A convolutional neural network for bleeding detection in capsule endoscopy using real clinical data.

Autor: Turck D; Department of Medicine, University of Cologne, Cologne, Germany., Dratsch T; Institute of Diagnostic and Interventional Radiology, University Hospital Cologne, Cologne, Germany., Schröder L; Department of Medicine, University of Cologne, Cologne, Germany., Lorenz F; Department of Gastroenterology and Hepatology, University Hospital Cologne, Cologne, Germany., Dinter J; Gastroenterologische Schwerpunktpraxis Stähler, Cologne, Germany., Bürger M; Department of Gastroenterology and Hepatology, University Hospital Cologne, Cologne, Germany., Schiffmann L; Department of General, Visceral, Cancer, and Transplant Surgery, University Hospital Cologne, Cologne, Germany., Kasper P; Department of Gastroenterology and Hepatology, University Hospital Cologne, Cologne, Germany., Allo G; Department of Gastroenterology and Hepatology, University Hospital Cologne, Cologne, Germany., Goeser T; Department of Gastroenterology and Hepatology, University Hospital Cologne, Cologne, Germany., Chon SH; Department of Gastroenterology and Hepatology, University Hospital Cologne, Cologne, Germany.; Department of General, Visceral, Cancer, and Transplant Surgery, University Hospital Cologne, Cologne, Germany., Nierhoff D; Department of Gastroenterology and Hepatology, University Hospital Cologne, Cologne, Germany.
Jazyk: angličtina
Zdroj: Minimally invasive therapy & allied technologies : MITAT : official journal of the Society for Minimally Invasive Therapy [Minim Invasive Ther Allied Technol] 2023 Dec; Vol. 32 (6), pp. 335-340. Date of Electronic Publication: 2023 Aug 28.
DOI: 10.1080/13645706.2023.2250445
Abstrakt: Background: The goal of the present study was to develop a convolutional neural network for the detection of bleedings in capsule endoscopy videos using realistic clinical data from one single-centre.
Methods: Capsule endoscopy videos from all 133 patients (79 male, 54 female; mean age = 53.73 years, SD age = 26.13) who underwent capsule endoscopy at our institution between January 2014 and August 2018 were screened for pathology. All videos were screened for pathology by two independent capsule experts and confirmed findings were checked again by a third capsule expert. From these videos, 125 pathological findings (individual episodes of bleeding spanning a total of 5696 images) and 103 non-pathological findings (sections of normal mucosal tissue without pathologies spanning a total of 7420 images) were used to develop and validate a neural network (Inception V3) using transfer learning.
Results: The overall accuracy of the model for the detection of bleedings was 90.6% [95%CI: 89.4%-91.7%], with a sensitivity of 89.4% [95%CI: 87.6%-91.2%] and a specificity of 91.7% [95%CI: 90.1%-93.2%].
Conclusion: Our results show that neural networks can detect bleedings in capsule endoscopy videos under realistic, clinical conditions with an accuracy of 90.6%, potentially reducing reading time per capsule and helping to improve diagnostic accuracy.
Databáze: MEDLINE
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje