Purpose: We evaluated the clinical usefulness of the Cho/Cr ratio of proton MR spectroscopy(1H-MRS) to differentiate residual/recurrent glioma from non-neoplastic lesions.
Patients and methods: 20 cases of glioma were involved in this study(astrocytoma grade I-II: 7, oligodendroglioma: 1, astrocytoma grade III: 2, glioblastoma: 10). Seven of the patients underwent surgical resection only, 4 underwent surgical resection and radiotherapy(40-60 Gy), and 9 underwent surgical resection and radiotherapy with concurrent chemotherapy(14-60 Gy). 1H-MRS was performed on a 1.5 Tesla clinical MR unit using a 3D-chemical shift imaging sequence(1500 msec/270 msec/1 (TR/TE/excitations), and the Cho/Cr ratio was calculated in the voxel where neoplastic lesion was most suspected on MRI. The presence of lactate + lipid peak was also evaluated. All spectra were obtained after the contrast enhanced study.
Results: Cho/Cr ratios were significantly higher in cases of residual/recurrent tumors(mean +/- SD = 1.70 +/- 0.96) than in non-neoplastic lesions(mean +/- SD = 1.04 +/- 1.16) (Mann-Whitney U-test p = 0.047). If a Cho/Cr ratio of more than 1.5 was used as a marker of tumor presence, its sensitivity was 64%, specificity 83%, and accuracy 70%. One false-positive case that of radiation necrosis whose spectrum showed a high Cho/Cr ratio with markedly elevated lactate + lipid peak.
Conclusion: The Cho/Cr ratio of 1H-MRS provides additional information to MRI in differentiating residual/recurrent gliomas from non-neoplastic lesions.