Automated segmentation of an intensity calibration phantom in clinical CT images using a convolutional neural network
Purpose In quantitative computed tomography (CT), manual selection of the intensity calibration phantom’s region of interest is necessary for calculating density (mg/cm 3 ) from the radiodensity values (Hounsfield units: HU). However, as this manual process requires effort and time, the purposes of...
Gespeichert in:
| Veröffentlicht in: | International journal for computer assisted radiology and surgery 2021-11, Vol.16 (11), p.1855-1864 |
|---|---|
| Hauptverfasser: | , , , , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 1864 |
|---|---|
| container_issue | 11 |
| container_start_page | 1855 |
| container_title | International journal for computer assisted radiology and surgery |
| container_volume | 16 |
| creator | Uemura, Keisuke Otake, Yoshito Takao, Masaki Soufi, Mazen Kawasaki, Akihiro Sugano, Nobuhiko Sato, Yoshinobu |
| description | Purpose
In quantitative computed tomography (CT), manual selection of the intensity calibration phantom’s region of interest is necessary for calculating density (mg/cm
3
) from the radiodensity values (Hounsfield units: HU). However, as this manual process requires effort and time, the purposes of this study were to develop a system that applies a convolutional neural network (CNN) to automatically segment intensity calibration phantom regions in CT images and to test the system in a large cohort to evaluate its robustness.
Methods
This cross-sectional, retrospective study included 1040 cases (520 each from two institutions) in which an intensity calibration phantom (B-MAS200, Kyoto Kagaku, Kyoto, Japan) was used. A training dataset was created by manually segmenting the phantom regions for 40 cases (20 cases for each institution). The CNN model’s segmentation accuracy was assessed with the Dice coefficient, and the average symmetric surface distance was assessed through fourfold cross-validation. Further, absolute difference of HU was compared between manually and automatically segmented regions. The system was tested on the remaining 1000 cases. For each institution, linear regression was applied to calculate the correlation coefficients between HU and phantom density.
Results
The source code and the model used for phantom segmentation can be accessed at
https://github.com/keisuke-uemura/CT-Intensity-Calibration-Phantom-Segmentation
. The median Dice coefficient was 0.977, and the median average symmetric surface distance was 0.116 mm. The median absolute difference of the segmented regions between manual and automated segmentation was 0.114 HU. For the test cases, the median correlation coefficients were 0.9998 and 0.999 for the two institutions, with a minimum value of 0.9863.
Conclusion
The proposed CNN model successfully segmented the calibration phantom regions in CT images with excellent accuracy. |
| doi_str_mv | 10.1007/s11548-021-02345-w |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_miscellaneous_2502805813</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2502805813</sourcerecordid><originalsourceid>FETCH-LOGICAL-c441t-706f770aa34501197e2ab12aeffe008287a3692d3ace16fa07bda2f6247889b93</originalsourceid><addsrcrecordid>eNp9kUtP3DAUha0KVB7tH2BRWWLTTeDaTmJniUa0RUJiA2vrJuNMDYk99aMj_n09hFKpCxbWtXS-c-yrQ8gZgwsGIC8jY02tKuCsHFE31e4DOWaqZVVb8-7g7c7giJzE-AhQN1I0H8mREFKAaPgxyVc5-RmTWdNoNrNxCZP1jvqRoqPWJeOiTc90wMn2YdG2P9EVU1HpMFlni0ZX99TOuDGR5mjdhiIdvPvtp7x3FN2ZHF5G2vnw9IkcjjhF8_l1npKHb9f3qx_V7d33m9XVbTXUNUuVhHaUEhDLcsBYJw3HnnE042gAFFcSRdvxtcDBsHZEkP0a-djyWirV9Z04JV-X3G3wv7KJSc82Dmaa0Bmfo-YNcAWNYqKg5_-hjz6H8vU91bVdA63aB_KFGoKPMZhRb0PZOzxrBnpfil5K0aUU_VKK3hXTl9fo3M9m_Wb520IBxALEIrmNCf_efif2D1XUmTc</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2596950689</pqid></control><display><type>article</type><title>Automated segmentation of an intensity calibration phantom in clinical CT images using a convolutional neural network</title><source>Springer Nature - Complete Springer Journals</source><creator>Uemura, Keisuke ; Otake, Yoshito ; Takao, Masaki ; Soufi, Mazen ; Kawasaki, Akihiro ; Sugano, Nobuhiko ; Sato, Yoshinobu</creator><creatorcontrib>Uemura, Keisuke ; Otake, Yoshito ; Takao, Masaki ; Soufi, Mazen ; Kawasaki, Akihiro ; Sugano, Nobuhiko ; Sato, Yoshinobu</creatorcontrib><description>Purpose
In quantitative computed tomography (CT), manual selection of the intensity calibration phantom’s region of interest is necessary for calculating density (mg/cm
3
) from the radiodensity values (Hounsfield units: HU). However, as this manual process requires effort and time, the purposes of this study were to develop a system that applies a convolutional neural network (CNN) to automatically segment intensity calibration phantom regions in CT images and to test the system in a large cohort to evaluate its robustness.
