Fully Automated, Semantic Segmentation of Whole-Body 18F-FDG PET/CT Images Based on Data-Centric Artificial Intelligence
We introduce multiple-organ objective segmentation (MOOSE) software that generates subject-specific, multiorgan segmentation using data-centric artificial intelligence principles to facilitate high-throughput systemic investigations of the human body via whole-body PET imaging. Methods: Image data f...
Uloženo v:
| Vydáno v: | The Journal of nuclear medicine (1978) Ročník 63; číslo 12; s. 1941 |
|---|---|
| Hlavní autoři: | , , , , , , , , , , , , , , , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
01.12.2022
|
| ISSN: | 1535-5667, 1535-5667 |
| On-line přístup: | Zjistit podrobnosti o přístupu |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstract | We introduce multiple-organ objective segmentation (MOOSE) software that generates subject-specific, multiorgan segmentation using data-centric artificial intelligence principles to facilitate high-throughput systemic investigations of the human body via whole-body PET imaging. Methods: Image data from 2 PET/CT systems were used in training MOOSE. For noncerebral structures, 50 whole-body CT images were used, 30 of which were acquired from healthy controls (14 men and 16 women), and 20 datasets were acquired from oncology patients (14 men and 6 women). Noncerebral tissues consisted of 13 abdominal organs, 20 bone segments, subcutaneous fat, visceral fat, psoas muscle, and skeletal muscle. An expert panel manually segmented all noncerebral structures except for subcutaneous fat, visceral fat, and skeletal muscle, which were semiautomatically segmented using thresholding. A majority-voting algorithm was used to generate a reference-standard segmentation. From the 50 CT datasets, 40 were used for training and 10 for testing. For cerebral structures, 34 18F-FDG PET/MRI brain image volumes were used from 10 healthy controls (5 men and 5 women imaged twice) and 14 nonlesional epilepsy patients (7 men and 7 women). Only 18F-FDG PET images were considered for training: 24 and 10 of 34 volumes were used for training and testing, respectively. The Dice score coefficient (DSC) was used as the primary metric, and the average symmetric surface distance as a secondary metric, to evaluate the automated segmentation performance. Results: An excellent overlap between the reference labels and MOOSE-derived organ segmentations was observed: 92% of noncerebral tissues showed DSCs of more than 0.90, whereas a few organs exhibited lower DSCs (e.g., adrenal glands [0.72], pancreas [0.85], and bladder [0.86]). The median DSCs of brain subregions derived from PET images were lower. Only 29% of the brain segments had a median DSC of more than 0.90, whereas segmentation of 60% of regions yielded a median DSC of 0.80-0.89. The results of the average symmetric surface distance analysis demonstrated that the average distance between the reference standard and the automatically segmented tissue surfaces (organs, bones, and brain regions) lies within the size of image voxels (2 mm). Conclusion: The proposed segmentation pipeline allows automatic segmentation of 120 unique tissues from whole-body 18F-FDG PET/CT images with high accuracy.We introduce multiple-organ objective segmentation (MOOSE) software that generates subject-specific, multiorgan segmentation using data-centric artificial intelligence principles to facilitate high-throughput systemic investigations of the human body via whole-body PET imaging. Methods: Image data from 2 PET/CT systems were used in training MOOSE. For noncerebral structures, 50 whole-body CT images were used, 30 of which were acquired from healthy controls (14 men and 16 women), and 20 datasets were acquired from oncology patients (14 men and 6 women). Noncerebral tissues consisted of 13 abdominal organs, 20 bone segments, subcutaneous fat, visceral fat, psoas muscle, and skeletal muscle. An expert panel manually segmented all noncerebral structures except for subcutaneous fat, visceral fat, and skeletal