In vivo single-voxel proton MR spectroscopy in the differentiation of high-grade gliomas and solitary metastases

Clin Radiol. 2004 Jan;59(1):77-85. doi: 10.1016/j.crad.2003.08.006.


Aim: To determine whether single-voxel proton magnetic resonance spectroscopy (1HMRS) could be used to differentiate gliomas from metastases on the basis of differences in metabolite levels in the different involved regions.

Materials and methods: Twenty-two patients (age range from 32 to 62 years, with a median age of 46.7 years) with a solitary brain tumour (14 gliomas, eight metastases) underwent conventional, gadolinium-DTPA enhanced T1-weighted images, and 1HMRS before surgical resection. Spectra from the enhancing tumour, the peritumoural region, and normal brain were obtained from 1HMRS. A point resolved spectroscopy sequence was required for 1HMRS. The metabolites in the spectra include: N-acetylaspartate (NAA), choline (CHO), creatine compounds (CR), myo-inositol (MI), lactate (LAC), glutamate and glutamine (Glu-n). Relative concentrations of metabolites were related to the peak area, and expressed with reference to CR. Student's t-test was used to determine whether there was a statistically significant difference in relative metabolic ratios between high-grade gliomas and metastases. Meanwhile, 16 of all 22 patients were re-examined using magnetic resonance imaging (MRI) within 6 months of surgical resection. Recurrence was present in three patients (two gliomas, one metastasis).

Results: Of the 14 patients with gliomas, the peaks of NAA were reduced in three cases; the peaks of LAC, which were elevated, appeared as typical double-peaks in the peritumoural region in nine cases; the peaks of Glu-n, which were also elevated, had a zigzag appearance in seven cases. The peaks of MI were increased in the tumoural region in eight cases, and CHO levels were elevated in all 14 cases. Of the eight patients with metastases, Glu-n peaks in the tumoural region in three cases and CHO peaks in the tumoural region in four cases were elevated, respectively, while the peaks of CR were reduced in three cases, and the peaks of NAA were markedly reduced in four cases within the enhancing tumoural region. Elevated CHO levels (CHO-to-CR ratio was 4.98:1.46, 2.65:0.32) in both the tumoural and peritumoural regions of gliomas but not in the metastases (CHO-to-CR ratio was 1.37:0.92, 1.22:0.38), and elevated MI levels were present (MI-to-CR ratio was 1.67:0.35) within the enhancing foci of gliomas but not in the metastases (MI-to-CR ratio was 0.89:0.31). The difference was statistically significant (p<0.01). Elevated Glu-n and lipid levels were present in all three patients with recurrences.

Conclusion: 1HMRS is a useful method in the distinction of these two kinds of tumours. It may also may provide useful prognostic information.

MeSH terms

  • Adult
  • Brain Neoplasms / diagnosis*
  • Brain Neoplasms / secondary
  • Contrast Media
  • Female
  • Gadolinium DTPA
  • Glioma / diagnosis*
  • Humans
  • Magnetic Resonance Imaging / methods
  • Magnetic Resonance Spectroscopy / methods*
  • Male
  • Middle Aged
  • Patient Selection
  • Protons


  • Contrast Media
  • Protons
  • Gadolinium DTPA