Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network

Objective Capsule endoscopy (CE) is pivotal for evaluation of small bowel disease. Obscure gastrointestinal bleeding most often originates from the small bowel. CE frequently identifies a wide range of lesions with different bleeding potentials in these patients. However, reading CE examinations is...

Full description

Saved in:
Bibliographic Details
Main Authors: Guilherme Macedo, Miguel José Mascarenhas Saraiva, João Afonso, Tiago Ribeiro, João Ferreira, Helder Cardoso, Ana Patricia Andrade, Marco Parente, Renato Natal, Miguel Mascarenhas Saraiva
Format: Article
Language:English
Published: BMJ Publishing Group 2021-10-01
Series:BMJ Open Gastroenterology
Online Access:https://bmjopengastro.bmj.com/content/8/1/e000753.full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846138028615008256
author Guilherme Macedo
Miguel José Mascarenhas Saraiva
João Afonso
Tiago Ribeiro
João Ferreira
Helder Cardoso
Ana Patricia Andrade
Marco Parente
Renato Natal
Miguel Mascarenhas Saraiva
author_facet Guilherme Macedo
Miguel José Mascarenhas Saraiva
João Afonso
Tiago Ribeiro
João Ferreira
Helder Cardoso
Ana Patricia Andrade
Marco Parente
Renato Natal
Miguel Mascarenhas Saraiva
author_sort Guilherme Macedo
collection DOAJ
description Objective Capsule endoscopy (CE) is pivotal for evaluation of small bowel disease. Obscure gastrointestinal bleeding most often originates from the small bowel. CE frequently identifies a wide range of lesions with different bleeding potentials in these patients. However, reading CE examinations is a time-consuming task. Convolutional neural networks (CNNs) are highly efficient artificial intelligence tools for image analysis. This study aims to develop a CNN-based model for identification and differentiation of multiple small bowel lesions with distinct haemorrhagic potential using CE images.Design We developed, trained, and validated a denary CNN based on CE images. Each frame was labelled according to the type of lesion (lymphangiectasia, xanthomas, ulcers, erosions, vascular lesions, protruding lesions, and blood). The haemorrhagic potential was assessed by Saurin’s classification. The entire dataset was divided into training and validation sets. The performance of the CNN was measured by the area under the receiving operating characteristic curve, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).Results A total of 53 555 CE images were included. The model had an overall accuracy of 99%, a sensitivity of 88%, a specificity of 99%, a PPV of 87%, and an NPV of 99% for detection of multiple small bowel abnormalities and respective classification of bleeding potential.Conclusion We developed and tested a CNN-based model for automatic detection of multiple types of small bowel lesions and classification of the respective bleeding potential. This system may improve the diagnostic yield of CE for these lesions and overall CE efficiency.
format Article
id doaj-art-69bd08440f494badb4d58a7fa3e002cb
institution Kabale University
issn 2054-4774
language English
publishDate 2021-10-01
publisher BMJ Publishing Group
record_format Article
series BMJ Open Gastroenterology
spelling doaj-art-69bd08440f494badb4d58a7fa3e002cb2024-12-07T12:30:10ZengBMJ Publishing GroupBMJ Open Gastroenterology2054-47742021-10-018110.1136/bmjgast-2021-000753Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural networkGuilherme Macedo0Miguel José Mascarenhas Saraiva1João Afonso2Tiago Ribeiro3João Ferreira4Helder Cardoso5Ana Patricia Andrade6Marco Parente7Renato Natal8Miguel Mascarenhas Saraiva92 Faculty of Medicine, University of Porto, Porto, PortugalDepartment of Gastroenterology, Hospital São João, Porto, PortugalDepartment of Gastroenterology, Hospital São João, Porto, PortugalGastroenterology and Hepatology, WGO Gastroenterology and Hepatology Training Centre, Porto, PortugalElectrical and Computer Engineering, University of Porto Faculty of Engineering, Porto, PortugalUniversity of Porto Faculty of Medicine, Porto, Porto, PortugalDepartment of Gastroenterology, Hospital São João, Porto, PortugalDepartment of Mechanical Engineering, Faculty of Engineering of the University of Porto, Porto, PortugalDepartment of Mechanical Engineering, Faculty of Engineering of the University of Porto, Porto, PortugalEndoscopy and Digestive Motility Laboratory, ManopH, Porto, PortugalObjective Capsule endoscopy (CE) is pivotal for evaluation of small bowel disease. Obscure gastrointestinal bleeding most often originates from the small bowel. CE frequently identifies a wide range of lesions with different bleeding potentials in these patients. However, reading CE examinations is a time-consuming task. Convolutional neural networks (CNNs) are highly efficient artificial intelligence tools for image analysis. This study aims to develop a CNN-based model for identification and differentiation of multiple small bowel lesions with distinct haemorrhagic potential using CE images.Design We developed, trained, and validated a denary CNN based on CE images. Each frame was labelled according to the type of lesion (lymphangiectasia, xanthomas, ulcers, erosions, vascular lesions, protruding lesions, and blood). The haemorrhagic potential was assessed by Saurin’s classification. The entire dataset was divided into training and validation sets. The performance of the CNN was measured by the area under the receiving operating characteristic curve, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).Results A total of 53 555 CE images were included. The model had an overall accuracy of 99%, a sensitivity of 88%, a specificity of 99%, a PPV of 87%, and an NPV of 99% for detection of multiple small bowel abnormalities and respective classification of bleeding potential.Conclusion We developed and tested a CNN-based model for automatic detection of multiple types of small bowel lesions and classification of the respective bleeding potential. This system may improve the diagnostic yield of CE for these lesions and overall CE efficiency.https://bmjopengastro.bmj.com/content/8/1/e000753.full
spellingShingle Guilherme Macedo
Miguel José Mascarenhas Saraiva
João Afonso
Tiago Ribeiro
João Ferreira
Helder Cardoso
Ana Patricia Andrade
Marco Parente
Renato Natal
Miguel Mascarenhas Saraiva
Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network
BMJ Open Gastroenterology
title Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network
title_full Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network
title_fullStr Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network
title_full_unstemmed Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network
title_short Deep learning and capsule endoscopy: automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network
title_sort deep learning and capsule endoscopy automatic identification and differentiation of small bowel lesions with distinct haemorrhagic potential using a convolutional neural network
url https://bmjopengastro.bmj.com/content/8/1/e000753.full
work_keys_str_mv AT guilhermemacedo deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork
AT migueljosemascarenhassaraiva deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork
AT joaoafonso deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork
AT tiagoribeiro deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork
AT joaoferreira deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork
AT heldercardoso deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork
AT anapatriciaandrade deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork
AT marcoparente deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork
AT renatonatal deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork
AT miguelmascarenhassaraiva deeplearningandcapsuleendoscopyautomaticidentificationanddifferentiationofsmallbowellesionswithdistincthaemorrhagicpotentialusingaconvolutionalneuralnetwork