A Computed Tomography Vertebral Segmentation Dataset with Anatomical Variations and Multi-vendor Scanner Data

Overview

Journal Sci Data

Specialty Science

Date 2021 Oct 29

PMID 34711848

Citations 16

Authors

Hans Liebl

David Schinz

Anjany Sekuboyina

Luca Malagutti

Maximilian T Loffler

Amirhossein Bayat

Malek El Husseini

Giles Tetteh

Katharina Grau

Eva Niederreiter

Thomas Baum

Benedikt Wiestler

Bjoern Menze

Rickmer Braren

Claus Zimmer

Jan S Kirschke

Affiliations

Soon will be listed here.

Abstract

With the advent of deep learning algorithms, fully automated radiological image analysis is within reach. In spine imaging, several atlas- and shape-based as well as deep learning segmentation algorithms have been proposed, allowing for subsequent automated analysis of morphology and pathology. The first "Large Scale Vertebrae Segmentation Challenge" (VerSe 2019) showed that these perform well on normal anatomy, but fail in variants not frequently present in the training dataset. Building on that experience, we report on the largely increased VerSe 2020 dataset and results from the second iteration of the VerSe challenge (MICCAI 2020, Lima, Peru). VerSe 2020 comprises annotated spine computed tomography (CT) images from 300 subjects with 4142 fully visualized and annotated vertebrae, collected across multiple centres from four different scanner manufacturers, enriched with cases that exhibit anatomical variants such as enumeration abnormalities (n = 77) and transitional vertebrae (n = 161). Metadata includes vertebral labelling information, voxel-level segmentation masks obtained with a human-machine hybrid algorithm and anatomical ratings, to enable the development and benchmarking of robust and accurate segmentation algorithms.

Citing Articles

Enhanced Detection Performance of Acute Vertebral Compression Fractures Using a Hybrid Deep Learning and Traditional Quantitative Measurement Approach: Beyond the Limitations of Genant Classification.

Lee J, Kim M, Park H, Yang Z, Woo O, Kang W Bioengineering (Basel). 2025; 12(1).

PMID: 39851338 PMC: 11761558. DOI: 10.3390/bioengineering12010064.

Can back exosuits simultaneously increase lifting endurance and reduce musculoskeletal disorder risk?.

Rodzak K, Slaughter P, Wolf D, Ice C, Fine S, Zelik K Wearable Technol. 2025; 5():e17.

PMID: 39811480 PMC: 11729484. DOI: 10.1017/wtc.2024.8.

Integrating radiomics with clinical data for enhanced prediction of vertebral fracture risk.

Saravi B, Zink A, Tabukashvili E, Guzel H, Ulkumen S, Couillard-Despres S Front Bioeng Biotechnol. 2024; 12:1485364.

PMID: 39650236 PMC: 11620855. DOI: 10.3389/fbioe.2024.1485364.

Improved Detection Accuracy of Chronic Vertebral Compression Fractures by Integrating Height Loss Ratio and Deep Learning Approaches.

Lee J, Park H, Yang Z, Woo O, Kang W, Kim J Diagnostics (Basel). 2024; 14(22).

PMID: 39594143 PMC: 11593039. DOI: 10.3390/diagnostics14222477.

An open-access lumbosacral spine MRI dataset with enhanced spinal nerve root structure resolution.

Liu J, Zhang W, Zhou Y, Xu L, Chu Y, Jia F Sci Data. 2024; 11(1):1131.

PMID: 39406785 PMC: 11480038. DOI: 10.1038/s41597-024-03919-4.

References

Paik N, Lim C, Jang H . Numeric and morphological verification of lumbosacral segments in 8280 consecutive patients. Spine (Phila Pa 1976). 2013; 38(10):E573-8. DOI: 10.1097/BRS.0b013e31828b7195. View

Loffler M, Sekuboyina A, Jacob A, Grau A, Scharr A, El Husseini M . A Vertebral Segmentation Dataset with Fracture Grading. Radiol Artif Intell. 2021; 2(4):e190138. PMC: 8082364. DOI: 10.1148/ryai.2020190138. View

Hosny A, Parmar C, Quackenbush J, Schwartz L, Aerts H . Artificial intelligence in radiology. Nat Rev Cancer. 2018; 18(8):500-510. PMC: 6268174. DOI: 10.1038/s41568-018-0016-5. View

Sim Y, Chung M, Kotter E, Yune S, Kim M, Do S . Deep Convolutional Neural Network-based Software Improves Radiologist Detection of Malignant Lung Nodules on Chest Radiographs. Radiology. 2019; 294(1):199-209. DOI: 10.1148/radiol.2019182465. View

