Fast meningioma segmentation in T1-weighted magnetic resonance imaging volumes using a lightweight 3D deep learning architecture

Automatic and consistent meningioma segmentation in T1-weighted magnetic resonance (MR) imaging volumes and corresponding volumetric assessment is of use for diagnosis, treatment planning, and tumor growth evaluation. We optimized the segmentation and processing speed performances using a large numb...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of medical imaging (Bellingham, Wash.) Ročník 8; číslo 2; s. 024002
Hlavní autoři: Bouget, David, Pedersen, André, Hosainey, Sayied Abdol Mohieb, Vanel, Johanna, Solheim, Ole, Reinertsen, Ingerid
Médium: Journal Article
Jazyk:angličtina
Vydáno: United States 01.03.2021
Témata:
ISSN:2329-4302
On-line přístup:Zjistit podrobnosti o přístupu
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract Automatic and consistent meningioma segmentation in T1-weighted magnetic resonance (MR) imaging volumes and corresponding volumetric assessment is of use for diagnosis, treatment planning, and tumor growth evaluation. We optimized the segmentation and processing speed performances using a large number of both surgically treated meningiomas and untreated meningiomas followed at the outpatient clinic. We studied two different three-dimensional (3D) neural network architectures: (i) a simple encoder-decoder similar to a 3D U-Net, and (ii) a lightweight multi-scale architecture [Pulmonary Lobe Segmentation Network (PLS-Net)]. In addition, we studied the impact of different training schemes. For the validation studies, we used 698 T1-weighted MR volumes from St. Olav University Hospital, Trondheim, Norway. The models were evaluated in terms of detection accuracy, segmentation accuracy, and training/inference speed. While both architectures reached a similar Dice score of 70% on average, the PLS-Net was more accurate with an -score of up to 88%. The highest accuracy was achieved for the largest meningiomas. Speed-wise, the PLS-Net architecture tended to converge in about 50 h while 130 h were necessary for U-Net. Inference with PLS-Net takes less than a second on GPU and about 15 s on CPU. Overall, with the use of mixed precision training, it was possible to train competitive segmentation models in a relatively short amount of time using the lightweight PLS-Net architecture. In the future, the focus should be brought toward the segmentation of small meningiomas ( ) to improve clinical relevance for automatic and early diagnosis and speed of growth estimates.
AbstractList Purpose: Automatic and consistent meningioma segmentation in T1-weighted magnetic resonance (MR) imaging volumes and corresponding volumetric assessment is of use for diagnosis, treatment planning, and tumor growth evaluation. We optimized the segmentation and processing speed performances using a large number of both surgically treated meningiomas and untreated meningiomas followed at the outpatient clinic. Approach: We studied two different three-dimensional (3D) neural network architectures: (i) a simple encoder-decoder similar to a 3D U-Net, and (ii) a lightweight multi-scale architecture [Pulmonary Lobe Segmentation Network (PLS-Net)]. In addition, we studied the impact of different training schemes. For the validation studies, we used 698 T1-weighted MR volumes from St. Olav University Hospital, Trondheim, Norway. The models were evaluated in terms of detection accuracy, segmentation accuracy, and training/inference speed. Results: While both architectures reached a similar Dice score of 70% on average, the PLS-Net was more accurate with an F 1 -score of up to 88%. The highest accuracy was achieved for the largest meningiomas. Speed-wise, the PLS-Net architecture tended to converge in about 50 h while 130 h were necessary for U-Net. Inference with PLS-Net takes less than a second on GPU and about 15 s on CPU. Conclusions: Overall, with the use of mixed precision training, it was possible to train competitive segmentation models in a relatively short amount of time using the lightweight PLS-Net architecture. In the future, the focus should be brought toward the segmentation of small meningiomas ( < 2    ml ) to improve clinical relevance for automatic and early diagnosis and speed of growth estimates.Purpose: Automatic and consistent meningioma segmentation in T1-weighted magnetic resonance (MR) imaging volumes and corresponding volumetric assessment is of use for diagnosis, treatment planning, and tumor growth evaluation. We optimized the segmentation and processing speed performances using a large number of both surgically treated meningiomas and untreated meningiomas followed at the outpatient clinic. Approach: We studied two different three-dimensional (3D) neural network architectures: (i) a simple encoder-decoder similar to a 3D U-Net, and (ii) a lightweight multi-scale architecture [Pulmonary Lobe Segmentation Network (PLS-Net)]. In addition, we studied the impact of different training schemes. For the validation studies, we used 698 T1-weighted MR volumes from St. Olav University Hospital, Trondheim, Norway. The models were evaluated in terms of detection accuracy, segmentation accuracy, and training/inference speed. Results: While both architectures reached a similar Dice score of 70% on average, the PLS-Net was more accurate with an F 1 -score of up to 88%. The highest accuracy was achieved for the largest meningiomas. Speed-wise, the PLS-Net architecture tended to converge in about 50 h while 130 h were necessary for U-Net. Inference with PLS-Net takes less than a second on GPU and about 15 s on CPU. Conclusions: Overall, with the use of mixed precision training, it was possible to train competitive segmentation models in a relatively short amount of time using the lightweight PLS-Net architecture. In the future, the focus should be brought toward the segmentation of small meningiomas ( < 2    ml ) to improve clinical relevance for automatic and early diagnosis and speed of growth estimates.
