Tissue Type Differentiation for Brain Glioma using Non-negative Matrix Factorization

The purpose of this paper is to introduce a hierarchical Non-negative Matrix Factorization (NMF) approach, customized for the problem of blindly separating brain glioma tumor tissue types using short-echo time proton magnetic resonance spectroscopic imaging (H MRSI) signals. The proposed algorithm consists of two levels of NMF, where two constituent spectra are computed in each level. The first level is able to correctly detect the spectral profile corresponding to the most predominant tissue type, i.e., normal tissue, while the second level is optimized in order to detect two ‘abnormal’ spectral profiles so that the 3 recovered spectral profiles are least correlated with each other. The two-level decomposition is followed by the reestimation of the overall spatial distribution of each tissue type via standard Non-negative Least Square (NNLS). This method is demonstrated on in vivo short-TE H MRSI brain data of a glioblastoma multiforme patient and a grade II-III glioma patient. The results show the possibility of differentiating normal tissue, tumor tissue and necrotic tissue in the form of recovered tissue-specific spectra with accurate spatial distributions.

[1]  V. L. Doyle,et al.  Metabolic profiles of human brain tumors using quantitative in vivo 1H magnetic resonance spectroscopy , 2003, Magnetic resonance in medicine.

[2]  Andrzej Cichocki,et al.  Hierarchical ALS Algorithms for Nonnegative Matrix and 3D Tensor Factorization , 2007, ICA.

[3]  P. Sajda,et al.  RECOVERY OF CONSTITUENT SPECTRA IN 3D CHEMICAL SHIFT IMAGING USING NON-NEGATIVE MATRIX FACTORIZATION , 2003 .

[4]  Nicolas Gillis.  Nonnegative matrix factorization : complexity, algorithms and applications , 2011 .

[5]  B R Rosen,et al.  NMR chemical shift imaging. , 1989, Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society.

[6]  H. Sebastian Seung,et al.  Learning the parts of objects by non-negative matrix factorization , 1999, Nature.

[7]  P. Bottomley.  Spatial Localization in NMR Spectroscopy in Vivo , 1987, Annals of the New York Academy of Sciences.

[8]  K. Uğurbil,et al.  NMR chemical shift imaging in three dimensions. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[9]  S. Van Huffel,et al.  Ex vivo high resolution magic angle spinning metabolic profiles describe intratumoral histopathological tissue properties in adult human gliomas , 2011, Magnetic resonance in medicine.

[10]  Faculteit Ingenieurswetenschappen,et al.  QUANTIFICATION AND CLASSIFICATION OF MAGNETIC RESONANCE SPECTROSCOPIC DATA FOR BRAIN TUMOR DIAGNOSIS , 2008 .

[11]  Andrzej Cichocki,et al.  Fast Local Algorithms for Large Scale Nonnegative Matrix and Tensor Factorizations , 2009, IEICE Trans. Fundam. Electron. Commun. Comput. Sci..

[12]  S Van Huffel,et al.  Improved Lanczos algorithms for blackbox MRS data quantitation. , 2002, Journal of magnetic resonance.

[13]  Paul Sajda,et al.  Automated tissue segmentation and blind recovery of 1H MRS imaging spectral patterns of normal and diseased human brain , 2008, NMR in biomedicine.

[14]  P. Paatero,et al.  Positive matrix factorization: A non-negative factor model with optimal utilization of error estimates of data values† , 1994 .

[15]  Lucas C. Parra,et al.  Nonnegative matrix factorization for rapid recovery of constituent spectra in magnetic resonance chemical shift imaging of the brain , 2004, IEEE Transactions on Medical Imaging.

[16]  Yuzhuo Su,et al.  Spectrum separation resolves partial‐volume effect of MRSI as demonstrated on brain tumor scans , 2008, NMR in biomedicine.