell turnover. Brain tumors mostly demonstrate elevation in the ratio of Cho/Creatine (Cr); however, glial neoplasms without elevated Cho/Cr ratios also have been reported (11–13). N-acetyl aspartate (NAA) decreases in any disease associated with loss of neurons. Cr metabolite reflects information on energy metabolism. Results of several studies suggest, surprisingly, that levels of Cr are the same in both lowand high-grade gliomas (14, 15). Nonetheless, previous studies have pointed out that decreased Cr levels can be seen in brain tumors (16, 17). Additionally, the amount of Cr may be variable in different regions of the same tumor. Elevated Cr may be seen in the hypometabolic areas and decreased Cr may be observed in the hypermetabolic areas of the same tumor (4, 18). The importance of Cr levels is not clear in the differentiation of lowand high-grade gliomas. If the Cr signal in the tumor area is used as an internal standard, it may be difficult to identify differences in metabolite ratios and it also might cause errors in predicting tumor grade. In this study, we used the metabolite of Cr from the symmetrical normal brain area as a normal reference peak. We calculated both the ratios of tumor main metabolites to Cr from contralateral normal brain parenchyma and the ratios of tumor main metabolites to each other, semiquantitatively. Our purpose was to evaluate cerebral glioma grade by using normal side Cr as an internal reference in multi-voxel 1H-MRS to avoid the problems of using Cr as a reference peak from the tumor area.