Proton Magnetic Resonance Spectroscopy in Brain Tumors - Clinical Applications

Sammanfattning: Aims: · to present a test phantom and test method for evaluation of proton magnetic resonance spectroscopy (1H-MRS) single volume localization techniques regarding signal contamination; · to describe the introduction of 1H-MRS in the daily routine of a standard MR-unit; · to evaluate the possibilities offered by 1H-MRS to improve preoperative diagnostic accuracy in patients scheduled for stereotactic biopsy of intracranial mass lesions; · to evaluate the possibilities offered by 1H-MRS to improve diagnostics and therapy control in patients with brain abscesses; · to test the hypothesis that 1H-MRS might be useful for differentiation of brain tumors, based on metabolic changes in tissue across brain lesions or surrounding tissue. Results and conclusions: · The test phantom and test method allow quality assessment with good reproducibility. Contamination values for the tested sequence types and sizes of volume of interest (VOI) differed between 0.7% and 13.8% in one dimension and between 7% and 19.1% in three dimensions. · Of the first 102 MRS measurements performed 85% were of high spectral quality. Main reasons for poor quality were mal-positioning of the VOI, susceptibility problems due to hemorrhage, and post operative changes. · Metabolite ratios of normal tissue differed from those representing stereotactic target points. No significant difference between metabolite ratios representing different histopathologic tumor types could be seen. MRS could improve diagnostic accuracy in differentiation between infiltrating and circumscribed brain lesions based on different spatial distribution of pathologic spectra across the lesion. · MR spectra from abscesses showed resonances, not found in other intracranial pathologies and vanishing under antibiotic treatment, and can therefore be used to differentiate abscesses from other ring-enhancing lesions and might be used for therapy control. · Metabolite ratios representing different components of brain lesions or adjacent tissue differed significantly for the following differential diagnoses: a) high-grade glioma versus metastasis: NAA/Cho and Cr/Cho in contrast-enhancing tissue, and NAA/Cho in tissue adjacent to and at one voxel distance to the lesion; b) meningeoma versus low-grade glioma and metastasis: NAA/Cho in contrast-enhancing tissue; c) low-grade glioma versus metastasis: NAA/Cr in tissue adjacent to the lesion. Our data support the hypothesis that proton MRS can be useful for the reliable differentiation between brain tumors, tending to infiltrate adjacent tissue, and circumscribed, non-infiltrating brain lesions.