The Argos project: the development of a computer-aided detection system to improve detection of Barrett's neoplasia on white light endoscopy

Jeroen de Groof (Corresponding author), Fons van der Sommen, Joost van der Putten, Maarten R. Struyvenberg, Sveta Zinger, Wouter L. Curvers, Oliver Pech, Alexander Meining, Horst Neuhaus, Raf Bisschops, Erik J. Schoon, Peter H. de With, Jacques J. Bergman

Research output: Contribution to journalArticleAcademicpeer-review

3 Citations (Scopus)
43 Downloads (Pure)

Abstract

Background: Computer-aided detection (CAD) systems might assist endoscopists in the recognition of Barrett's neoplasia. Aim: To develop a CAD system using endoscopic images of Barrett's neoplasia. Methods: White light endoscopy (WLE) overview images of 40 neoplastic Barrett's lesions and 20 non-dysplastic Barret's oesophagus (NDBO) patients were prospectively collected. Experts delineated all neoplastic images. The overlap area of at least four delineations was labelled as the ‘sweet spot’. The area with at least one delineation was labelled as the ‘soft spot’. The CAD system was trained on colour and texture features. Positive features were taken from the sweet spot and negative features from NDBO images. Performance was evaluated using leave-one-out cross-validation. Outcome parameters were diagnostic accuracy of the CAD system per image, and localization of the expert soft spot by CAD delineation (localization score) and its indication of preferred biopsy location (red-flag indication score). Results: Accuracy, sensitivity and specificity for detection were 92, 95 and 85%, respectively. The system localized and red-flagged the soft spot in 100 and 90%, respectively. Conclusion: This uniquely trained and validated CAD system detected and localized early Barrett's neoplasia on WLE images with high accuracy. This is an important step towards real-time automated detection of Barrett's neoplasia.

Original languageEnglish
Pages (from-to)538-547
Number of pages10
JournalUnited European Gastroenterology Journal
Volume7
Issue number4
Early online date1 Jan 2019
DOIs
Publication statusPublished - 1 May 2019

Fingerprint

Endoscopy
Light
Neoplasms
Esophagus
Color
Biopsy
Sensitivity and Specificity

Keywords

  • artificial intelligence
  • Barrett's neoplasia
  • Barrett's oesophagus
  • computer-aided detection
  • endoscopy

Cite this

Groof, Jeroen de ; van der Sommen, Fons ; van der Putten, Joost ; Struyvenberg, Maarten R. ; Zinger, Sveta ; Curvers, Wouter L. ; Pech, Oliver ; Meining, Alexander ; Neuhaus, Horst ; Bisschops, Raf ; Schoon, Erik J. ; de With, Peter H. ; Bergman, Jacques J. / The Argos project : the development of a computer-aided detection system to improve detection of Barrett's neoplasia on white light endoscopy. In: United European Gastroenterology Journal. 2019 ; Vol. 7, No. 4. pp. 538-547.
@article{56395d14c80f4bcf86d5a481e77e6b4e,
title = "The Argos project: the development of a computer-aided detection system to improve detection of Barrett's neoplasia on white light endoscopy",
abstract = "Background: Computer-aided detection (CAD) systems might assist endoscopists in the recognition of Barrett's neoplasia. Aim: To develop a CAD system using endoscopic images of Barrett's neoplasia. Methods: White light endoscopy (WLE) overview images of 40 neoplastic Barrett's lesions and 20 non-dysplastic Barret's oesophagus (NDBO) patients were prospectively collected. Experts delineated all neoplastic images. The overlap area of at least four delineations was labelled as the ‘sweet spot’. The area with at least one delineation was labelled as the ‘soft spot’. The CAD system was trained on colour and texture features. Positive features were taken from the sweet spot and negative features from NDBO images. Performance was evaluated using leave-one-out cross-validation. Outcome parameters were diagnostic accuracy of the CAD system per image, and localization of the expert soft spot by CAD delineation (localization score) and its indication of preferred biopsy location (red-flag indication score). Results: Accuracy, sensitivity and specificity for detection were 92, 95 and 85{\%}, respectively. The system localized and red-flagged the soft spot in 100 and 90{\%}, respectively. Conclusion: This uniquely trained and validated CAD system detected and localized early Barrett's neoplasia on WLE images with high accuracy. This is an important step towards real-time automated detection of Barrett's neoplasia.",
keywords = "artificial intelligence, Barrett's neoplasia, Barrett's oesophagus, computer-aided detection, endoscopy",
author = "Groof, {Jeroen de} and {van der Sommen}, Fons and {van der Putten}, Joost and Struyvenberg, {Maarten R.} and Sveta Zinger and Curvers, {Wouter L.} and Oliver Pech and Alexander Meining and Horst Neuhaus and Raf Bisschops and Schoon, {Erik J.} and {de With}, {Peter H.} and Bergman, {Jacques J.}",
year = "2019",
month = "5",
day = "1",
doi = "10.1177/2050640619837443",
language = "English",
volume = "7",
pages = "538--547",
journal = "United European Gastroenterology Journal",
issn = "2050-6406",
publisher = "SAGE Publications Ltd",
number = "4",

