View in EDS HTML Full Text PDF Full Text

Autosegmentation for thoracic radiation treatment planning: A grand challenge at AAPM 2017.

Saved in:
Bibliographic Details
Title: Autosegmentation for thoracic radiation treatment planning: A grand challenge at AAPM 2017.
Authors: Yang, Jinzhong1 jyang4@mdanderson.org, Veeraraghavan, Harini2, Armato, Samuel G.3, Farahani, Keyvan4, Kirby, Justin S.5, Kalpathy‐Kramer, Jayashree6,7, van Elmpt, Wouter8, Dekker, Andre8, Han, Xiao9, Feng, Xue10, Aljabar, Paul11, Oliveira, Bruno12,13, van der Heyden, Brent8, Zamdborg, Leonid14, Lam, Dao15, Gooding, Mark11, Sharp, Gregory C.7
Source: Medical Physics. Oct2018, Vol. 45 Issue 10, p4568-4581. 14p.
Subjects: Radiotherapy treatment planning, Image segmentation, Lung cancer, Computed tomography, Deep learning, Machine learning
Abstract: Purpose: This report presents the methods and results of the Thoracic Auto‐Segmentation Challenge organized at the 2017 Annual Meeting of American Association of Physicists in Medicine. The purpose of the challenge was to provide a benchmark dataset and platform for evaluating performance of autosegmentation methods of organs at risk (OARs) in thoracic CT images. Methods : Sixty thoracic CT scans provided by three different institutions were separated into 36 training, 12 offline testing, and 12 online testing scans. Eleven participants completed the offline challenge, and seven completed the online challenge. The OARs were left and right lungs, heart, esophagus, and spinal cord. Clinical contours used for treatment planning were quality checked and edited to adhere to the RTOG 1106 contouring guidelines. Algorithms were evaluated using the Dice coefficient, Hausdorff distance, and mean surface distance. A consolidated score was computed by normalizing the metrics against interrater variability and averaging over all patients and structures. Results : The interrater study revealed highest variability in Dice for the esophagus and spinal cord, and in surface distances for lungs and heart. Five out of seven algorithms that participated in the online challenge employed deep‐learning methods. Although the top three participants using deep learning produced the best segmentation for all structures, there was no significant difference in the performance among them. The fourth place participant used a multi‐atlas‐based approach. The highest Dice scores were produced for lungs, with averages ranging from 0.95 to 0.98, while the lowest Dice scores were produced for esophagus, with a range of 0.55–0.72. Conclusion : The results of the challenge showed that the lungs and heart can be segmented fairly accurately by various algorithms, while deep‐learning methods performed better on the esophagus. Our dataset together with the manual contours for all training cases continues to be available publicly as an ongoing benchmarking resource. [ABSTRACT FROM AUTHOR]
Copyright of Medical Physics is the property of Wiley-Blackwell 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: 132313234
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Autosegmentation for thoracic radiation treatment planning: A grand challenge at AAPM 2017.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Yang%2C+Jinzhong%22">Yang, Jinzhong</searchLink><relatesTo>1</relatesTo><i> jyang4@mdanderson.org</i><br /><searchLink fieldCode="AR" term="%22Veeraraghavan%2C+Harini%22">Veeraraghavan, Harini</searchLink><relatesTo>2</relatesTo><br /><searchLink fieldCode="AR" term="%22Armato%2C+Samuel+G%2E%22">Armato, Samuel G.</searchLink><relatesTo>3</relatesTo><br /><searchLink fieldCode="AR" term="%22Farahani%2C+Keyvan%22">Farahani, Keyvan</searchLink><relatesTo>4</relatesTo><br /><searchLink fieldCode="AR" term="%22Kirby%2C+Justin+S%2E%22">Kirby, Justin S.</searchLink><relatesTo>5</relatesTo><br /><searchLink fieldCode="AR" term="%22Kalpathy‐Kramer%2C+Jayashree%22">Kalpathy‐Kramer, Jayashree</searchLink><relatesTo>6,7</relatesTo><br /><searchLink fieldCode="AR" term="%22van+Elmpt%2C+Wouter%22">van Elmpt, Wouter</searchLink><relatesTo>8</relatesTo><br /><searchLink fieldCode="AR" term="%22Dekker%2C+Andre%22">Dekker, Andre</searchLink><relatesTo>8</relatesTo><br /><searchLink fieldCode="AR" term="%22Han%2C+Xiao%22">Han, Xiao</searchLink><relatesTo>9</relatesTo><br /><searchLink fieldCode="AR" term="%22Feng%2C+Xue%22">Feng, Xue</searchLink><relatesTo>10</relatesTo><br /><searchLink fieldCode="AR" term="%22Aljabar%2C+Paul%22">Aljabar, Paul</searchLink><relatesTo>11</relatesTo><br /><searchLink fieldCode="AR" term="%22Oliveira%2C+Bruno%22">Oliveira, Bruno</searchLink><relatesTo>12,13</relatesTo><br /><searchLink fieldCode="AR" term="%22van+der+Heyden%2C+Brent%22">van der Heyden, Brent</searchLink><relatesTo>8</relatesTo><br /><searchLink fieldCode="AR" term="%22Zamdborg%2C+Leonid%22">Zamdborg, Leonid</searchLink><relatesTo>14</relatesTo><br /><searchLink fieldCode="AR" term="%22Lam%2C+Dao%22">Lam, Dao</searchLink><relatesTo>15</relatesTo><br /><searchLink fieldCode="AR" term="%22Gooding%2C+Mark%22">Gooding, Mark</searchLink><relatesTo>11</relatesTo><br /><searchLink fieldCode="AR" term="%22Sharp%2C+Gregory+C%2E%22">Sharp, Gregory C.</searchLink><relatesTo>7</relatesTo>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Medical+Physics%22">Medical Physics</searchLink>. Oct2018, Vol. 45 Issue 10, p4568-4581. 