muscle, which were semiautomatically segmented using thresholding. A majority-voting algorithm was used to generate a reference-standard segmentation. From the 50 CT datasets, 40 were used for training and 10 for testing. For cerebral structures, 34 18F-FDG PET/MRI brain image volumes were used from 10 healthy controls (5 men and 5 women imaged twice) and 14 nonlesional epilepsy patients (7 men and 7 women). Only 18F-FDG PET images were considered for training: 24 and 10 of 34 volumes were used for training and testing, respectively. The Dice score coefficient (DSC) was used as the primary metric, and the average symmetric surface distance as a secondary metric, to evaluate the automated segmentation performance. Results: An excellent overlap between the reference labels and MOOSE-derived organ segmentations was observed: 92% of noncerebral tissues showed DSCs of more than 0.90, whereas a few organs exhibited lower DSCs (e.g., adrenal glands [0.72], pancreas [0.85], and bladder [0.86]). The median DSCs of brain subregions derived from PET images were lower. Only 29% of the brain segments had a median DSC of more than 0.90, whereas segmentation of 60% of regions yielded a median DSC of 0.80-0.89. The results of the average symmetric surface distance analysis demonstrated that the average distance between the reference standard and the automatically segmented tissue surfaces (organs, bones, and brain regions) lies within the size of image voxels (2 mm). Conclusion: The proposed segmentation pipeline allows automatic segmentation of 120 unique tissues from whole-body 18F-FDG PET/CT images with high accuracy. |
|---|---|
| AbstractList | We introduce multiple-organ objective segmentation (MOOSE) software that generates subject-specific, multiorgan segmentation using data-centric artificial intelligence principles to facilitate high-throughput systemic investigations of the human body via whole-body PET imaging. Methods: Image data from 2 PET/CT systems were used in training MOOSE. For noncerebral structures, 50 whole-body CT images were used, 30 of which were acquired from healthy controls (14 men and 16 women), and 20 datasets were acquired from oncology patients (14 men and 6 women). Noncerebral tissues consisted of 13 abdominal organs, 20 bone segments, subcutaneous fat, visceral fat, psoas muscle, and skeletal muscle. An expert panel manually segmented all noncerebral structures except for subcutaneous fat, visceral fat, and skeletal muscle, which were semiautomatically segmented using thresholding. A majority-voting algorithm was used to generate a reference-standard segmentation. From the 50 CT datasets, 40 were used for training and 10 for testing. For cerebral structures, 34 18F-FDG PET/MRI brain image volumes were used from 10 healthy controls (5 men and 5 women imaged twice) and 14 nonlesional epilepsy patients (7 men and 7 women). Only 18F-FDG PET images were considered for training: 24 and 10 of 34 volumes were used for training and testing, respectively. The Dice score coefficient (DSC) was used as the primary metric, and the average symmetric surface distance as a secondary metric, to evaluate the automated segmentation performance. Results: An excellent overlap between the reference labels and MOOSE-derived organ segmentations was observed: 92% of noncerebral tissues showed DSCs of more than 0.90, whereas a few organs exhibited lower DSCs (e.g., adrenal glands [0.72], pancreas [0.85], and bladder [0.86]). The median DSCs of brain subregions derived from PET images were lower. Only 29% of the brain segments had a median DSC of more than 0.90, whereas segmentation of 60% of regions yielded a median DSC of 0.80-0.89. The results of the average symmetric surface distance analysis demonstrated that the average distance between the reference standard and the automatically segmented tissue surfaces (organs, bones, and brain regions) lies within the size of image voxels (2 mm). Conclusion: The proposed segmentation pipeline allows automatic segmentation of 120 unique tissues