UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Polyp characterization using deep learning and a publicly accessible polyp video database

Kader, Rawen; Cid-Mejias, Anton; Brandao, Patrick; Islam, Shahraz; Hebbar, Sanjith; González-Bueno Puyal, Juana; Ahmad, Omer F; ... Lovat, Laurence B; + view all (2022) Polyp characterization using deep learning and a publicly accessible polyp video database. Digestive Endoscopy 10.1111/den.14500. (In press). Green open access

[thumbnail of Kader_Digestive Endoscopy - 2022 - Kader - Polyp characterization using deep learning and a publicly accessible polyp video.pdf]
Preview
Text
Kader_Digestive Endoscopy - 2022 - Kader - Polyp characterization using deep learning and a publicly accessible polyp video.pdf

Download (5MB) | Preview

Abstract

OBJECTIVE: Convolutional neural networks (CNN) for computer-aided diagnosis (CADx) of polyps are often trained using high-quality still images in a single chromoendoscopy imaging modality with sessile serrated lesions (SSL) often excluded. This study developed a CNN from videos to classify polyps as adenomatous or non-adenomatous using standard narrow-band imaging (NBI) and NBI-near focus (NBI-NF) and created a publicly accessible polyp video database. METHODS: We trained a CNN with 16,832 high and moderate-quality frames from 229 polyp videos (56 SSLs). It was evaluated with 222 polyp videos (36 SSLs) across two test-sets. Test-set I consists of 14,320 frames (157 polyps, 111 diminutive). Test-set II, which is publicly accessible, 3,317 video frames (65 polyps, 41 diminutive) which was benchmarked with three expert and three non-expert endoscopists. RESULTS: Sensitivity for adenoma characterisation was 91.6% in test-set I and 89.7% in test-set II. Specificity was 91.9% and 88.5%. Sensitivity for diminutive polyps was 89.9% and 87.5%; specificity 90.5% and 88.2%. In NBI-NF, sensitivity was 89.4% and 89.5%, with a specificity of 94.7% and 83.3%. In NBI, sensitivity was 85.3 % and 91.7%, with a specificity of 87.5% and 90.0%, respectively. The CNN achieved PIVI-1 and PIVI-2 thresholds for each test-set. In the benchmarking of test-set II, the CNN was significantly more accurate than non-experts (13.8% difference (95%CI 3.2-23.6), p=0.01) with no significant difference with experts. CONCLUSIONS: A single CNN can differentiate adenomas from SSL and hyperplastic polyps in both NBI and NBI-NF. A publicly accessible NBI polyp video database was created and benchmarked.

Type: Article
Title: Polyp characterization using deep learning and a publicly accessible polyp video database
Location: Australia
Open access status: An open access version is available from UCL Discovery
DOI: 10.1111/den.14500
Publisher version: https://doi.org/10.1111/den.14500
Language: English
Additional information: © 2023 The Authors. Digestive Endoscopy published by John Wiley & Sons Australia, Ltd on behalf of Japan Gastroenterological Endoscopy Society. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
Keywords: Artificial intelligence, colonic polyps, colonoscopy, colorectal neoplasms, deep learning
UCL classification: UCL
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences
UCL > Provost and Vice Provost Offices > UCL BEAMS
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Medical Sciences
UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Medical Sciences > Div of Surgery and Interventional Sci
UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Computer Science
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Medical Sciences > Div of Surgery and Interventional Sci > Department of Targeted Intervention
URI: https://discovery.ucl.ac.uk/id/eprint/10162054
Downloads since deposit
92Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item