Methods
This cross-sectional, retrospective study included 1040 cases (520 each from two institutions) in which an intensity calibration phantom (B-MAS200, Kyoto Kagaku, Kyoto, Japan) was used. A training dataset was created by manually segmenting the phantom regions for 40 cases (20 cases for each institution). The CNN model’s segmentation accuracy was assessed with the Dice coefficient, and the average symmetric surface distance was assessed through fourfold cross-validation. Further, absolute difference of HU was compared between manually and automatically segmented regions. The system was tested on the remaining 1000 cases. For each institution, linear regression was applied to calculate the correlation coefficients between HU and phantom density.
Results
The source code and the model used for phantom segmentation can be accessed at
https://github.com/keisuke-uemura/CT-Intensity-Calibration-Phantom-Segmentation
. The median Dice coefficient was 0.977, and the median average symmetric surface distance was 0.116 mm. The median absolute difference of the segmented regions between manual and automated segmentation was 0.114 HU. For the test cases, the median correlation coefficients were 0.9998 and 0.999 for the two institutions, with a minimum value of 0.9863.
Conclusion
The proposed CNN model successfully segmented the calibration phantom regions in CT images with excellent accuracy.</description><identifier>ISSN: 1861-6410</identifier><identifier>EISSN: 1861-6429</identifier><identifier>DOI: 10.1007/s11548-021-02345-w</identifier><identifier>PMID: 33730352</identifier><language>eng</language><publisher>Cham: Springer International Publishing</publisher><subject>Artificial neural networks ; Automation ; Calibration ; Computed tomography ; Computer Imaging ; Computer Science ; Correlation coefficients ; Density ; Health Informatics ; Image segmentation ; Imaging ; Mathematical models ; Medical imaging ; Medicine ; Medicine & Public Health ; Model accuracy ; Neural networks ; Original Article ; Pattern Recognition and Graphics ; Radiology ; Source code ; Surgery ; Vision</subject><ispartof>International journal for computer assisted radiology and surgery, 2021-11, Vol.16 (11), p.1855-1864</ispartof><rights>CARS 2021</rights><rights>CARS 2021.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c441t-706f770aa34501197e2ab12aeffe008287a3692d3ace16fa07bda2f6247889b93</citedby><cites>FETCH-LOGICAL-c441t-706f770aa34501197e2ab12aeffe008287a3692d3ace16fa07bda2f6247889b93</cites><orcidid>0000-0002-9245-1743</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://link.springer.com/content/pdf/10.1007/s11548-021-02345-w$$EPDF$$P50$$Gspringer$$H</linktopdf><linktohtml>$$Uhttps://link.springer.com/10.1007/s11548-021-02345-w$$EHTML$$P50$$Gspringer$$H</linktohtml><link.rule.ids>314,776,780,27901,27902,41464,42533,51294</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/33730352$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Uemura, Keisuke</creatorcontrib><creatorcontrib>Otake, Yoshito</creatorcontrib><creatorcontrib>Takao, Masaki</creatorcontrib><creatorcontrib>Soufi, Mazen</creatorcontrib><creatorcontrib>Kawasaki, Akihiro</creatorcontrib><creatorcontrib>Sugano, Nobuhiko</creatorcontrib><creatorcontrib>Sato, Yoshinobu</creatorcontrib><title>Automated segmentation of an intensity calibration phantom in clinical CT images using a convolutional neural network</title><title>International journal for computer assisted radiology and surgery</title><addtitle>Int J CARS</addtitle><addtitle>Int J Comput Assist Radiol Surg</addtitle><description>Purpose
In quantitative computed tomography (CT), manual selection of the intensity calibration phantom’s region of interest is necessary for calculating density (mg/cm
3
) from the radiodensity values (Hounsfield units: HU). However, as this manual process requires effort and time, the purposes of this study were to develop a system that applies a convolutional neural network (CNN) to automatically segment intensity calibration phantom regions in CT images and to test the system in a large cohort to evaluate its robustness.