from whole-body 18F-FDG PET/CT images with high accuracy.We introduce multiple-organ objective segmentation (MOOSE) software that generates subject-specific, multiorgan segmentation using data-centric artificial intelligence principles to facilitate high-throughput systemic investigations of the human body via whole-body PET imaging. Methods: Image data from 2 PET/CT systems were used in training MOOSE. For noncerebral structures, 50 whole-body CT images were used, 30 of which were acquired from healthy controls (14 men and 16 women), and 20 datasets were acquired from oncology patients (14 men and 6 women). Noncerebral tissues consisted of 13 abdominal organs, 20 bone segments, subcutaneous fat, visceral fat, psoas muscle, and skeletal muscle. An expert panel manually segmented all noncerebral structures except for subcutaneous fat, visceral fat, and skeletal muscle, which were semiautomatically segmented using thresholding. A majority-voting algorithm was used to generate a reference-standard segmentation. From the 50 CT datasets, 40 were used for training and 10 for testing. For cerebral structures, 34 18F-FDG PET/MRI brain image volumes were used from 10 healthy controls (5 men and 5 women imaged twice) and 14 nonlesional epilepsy patients (7 men and 7 women). Only 18F-FDG PET images were considered for training: 24 and 10 of 34 volumes were used for training and testing, respectively. The Dice score coefficient (DSC) was used as the primary metric, and the average symmetric surface distance as a secondary metric, to evaluate the automated segmentation performance. Results: An excellent overlap between the reference labels and MOOSE-derived organ segmentations was observed: 92% of noncerebral tissues showed DSCs of more than 0.90, whereas a few organs exhibited lower DSCs (e.g., adrenal glands [0.72], pancreas [0.85], and bladder [0.86]). The median DSCs of brain subregions derived from PET images were lower. Only 29% of the brain segments had a median DSC of more than 0.90, whereas segmentation of 60% of regions yielded a median DSC of 0.80-0.89. The results of the average symmetric surface distance analysis demonstrated that the average distance between the reference standard and the automatically segmented tissue surfaces (organs, bones, and brain regions) lies within the size of image voxels (2 mm). Conclusion: The proposed segmentation pipeline allows automatic segmentation of 120 unique tissues from whole-body 18F-FDG PET/CT images with high accuracy. |
| Author | Kulterer, Oana C Kifjak, Daria Shin, Hyung Min Muzik, Otto Nakuz, Thomas Hacker, Marcus Nardo, Lorenzo Cherry, Simon R Shiyam Sundar, Lalith Kumar Kitzmantl, Daniela Beyer, Thomas Yu, Josef Sima, Annika Katharina Badawi, Ramsey D Fueger, Barbara Spencer, Benjamin A |
| Author_xml | – sequence: 1 givenname: Lalith Kumar surname: Shiyam Sundar fullname: Shiyam Sundar, Lalith Kumar – sequence: 2 givenname: Josef surname: Yu fullname: Yu, Josef – sequence: 3 givenname: Otto surname: Muzik fullname: Muzik, Otto – sequence: 4 givenname: Oana C surname: Kulterer fullname: Kulterer, Oana C – sequence: 5 givenname: Barbara surname: Fueger fullname: Fueger, Barbara – sequence: 6 givenname: Daria surname: Kifjak fullname: Kifjak, Daria – sequence: 7 givenname: Thomas surname: Nakuz fullname: Nakuz, Thomas – sequence: 8 givenname: Hyung Min surname: Shin fullname: Shin, Hyung Min – sequence: 9 givenname: Annika Katharina surname: Sima fullname: Sima, Annika Katharina – sequence: 10 givenname: Daniela surname: Kitzmantl fullname: Kitzmantl, Daniela – sequence: 11 givenname: Ramsey D surname: Badawi fullname: Badawi, Ramsey D – sequence: 12 givenname: Lorenzo surname: Nardo fullname: Nardo, Lorenzo – sequence: 13 givenname: Simon R surname: Cherry fullname: Cherry, Simon R – sequence: 14 givenname: Benjamin A surname: Spencer fullname: Spencer, Benjamin A – sequence: 15 givenname: Marcus surname: Hacker fullname: Hacker, Marcus – sequence: 16 givenname: Thomas surname: Beyer fullname: Beyer, Thomas |
