View in EDS HTML Full Text PDF Full Text

SPINEPS—automatic whole spine segmentation of T2-weighted MR images using a two-phase approach to multi-class semantic and instance segmentation.

Saved in:
Bibliographic Details
Title: SPINEPS—automatic whole spine segmentation of T2-weighted MR images using a two-phase approach to multi-class semantic and instance segmentation.
Authors: Möller, Hendrik1,2 (AUTHOR) hendrik.moeller@tum.de, Graf, Robert1,2 (AUTHOR), Schmitt, Joachim1 (AUTHOR), Keinert, Benjamin1 (AUTHOR), Schön, Hanna1,3 (AUTHOR), Atad, Matan1,2 (AUTHOR), Sekuboyina, Anjany1,4 (AUTHOR), Streckenbach, Felix3 (AUTHOR), Kofler, Florian1,2,5,6 (AUTHOR), Kroencke, Thomas7,8 (AUTHOR), Bette, Stefanie7 (AUTHOR), Willich, Stefan N.9 (AUTHOR), Keil, Thomas9,10,11 (AUTHOR), Niendorf, Thoralf12 (AUTHOR), Pischon, Tobias13,14,15 (AUTHOR), Endemann, Beate12 (AUTHOR), Menze, Bjoern4 (AUTHOR), Rueckert, Daniel2,16 (AUTHOR), Kirschke, Jan S.1 (AUTHOR)
Source: European Radiology. Mar2025, Vol. 35 Issue 3, p1178-1189. 12p.
Subjects: Intervertebral disk, Spinal canal, Wilcoxon signed-rank test, Magnetic resonance imaging, Deep learning
Abstract: Objectives: 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. Material and methods: 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. Results: 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). Conclusion: 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. Key Points: QuestionNo publicly available automatic approach can yield semantic and instance segmentation masks for the whole spine (including posterior elements) in T2-weighted sagittal TSE images. FindingsSegmenting 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 relevanceThis 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. [ABSTRACT FROM AUTHOR]
Copyright of European Radiology is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
Text:
  Availability: 1
Header DbId: egs
DbLabel: Engineering Source
An: 183110198
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: SPINEPS—automatic whole spine segmentation of T2-weighted MR images using a two-phase approach to multi-class semantic and instance segmentation.
– Name: Author
  Label: Authors
  Group: Au
  Data: &lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22M&#246;ller%2C+Hendrik%22&quot;&gt;M&#246;ller, Hendrik&lt;/searchLink&gt;&lt;relatesTo&gt;1,2&lt;/relatesTo&gt; (AUTHOR)&lt;i&gt; hendrik.moeller@tum.de&lt;/i&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Graf%2C+Robert%22&quot;&gt;Graf, Robert&lt;/searchLink&gt;&lt;relatesTo&gt;1,2&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Schmitt%2C+Joachim%22&quot;&gt;Schmitt, Joachim&lt;/searchLink&gt;&lt;relatesTo&gt;1&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Keinert%2C+Benjamin%22&quot;&gt;Keinert, Benjamin&lt;/searchLink&gt;&lt;relatesTo&gt;1&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Sch&#246;n%2C+Hanna%22&quot;&gt;Sch&#246;n, Hanna&lt;/searchLink&gt;&lt;relatesTo&gt;1,3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Atad%2C+Matan%22&quot;&gt;Atad, Matan&lt;/searchLink&gt;&lt;relatesTo&gt;1,2&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Sekuboyina%2C+Anjany%22&quot;&gt;Sekuboyina, Anjany&lt;/searchLink&gt;&lt;relatesTo&gt;1,4&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Streckenbach%2C+Felix%22&quot;&gt;Streckenbach, Felix&lt;/searchLink&gt;&lt;relatesTo&gt;3&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Kofler%2C+Florian%22&quot;&gt;Kofler, Florian&lt;/searchLink&gt;&lt;relatesTo&gt;1,2,5,6&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Kroencke%2C+Thomas%22&quot;&gt;Kroencke, Thomas&lt;/searchLink&gt;&lt;relatesTo&gt;7,8&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Bette%2C+Stefanie%22&quot;&gt;Bette, Stefanie&lt;/searchLink&gt;&lt;relatesTo&gt;7&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Willich%2C+Stefan+N%2E%22&quot;&gt;Willich, Stefan N.&lt;/searchLink&gt;&lt;relatesTo&gt;9&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Keil%2C+Thomas%22&quot;&gt;Keil, Thomas&lt;/searchLink&gt;&lt;relatesTo&gt;9,10,11&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Niendorf%2C+Thoralf%22&quot;&gt;Niendorf, Thoralf&lt;/searchLink&gt;&lt;relatesTo&gt;12&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Pischon%2C+Tobias%22&quot;&gt;Pischon, Tobias&lt;/searchLink&gt;&lt;relatesTo&gt;13,14,15&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Endemann%2C+Beate%22&quot;&gt;Endemann, Beate&lt;/searchLink&gt;&lt;relatesTo&gt;12&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Menze%2C+Bjoern%22&quot;&gt;Menze, Bjoern&lt;/searchLink&gt;&lt;relatesTo&gt;4&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Rueckert%2C+Daniel%22&quot;&gt;Rueckert, Daniel&lt;/searchLink&gt;&lt;relatesTo&gt;2,16&lt;/relatesTo&gt; (AUTHOR)&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Kirschke%2C+Jan+S%2E%22&quot;&gt;Kirschke, Jan S.&lt;/searchLink&gt;&lt;relatesTo&gt;1&lt;/relatesTo&gt; (AUTHOR)
