Brain lesion segmentation through image synthesis and outlier detection

Bowles, C., Qin, C., Guerrero, R., Gunn, R., Hammers, A., Dickie, D. A. , Valdés Hernández, M., Wardlaw, J. and Rueckert, D. (2017) Brain lesion segmentation through image synthesis and outlier detection. NeuroImage: Clinical, 16, pp. 643-658. (doi: 10.1016/j.nicl.2017.09.003) (PMID:29868438) (PMCID:PMC5984574)

186949.pdf - Published Version
Available under License Creative Commons Attribution.



Cerebral small vessel disease (SVD) can manifest in a number of ways. Many of these result in hyperintense regions visible on T2-weighted magnetic resonance (MR) images. The automatic segmentation of these lesions has been the focus of many studies. However, previous methods tended to be limited to certain types of pathology, as a consequence of either restricting the search to the white matter, or by training on an individual pathology. Here we present an unsupervised abnormality detection method which is able to detect abnormally hyperintense regions on FLAIR regardless of the underlying pathology or location. The method uses a combination of image synthesis, Gaussian mixture models and one class support vector machines, and needs only be trained on healthy tissue. We evaluate our method by comparing segmentation results from 127 subjects with SVD with three established methods and report significantly superior performance across a number of metrics.

Item Type:Articles
Additional Information:This work is funded by the King's College London & Imperial College London EPSRC Centre for Doctoral Training in Medical Imaging (EP/L015226/1), Row Fogo Charitable Trust (grant no. BRO-D.FID3668413) and Innovate UK (Ref. 46917-348146). The generation of the reference data received funds from Age UK with additional funding from the UK Medical Research Council (Grant nos. G0701120, G1001245 and MR/M013111/1). Magnetic Resonance Image acquisition and analyses were conducted at the Brain Research Imaging Centre, Neuroimaging Sciences, University of Edinburgh ( which is part of SINAPSE (Scottish Imaging Network–A Platform for Scientific Excellence) collaboration ( funded by the Scottish Funding Council and the Chief Scientist Office. Support from the Fondation Leducq Network for the Study of Perivascular Spaces in Small Vessel Disease (ref no. 16 CVD 05) and European Union Horizon 2020, PHC-03-15 (project no. 666881, ‘SVDs@Target’).
Glasgow Author(s) Enlighten ID:Dickie, Dr David Alexander
Authors: Bowles, C., Qin, C., Guerrero, R., Gunn, R., Hammers, A., Dickie, D. A., Valdés Hernández, M., Wardlaw, J., and Rueckert, D.
College/School:College of Medical Veterinary and Life Sciences > School of Cardiovascular & Metabolic Health
Journal Name:NeuroImage: Clinical
ISSN (Online):2213-1582
Published Online:08 September 2017
Copyright Holders:Copyright © 2017 The Authors
First Published:First published in NeuroImage: Clinical 16:643-658
Publisher Policy:Reproduced under a Creative Commons License

University Staff: Request a correction | Enlighten Editors: Update this record