Output details

Non-negative matrix factorisation methods for the spectral decomposition of MRS data from human brain tumours.

<28> Bioinformatics. The originality of this paper is to rigorously apply Convex Non-negative Matrix Factorisation (CNMF) methods for accurate labelling of Magnetic Resonance Spectra (MRS), especially where differential diagnosis is most difficult, e.g. grading astrocytic tumours and separating them from metastatic lesions. The significance of the paper is to demonstrate the potential of this particularly interpretable approach for tissue characterisation in vivo in a rigorous clinical study. The method was successfully translated to brain tumour delineation in animal models (Ortega-Martorell, Lisboa et al. ‘Convex Non-Negative Matrix Factorization for Brain Tumor Delimitation from MRSI Data’ PLoS ONE, 2012).