SPINEPS-automatic whole spine segmentation of T2-weighted MR images using a two-phase approach to multi-class semantic and instance segmentation.
Deep learning
Intervertebral disc
Magnetic resonance imaging
Spine
Vertebral body
Journal
European radiology
ISSN: 1432-1084
Titre abrégé: Eur Radiol
Pays: Germany
ID NLM: 9114774
Informations de publication
Date de publication:
29 Oct 2024
29 Oct 2024
Historique:
received:
25
04
2024
accepted:
14
10
2024
revised:
02
09
2024
medline:
29
10
2024
pubmed:
29
10
2024
entrez:
29
10
2024
Statut:
aheadofprint
Résumé
Introducing SPINEPS, a deep learning method for semantic and instance segmentation of 14 spinal structures (ten vertebra substructures, intervertebral discs, spinal cord, spinal canal, and sacrum) in whole-body sagittal T2-weighted turbo spin echo images. This local ethics committee-approved study utilized a public dataset (train/test 179/39 subjects, 137 female), a German National Cohort (NAKO) subset (train/test 1412/65 subjects, mean age 53, 694 female), and an in-house dataset (test 10 subjects, mean age 70, 5 female). SPINEPS is a semantic segmentation model, followed by a sliding window approach utilizing a second model to create instance masks from the semantic ones. Segmentation evaluation metrics included the Dice score and average symmetrical surface distance (ASSD). Statistical significance was assessed using the Wilcoxon signed-rank test. On the public dataset, SPINEPS outperformed a nnUNet baseline on every structure and metric (e.g., an average over vertebra instances: dice 0.933 vs 0.911, p < 0.001, ASSD 0.21 vs 0.435, p < 0.001). SPINEPS trained on automated annotations of the NAKO achieves an average global Dice score of 0.918 on the combined NAKO and in-house test split. Adding the training data from the public dataset outperforms this (average instance-wise Dice score over the vertebra substructures 0.803 vs 0.778, average global Dice score 0.931 vs 0.918). SPINEPS offers segmentation of 14 spinal structures in T2w sagittal images. It provides a semantic mask and an instance mask separating the vertebrae and intervertebral discs. This is the first publicly available algorithm to enable this segmentation. Question No publicly available automatic approach can yield semantic and instance segmentation masks for the whole spine (including posterior elements) in T2-weighted sagittal TSE images. Findings Segmenting semantically first and then instance-wise outperforms a baseline trained directly on instance segmentation. The developed model produces high-resolution MRI segmentations for the whole spine. Clinical relevance This study introduces an automatic approach to whole spine segmentation, including posterior elements, in arbitrary fields of view T2w sagittal MR images, enabling easy biomarker extraction, automatic localization of pathologies and degenerative diseases, and quantifying analyses as downstream research.
Identifiants
pubmed: 39470797
doi: 10.1007/s00330-024-11155-y
pii: 10.1007/s00330-024-11155-y
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Subventions
Organisme : HORIZON EUROPE European Research Council
ID : 101045128
Informations de copyright
© 2024. The Author(s).
Références
Tamagawa S, Sakai D, Nojiri H, Sato M, Ishijima M, Watanabe M (2022) Imaging evaluation of intervertebral disc degeneration and painful discs—advances and challenges in quantitative MRI. Diagnostics 12:707. https://doi.org/10.3390/diagnostics12030707
doi: 10.3390/diagnostics12030707
Chen X, Sun S, Bai N et al (2021) A deep learning-based auto-segmentation system for organs-at-risk on whole-body computed tomography images for radiation therapy. Radiother Oncol 160:175–184. https://doi.org/10.1016/j.radonc.2021.04.019
doi: 10.1016/j.radonc.2021.04.019
Zheng H-D, Sun Y-L, Kong D-W et al (2022) Deep learning-based high-accuracy quantitation for lumbar intervertebral disc degeneration from MRI. Nat Commun 13:841. https://doi.org/10.1038/s41467-022-28387-5