Automatic and consistent meningioma segmentation in T1-weighted magnetic resonance (MR) imaging volumes and corresponding volumetric assessment is of use for diagnosis, treatment planning, and tumor growth evaluation. We optimized the segmentation and processing speed performances using a large number of both surgically treated meningiomas and untreated meningiomas followed at the outpatient clinic. We studied two different three-dimensional (3D) neural network architectures: (i) a simple encoder-decoder similar to a 3D U-Net, and (ii) a lightweight multi-scale architecture [Pulmonary Lobe Segmentation Network (PLS-Net)]. In addition, we studied the impact of different training schemes. For the validation studies, we used 698 T1-weighted MR volumes from St. Olav University Hospital, Trondheim, Norway. The models were evaluated in terms of detection accuracy, segmentation accuracy, and training/inference speed. While both architectures reached a similar Dice score of 70% on average, the PLS-Net was more accurate with an -score of up to 88%. The highest accuracy was achieved for the largest meningiomas. Speed-wise, the PLS-Net architecture tended to converge in about 50 h while 130 h were necessary for U-Net. Inference with PLS-Net takes less than a second on GPU and about 15 s on CPU. Overall, with the use of mixed precision training, it was possible to train competitive segmentation models in a relatively short amount of time using the lightweight PLS-Net architecture. In the future, the focus should be brought toward the segmentation of small meningiomas ( ) to improve clinical relevance for automatic and early diagnosis and speed of growth estimates.
Author Bouget, David
Solheim, Ole
Pedersen, André
Vanel, Johanna
Reinertsen, Ingerid
Hosainey, Sayied Abdol Mohieb
Author_xml – sequence: 1
  givenname: David
  orcidid: 0000-0002-5669-9514
  surname: Bouget
  fullname: Bouget, David
  organization: SINTEF, Medical Technology Department, Trondheim, Norway
– sequence: 2
  givenname: André
  orcidid: 0000-0002-3637-953X
  surname: Pedersen
  fullname: Pedersen, André
  organization: SINTEF, Medical Technology Department, Trondheim, Norway
– sequence: 3
  givenname: Sayied Abdol Mohieb
  surname: Hosainey
  fullname: Hosainey, Sayied Abdol Mohieb
  organization: Bristol Royal Hospital for Children, Department of Neurosurgery, Bristol, United Kingdom
– sequence: 4
  givenname: Johanna
  surname: Vanel
  fullname: Vanel, Johanna
  organization: SINTEF, Medical Technology Department, Trondheim, Norway
– sequence: 5
  givenname: Ole
  surname: Solheim
  fullname: Solheim, Ole
  organization: St. Olavs Hospital, Department of Neurosurgery, Trondheim, Norway
– sequence: 6
  givenname: Ingerid
  orcidid: 0000-0003-0999-3849
  surname: Reinertsen
  fullname: Reinertsen, Ingerid
  organization: SINTEF, Medical Technology Department, Trondheim, Norway
BackLink https://www.ncbi.nlm.nih.gov/pubmed/33778095$$D View this record in MEDLINE/PubMed
BookMark eNo1kL1PwzAQxT0U0VK6MyGPLAn-SOJkRIVCURFLmaOLc0mNEqfEDoiNP52Ulun07t7vSe8uyMR2Fgm54izknKtbHj6_rMM0FCETEWNiQmZCiiyIJBNTsnDunTHGOYsFj87JVEqlUpbFM_KzAudpi9bY2nQtUIf1qDx401lqLN3y4AtNvfNY0hZqi95o2qPrLFiN1Iy7EaWfXTO06OjgDgpoc0COIJX3tETc0waht3_nXu-MR-2HHi_JWQWNw8Vpzsnb6mG7fAo2r4_r5d0m0DITPkixTLGSseJpFSUQV4lWScGzNFaqBIiElCxJi4oBsKIErmSkNFRSJylyWYGYk5tj7r7vPgZ0Pm-N09g0YLEbXC5ilsQszjIxWq9P1qFoscz3_diy_87_vyZ-AYD7ccU
CitedBy_id crossref_primary_10_3390_cancers13184674
crossref_primary_10_1177_19714009231224442
crossref_primary_10_3390_cancers14112676
crossref_primary_10_3390_jimaging8120327
crossref_primary_10_3389_fneur_2022_932219
crossref_primary_10_3389_fneur_2025_1536751
ContentType Journal Article
Copyright 2021 Society of Photo-Optical Instrumentation Engineers (SPIE).
Copyright_xml – notice: 2021 Society of Photo-Optical Instrumentation Engineers (SPIE).
