Title: Classification of individual disability progression trajectories of multiple sclerosis patients
Authors: Ceren Tozlu - University of Lyon (France) [presenting]
Gabriel Kocevar - Claude Bernard University-Lyon (France)
Francoise Durand-Dubief - Claude Bernard University-Lyon (France)
Sandra Vukusic - Hospices Civils de Lyon (France)
Dominique Sappey-Marinier - Claude Bernard University-Lyon (France)
Delphine Maucort-Boulch - Claude Bernard University-Lyon (France)
Abstract: Multiple sclerosis is a demyelinating, inflammatory, chronic disease of the central nervous system. MS occurs in 4 subtypes such as the clinically isolated syndrome, remitting-relapsing, secondary progressive and primary-progressive. The course of the disease is very different between patients. Therefore, the major challenge of to-days neurologist is to classify patients. The diffusion tensor imaging is an effective mean for the quantification of demyelination and axonal loss measured with fractional anisotropy, axonal, radial and mean diffusivity markers. 80 MS patients divided in 4 subtypes were included in the study. The patients were followed up with standardized clinical and MRI examination every six months during the first 3 years and every year during the last 2 years. The imaging data is obtained in 5 regions of Corpus Callosum. The kml and kml3d packages are proposed to cluster and to choose the best number of clusters for respectively the clinical and imaging trajectories. The imaging trajectories are classified for all couple of markers at each CC region. The best number of clusters is obtained as 3 for clinical data as well as for DTI data which is generated with fractional anisotropy and mean diffusivity markers at 4 regions of CC. Although 4 subtypes are predefined, 3 clusters of clinical data may be explained by the evolution of CIS patients afterwards as RR. The packages provided promising classifications which correspond well to the clinical classifications.