Abstract #0488

Toward Fully Automated Assessment of the Central Vein Sign Using Deep Learning

Till Huelnhagen^1,2,3, Omar Al Louzi⁴, Mário João Fartaria^1,2,3, Lynn Daboul⁴, Pietro Maggi^5,6, Cristina Granziera^7,8,9, Meritxell Bach Cuadra^2,3,10, Jonas Richiardi², Daniel S Reich⁴, Tobias Kober^1,2,3, and Pascal Sati^4,11

¹Advanced Clinical Imaging Technology, Siemens Healthcare AG, Lausanne, Switzerland, ²Department of Radiology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland, ³Signal Processing Laboratory (LTS 5), Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland, ⁴Translational Neuroradiology Section, National Institute of Neurological Disorders and Stroke, National Institutes of Health (NIH), Bethesda, MD, United States, ⁵Department of Neurology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland, ⁶Cliniques universitaires Saint-Luc, Université catholique de Louvain, Brussels, Belgium, ⁷Neurologic Clinic and Policlinic, Departments of Medicine, Clinical Research and Biomedical Engineering, University Hospital Basel and University of Basel, Basel, Switzerland, ⁸Translational Imaging in Neurology (ThINk) Basel, Department of Medicine and Biomedical Engineering, University Hospital Basel and University of Basel, Basel, Switzerland, ⁹Research Center for Clinical Neuroimmunology and Neuroscience (RC2NB) Basel, University Hospital Basel and University of Basel, Basel, Switzerland, ¹⁰Medical Image Analysis Laboratory (MIAL), Centre d'Imagerie BioMédicale (CIBM), University of Lausanne, Lausanne, Switzerland, ¹¹Department of Neurology, Cedars-Sinai Medical Center, Los Angeles, CA, United States

The fraction of white matter lesions exhibiting the central vein sign (CVS) has shown promise as a biomarker in the diagnosis of multiple sclerosis. As manual CVS assessment is not clinically feasible, automated solutions have been proposed to perform this task. A deep-learning-based method called “CVSnet” demonstrated effective and accurate discrimination of MS from its mimics but required manual pre-selection. This work extends CVSnet to allow fully automated CVS assessment without manual interaction. High-quality, expert-reviewed segmentations of almost 6300 lesions were used for training and testing. The proposed method achieved accuracies between 75% and 80% in an unseen testing set.

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.