Abstract #0624
TotalSpineSeg: Robust Spine Segmentation and Labeling Across Multiple MRI Contrasts
Yehuda Warszawer1,2,3, Nathan Molinier4,5, Jan Valosek4,5, Emanuel Shirbint1, Pierre-louis Benveniste4,5, Tobias Granberg6,7, Russell Ouellette6,7, Charidimos Tsagkas8,9,10, Virginie Callot11,12, Feroze Mohamed13, Josef Bednarik14,15, Kristin Poole O'Grady16,17, Anat Achiron1,18, and Julien Cohen-Adad4,5,19
1Multiple Sclerosis Center, Sheba Medical Center, Ramat Gan, Israel, 2Arrow Program for Medical Research Education, Sheba Medical Center, Ramat Gan, Israel, 3Adelson School of Medicine, Ariel University, Ariel, Israel, 4NeuroPoly Lab, Institute of Biomedical Engineering, Polytechnique Montreal, Montreal, QC, Canada, 5Mila - Quebec AI Institute, Montreal, QC, Canada, 6Department of Neuroradiology, Karolinska University Hospital, Stockholm, Sweden, 7Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden, 8Translational Neuroradiology Section, National Institute of Neurological Disorders and Stroke, National Institutes of Health (NIH), Bethesda, MD, United States, 9Translational Imaging in Neurology (ThINk) Basel, Department of Biomedical Engineering, University of Basel, Allschwil, Switzerland, 10University Hospital Basel and University of Basel, Basel, Switzerland, 11CRMBM,UMR 7339, CNRS, Aix-Marseille Université, Marseille, France, 12CEMEREM - Hôpital de la Timone, CHU de Marseille, Marseille, France, 13Department of Radiology, Thomas Jefferson University, Philadelphia, PA, United States, 14Department of Neurology, University Hospital Brno, Brno, Czech Republic, 15Faculty of Medicine, Masaryk University Brno, Brno, Czech Republic, 16Vanderbilt University Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, TN, United States, 17Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, United States, 18Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel, 19Functional Neuroimaging Unit, CRIUGM, Université de Montréal, Montreal, QC, Canada
Synopsis
Keywords: Analysis/Processing, Multi-Contrast, Spine, Deep learning, Segmentation
Motivation: Precise segmentation and labeling of vertebrae is essential to properly assess spine and spinal cord diseases such as spinal cord injury, degenerative diseases and scoliosis.
Goal(s): Develop an automatic tool to segment and identify vertebrae on a wide range of MRI protocols, including different MRI sequences, contrasts, fields-of-view and resolutions.
Approach: A three-step process including two nnUNetV21 models and an iterative labeling algorithm was used to segment and identify vertebrae using anatomical landmarks.
Results: The model generalizes well across MRI sequences, with a segmentation Dice score of 0.84 and an average labeling accuracy of 0.99.
Impact: TotalSpineSeg could enhance clinical workflows by providing automatic vertebrae segmentation, improving the diagnosis of various spinal pathologies and supporting informed clinical decision-making. It is available on GitHub (https://github.com/neuropoly/totalspineseg) and in Spinal Cord Toolbox v.6.514.
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.
