Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm

Detalhes bibliográficos
Autor(a) principal: Ebigbo, Alanna
Data de Publicação: 2022
Outros Autores: Mendel, Robert, Scheppach, Markus W., Probst, Andreas, Shahidi, Neal, Prinz, Friederike, Fleischmann, Carola, Roemmele, Christoph, Goelder, Stefan Karl, Braun, Georg, Rauber, David, Rueckert, Tobias, Souza Jr, Luis A. de, Papa, Joao [UNESP], Byrne, Michael, Palm, Christoph, Messmann, Helmut
Tipo de documento: Artigo
Idioma: eng
Título da fonte: Repositório Institucional da UNESP
Texto Completo: http://dx.doi.org/10.1136/gutjnl-2021-326470
http://hdl.handle.net/11449/237698
Resumo: In this study, we aimed to develop an artificial intelligence clinical decision support solution to mitigate operator-dependent limitations during complex endoscopic procedures such as endoscopic submucosal dissection and peroral endoscopic myotomy, for example, bleeding and perforation. A DeepLabv3-based model was trained to delineate vessels, tissue structures and instruments on endoscopic still images from such procedures. The mean cross-validated Intersection over Union and Dice Score were 63% and 76%, respectively. Applied to standardised video clips from third-space endoscopic procedures, the algorithm showed a mean vessel detection rate of 85% with a false-positive rate of 0.75/min. These performance statistics suggest a potential clinical benefit for procedure safety, time and also training.
id UNSP_f0125bdefea2543978940e5a5d0d547e
oai_identifier_str oai:repositorio.unesp.br:11449/237698
network_acronym_str UNSP
network_name_str Repositório Institucional da UNESP
repository_id_str 2946
spelling Vessel and tissue recognition during third-space endoscopy using a deep learning algorithmEndoscopic proceduresEndoscopySurgical oncologyIn this study, we aimed to develop an artificial intelligence clinical decision support solution to mitigate operator-dependent limitations during complex endoscopic procedures such as endoscopic submucosal dissection and peroral endoscopic myotomy, for example, bleeding and perforation. A DeepLabv3-based model was trained to delineate vessels, tissue structures and instruments on endoscopic still images from such procedures. The mean cross-validated Intersection over Union and Dice Score were 63% and 76%, respectively. Applied to standardised video clips from third-space endoscopic procedures, the algorithm showed a mean vessel detection rate of 85% with a false-positive rate of 0.75/min. These performance statistics suggest a potential clinical benefit for procedure safety, time and also training.Univ Klinikum Augsburg, Dept Gastroenterol, D-86156 Augsburg, Bayern, GermanyOstbayer TH Regensburg, Regensburg Med Image Comp ReMIC, Regensburg, GermanyUniv British Columbia, Dept Med, Vancouver, BC, CanadaOstalb Klinikum Aalen, Dept Gastroenterol, Aalen, GermanyUniv Fed Sao Carlos, Dept Comp, Sao Carlos, BrazilSao Paulo State Univ, Dept Comp, Botucatu, SP, BrazilUniv British Columbia, Vancouver Gen Hosp, Vancouver, BC, CanadaSao Paulo State Univ, Dept Comp, Botucatu, SP, BrazilBmj Publishing GroupUniv Klinikum AugsburgOstbayer TH RegensburgUniv British ColumbiaOstalb Klinikum AalenUniversidade Federal de São Carlos (UFSCar)Universidade Estadual Paulista (UNESP)Ebigbo, AlannaMendel, RobertScheppach, Markus W.Probst, AndreasShahidi, NealPrinz, FriederikeFleischmann, CarolaRoemmele, ChristophGoelder, Stefan KarlBraun, GeorgRauber, DavidRueckert, TobiasSouza Jr, Luis A. dePapa, Joao [UNESP]Byrne, MichaelPalm, ChristophMessmann, Helmut2022-11-30T13:42:13Z2022-11-30T13:42:13Z2022-09-15info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/article3http://dx.doi.org/10.1136/gutjnl-2021-326470Gut. London: Bmj Publishing Group, 3 p., 2022.0017-5749http://hdl.handle.net/11449/23769810.1136/gutjnl-2021-326470WOS:000855856700001Web of Sciencereponame:Repositório Institucional da UNESPinstname:Universidade Estadual Paulista (UNESP)instacron:UNESPengGutinfo:eu-repo/semantics/openAccess2022-11-30T13:42:14Zoai:repositorio.unesp.br:11449/237698Repositório InstitucionalPUBhttp://repositorio.unesp.br/oai/requestopendoar:29462024-08-05T14:26:06.777150Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP)false
dc.title.none.fl_str_mv Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm
title Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm
spellingShingle Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm
Ebigbo, Alanna
Endoscopic procedures
Endoscopy
Surgical oncology
title_short Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm
title_full Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm
title_fullStr Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm
title_full_unstemmed Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm
title_sort Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm
author Ebigbo, Alanna
author_facet Ebigbo, Alanna
Mendel, Robert
Scheppach, Markus W.