Methods
This cross-sectional, retrospective study included 1040 cases (520 each from two institutions) in which an intensity calibration phantom (B-MAS200, Kyoto Kagaku, Kyoto, Japan) was used. A training dataset was created by manually segmenting the phantom regions for 40 cases (20 cases for each institution). The CNN model’s segmentation accuracy was assessed with the Dice coefficient, and the average symmetric surface distance was assessed through fourfold cross-validation. Further, absolute difference of HU was compared between manually and automatically segmented regions. The system was tested on the remaining 1000 cases. For each institution, linear regression was applied to calculate the correlation coefficients between HU and phantom density.
Results
The source code and the model used for phantom segmentation can be accessed at
https://github.com/keisuke-uemura/CT-Intensity-Calibration-Phantom-Segmentation
. The median Dice coefficient was 0.977, and the median average symmetric surface distance was 0.116 mm. The median absolute difference of the segmented regions between manual and automated segmentation was 0.114 HU. For the test cases, the median correlation coefficients were 0.9998 and 0.999 for the two institutions, with a minimum value of 0.9863.
Conclusion
The proposed CNN model successfully segmented the calibration phantom regions in CT images with excellent accuracy.</description><subject>Artificial neural networks</subject><subject>Automation</subject><subject>Calibration</subject><subject>Computed tomography</subject><subject>Computer Imaging</subject><subject>Computer Science</subject><subject>Correlation coefficients</subject><subject>Density</subject><subject>Health Informatics</subject><subject>Image segmentation</subject><subject>Imaging</subject><subject>Mathematical models</subject><subject>Medical imaging</subject><subject>Medicine</subject><subject>Medicine & Public Health</subject><subject>Model accuracy</subject><subject>Neural networks</subject><subject>Original Article</subject><subject>Pattern Recognition and Graphics</subject><subject>Radiology</subject><subject>Source code</subject><subject>Surgery</subject><subject>Vision</subject><issn>1861-6410</issn><issn>1861-6429</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><recordid>eNp9kUtP3DAUha0KVB7tH2BRWWLTTeDaTmJniUa0RUJiA2vrJuNMDYk99aMj_n09hFKpCxbWtXS-c-yrQ8gZgwsGIC8jY02tKuCsHFE31e4DOWaqZVVb8-7g7c7giJzE-AhQN1I0H8mREFKAaPgxyVc5-RmTWdNoNrNxCZP1jvqRoqPWJeOiTc90wMn2YdG2P9EVU1HpMFlni0ZX99TOuDGR5mjdhiIdvPvtp7x3FN2ZHF5G2vnw9IkcjjhF8_l1npKHb9f3qx_V7d33m9XVbTXUNUuVhHaUEhDLcsBYJw3HnnE042gAFFcSRdvxtcDBsHZEkP0a-djyWirV9Z04JV-X3G3wv7KJSc82Dmaa0Bmfo-YNcAWNYqKg5_-hjz6H8vU91bVdA63aB_KFGoKPMZhRb0PZOzxrBnpfil5K0aUU_VKK3hXTl9fo3M9m_Wb520IBxALEIrmNCf_efif2D1XUmTc</recordid><startdate>20211101</startdate><enddate>20211101</enddate><creator>Uemura, Keisuke</creator><creator>Otake, Yoshito</creator><creator>Takao, Masaki</creator><creator>Soufi, Mazen</creator><creator>Kawasaki, Akihiro</creator><creator>Sugano, Nobuhiko</creator><creator>Sato, Yoshinobu</creator><general>Springer International Publishing</general><general>Springer Nature B.V</general><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7X8</scope><orcidid>https://orcid.org/0000-0002-9245-1743</orcidid></search><sort><creationdate>20211101</creationdate><title>Automated segmentation of an intensity calibration phantom in clinical CT images using a convolutional neural network</title><author>Uemura, Keisuke ; Otake, Yoshito ; Takao, Masaki ; Soufi, Mazen ; Kawasaki, Akihiro ; Sugano, Nobuhiko ; Sato, Yoshinobu</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c441t-706f770aa34501197e2ab12aeffe008287a3692d3ace16fa07bda2f6247889b93</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Artificial neural networks</topic><topic>Automation</topic><topic>Calibration</topic><topic>Computed tomography</topic><topic>Computer Imaging</topic><topic>Computer Science</topic><topic>Correlation coefficients</topic><topic>Density</topic><topic>Health Informatics</topic><topic>Image segmentation</topic><topic>Imaging</topic><topic>Mathematical models</topic><topic>Medical imaging</topic><topic>Medicine</topic><topic>Medicine & Public Health</topic><topic>Model accuracy</topic><topic>Neural networks</topic><topic>Original Article</topic><topic>Pattern Recognition and Graphics</topic><topic>Radiology</topic><topic>Source code</topic><topic>Surgery</topic><topic>Vision</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Uemura, Keisuke</creatorcontrib><creatorcontrib>Otake, Yoshito</creatorcontrib><creatorcontrib>Takao, Masaki</creatorcontrib><creatorcontrib>Soufi, Mazen</creatorcontrib><creatorcontrib>Kawasaki, Akihiro</creatorcontrib><creatorcontrib>Sugano, Nobuhiko</creatorcontrib><creatorcontrib>Sato, Yoshinobu</creatorcontrib><collection>PubMed</collection><collection>CrossRef</collection><collection>MEDLINE - Academic</collection><jtitle>International journal for computer assisted radiology and surgery</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Uemura, Keisuke</au><au>Otake, Yoshito</au><au>Takao, Masaki</au><au>Soufi, Mazen</au><au>Kawasaki, Akihiro</au><au>Sugano, Nobuhiko</au><au>Sato, Yoshinobu</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Automated segmentation of an intensity calibration phantom in clinical CT images using a convolutional neural network</atitle><jtitle>International journal for computer assisted radiology and surgery</jtitle><stitle>Int J CARS</stitle><addtitle>Int J Comput Assist Radiol Surg</addtitle><date>2021-11-01</date><risdate>2021</risdate><volume>16</volume><issue>11</issue><spage>1855</spage><epage>1864</epage><pages>1855-1864</pages><issn>1861-6410</issn><eissn>1861-6429</eissn><abstract>Purpose
In quantitative computed tomography (CT), manual selection of the intensity calibration phantom’s region of interest is necessary for calculating density (mg/cm
3
) from the radiodensity values (Hounsfield units: HU). However, as this manual process requires effort and time, the purposes of this study were to develop a system that applies a convolutional neural network (CNN) to automatically segment intensity calibration phantom regions in CT images and to test the system in a large cohort to evaluate its robustness.
Methods
This cross-sectional, retrospective study included 1040 cases (520 each from two institutions) in which an intensity calibration phantom (B-MAS200, Kyoto Kagaku, Kyoto, Japan) was used. A training dataset was created by manually segmenting the phantom regions for 40 cases (20 cases for each institution). The CNN model’s segmentation accuracy was assessed with the Dice coefficient, and the average symmetric surface distance was assessed through fourfold cross-validation. Further, absolute difference of HU was compared between manually and automatically segmented regions. The system was tested on the remaining 1000 cases. For each institution, linear regression was applied to calculate the correlation coefficients between HU and phantom density.
Results
The source code and the model used for phantom segmentation can be accessed at
https://github.com/keisuke-uemura/CT-Intensity-Calibration-Phantom-Segmentation
. The median Dice coefficient was 0.977, and the median average symmetric surface distance was 0.116 mm. The median absolute difference of the segmented regions between manual and automated segmentation was 0.114 HU. For the test cases, the median correlation coefficients were 0.9998 and 0.999 for the two institutions, with a minimum value of 0.9863.
Conclusion
The proposed CNN model successfully segmented the calibration phantom regions in CT images with excellent accuracy.</abstract><cop>Cham</cop><pub>Springer International Publishing</pub><pmid>33730352</pmid><doi>10.1007/s11548-021-02345-w</doi><tpages>10</tpages><orcidid>https://orcid.org/0000-0002-9245-1743</orcidid></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 1861-6410 |
| ispartof | International journal for computer assisted radiology and surgery, 2021-11, Vol.16 (11), p.1855-1864 |
| issn | 1861-6410 1861-6429 |
| language | eng |
| recordid | cdi_proquest_miscellaneous_2502805813 |
| source | Springer Nature - Complete Springer Journals |
| subjects | Artificial neural networks Automation Calibration Computed tomography Computer Imaging Computer Science Correlation coefficients Density Health Informatics Image segmentation Imaging Mathematical models Medical imaging Medicine Medicine & Public Health Model accuracy Neural networks Original Article Pattern Recognition and Graphics Radiology Source code Surgery Vision |
| title | Automated segmentation of an intensity calibration phantom in clinical CT images using a convolutional neural network |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-01T18%3A53%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Automated%20segmentation%20of%20an%20intensity%20calibration%20phantom%20in%20clinical%20CT%20images%20using%20a%20convolutional%20neural%20network&rft.jtitle=International%20journal%20for%20computer%20assisted%20radiology%20and%20surgery&rft.au=Uemura,%20Keisuke&rft.date=2021-11-01&rft.volume=16&rft.issue=11&rft.spage=1855&rft.epage=1864&rft.pages=1855-1864&rft.issn=1861-6410&rft.eissn=1861-6429&rft_id=info:doi/10.1007/s11548-021-02345-w&rft_dat=%3Cproquest_cross%3E2502805813%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2596950689&rft_id=info:pmid/33730352&rfr_iscdi=true |