– Name: TitleSource
  Label: Source
  Group: Src
  Data: &lt;searchLink fieldCode=&quot;JN&quot; term=&quot;%22European+Radiology%22&quot;&gt;European Radiology&lt;/searchLink&gt;. Mar2025, Vol. 35 Issue 3, p1178-1189. 12p.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: &lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Intervertebral+disk%22&quot;&gt;Intervertebral disk&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Spinal+canal%22&quot;&gt;Spinal canal&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Wilcoxon+signed-rank+test%22&quot;&gt;Wilcoxon signed-rank test&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Magnetic+resonance+imaging%22&quot;&gt;Magnetic resonance imaging&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Deep+learning%22&quot;&gt;Deep learning&lt;/searchLink&gt;
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Objectives: 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. Material and methods: 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. Results: 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 &lt; 0.001, ASSD 0.21 vs 0.435, p &lt; 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). Conclusion: 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. Key Points: QuestionNo publicly available automatic approach can yield semantic and instance segmentation masks for the whole spine (including posterior elements) in T2-weighted sagittal TSE images. FindingsSegmenting 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 relevanceThis 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. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: &lt;i&gt;Copyright of European Radiology is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder&#39;s express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.&lt;/i&gt; (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=183110198
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1007/s00330-024-11155-y
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 12
        StartPage: 1178
    Subjects:
      – SubjectFull: Intervertebral disk
        Type: general
      – SubjectFull: Spinal canal
        Type: general
      – SubjectFull: Wilcoxon signed-rank test
        Type: general
      – SubjectFull: Magnetic resonance imaging
        Type: general
      – SubjectFull: Deep learning
        Type: general
    Titles:
      – TitleFull: SPINEPS—automatic whole spine segmentation of T2-weighted MR images using a two-phase approach to multi-class semantic and instance segmentation.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Möller, Hendrik
      – PersonEntity:
          Name:
            NameFull: Graf, Robert
      – PersonEntity:
          Name:
            NameFull: Schmitt, Joachim
      – PersonEntity:
          Name:
            NameFull: Keinert, Benjamin
      – PersonEntity:
          Name:
            NameFull: Schön, Hanna
      – PersonEntity:
          Name:
            NameFull: Atad, Matan
      – PersonEntity:
          Name:
            NameFull: Sekuboyina, Anjany
      – PersonEntity:
          Name:
            NameFull: Streckenbach, Felix
      – PersonEntity:
          Name:
            NameFull: Kofler, Florian
      – PersonEntity:
          Name:
            NameFull: Kroencke, Thomas
      – PersonEntity:
          Name:
            NameFull: Bette, Stefanie
      – PersonEntity:
          Name:
            NameFull: Willich, Stefan N.
      – PersonEntity:
          Name:
            NameFull: Keil, Thomas
      – PersonEntity:
          Name:
            NameFull: Niendorf, Thoralf
      – PersonEntity:
          Name:
            NameFull: Pischon, Tobias
      – PersonEntity:
          Name:
            NameFull: Endemann, Beate
      – PersonEntity:
          Name:
            NameFull: Menze, Bjoern
      – PersonEntity:
          Name:
            NameFull: Rueckert, Daniel
      – PersonEntity:
          Name:
            NameFull: Kirschke, Jan S.
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 03
              Text: Mar2025
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 09387994
          Numbering:
            – Type: volume
              Value: 35
            – Type: issue
              Value: 3
          Titles:
            – TitleFull: European Radiology
              Type: main
ResultId 1