Probst, Andreas
Shahidi, Neal
Prinz, Friederike
Fleischmann, Carola
Roemmele, Christoph
Goelder, Stefan Karl
Braun, Georg
Rauber, David
Rueckert, Tobias
Souza Jr, Luis A. de
Papa, Joao [UNESP]
Byrne, Michael
Palm, Christoph
Messmann, Helmut
author_role author
author2 Mendel, Robert
Scheppach, Markus W.
Probst, Andreas
Shahidi, Neal
Prinz, Friederike
Fleischmann, Carola
Roemmele, Christoph
Goelder, Stefan Karl
Braun, Georg
Rauber, David
Rueckert, Tobias
Souza Jr, Luis A. de
Papa, Joao [UNESP]
Byrne, Michael
Palm, Christoph
Messmann, Helmut
author2_role author
author
author
author
author
author
author
author
author
author
author
author
author
author
author
author
dc.contributor.none.fl_str_mv Univ Klinikum Augsburg
Ostbayer TH Regensburg
Univ British Columbia
Ostalb Klinikum Aalen
Universidade Federal de São Carlos (UFSCar)
Universidade Estadual Paulista (UNESP)
dc.contributor.author.fl_str_mv Ebigbo, Alanna
Mendel, Robert
Scheppach, Markus W.
Probst, Andreas
Shahidi, Neal
Prinz, Friederike
Fleischmann, Carola
Roemmele, Christoph
Goelder, Stefan Karl
Braun, Georg
Rauber, David
Rueckert, Tobias
Souza Jr, Luis A. de
Papa, Joao [UNESP]
Byrne, Michael
Palm, Christoph
Messmann, Helmut
dc.subject.por.fl_str_mv Endoscopic procedures
Endoscopy
Surgical oncology
topic Endoscopic procedures
Endoscopy
Surgical oncology
description In this study, we aimed to develop an artificial intelligence clinical decision support solution to mitigate operator-dependent limitations during complex endoscopic procedures such as endoscopic submucosal dissection and peroral endoscopic myotomy, for example, bleeding and perforation. A DeepLabv3-based model was trained to delineate vessels, tissue structures and instruments on endoscopic still images from such procedures. The mean cross-validated Intersection over Union and Dice Score were 63% and 76%, respectively. Applied to standardised video clips from third-space endoscopic procedures, the algorithm showed a mean vessel detection rate of 85% with a false-positive rate of 0.75/min. These performance statistics suggest a potential clinical benefit for procedure safety, time and also training.
publishDate 2022
dc.date.none.fl_str_mv 2022-11-30T13:42:13Z
2022-11-30T13:42:13Z
2022-09-15
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/article
format article
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://dx.doi.org/10.1136/gutjnl-2021-326470
Gut. London: Bmj Publishing Group, 3 p., 2022.
0017-5749
http://hdl.handle.net/11449/237698
10.1136/gutjnl-2021-326470
WOS:000855856700001
url http://dx.doi.org/10.1136/gutjnl-2021-326470
http://hdl.handle.net/11449/237698
identifier_str_mv Gut. London: Bmj Publishing Group, 3 p., 2022.
0017-5749
10.1136/gutjnl-2021-326470
WOS:000855856700001
dc.language.iso.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv Gut
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv 3
dc.publisher.none.fl_str_mv Bmj Publishing Group
publisher.none.fl_str_mv Bmj Publishing Group
dc.source.none.fl_str_mv Web of Science
reponame:Repositório Institucional da UNESP
instname:Universidade Estadual Paulista (UNESP)
instacron:UNESP
instname_str Universidade Estadual Paulista (UNESP)
instacron_str UNESP
institution UNESP
reponame_str Repositório Institucional da UNESP
collection Repositório Institucional da UNESP
repository.name.fl_str_mv Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP)
repository.mail.fl_str_mv
_version_ 1808128359437172736

Registros relacionados