}

The Argos project : the development of a computer-aided detection system to improve detection of Barrett's neoplasia on white light endoscopy. / Groof, Jeroen de (Corresponding author); van der Sommen, Fons; van der Putten, Joost; Struyvenberg, Maarten R.; Zinger, Sveta; Curvers, Wouter L.; Pech, Oliver; Meining, Alexander; Neuhaus, Horst; Bisschops, Raf; Schoon, Erik J.; de With, Peter H.; Bergman, Jacques J.

In: United European Gastroenterology Journal, Vol. 7, No. 4, 01.05.2019, p. 538-547.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - The Argos project

T2 - the development of a computer-aided detection system to improve detection of Barrett's neoplasia on white light endoscopy

AU - Groof, Jeroen de

AU - van der Sommen, Fons

AU - van der Putten, Joost

AU - Struyvenberg, Maarten R.

AU - Zinger, Sveta

AU - Curvers, Wouter L.

AU - Pech, Oliver

AU - Meining, Alexander

AU - Neuhaus, Horst

AU - Bisschops, Raf

AU - Schoon, Erik J.

AU - de With, Peter H.

AU - Bergman, Jacques J.

PY - 2019/5/1

Y1 - 2019/5/1

N2 - Background: Computer-aided detection (CAD) systems might assist endoscopists in the recognition of Barrett's neoplasia. Aim: To develop a CAD system using endoscopic images of Barrett's neoplasia. Methods: White light endoscopy (WLE) overview images of 40 neoplastic Barrett's lesions and 20 non-dysplastic Barret's oesophagus (NDBO) patients were prospectively collected. Experts delineated all neoplastic images. The overlap area of at least four delineations was labelled as the ‘sweet spot’. The area with at least one delineation was labelled as the ‘soft spot’. The CAD system was trained on colour and texture features. Positive features were taken from the sweet spot and negative features from NDBO images. Performance was evaluated using leave-one-out cross-validation. Outcome parameters were diagnostic accuracy of the CAD system per image, and localization of the expert soft spot by CAD delineation (localization score) and its indication of preferred biopsy location (red-flag indication score). Results: Accuracy, sensitivity and specificity for detection were 92, 95 and 85%, respectively. The system localized and red-flagged the soft spot in 100 and 90%, respectively. Conclusion: This uniquely trained and validated CAD system detected and localized early Barrett's neoplasia on WLE images with high accuracy. This is an important step towards real-time automated detection of Barrett's neoplasia.

AB - Background: Computer-aided detection (CAD) systems might assist endoscopists in the recognition of Barrett's neoplasia. Aim: To develop a CAD system using endoscopic images of Barrett's neoplasia. Methods: White light endoscopy (WLE) overview images of 40 neoplastic Barrett's lesions and 20 non-dysplastic Barret's oesophagus (NDBO) patients were prospectively collected. Experts delineated all neoplastic images. The overlap area of at least four delineations was labelled as the ‘sweet spot’. The area with at least one delineation was labelled as the ‘soft spot’. The CAD system was trained on colour and texture features. Positive features were taken from the sweet spot and negative features from NDBO images. Performance was evaluated using leave-one-out cross-validation. Outcome parameters were diagnostic accuracy of the CAD system per image, and localization of the expert soft spot by CAD delineation (localization score) and its indication of preferred biopsy location (red-flag indication score). Results: Accuracy, sensitivity and specificity for detection were 92, 95 and 85%, respectively. The system localized and red-flagged the soft spot in 100 and 90%, respectively. Conclusion: This uniquely trained and validated CAD system detected and localized early Barrett's neoplasia on WLE images with high accuracy. This is an important step towards real-time automated detection of Barrett's neoplasia.

KW - artificial intelligence

KW - Barrett's neoplasia

KW - Barrett's oesophagus

KW - computer-aided detection

KW - endoscopy

UR - http://www.scopus.com/inward/record.url?scp=85062656145&partnerID=8YFLogxK

U2 - 10.1177/2050640619837443

DO - 10.1177/2050640619837443

M3 - Article

C2 - 31065371

AN - SCOPUS:85062656145

VL - 7

SP - 538

EP - 547

JO - United European Gastroenterology Journal

JF - United European Gastroenterology Journal

SN - 2050-6406

IS - 4

ER -