14p.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Radiotherapy+treatment+planning%22">Radiotherapy treatment planning</searchLink><br /><searchLink fieldCode="DE" term="%22Image+segmentation%22">Image segmentation</searchLink><br /><searchLink fieldCode="DE" term="%22Lung+cancer%22">Lung cancer</searchLink><br /><searchLink fieldCode="DE" term="%22Computed+tomography%22">Computed tomography</searchLink><br /><searchLink fieldCode="DE" term="%22Deep+learning%22">Deep learning</searchLink><br /><searchLink fieldCode="DE" term="%22Machine+learning%22">Machine learning</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Purpose: This report presents the methods and results of the Thoracic Auto‐Segmentation Challenge organized at the 2017 Annual Meeting of American Association of Physicists in Medicine. The purpose of the challenge was to provide a benchmark dataset and platform for evaluating performance of autosegmentation methods of organs at risk (OARs) in thoracic CT images. Methods : Sixty thoracic CT scans provided by three different institutions were separated into 36 training, 12 offline testing, and 12 online testing scans. Eleven participants completed the offline challenge, and seven completed the online challenge. The OARs were left and right lungs, heart, esophagus, and spinal cord. Clinical contours used for treatment planning were quality checked and edited to adhere to the RTOG 1106 contouring guidelines. Algorithms were evaluated using the Dice coefficient, Hausdorff distance, and mean surface distance. A consolidated score was computed by normalizing the metrics against interrater variability and averaging over all patients and structures. Results : The interrater study revealed highest variability in Dice for the esophagus and spinal cord, and in surface distances for lungs and heart. Five out of seven algorithms that participated in the online challenge employed deep‐learning methods. Although the top three participants using deep learning produced the best segmentation for all structures, there was no significant difference in the performance among them. The fourth place participant used a multi‐atlas‐based approach. The highest Dice scores were produced for lungs, with averages ranging from 0.95 to 0.98, while the lowest Dice scores were produced for esophagus, with a range of 0.55–0.72. Conclusion : The results of the challenge showed that the lungs and heart can be segmented fairly accurately by various algorithms, while deep‐learning methods performed better on the esophagus. Our dataset together with the manual contours for all training cases continues to be available publicly as an ongoing benchmarking resource. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Medical Physics is the property of Wiley-Blackwell 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.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=132313234
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1002/mp.13141
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 14
        StartPage: 4568
    Subjects:
      – SubjectFull: Radiotherapy treatment planning
        Type: general
      – SubjectFull: Image segmentation
        Type: general
      – SubjectFull: Lung cancer
        Type: general
      – SubjectFull: Computed tomography
        Type: general
      – SubjectFull: Deep learning
        Type: general
      – SubjectFull: Machine learning
        Type: general
    Titles:
      – TitleFull: Autosegmentation for thoracic radiation treatment planning: A grand challenge at AAPM 2017.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Yang, Jinzhong
      – PersonEntity:
          Name:
            NameFull: Veeraraghavan, Harini
      – PersonEntity:
          Name:
            NameFull: Armato, Samuel G.
      – PersonEntity:
          Name:
            NameFull: Farahani, Keyvan
      – PersonEntity:
          Name:
            NameFull: Kirby, Justin S.
      – PersonEntity:
          Name:
            NameFull: Kalpathy‐Kramer, Jayashree
      – PersonEntity:
          Name:
            NameFull: van Elmpt, Wouter
      – PersonEntity:
          Name:
            NameFull: Dekker, Andre
      – PersonEntity:
          Name:
            NameFull: Han, Xiao
      – PersonEntity:
          Name:
            NameFull: Feng, Xue
      – PersonEntity:
          Name:
            NameFull: Aljabar, Paul
      – PersonEntity:
          Name:
            NameFull: Oliveira, Bruno
      – PersonEntity:
          Name:
            NameFull: van der Heyden, Brent
      – PersonEntity:
          Name:
            NameFull: Zamdborg, Leonid
      – PersonEntity:
          Name:
            NameFull: Lam, Dao
      – PersonEntity:
          Name:
            NameFull: Gooding, Mark
      – PersonEntity:
          Name:
            NameFull: Sharp, Gregory C.
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 10
              Text: Oct2018
              Type: published
              Y: 2018
          Identifiers:
            – Type: issn-print
              Value: 00942405
          Numbering:
            – Type: volume
              Value: 45
            – Type: issue
              Value: 10
          Titles:
            – TitleFull: Medical Physics
              Type: main
ResultId 1