Machine Learning Algorithm Detection of Confluent B-Lines

B-lines are a ring-down artifact of lung ultrasound that arise with increased alveolar water in conditions such as pulmonary edema and infectious pneumonitis. Confluent B-line presence may signify a different level of pathology compared with single B-lines. Existing algorithms aimed at B-line counti...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Ultrasound in medicine & biology Ročník 49; číslo 9; s. 2095
Hlavní autoři: Baloescu, Cristiana, Rucki, Agnieszka A, Chen, Alvin, Zahiri, Mohsen, Ghoshal, Goutam, Wang, Jing, Chew, Rita, Kessler, David, Chan, Daniela K I, Hicks, Bryson, Schnittke, Nikolai, Shupp, Jeffrey, Gregory, Kenton, Raju, Balasundar, Moore, Christopher
Médium: Journal Article
Jazyk:angličtina
Vydáno: England 01.09.2023
Témata:
ISSN:1879-291X, 1879-291X
On-line přístup:Zjistit podrobnosti o přístupu
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract B-lines are a ring-down artifact of lung ultrasound that arise with increased alveolar water in conditions such as pulmonary edema and infectious pneumonitis. Confluent B-line presence may signify a different level of pathology compared with single B-lines. Existing algorithms aimed at B-line counting do not distinguish between single and confluent B-lines. The objective of this study was to test a machine learning algorithm for confluent B-line identification. This study used a subset of 416 clips from 157 subjects, previously acquired in a prospective study enrolling adults with shortness of breath at two academic medical centers, using a hand-held tablet and a 14-zone protocol. After exclusions, random sampling generated a total of 416 clips (146 curvilinear, 150 sector and 120 linear) for review. A group of five experts in point-of-care ultrasound blindly evaluated the clips for presence/absence of confluent B-lines. Ground truth was defined as majority agreement among the experts and used for comparison with the algorithm. Confluent B-lines were present in 206 of 416 clips (49.5%). Sensitivity and specificity of confluent B-line detection by algorithm compared with expert determination were 83% (95% confidence interval [CI]: 0.77-0.88) and 92% (95% CI: 0.88-0.96). Sensitivity and specificity did not statistically differ between transducers. Agreement between algorithm and expert for confluent B-lines measured by unweighted κ was 0.75 (95% CI: 0.69-0.81) for the overall set. The confluent B-line detection algorithm had high sensitivity and specificity for detection of confluent B-lines in lung ultrasound point-of-care clips, compared with expert determination.
AbstractList B-lines are a ring-down artifact of lung ultrasound that arise with increased alveolar water in conditions such as pulmonary edema and infectious pneumonitis. Confluent B-line presence may signify a different level of pathology compared with single B-lines. Existing algorithms aimed at B-line counting do not distinguish between single and confluent B-lines. The objective of this study was to test a machine learning algorithm for confluent B-line identification. This study used a subset of 416 clips from 157 subjects, previously acquired in a prospective study enrolling adults with shortness of breath at two academic medical centers, using a hand-held tablet and a 14-zone protocol. After exclusions, random sampling generated a total of 416 clips (146 curvilinear, 150 sector and 120 linear) for review. A group of five experts in point-of-care ultrasound blindly evaluated the clips for presence/absence of confluent B-lines. Ground truth was defined as majority agreement among the experts and used for comparison with the algorithm. Confluent B-lines were present in 206 of 416 clips (49.5%). Sensitivity and specificity of confluent B-line detection by algorithm compared with expert determination were 83% (95% confidence interval [CI]: 0.77-0.88) and 92% (95% CI: 0.88-0.96). Sensitivity and specificity did not statistically differ between transducers. Agreement between algorithm and expert for confluent B-lines measured by unweighted κ was 0.75 (95% CI: 0.69-0.81) for the overall set. The confluent B-line detection algorithm had high sensitivity and specificity for detection of confluent B-lines in lung ultrasound point-of-care clips, compared with expert determination.
