Meeting Banner
Abstract #2809

Relevance analysis of identifying multiple sclerosis patients based on diffusion imaging data using CNN

Alina Lopatina1,2, Stefan Ropele3, Renat Sibgatulin1, Jürgen R Reichenbach1,2,4, and Daniel Güllmar1
1Medical Physics Group / IDIR, Jena University Hospital, Jena, Germany, 2Michael-Stifel-Center for Data-Driven and Simulation Science, Jena, Germany, 3Department of Neurology, Medical University of Graz, Graz, Austria, 4Center of Medical Optics and Photonics Jena, Jena, Germany

To analyze the classification procedure of identifying multiple sclerosis (MS) based on diffusion-weighted imaging data by using convolutional neural networks (CNNs), we generated relevance maps. The relevance maps indicate the contribution of each input voxel to the final classification score and may facilitate new findings regarding MS-specific biomarkers. The study showed that voxels in the central brain area including some of the lesion voxels are important for correct classification. This information may be used in the future to perform a more detailed analysis in order to classify different MS-phenotypes or predict disease progression.

This abstract and the presentation materials are available to members only; a login is required.

Join Here