doi: 10.1038/s41467-022-28387-5
Kowlagi N, Nguyen HH, McSweeney T et al (2022) A stronger baseline for automatic Pfirrmann grading of lumbar spine MRI using deep learning. Preprint at https://doi.org/10.48550/arXiv.2210.14597
Soydan Z, Bayramoglu E, Karasu R, Sayin I, Salturk S, Uvet H (2023) An automatized deep segmentation and classification model for lumbar disk degeneration and clarification of its impact on clinical decisions. Global Spine J 12:21925682231200783. https://doi.org/10.1177/21925682231200783
Almeida SD, Santinha J, Oliveira FPM et al (2020) Quantification of tumor burden in multiple myeloma by atlas-based semi-automatic segmentation of WB-DWI. Cancer Imaging 20:6. https://doi.org/10.1186/s40644-020-0286-5
doi: 10.1186/s40644-020-0286-5
van der Graaf JW, van Hooff ML, van Ginneken B et al (2024) Development and validation of AI-based automatic measurement of coronal Cobb angles in degenerative scoliosis using sagittal lumbar MRI. Eur Radiol. https://doi.org/10.1007/s00330-024-10616-8
Yao W, Bai J, Liao W, Chen Y, Liu M, Xie Y (2024) From CNN to Transformer:A Review of Medical Image Segmentation Models. J Imaging Inform Med 37:1529–1547. https://doi.org/10.1007/s10278-024-00981-7
Al Kafri AS, Sudirman S, Hussain AJ et al (2018) Segmentation of lumbar spine MRI images for stenosis detection using patch-based pixel classification neural network. 2018 IEEE Congress on evolutionary computation, CEC. IEEE, Rio de Janeiro. https://doi.org/10.1109/CEC.2018.8477893
Badarneh A, Abu-Qasmieh I, Otoom M, Alzubaidi MA (2021) Semi-automated spine and intervertebral disk detection and segmentation from whole spine MR images. Inform Med Unlocked 27:100810. https://doi.org/10.1016/j.imu.2021.100810
doi: 10.1016/j.imu.2021.100810
Khalil YA, Becherucci EA, Kirschke JS et al (2022) Multi-scanner and multi-modal lumbar vertebral body and intervertebral disc segmentation database. Sci Data 9:97. https://doi.org/10.1038/s41597-022-01222-8
doi: 10.1038/s41597-022-01222-8
Sáenz-Gamboa JJ, Domenech J, Alonso-Manjarrés A, Gómez JA, de la Iglesia-Vayá M (2023) Automatic semantic segmentation of the lumbar spine: clinical applicability in a multi-parametric and multi-center study on magnetic resonance images. Artif Intell Med 140:102559. https://doi.org/10.1016/j.artmed.2023.102559
doi: 10.1016/j.artmed.2023.102559
Shi R, Sun D, Qiu Z, Weiss K (2007) An efficient method for segmentation of MRI spine images. IEEE Xplore. https://doi.org/10.1109/ICCME.2007.4381830
Li H, Luo H, Huan W et al (2021) Automatic lumbar spinal MRI image segmentation with a multi-scale attention network. Neural Comput Appl 33:11589–11602. https://doi.org/10.1007/s00521-021-05856-4
doi: 10.1007/s00521-021-05856-4
Sekuboyina A, Husseini ME, Bayat A et al (2021) VerSe: a vertebrae labelling and segmentation benchmark for multi-detector CT images. Med Image Anal 73:102166. https://doi.org/10.1016/j.media.2021.102166
doi: 10.1016/j.media.2021.102166
Payer C, Štern D, Bischof H, Urschler M (2023) Coarse to fine vertebrae localization and segmentation with spatial configuration-Net and U-Net. https://www.scitepress.org/Link.aspx?doi=10.5220/0008975201240133 . Accessed 22 Nov 2023
Chen H, Shen C, Qin J et al (2015) Automatic localization and identification of vertebrae in spine CT via a joint learning model with deep neural networks. In: Navab N, Hornegger J, Wells WM, Frangi A (eds) Medical image computing and computer-assisted intervention—MICCAI 2015. Springer, Cham, pp 515–522. https://doi.org/10.1007/978-3-319-24553-9_63
Graf R, Schmitt J, Schlaeger S et al (2023) Denoising diffusion-based MRI to CT image translation enables automated spinal segmentation. Eur Radiol Exp 7:70. https://doi.org/10.1186/s41747-023-00385-2
doi: 10.1186/s41747-023-00385-2