B-lines are a ring-down artifact of lung ultrasound that arise with increased alveolar water in conditions such as pulmonary edema and infectious pneumonitis. Confluent B-line presence may signify a different level of pathology compared with single B-lines. Existing algorithms aimed at B-line counting do not distinguish between single and confluent B-lines. The objective of this study was to test a machine learning algorithm for confluent B-line identification.OBJECTIVEB-lines are a ring-down artifact of lung ultrasound that arise with increased alveolar water in conditions such as pulmonary edema and infectious pneumonitis. Confluent B-line presence may signify a different level of pathology compared with single B-lines. Existing algorithms aimed at B-line counting do not distinguish between single and confluent B-lines. The objective of this study was to test a machine learning algorithm for confluent B-line identification.This study used a subset of 416 clips from 157 subjects, previously acquired in a prospective study enrolling adults with shortness of breath at two academic medical centers, using a hand-held tablet and a 14-zone protocol. After exclusions, random sampling generated a total of 416 clips (146 curvilinear, 150 sector and 120 linear) for review. A group of five experts in point-of-care ultrasound blindly evaluated the clips for presence/absence of confluent B-lines. Ground truth was defined as majority agreement among the experts and used for comparison with the algorithm.METHODSThis study used a subset of 416 clips from 157 subjects, previously acquired in a prospective study enrolling adults with shortness of breath at two academic medical centers, using a hand-held tablet and a 14-zone protocol. After exclusions, random sampling generated a total of 416 clips (146 curvilinear, 150 sector and 120 linear) for review. A group of five experts in point-of-care ultrasound blindly evaluated the clips for presence/absence of confluent B-lines. Ground truth was defined as majority agreement among the experts and used for comparison with the algorithm.Confluent B-lines were present in 206 of 416 clips (49.5%). Sensitivity and specificity of confluent B-line detection by algorithm compared with expert determination were 83% (95% confidence interval [CI]: 0.77-0.88) and 92% (95% CI: 0.88-0.96). Sensitivity and specificity did not statistically differ between transducers. Agreement between algorithm and expert for confluent B-lines measured by unweighted κ was 0.75 (95% CI: 0.69-0.81) for the overall set.RESULTSConfluent B-lines were present in 206 of 416 clips (49.5%). Sensitivity and specificity of confluent B-line detection by algorithm compared with expert determination were 83% (95% confidence interval [CI]: 0.77-0.88) and 92% (95% CI: 0.88-0.96). Sensitivity and specificity did not statistically differ between transducers. Agreement between algorithm and expert for confluent B-lines measured by unweighted κ was 0.75 (95% CI: 0.69-0.81) for the overall set.The confluent B-line detection algorithm had high sensitivity and specificity for detection of confluent B-lines in lung ultrasound point-of-care clips, compared with expert determination.CONCLUSIONThe confluent B-line detection algorithm had high sensitivity and specificity for detection of confluent B-lines in lung ultrasound point-of-care clips, compared with expert determination.
Author Zahiri, Mohsen
Gregory, Kenton
Baloescu, Cristiana
Raju, Balasundar
Chen, Alvin
Wang, Jing
Chan, Daniela K I
Chew, Rita
Moore, Christopher
Kessler, David
Schnittke, Nikolai
Hicks, Bryson
Ghoshal, Goutam
Shupp, Jeffrey
Rucki, Agnieszka A
Author_xml – sequence: 1
  givenname: Cristiana
  surname: Baloescu
  fullname: Baloescu, Cristiana
  email: Cristiana.Baloescu@yale.edu
  organization: Department of Emergency Medicine, Yale University School of Medicine, New Haven, CT, USA. Electronic address: Cristiana.Baloescu@yale.edu
– sequence: 2
  givenname: Agnieszka A
  surname: Rucki
  fullname: Rucki, Agnieszka A
  organization: Philips Research North America, Cambridge, MA, USA
– sequence: 3
