Output details
13 - Electrical and Electronic Engineering, Metallurgy and Materials
University of Birmingham : A - Electronic, Electrical and computer engineering
Identification and characterisation of childhood cerebellar tumours by in vivo proton MRS.
This paper reports machine learning-based analyses of 1H MRS datasets for childhood cerebellar tumours, a largely unexplored approach for this type of disease. Novel multivariate classifiers of patient datasets have been developed for a specific cohort of patients, in collaboration with Birmingham Children’s Hospital. Our novel quantitative approach to classifiers deals with small datasets, as opposed to qualitative approaches reported historically. The study is strongly advocating the community’s hypothesis that 1H MRS detects key differences in metabolite profiles for the main types of childhood tumours. The work was influential in securing a £2.1M (£643k to Birmingham) CRUK-EPSRC multi-centre programme grant.