Abstract #1981

Biophysically meaningful MRI features for accurate classification of multiple sclerosis phenotypes

Antonio Ricciardi^1,2,3, Francesco Grussu^1,3, Wallace Brownlee¹, Baris Kanber^1,4, Ferran Prados^1,4, Sara Collorone¹, Enrico Kaden³, Ahmed Toosy^1,5, Sebastien Ourselin⁴, Olga Ciccarelli^1,5, Daniel C Alexander³, and Claudia Angela Gandini Wheeler-Kingshott^1,6,7

¹Queen Square MS Centre, Department of Neuroinflammation, UCL Institute of Neurology, Faculty of Brain Sciences, University College London, London, United Kingdom, ²Department of Medical Physics and Biomedical Engineering, University College London, London, United Kingdom, ³Centre for Medical Image Computing, Department of Computer Science, University College London, London, United Kingdom, ⁴Translational Imaging Group, Centre for Medical Image Computing, Medical Physics and Biomedical Engineering, University College London, London, United Kingdom, ⁵National Hospital of Neurology and Neurosurgery, London, United Kingdom, ⁶Department of Brain and Behavioral Sciences, University of Pavia, Pavia, Italy, ⁷Brain MRI 3T Research Centre, C. Mondino National Neurological Institute, Pavia, Italy

Quantitative MRI can provide maps of biophysically meaningful features (BMFs) that can be exploited using machine learning techniques to better correlate MR alterations with multiple sclerosis (MS) severity, and improve our understanding of the disease. In this study, a random forest classifier was trained over a rich multi-modal quantitative MRI dataset of healthy controls and MS patients with different phenotypes, to find the BMFs that best characterise disease course. Inflammation and atrophy were the most significant BMFs in distinguishing between controls and patients, with microstructural alterations arising particularly when comparing subjects who only experienced a clinically isolated syndrome with patients and controls.

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

Join Here