  givenname: Alvin
  surname: Chen
  fullname: Chen, Alvin
  organization: Philips Research North America, Cambridge, MA, USA
– sequence: 4
  givenname: Mohsen
  surname: Zahiri
  fullname: Zahiri, Mohsen
  organization: Philips Research North America, Cambridge, MA, USA
– sequence: 5
  givenname: Goutam
  surname: Ghoshal
  fullname: Ghoshal, Goutam
  organization: Philips Research North America, Cambridge, MA, USA
– sequence: 6
  givenname: Jing
  surname: Wang
  fullname: Wang, Jing
  organization: Philips Research North America, Cambridge, MA, USA
– sequence: 7
  givenname: Rita
  surname: Chew
  fullname: Chew, Rita
  organization: Philips Research North America, Cambridge, MA, USA
– sequence: 8
  givenname: David
  surname: Kessler
  fullname: Kessler, David
  organization: Department of Emergency Medicine, Columbia University Vagelos College of Physicians & Surgeons, New York, NY, USA
– sequence: 9
  givenname: Daniela K I
  surname: Chan
  fullname: Chan, Daniela K I
  organization: Department of Emergency Medicine, Oregon Health & Science University, Portland, OR, USA; Center for Regenerative Medicine, Oregon Health & Science University, Portland, OR, USA
– sequence: 10
  givenname: Bryson
  surname: Hicks
  fullname: Hicks, Bryson
  organization: Department of Emergency Medicine, Oregon Health & Science University, Portland, OR, USA; Center for Regenerative Medicine, Oregon Health & Science University, Portland, OR, USA
– sequence: 11
  givenname: Nikolai
  surname: Schnittke
  fullname: Schnittke, Nikolai
  organization: Department of Emergency Medicine, Oregon Health & Science University, Portland, OR, USA; Center for Regenerative Medicine, Oregon Health & Science University, Portland, OR, USA
– sequence: 12
  givenname: Jeffrey
  surname: Shupp
  fullname: Shupp, Jeffrey
  organization: Departments of Surgery, Biochemistry and Molecular & Cellular Biology, Georgetown University School of Medicine | Medstar Health, Washington, DC, USA
– sequence: 13
  givenname: Kenton
  surname: Gregory
  fullname: Gregory, Kenton
  organization: Department of Emergency Medicine, Oregon Health & Science University, Portland, OR, USA; Center for Regenerative Medicine, Oregon Health & Science University, Portland, OR, USA
– sequence: 14
  givenname: Balasundar
  surname: Raju
  fullname: Raju, Balasundar
  organization: Philips Research North America, Cambridge, MA, USA
– sequence: 15
  givenname: Christopher
  surname: Moore
  fullname: Moore, Christopher
  organization: Department of Emergency Medicine, Yale University School of Medicine, New Haven, CT, USA
BackLink https://www.ncbi.nlm.nih.gov/pubmed/37365065$$D View this record in MEDLINE/PubMed
BookMark eNpNUMtKxDAUDTLiPPQXpLhy05p3k-U44wsqbhTclTS9nenQJmOTLvx7C47g6pzLecA9SzRz3gFCNwRnBBN5d8jGLg4m9FBXrc8opizDIpukM7QgKtcp1eRz9o_P0TKEA8Y4lyy_QHOWMymwFAukX43dtw6SAszgWrdL1t3OD23c98kWItjYepf4Jtl413QjuJjcp8UUCJfovDFdgKsTrtDH48P75jkt3p5eNusitVzzmKoa18JaIo2pIc8xU7ypamBEKmqVMU2lDGe2mk6utRCcaOASmGCKYAuartDtb-9x8F8jhFj2bbDQdcaBH0NJFcOU5ELyyXp9so7VNE55HNreDN_l37v0B0XAXEI
CitedBy_id crossref_primary_10_1177_01617346251330111
crossref_primary_10_3390_diagnostics14020155
crossref_primary_10_3390_jcm13123534
crossref_primary_10_3390_app15169126
crossref_primary_10_1007_s00063_025_01268_7
crossref_primary_10_1007_s40477_025_01077_w
ContentType Journal Article
Copyright Copyright © 2023 World Federation for Ultrasound in Medicine & Biology. All rights reserved.
Copyright_xml – notice: Copyright © 2023 World Federation for Ultrasound in Medicine & Biology. All rights reserved.
DBID CGR
CUY
CVF
ECM
EIF
NPM
7X8
DOI 10.1016/j.ultrasmedbio.2023.05.016
DatabaseName Medline
MEDLINE
MEDLINE (Ovid)
MEDLINE
MEDLINE
PubMed
MEDLINE - Academic
DatabaseTitle MEDLINE
Medline Complete
MEDLINE with Full Text
PubMed
MEDLINE (Ovid)
MEDLINE - Academic
DatabaseTitleList MEDLINE
MEDLINE - Academic
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
Biology
Physics
EISSN 1879-291X
ExternalDocumentID 37365065
Genre Research Support, U.S. Gov't, P.H.S
Journal Article
GroupedDBID ---
--K
-DZ
.1-
.55
.FO
.GJ
0R~
123
1B1
1P~
1RT
1~5
29Q
3O-
4.4
457
4G.
