Abstract #0625
Automatic Multiple Sclerosis Lesion Segmentation in the Spinal Cord on 3T and 7T MP2RAGE images
Nilser Laines Medina1,2,3,4,5, Samira Mchinda1,2,3, Benoit Testud1,2,6, Sarah Demortière7, Michelle Chen4, Govind Nair8, Daniel Reich9, Cristina Granziera10,11, Charidimos Tsagkas9,10,11, Virginie Callot1,2,3, and Julien Cohen-Adad4,5,12,13
1Aix Marseille Univ, CNRS, CRMBM, Marseille, France, 2APHM, CHU Timone, Pôle d’Imagerie Médicale, CEMEREM, Marseille, France, 3iLab-Spine, International Associated Laboratory, Marseille-Montreal, France, 4NeuroPoly Lab, Institute of Biomedical Engineering, Polytechnique Montreal, Montréal, QC, Canada, 5Mila - Quebec AI Institute, Montréal, QC, Canada, 6APHM, Hôpital Universitaire Timone, Department of Neuroradiology, Marseille, France, 7APHM, Hôpital Universitaire Timone, Department of Neurology, Marseille, France, 8Quantitative MRI Core Facility, National Institute of Neurological Disorders and Stroke, National Institutes of Health (NIH), Bethesda, MD, United States, 9Translational Neuroradiology Section, National Institute of Neurological Disorders and Stroke, National Institutes of Health (NIH), Bethesda, MD, United States, 10Neurologic Clinic and Policlinic, Departments of Medicine, Clinical Research and Biomedical Engineering, University Hospital Basel and University of Basel, Basel, Switzerland, 11Translational Imaging in Neurology (ThINk) Basel, Department of Biomedical Engineering, University Hospital Basel and University of Basel, Basel, Switzerland, 12Functional Neuroimaging Unit, CRIUGM, Université de Montréal, Montréal, QC, Canada, 13Centre de Recherche du CHU Sainte-Justine, Université de Montréal, Montréal, QC, Canada
Synopsis
Keywords: Multiple Sclerosis, Segmentation, Multiple sclerosis, Spinal cord, Segmentation, 7T
Motivation: Manual detection and segmentation of multiple sclerosis (MS) lesions in the spinal cord are subject to intra- and inter-rater variability and are time-consuming, impacting the efficiency of MS diagnosis and prognosis
Goal(s): To develop a robust, automated model for MS lesion segmentation using deep learning on MP2RAGE images and make the algorithm available in an open-source, maintained package (Spinal Cord Toolbox)
Approach: Deep learning-based algorithms for image segmentation, utilizing the nnU-Net framework on a multicenter database of 3T and 7T MP2RAGE images
Results: Our approach (UNIseg) outperforms the state-of-the-art method in MS lesion detection and segmentation
Impact: This study presents a deep-learning-based method for MS lesion segmentation in the SC, which enhances diagnostic accuracy, reduces segmentation time, and offers lower variability compared to manual approaches, demonstrating significant potential to impact clinical practice and improve routine MS diagnosis
How to access this content:
For one year after publication, abstracts and videos are only open to registrants of this annual meeting. Registrants should use their existing login information. Non-registrant access can be purchased via the ISMRM E-Library.
After one year, current ISMRM & ISMRT members get free access to both the abstracts and videos. Non-members and non-registrants must purchase access via the ISMRM E-Library.
After two years, the meeting proceedings (abstracts) are opened to the public and require no login information. Videos remain behind password for access by members, registrants and E-Library customers.
Click here for more information on becoming a member.
