Creating a training set for artificial intelligence from initial segmentations of airways

Abstract Airways segmentation is important for research about pulmonary disease but require a large amount of time by trained specialists. We used an openly available software to improve airways segmentations obtained from an artificial intelligence (AI) tool and retrained the tool to get a better p...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Ivan Dudurych, Antonio Garcia-Uceda, Zaigham Saghir, Harm A. W. M. Tiddens, Rozemarijn Vliegenthart, Marleen de Bruijne
Formato: article
Lenguaje:EN
Publicado: SpringerOpen 2021
Materias:
Acceso en línea:https://doaj.org/article/31bfd843b25f4eeaace52d7880566288
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:31bfd843b25f4eeaace52d7880566288
record_format dspace
spelling oai:doaj.org-article:31bfd843b25f4eeaace52d78805662882021-12-05T12:10:48ZCreating a training set for artificial intelligence from initial segmentations of airways10.1186/s41747-021-00247-92509-9280https://doaj.org/article/31bfd843b25f4eeaace52d78805662882021-11-01T00:00:00Zhttps://doi.org/10.1186/s41747-021-00247-9https://doaj.org/toc/2509-9280Abstract Airways segmentation is important for research about pulmonary disease but require a large amount of time by trained specialists. We used an openly available software to improve airways segmentations obtained from an artificial intelligence (AI) tool and retrained the tool to get a better performance. Fifteen initial airway segmentations from low-dose chest computed tomography scans were obtained with a 3D-Unet AI tool previously trained on Danish Lung Cancer Screening Trial and Erasmus-MC Sophia datasets. Segmentations were manually corrected in 3D Slicer. The corrected airway segmentations were used to retrain the 3D-Unet. Airway measurements were automatically obtained and included count, airway length and luminal diameter per generation from the segmentations. Correcting segmentations required 2–4 h per scan. Manually corrected segmentations had more branches (p < 0.001), longer airways (p < 0.001) and smaller luminal diameters (p = 0.004) than initial segmentations. Segmentations from retrained 3D-Unets trended towards more branches and longer airways compared to the initial segmentations. The largest changes were seen in airways from 6th generation onwards. Manual correction results in significantly improved segmentations and is potentially a useful and time-efficient method to improve the AI tool performance on a specific hospital or research dataset.Ivan DudurychAntonio Garcia-UcedaZaigham SaghirHarm A. W. M. TiddensRozemarijn VliegenthartMarleen de BruijneSpringerOpenarticleArtificial intelligenceImage processing (computer-assisted)Respiratory systemThoraxTomography (x-ray computed)Medical physics. Medical radiology. Nuclear medicineR895-920ENEuropean Radiology Experimental, Vol 5, Iss 1, Pp 1-7 (2021)
institution DOAJ
collection DOAJ
language EN
topic Artificial intelligence
Image processing (computer-assisted)
Respiratory system
Thorax
Tomography (x-ray computed)
Medical physics. Medical radiology. Nuclear medicine
R895-920
spellingShingle Artificial intelligence
Image processing (computer-assisted)
Respiratory system
Thorax
Tomography (x-ray computed)
Medical physics. Medical radiology. Nuclear medicine
R895-920
Ivan Dudurych
Antonio Garcia-Uceda
Zaigham Saghir
Harm A. W. M. Tiddens
Rozemarijn Vliegenthart
Marleen de Bruijne
Creating a training set for artificial intelligence from initial segmentations of airways
description Abstract Airways segmentation is important for research about pulmonary disease but require a large amount of time by trained specialists. We used an openly available software to improve airways segmentations obtained from an artificial intelligence (AI) tool and retrained the tool to get a better performance. Fifteen initial airway segmentations from low-dose chest computed tomography scans were obtained with a 3D-Unet AI tool previously trained on Danish Lung Cancer Screening Trial and Erasmus-MC Sophia datasets. Segmentations were manually corrected in 3D Slicer. The corrected airway segmentations were used to retrain the 3D-Unet. Airway measurements were automatically obtained and included count, airway length and luminal diameter per generation from the segmentations. Correcting segmentations required 2–4 h per scan. Manually corrected segmentations had more branches (p < 0.001), longer airways (p < 0.001) and smaller luminal diameters (p = 0.004) than initial segmentations. Segmentations from retrained 3D-Unets trended towards more branches and longer airways compared to the initial segmentations. The largest changes were seen in airways from 6th generation onwards. Manual correction results in significantly improved segmentations and is potentially a useful and time-efficient method to improve the AI tool performance on a specific hospital or research dataset.
format article
author Ivan Dudurych
Antonio Garcia-Uceda
Zaigham Saghir
Harm A. W. M. Tiddens
Rozemarijn Vliegenthart
Marleen de Bruijne
author_facet Ivan Dudurych
Antonio Garcia-Uceda
Zaigham Saghir
Harm A. W. M. Tiddens
Rozemarijn Vliegenthart
Marleen de Bruijne
author_sort Ivan Dudurych
title Creating a training set for artificial intelligence from initial segmentations of airways
title_short Creating a training set for artificial intelligence from initial segmentations of airways
title_full Creating a training set for artificial intelligence from initial segmentations of airways
title_fullStr Creating a training set for artificial intelligence from initial segmentations of airways
title_full_unstemmed Creating a training set for artificial intelligence from initial segmentations of airways
title_sort creating a training set for artificial intelligence from initial segmentations of airways
publisher SpringerOpen
publishDate 2021
url https://doaj.org/article/31bfd843b25f4eeaace52d7880566288
work_keys_str_mv AT ivandudurych creatingatrainingsetforartificialintelligencefrominitialsegmentationsofairways
AT antoniogarciauceda creatingatrainingsetforartificialintelligencefrominitialsegmentationsofairways
AT zaighamsaghir creatingatrainingsetforartificialintelligencefrominitialsegmentationsofairways
AT harmawmtiddens creatingatrainingsetforartificialintelligencefrominitialsegmentationsofairways
AT rozemarijnvliegenthart creatingatrainingsetforartificialintelligencefrominitialsegmentationsofairways
AT marleendebruijne creatingatrainingsetforartificialintelligencefrominitialsegmentationsofairways
_version_ 1718372182546972672