53G
5RE
5VS
7-5
AACTN
AAEDT
AAEDW
AALRI
AAQFI
AAQXK
AAWTL
AAXUO
ABDPE
ABJNI
ABLJU
ABMAC
ABNEU
ABOCM
ABTAH
ABWVN
ACGFS
ACIUM
ACRPL
ADBBV
ADMUD
ADNMO
AENEX
AEVXI
AFCTW
AFJKZ
AFRHN
AFTJW
AHHHB
AITUG
AJUYK
AKRWK
ALMA_UNASSIGNED_HOLDINGS
AMRAJ
ASPBG
AVWKF
AZFZN
BELOY
C5W
CGR
CS3
CUY
CVF
DU5
EBS
ECM
EFJIC
EIF
EJD
F5P
FDB
FEDTE
FGOYB
FIRID
G-2
GBLVA
HEI
HMK
HMO
HVGLF
HZ~
IHE
J1W
K-O
KOM
L7B
M29
M41
MO0
NPM
NQ-
O9-
OI~
OU0
P2P
R2-
RIG
ROL
RPZ
SAE
SDG
SEL
SES
SEW
SSZ
WUQ
X7M
XH2
Z5R
ZGI
ZXP
ZY4
~S-
7X8
AAYWO
APXCP
EFKBS
ID FETCH-LOGICAL-c494t-8d0d5cc16aade770384fbde31682c8aafb8a43cb68249955419e46e353810ce92
IEDL.DBID 7X8
ISICitedReferencesCount 7
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001051294200001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1879-291X
IngestDate Sat Sep 27 21:23:29 EDT 2025
Thu Apr 03 06:57:05 EDT 2025
IsPeerReviewed true
IsScholarly true
Issue 9
Keywords Point-of-care ultrasound
B-line
Lung ultrasound
Artificial intelligence
Machine learning
Language English
License Copyright © 2023 World Federation for Ultrasound in Medicine & Biology. All rights reserved.
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c494t-8d0d5cc16aade770384fbde31682c8aafb8a43cb68249955419e46e353810ce92
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
PMID 37365065
PQID 2830217564
PQPubID 23479
ParticipantIDs proquest_miscellaneous_2830217564
pubmed_primary_37365065
PublicationCentury 2000
PublicationDate 2023-09-01
PublicationDateYYYYMMDD 2023-09-01
PublicationDate_xml – month: 09
  year: 2023
  text: 2023-09-01
  day: 01
PublicationDecade 2020
PublicationPlace England
PublicationPlace_xml – name: England
PublicationTitle Ultrasound in medicine & biology
PublicationTitleAlternate Ultrasound Med Biol
PublicationYear 2023
SSID ssj0007637
Score 2.4433072
Snippet B-lines are a ring-down artifact of lung ultrasound that arise with increased alveolar water in conditions such as pulmonary edema and infectious pneumonitis....
SourceID proquest
pubmed
SourceType Aggregation Database
Index Database
StartPage 2095
SubjectTerms Adult
Algorithms
Humans
Lung - diagnostic imaging
Machine Learning
Prospective Studies
Pulmonary Edema
Ultrasonography - methods
Title Machine Learning Algorithm Detection of Confluent B-Lines
URI https://www.ncbi.nlm.nih.gov/pubmed/37365065
https://www.proquest.com/docview/2830217564
Volume 49
WOSCitedRecordID wos001051294200001&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/eLvHCXMwpV1bS8MwFD6oU_HFy7zNGxV8Da5JesmTzMvwwY09KPStpGk6B7Odayf47z1JO_YkCL4UWkgp6ck5X5Iv3wdw47mpJwNubVI9goi4S0KdKaIEBoifJYJ50ppNBMNhGEVi1Cy4lQ2tcpkTbaJOC2XWyG-tUBXWOp_fzT6JcY0yu6uNhcY6tBhCGRPVQbRSC8exY81VwkAQKtxoKTpq-V2LaTWXJZacZGIOAVJWK3j6v0NNW3L6e__92H3YbcCm06uj4wDWdN6Grdp-8rsN24NmYx0fWiaoKg9BDCy9UjuN8urY6U3H-O7q_cN51JVlbuVOkTnmrKDxN6mce_JiyPNH8NZ_en14Jo2_AlFc8IqEaTf1lHJ9KVMd4NAPeZak2lhZURVKmSWh5EwleIvzIsQdrtDc18wzqmBKC3oMG3mR61NwqE-pTASiAYYpARsHuqtFmnEXp0hUux24XnZUjPFrNiVkrotFGa-6qgMndW_Hs1poI2YBQwDpe2d_aH0OO-Yn1vSvC2hlOHr1JWyqr2pSzq9sYOB1OBr8AKmFwAI
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=Machine+Learning+Algorithm+Detection+of+Confluent+B-Lines&rft.jtitle=Ultrasound+in+medicine+%26+biology&rft.au=Baloescu%2C+Cristiana&rft.au=Rucki%2C+Agnieszka+A&rft.au=Chen%2C+Alvin&rft.au=Zahiri%2C+Mohsen&rft.date=2023-09-01&rft.eissn=1879-291X&rft.volume=49&rft.issue=9&rft.spage=2095&rft_id=info:doi/10.1016%2Fj.ultrasmedbio.2023.05.016&rft_id=info%3Apmid%2F37365065&rft_id=info%3Apmid%2F37365065&rft.externalDocID=37365065
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1879-291X&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1879-291X&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1879-291X&client=summon