AUTOMATIC LONGITUDINAL MULTIPLE SCLEROSIS LESION SEGMENTATION: MSmetrix

Accurate and consistent multiple sclerosis (MS) brain lesion segmentation and volumetry could be an added value to MS clinicians. In this paper, MSmetrix is presented, an automatic and reliable method, which uses 3D T1-weighted and FLAIR MR images in a probabilistic model to detect white matter lesions as an outlier with respect to the normal brain, while segmenting the brain tissue into grey matter, white matter and cerebrospinal fluid. The actual lesion segmentation is performed based on prior knowledge about the location (within white matter) and the appearance (hyperintense on FLAIR) of lesions. The randomness in longitudinal lesion segmentation for each subject is reduced by harmonising the trade-off between temporal consistency across time points and segmentation diversity at each time point. The method has been validated on the dataset available from the longitudinal MS lesion segmentation challenge 2015.