Abstract #1727
A generalized deep learning network for fractional anisotropy reconstruction: application to epilepsy and multiple sclerosis
Marta Gaviraghi1, Antonio Ricciardi2, Fulvia Palesi3, Wallace Brownlee2, Paolo Vitali4, Ferran Prados2, Baris Kanber2, and Claudia AM Gandini Wheeler-Kingshott2,3,5
1Department of Electrical, Computer and Biomedical Engineering, University of Pavia, Pavia, Italy, 2NMR Research Unit, Queen Square Multiple Sclerosis Centre, Department of Neuroinflammation, UCL Queen Square Institute of Neurology, UCL, London, United Kingdom, 3Department of Brain and Behavioural Sciences, University of Pavia, Pavia, Italy, 4IRCCS Policlinico San Donato, Milano, Italy, 5Brain Connectivity Centre, IRCCS Mondino Foundation, Pavia, Italy
Synopsis
Quantitative maps obtained from Diffusion Tensor are very useful for investigating microstructural changes that occur in brain diseases. However, the long acquisition times required for a fully sampled diffusion-weighted space makes their clinical use unfeasible. Here we have adapted a U-net that obtains reliable Fractional Anisotropy (FA) maps from a reduced set of 10 Diffusion Weighted volumes (that can be acquired in less than 1 min). Our network was applied to two independent, clinical datasets, without retraining, and produced FA that retained clinical sensitivity and characteristic FA value distributions in the brain white matter.
This abstract and the presentation materials are available to members only;
a login is required.
Join Here
