A Feature Selection (FS) process with a simple Machine Learning method, namely the Single-Layer Perceptron (SLP), is shown to discriminate metastases from glioblastomas with high accuracy using single voxel H-MRS from an international, multi-centre database of brain tumors. The method has low computational cost and its results are intuitively interpretable.
CitacióRomero, E. [et al.]. Discriminating glioblastomas from metastases in a SV1H-MRS brain tumour database. A: Annual Scientific Meeting of the European Society for Magnetic Resonance in Medicine and Biology. "Magnetic Resonance Materials in Physics, Biology and Medicine, Vol. 22, Sup.1: Book of Abstracts ESMRMB 2009". Antalya: 2009, p. 18-19.