EUVIMEDEUVIMED entdecken
Crossref

TotalSpineSeg: automated whole-spine vertebral segmentation with high robustness to labeling errors

Yehuda Warszawer, Nathan Molinier, Jan Valošek, Pierre-Louis Benveniste, Sandrine Bédard, Emanuel Shirbint, Feroze B Mohamed, Charidimos Tsagkas, Shannon Kolind, Larry Lynd, Jiwon Oh, Alexandre Prat und 6 weitere

European Radiology · 07.11.2026

DOI: 10.1007/s00330-026-13008-2 ↗

Kurz gefasst

Fachartikel in der Zeitschrift European Radiology, online erschienen am 07.11.2026. Verfasst von Yehuda Warszawer und 17 weiteren Autor:innen. Beteiligte Einrichtung: Department of Urology, Shaare Zedek Medical Center, Faculty of Medicine, Hebrew University, Jerusalem, Israel. Die Arbeit nennt 40 Literaturangaben. Gefördert von: Canada Research Chair in Quantitative Magnetic Resonance Imaging, Canadian Institute of Health Research, Canada Foundation for Innovation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

OBJECTIVES: To develop TotalSpineSeg, a robust tool for vertebrae and intervertebral disc instance segmentation, along with spinal cord and spinal canal segmentation, capable of processing any 3D medical image regardless of contrast, resolution, or field of view, while overcoming fusion and fragmentation errors common in standard segmentation. MATERIALS AND METHODS: A two-stage cascaded deep learning model was trained on 1404 MRIs. To explicitly prevent errors where adjacent vertebrae merge or a single vertebra splits, the architecture employs a novel "odd/even" representation. The second stage is conditioned on an odd-intervertebral disc mask from the first stage to assign distinct semantic classes to alternating vertebrae, enforcing accurate separation between neighboring vertebrae. Finally, a landmark-based logic uses key anatomical points (C2-C3, C7-T1, T12-L1, and L5-S) to anchor sequential labeling, while a localizer-based strategy ensures accurate identification even in truncated, short field-of-view scans. RESULTS: Evaluation across diverse contrasts and pathologies yielded a vertebral identification rate of 96.22% on localizer MRI and generalized to CT scans with a 92.87% identification rate. For intervertebral disc instance labeling, where vertebral separation is not mandatory, both model stages attained 99% labeling accuracy with a mean positional L2 error of 1.55 mm, surpassing existing state-of-the-art methods. CONCLUSION: TotalSpineSeg successfully generalizes to diverse 3D medical images, overcoming variations in contrast, resolution, and field of view. The proposed topological constraints effectively eliminate vertebral fusion and fragmentation errors, delivering a state-of-the-art identification rate for both vertebrae and intervertebral discs across MRI and CT modalities. KEY POINTS: Question Can a topology-constrained framework overcome heterogeneous contrasts and resolution limits, including localizers, to ensure robust, automated vertebral and disc labeling? Findings TotalSpineSeg achieved a high identification rate across diverse contrasts, localizers, and CT, surpassing state-of-the-art methods by eliminating fusion and fragmentation errors. Clinical relevance TotalSpineSeg enables fully automated biomarker extraction from clinical scans, allowing precise assessment of stenosis, lesions, and atrophy against normative ranges.

Schlagwörter

Stichwörter der Autor:innen

Intervertebral discLabelingMRISpineVertebrae

Autor:innen und Institutionen

  • Department of Urology, Shaare Zedek Medical Center, Faculty of Medicine, Hebrew University, Jerusalem, Israel.
  • Multiple Sclerosis Center, Sheba Medical Center, Ramat-Gan, Israel.
  • Arrow Program for Medical Research Education, Sheba Medical Center, Ramat-Gan, Israel.
  • NeuroPoly Lab, Institute of Biomedical Engineering, Polytechnique Montreal, Montreal, QC, Canada.
  • Mila-Quebec AI Institute, Montreal, QC, Canada.
  • Department of Radiology, Jefferson Integrated Magnetic Resonance Imaging Center, Thomas Jefferson University, Philadelphia, PA, USA.

Förderung

  • Canada Research Chair in Quantitative Magnetic Resonance Imaging
  • Canadian Institute of Health Research
  • Canada Foundation for Innovation
  • Fonds de Recherche du Québec - Santé
  • Natural Sciences and Engineering Research Council of Canada
  • Canada First Research Excellence Fund
  • Quebec BioImaging Network
  • National Institute of Neurological Disorders and Stroke (NINDS) of the National Institutes of Health (NIH)

Zitieren

APA

Warszawer, Y., Molinier, N., Valošek, J., Benveniste, P. L., Bédard, S., Shirbint, E., Mohamed, F. B., Tsagkas, C., Kolind, S., Lynd, L., Oh, J., Prat, A., Tam, R., Traboulsee, A., Patten, S., Lee, L. E., Achiron, A., & Cohen-Adad, J. (2026). TotalSpineSeg: automated whole-spine vertebral segmentation with high robustness to labeling errors. European Radiology. https://doi.org/10.1007/s00330-026-13008-2

Vancouver

Warszawer Y, Molinier N, Valošek J, Benveniste PL, Bédard S, Shirbint E, et al. TotalSpineSeg: automated whole-spine vertebral segmentation with high robustness to labeling errors. European Radiology. 2026. doi:10.1007/s00330-026-13008-2

BibTeX
@article{Warszawer2026,
  author = {Warszawer, Yehuda and Molinier, Nathan and Valošek, Jan and Benveniste, Pierre-Louis and Bédard, Sandrine and Shirbint, Emanuel and Mohamed, Feroze B and Tsagkas, Charidimos and Kolind, Shannon and Lynd, Larry and Oh, Jiwon and Prat, Alexandre and Tam, Roger and Traboulsee, Anthony and Patten, Scott and Lee, Lisa Eunyoung and Achiron, Anat and Cohen-Adad, Julien},
  title = {TotalSpineSeg: automated whole-spine vertebral segmentation with high robustness to labeling errors},
  journal = {European Radiology},
  year = {2026},
  doi = {10.1007/s00330-026-13008-2}
}
Kostenlos starten →