| BookMark | eNpNjsFPwjAchRuDiYD-Ad569OCg7bauPcJgSEKiiRiPpKy_YsnWIu0S-O9dogdP7zu89-WN0MB5Bwg9UjJhkhfTo-ta0BPK2ITxjPD0Bg1pnuZJznkx-Md3aBTCkRDChRBDdKm6prniWRd9qyLoZ_wOrXLR1j0cWnBRResd9gZ_fvkGkrnXV0xFlVSLFX5bbqflFq9bdYCA5yqAxn15oaJKyn577jWzc7TG1lY1eO0iNI09gKvhHt0a1QR4-Msx-qiW2_Il2byu1uVsk5xYQWPCaAGai4wbqDnLMyIhTTMmjJSS6GwPQmpNM8FrI4nMjdlLoDmwNFcKTJ2yMXr69Z7O_ruDEHetDXV_QznwXdixXk4JFYVgP2iVYpo |
| ContentType | Journal Article |
| Copyright | 2022 by the Society of Nuclear Medicine and Molecular Imaging. |
| Copyright_xml | – notice: 2022 by the Society of Nuclear Medicine and Molecular Imaging. |
| DBID | 7X8 |
| DOI | 10.2967/jnumed.122.264063 |
| DatabaseName | MEDLINE - Academic |
| DatabaseTitle | MEDLINE - Academic |
| DatabaseTitleList | MEDLINE - Academic |
| Database_xml | – sequence: 1 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | no_fulltext_linktorsrc |
| Discipline | Medicine |
| EISSN | 1535-5667 |
| GroupedDBID | --- -~X .55 29L 2WC 53G 5RE 7X8 8R4 8R5 8WZ A6W ABSQV ACGOD ACIWK ACPRK ADDZX ADMOG AENEX AFRAH AHMBA ALMA_UNASSIGNED_HOLDINGS CS3 DIK DU5 E3Z EBD EBS EJD EMOBN F5P F9R H13 I-F INIJC KQ8 L7B OK1 P2P Q2X R0Z RHI RNS RWL SJN SV3 TAE TR2 TSM TUS W8F WH7 WOQ X7M YHG |
| ID | FETCH-LOGICAL-p271t-217ed6846fec625409e33428f9990d4be89dd1486cf9095ffb9e15e235aaefc32 |
| IEDL.DBID | 7X8 |
| ISICitedReferencesCount | 90 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000908018100023&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 1535-5667 |
| IngestDate | Thu Sep 04 16:30:29 EDT 2025 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 12 |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-p271t-217ed6846fec625409e33428f9990d4be89dd1486cf9095ffb9e15e235aaefc32 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| PQID | 2684101878 |
| PQPubID | 23479 |
| ParticipantIDs | proquest_miscellaneous_2684101878 |
| PublicationCentury | 2000 |
| PublicationDate | 2022-12-01 |
| PublicationDateYYYYMMDD | 2022-12-01 |
| PublicationDate_xml | – month: 12 year: 2022 text: 2022-12-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationTitle | The Journal of nuclear medicine (1978) |
| PublicationYear | 2022 |
| SSID | ssj0006888 |
| Score | 2.6529708 |
| Snippet | We introduce multiple-organ objective segmentation (MOOSE) software that generates subject-specific, multiorgan segmentation using data-centric artificial... |
| SourceID | proquest |
| SourceType | Aggregation Database |
| StartPage | 1941 |
| Title | Fully Automated, Semantic Segmentation of Whole-Body 18F-FDG PET/CT Images Based on Data-Centric Artificial Intelligence |
| URI | https://www.proquest.com/docview/2684101878 |
| Volume | 63 |
| WOSCitedRecordID | wos000908018100023&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV3JTsMwELWAIsSFHbHLSBxxm9hZnBPqFqgEVSUK9FY5sY2K1KSQFNG_Z5ymKuKExC0HK8tkljcz9huErnwtuIyVQW4sIo6OLcIh8BHmKiqF0BEv6Bie7_1ulw8GQa8suGXltsqFTywctUxjUyOvGVYSwy7l85vJOzFTo0x3tRyhsYoqDKCM0Wp_sGQL93gxdxKM2iUAW_x5V5MGnl97S8D0ZdWmtAqQwDIsoL98cRFgwu3_vtoO2iqhJa7PdWEXrahkD208lM3zffRl0s0Zrk_zFGCqktf4UY1BsKMYLl7H5SGkBKcav5ipuaSRyhm2eUjC1i3utfu1Zh93xuCAMtyA4CcxLG6JXJCiRAy3MY-eM1Lgzg-qzwP0FLb7zTtSDl4gE-rbOYE0RUn4KE-rGPIjSAEVY5CnaECTlnQixQMpIY_yYh0ARNM6CpQNP5e5QigdM3qI1pI0UUcIC0FtSZkPONFynEBGknEaGIYfqZlwg2N0uZDpEBTbdCtEotJpNlxK9eQPa07RJjUHE4qNJmeoosF41Tlajz_zUfZxUejFN-PDwWE |
| linkProvider | ProQuest |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Fully+Automated%2C+Semantic+Segmentation+of+Whole-Body+18F-FDG+PET%2FCT+Images+Based+on+Data-Centric+Artificial+Intelligence&rft.jtitle=The+Journal+of+nuclear+medicine+%281978%29&rft.au=Shiyam+Sundar%2C+Lalith+Kumar&rft.au=Yu%2C+Josef&rft.au=Muzik%2C+Otto&rft.au=Kulterer%2C+Oana+C&rft.date=2022-12-01&rft.issn=1535-5667&rft.eissn=1535-5667&rft.volume=63&rft.issue=12&rft.spage=1941&rft_id=info:doi/10.2967%2Fjnumed.122.264063&rft.externalDBID=NO_FULL_TEXT |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1535-5667&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1535-5667&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1535-5667&client=summon |