Toward accurate quantification of metabolites, lipids, and macromolecules in HRMAS spectra of human brain tumor biopsies using LCModel.

High-resolution magic angle spinning (HRMAS) (1)H MR spectroscopy of biopsy samples provides detailed biochemical profiles that can be related to the lower-resolution spectra obtained in vivo. Nevertheless, there is still significant overlap of many resonance peaks and contributions from broad lipid and macromolecule resonances that impede accurate quantification. We determined a minimum set of in vitro metabolite and simulated lipid and macromolecule resonances needed for LCModel analysis and quantification of brain tumor biopsy HRMAS spectra. We also demonstrate the quality of the LCModel fit for the four main brain tumor types (astrocytoma grade II, glioblastoma, metastasis, and meningioma). Our data suggest that when fitting resonances of coupled spins systems in high-resolution spectra, interactions between metabolites and the macromolecular environment of the biopsy may cause small peak shifts not found in the solution spectra. However, LCModel is shown to provide a user-independent method of analyzing HRMAS brain tumor spectra.