Ronneberger O, Fischer P, Brox T (2015) U-Net: convolutional networks for biomedical image segmentation. In: Navab N, Hornegger J, Wells WM, Frangi AF (eds) Medical image computing and computer-assisted intervention—MICCAI 2015. Springer, Cham, pp 234–241. https://doi.org/10.1007/978-3-319-24574-4_28
Isensee F, Ulrich C, Wald T, Maier-Hein KH (2023) Extending nnU-Net is all you need. In: Deserno TM, Handels H, Maier A, Maier-Hein K, Palm C, Tolxdorff T (eds) Bildverarb Für Med 2023. Springer Fachmedien, Wiesbaden, pp 12–17. https://doi.org/10.1007/978-3-658-41657-7_7
Husseini M, Sekuboyina A, Bayat A, Menze BH, Loeffler M, Kirschke JS (2020) Conditioned variational auto-encoder for detecting osteoporotic vertebral fractures. In: Cai Y, Wang L, Audette M, Zheng G, Li S (eds) Computational methods and clinical applications for spine imaging. Springer, Cham, pp 29–38. https://doi.org/10.1007/978-3-030-39752-4_3
van der Graaf JW, van Hooff ML, Buckens CFM et al (2023) Lumbar spine segmentation in MR images: a dataset and a public benchmark. http://arxiv.org/abs/2306.12217 . Accessed 8 Jan 2024
Bamberg F, Kauczor H-U, Weckbach S et al (2015) Whole-body MR imaging in the German National Cohort: rationale, design, and technical background. Radiology 277:206–220. https://doi.org/10.1148/radiol.2015142272
doi: 10.1148/radiol.2015142272
Yushkevich PA, Piven J, Hazlett HC et al (2006) User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. Neuroimage 31:1116–1128. https://doi.org/10.1016/j.neuroimage.2006.01.015
doi: 10.1016/j.neuroimage.2006.01.015
Streckenbach F, Leifert G, Beyer T et al (2022) Application of a deep learning approach to analyze large-scale MRI data of the spine. Healthcare 10:2132. https://doi.org/10.3390/healthcare10112132
doi: 10.3390/healthcare10112132
Isensee F, Jaeger PF, Kohl SAA, Petersen J, Maier-Hein KH (2021) nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods 18:203–211. https://doi.org/10.1038/s41592-020-01008-z
doi: 10.1038/s41592-020-01008-z
De Leener B, Lévy S, Dupont SM et al (2017) SCT: spinal cord toolbox, an open-source software for processing spinal cord MRI data. Neuroimage 145:24–43. https://doi.org/10.1016/j.neuroimage.2016.10.009
doi: 10.1016/j.neuroimage.2016.10.009
Sekuboyina A, Rempfler M, Kukačka J et al (2018) Btrfly net: vertebrae labelling with energy-based adversarial learning of local spine prior. In: Frangi AF, Schnabel JA, Davatzikos C, Alberola-López C, Fichtinger G (eds) Medical image computing and computer assisted intervention—MICCAI 2018—21st international conference. Springer, Cham, pp 649–657. https://doi.org/10.1007/978-3-030-00937-3_74
Kirillov A, He K, Girshick R, Rother C, Dollár P (2019) Panoptic segmentation. In: Proceedings of 2019 IEEE/CVF conference on computer vision and pattern recognition (CVPR). IEEE, California, pp 9396–9405. https://doi.org/10.1109/CVPR.2019.00963
Kofler F, Möller H, Buchner JA et al (2023) Panoptica—instance-wise evaluation of 3D semantic and instance segmentation maps. http://arxiv.org/abs/2312.02608 . Accessed 6 Dec 2023
Wasserthal J, Breit H-C, Meyer MT et al (2023) TotalSegmentator: robust segmentation of 104 anatomic structures in CT images. Radiol Artif Intell 5:e230024. https://doi.org/10.1148/ryai.230024
doi: 10.1148/ryai.230024
Georgiev R, Novakova M, Bliznakova K (2023) Clinical Assessment of CoLumbo Deep Learning System for CentralCanal Stenosis Diagnostics. EJMO 7:42–48. https://doi.org/10.14744/ejmo.2023.59207
Du Plessis AM, Greyling LM, Page BJ (2018) Differentiation and classification of thoracolumbar transitional vertebrae. J Anat 232:850–856. https://doi.org/10.1111/joa.12781
doi: 10.1111/joa.12781
Meng D, Mohammed E, Boyer E, Pujades S (2022) 2022 Vertebrae localization, segmentation and identification using a graph optimization and an anatomic consistency cycle. In: Lian C, Cao X, Rekik I, Xu X, Cui Z (eds) MLMI 2022 machine learning in medical imaging, vol 13583. Springer, pp 307–317. https://doi.org/10.1007/978-3-031-21014-3_32