Classification of single-voxel 1H spectra of brain tumours using LCModel

• Published in: NMR in biomedicine 2012-02, Vol.25 (2), p.322-331
• Main Authors: Raschke, F., Fuster-Garcia, E., Opstad, K. S., Howe, F. A.
• Format: Article
• Language: English
• Subjects: Adult; Algorithms; brain; Brain - pathology; Brain Neoplasms - pathology; classification; Decision Support Systems, Clinical; Glioma - pathology; Humans; LCModel; Magnetic Resonance Spectroscopy - classification; Magnetic Resonance Spectroscopy - methods; MRS; Neoplasm Grading; Organ Specificity; pattern; Protons; recognition; tumour
• ISSN: 0952-3480; 1099-1492
• DOI: 10.1002/nbm.1753

This study presents a novel method for the direct classification of 1H single‐voxel MR brain tumour spectra using the widespread analysis tool LCModel. LCModel is designed to estimate individual metabolite proportions by fitting a linear combination of in vitro metabolite spectra to an in vivo MR spectrum. In this study, it is used to fit representations of complete tumour spectra and to perform a classification according to the highest estimated tissue proportion. Each tumour type is represented by two spectra, a mean component and a variability term, as calculated using a principal component analysis of a training dataset. In the same manner, a mean component and a variability term for normal white matter are also added into the analysis to allow a mixed tissue approach. An unbiased evaluation of the method is carried out through the automatic selection of training and test sets using the Kennard and Stone algorithm, and a comparison of LCModel classification results with those of the INTERPRET Decision Support System (IDSS) which incorporates an advanced pattern recognition method. In a test set of 46 spectra comprising glioblastoma multiforme, low‐grade gliomas and meningiomas, LCModel gives a classification accuracy of 90% compared with an accuracy of 95% by IDSS. Copyright © 2011 John Wiley & Sons, Ltd. This study presents a novel method for the direct classification of single‐voxel 1H MR spectra of brain tumours using the widespread analysis tool LCModel, estimating tumour proportions by fitting representations of complete tumour spectra to in vivo MR spectra. Classification according to the highest estimated tumour proportion in a test set of 46 spectra, comprising glioblastoma multiforme, low‐grade gliomas and meningiomas, LCModel gives a classification accuracy of 90%, compared with 95% by the INTERPRET Decision Support System.