DBID NPM
7X8
DOI 10.1117/1.JMI.8.2.024002
DatabaseName PubMed
MEDLINE - Academic
DatabaseTitle PubMed
MEDLINE - Academic
DatabaseTitleList MEDLINE - Academic
PubMed
Database_xml – sequence: 1
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 2
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod no_fulltext_linktorsrc
Discipline Medicine
ExternalDocumentID 33778095
Genre Journal Article
GroupedDBID 0R~
4.4
ABJNI
ACGFS
ADMLS
ALMA_UNASSIGNED_HOLDINGS
EBS
EJD
FQ0
HYE
M4X
NPM
O9-
OK1
RPM
SPBNH
7X8
AKROS
ID FETCH-LOGICAL-c392t-8ed8ef35718f46a5f6c76b198577daa4233068bf0aa0bda17347caf3c68e13fa2
IEDL.DBID 7X8
ISICitedReferencesCount 14
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000647159200007&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 2329-4302
IngestDate Sun Nov 09 13:21:58 EST 2025
Wed Feb 19 02:26:17 EST 2025
IsPeerReviewed true
IsScholarly true
Issue 2
Keywords clinical diagnosis
meningioma
deep learning
magnetic resonance imaging
three-dimensional segmentation
Language English
License 2021 Society of Photo-Optical Instrumentation Engineers (SPIE).
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c392t-8ed8ef35718f46a5f6c76b198577daa4233068bf0aa0bda17347caf3c68e13fa2
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ORCID 0000-0003-0999-3849
0000-0002-3637-953X
0000-0002-5669-9514
PMID 33778095
PQID 2506505992
PQPubID 23479
ParticipantIDs proquest_miscellaneous_2506505992
pubmed_primary_33778095
PublicationCentury 2000
PublicationDate 2021-03-01
PublicationDateYYYYMMDD 2021-03-01
PublicationDate_xml – month: 03
  year: 2021
  text: 2021-03-01
  day: 01
PublicationDecade 2020
PublicationPlace United States
PublicationPlace_xml – name: United States
PublicationTitle Journal of medical imaging (Bellingham, Wash.)
PublicationTitleAlternate J Med Imaging (Bellingham)
PublicationYear 2021
SSID ssj0001105214
Score 2.2579565
Snippet Automatic and consistent meningioma segmentation in T1-weighted magnetic resonance (MR) imaging volumes and corresponding volumetric assessment is of use for...
Purpose: Automatic and consistent meningioma segmentation in T1-weighted magnetic resonance (MR) imaging volumes and corresponding volumetric assessment is of...
SourceID proquest
pubmed
SourceType Aggregation Database
Index Database
StartPage 024002
Title Fast meningioma segmentation in T1-weighted magnetic resonance imaging volumes using a lightweight 3D deep learning architecture
URI https://www.ncbi.nlm.nih.gov/pubmed/33778095
https://www.proquest.com/docview/2506505992
Volume 8
WOSCitedRecordID wos000647159200007&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/eLvHCXMwpV1La9wwEBZtt5RemqbPTZOiQK5y1pZsyacQmixp6S45bGBvZiyNl4XY3tbb9JqfnpHssL0UCr34YgaMNJ755vkxdqISdCUFykJZyITSKhWlrIxQJQKpC4GOKqzM_67nc7Nc5tdDwq0b2iofbWIw1K61Pkd-Sq6awESa58nZ5ofwrFG-ujpQaDxlI0lQxrd06aXZ5VhiP5qqAr9ckgslfTNPX6mM9WkcfZt9jUyURH7PV8ir_AVjBl8z3fvfr3zNXg0ok5_3arHPnmDzhr2YDXX0t-x-Ct2W1xjoitoaeIerephCavi64YtY_A5JU3S8hlXjRx05Reat38-BfF0HciPe27aO--75FQd-60V6QS4vuEPc8IGXgl7_UbN4x26ml4svV2LgYhCWENRWGHQGK5mSK6tUBmmVWZ2VcW5SrR0AgTKKPUxZTQAmpYNYS6UtVNJmBum-IXnPnjVtgx8Zl8qSnEvKDI3KwYKZGBvDxDojESEfs-PHsy1I130BAxpsf3XF7nTH7EN_QcWmX8pRSKm1Ibx48A_Sn9jLxLemhFayQzaq6E_HI_bc3m3X3c_PQYnoOb-ePQDV2NLb
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=Fast+meningioma+segmentation+in+T1-weighted+magnetic+resonance+imaging+volumes+using+a+lightweight+3D+deep+learning+architecture&rft.jtitle=Journal+of+medical+imaging+%28Bellingham%2C+Wash.%29&rft.au=Bouget%2C+David&rft.au=Pedersen%2C+Andr%C3%A9&rft.au=Hosainey%2C+Sayied+Abdol+Mohieb&rft.au=Vanel%2C+Johanna&rft.date=2021-03-01&rft.issn=2329-4302&rft.volume=8&rft.issue=2&rft.spage=024002&rft_id=info:doi/10.1117%2F1.JMI.8.2.024002&rft.externalDBID=NO_FULL_TEXT
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2329-4302&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2329-4302&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2329-4302&client=summon