Rebsamen M, Knecht U, Reyes M, Wiest R, Meier R, McKinley R . Divide and Conquer: Stratifying Training Data by Tumor Grade Improves Deep Learning-Based Brain Tumor Segmentation. Front Neurosci. 2019; 13:1182. PMC: 6848279. DOI: 10.3389/fnins.2019.01182. View

Lessmann N, van Ginneken B, de Jong P, Isgum I . Iterative fully convolutional neural networks for automatic vertebra segmentation and identification. Med Image Anal. 2019; 53:142-155. DOI: 10.1016/j.media.2019.02.005. View

Rajpurkar P, OConnell C, Schechter A, Asnani N, Li J, Kiani A . CheXaid: deep learning assistance for physician diagnosis of tuberculosis using chest x-rays in patients with HIV. NPJ Digit Med. 2020; 3:115. PMC: 7481246. DOI: 10.1038/s41746-020-00322-2. View

Burns J, Yao J, Munoz H, Summers R . Automated Detection, Localization, and Classification of Traumatic Vertebral Body Fractures in the Thoracic and Lumbar Spine at CT. Radiology. 2015; 278(1):64-73. PMC: 4699497. DOI: 10.1148/radiol.2015142346. View

Glocker B, Zikic D, Konukoglu E, Haynor D, Criminisi A . Vertebrae localization in pathological spine CT via dense classification from sparse annotations. Med Image Comput Comput Assist Interv. 2014; 16(Pt 2):262-70. DOI: 10.1007/978-3-642-40763-5_33. View

10.

Plessis A, Greyling L, Page B . Differentiation and classification of thoracolumbar transitional vertebrae. J Anat. 2018; 232(5):850-856. PMC: 5879990. DOI: 10.1111/joa.12781. View

11.

Tins B, Balain B . Incidence of numerical variants and transitional lumbosacral vertebrae on whole-spine MRI. Insights Imaging. 2016; 7(2):199-203. PMC: 4805617. DOI: 10.1007/s13244-016-0468-7. View

12.

Yushkevich P, Piven J, Hazlett H, Smith R, Ho S, Gee J . User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. Neuroimage. 2006; 31(3):1116-28. DOI: 10.1016/j.neuroimage.2006.01.015. View

13.

Wigh R . The thoracolumbar and lumbosacral transitional junctions. Spine (Phila Pa 1976). 1980; 5(3):215-22. DOI: 10.1097/00007632-198005000-00003. View

14.

Bauer J, Muller D, Ambekar A, Dobritz M, Matsuura M, Eckstein F . Detection of osteoporotic vertebral fractures using multidetector CT. Osteoporos Int. 2006; 17(4):608-15. DOI: 10.1007/s00198-005-0023-8. View

15.

Hsieh C, Vanderford J, Moreau S, Prong T . Lumbosacral transitional segments: classification, prevalence, and effect on disk height. J Manipulative Physiol Ther. 2000; 23(7):483-9. DOI: 10.1067/mmt.2000.108817. View

16.

Burns J, Yao J, Summers R . Vertebral Body Compression Fractures and Bone Density: Automated Detection and Classification on CT Images. Radiology. 2017; 284(3):788-797. PMC: 5584647. DOI: 10.1148/radiol.2017162100. View

17.

Valentinitsch A, Trebeschi S, Alarcon E, Baum T, Kaesmacher J, Zimmer C . Regional analysis of age-related local bone loss in the spine of a healthy population using 3D voxel-based modeling. Bone. 2017; 103:233-240. DOI: 10.1016/j.bone.2017.06.013. View

18.

Sekuboyina A, Rempfler M, Valentinitsch A, Menze B, Kirschke J . Labeling Vertebrae with Two-dimensional Reformations of Multidetector CT Images: An Adversarial Approach for Incorporating Prior Knowledge of Spine Anatomy. Radiol Artif Intell. 2021; 2(2):e190074. PMC: 8017405. DOI: 10.1148/ryai.2020190074. View

19.

Ruiz-Espana S, Diaz-Parra A, Arana E, Moratal D . A fully automated level-set based segmentation method of thoracic and lumbar vertebral bodies in Computed Tomography images. Annu Int Conf IEEE Eng Med Biol Soc. 2016; 2015:3049-52. DOI: 10.1109/EMBC.2015.7319035. View

20.

Hou L, Bai X, Li H, Gao T, Li W, Wen T . Lumbar plain radiograph is not reliable to identify lumbosacral transitional vertebra types according to Castellvi classification principle. BMC Musculoskelet Disord. 2020; 21(1):333. PMC: 7260820. DOI: 10.1186/s12891-020-03358-3. View