Lily L Lai1, Andrew Blakely2, Marta Invernizzi1, James Lin3, Trilokesh Kidambi3, Kurt A Melstrom1, Kevin Yu4, Thomas Lu4. 1. City of Hope, Department of Surgery, Duarte, California, United States. 2. National Cancer Institute, National Institutes of Health Campus, Department of Surgery, Bethesda, Ma, United States. 3. City of Hope, Division of Gastroenterology, Duarte, California, United States. 4. Jet Propulsion Labarotory, Pasadena, California, United States.
Abstract
SIGNIFICANCE: Colorectal cancer incidence has decreased largely due to detection and removal of polyps. Computer-aided diagnosis development may improve on polyp detection and discrimination. AIM: To advance detection and discrimination using currently available commercial colonoscopy systems, we developed a deep neural network (DNN) separating the color channels from images acquired under narrow-band imaging (NBI) and white-light endoscopy (WLE). APPROACH: Images of normal colon mucosa and polyps from colonoscopies were studied. Each color image was extracted based on the color channel: red/green/blue. A multilayer DNN was trained using one-channel, two-channel, and full-color images. The trained DNN was then tested for performance in detection of polyps. RESULTS: The DNN performed better using full-colored NBI over WLE images in the detection of polyps. Furthermore, the DNN performed better using the two-channel red + green images when compared to full-color WLE images. CONCLUSIONS: The separation of color channels from full-color NBI and WLE images taken from commercially available colonoscopes may improve the ability of the DNN to detect and discriminate polyps. Further studies are needed to better determine the color channels and combination of channels to include and exclude in DNN development for clinical use.
SIGNIFICANCE: Colorectal cancer incidence has decreased largely due to detection and removal of polyps. Computer-aided diagnosis development may improve on polyp detection and discrimination. AIM: To advance detection and discrimination using currently available commercial colonoscopy systems, we developed a deep neural network (DNN) separating the color channels from images acquired under narrow-band imaging (NBI) and white-light endoscopy (WLE). APPROACH: Images of normal colon mucosa and polyps from colonoscopies were studied. Each color image was extracted based on the color channel: red/green/blue. A multilayer DNN was trained using one-channel, two-channel, and full-color images. The trained DNN was then tested for performance in detection of polyps. RESULTS: The DNN performed better using full-colored NBI over WLE images in the detection of polyps. Furthermore, the DNN performed better using the two-channel red + green images when compared to full-color WLE images. CONCLUSIONS: The separation of color channels from full-color NBI and WLE images taken from commercially available colonoscopes may improve the ability of the DNN to detect and discriminate polyps. Further studies are needed to better determine the color channels and combination of channels to include and exclude in DNN development for clinical use.
Entities:
Keywords:
artificial intelligence algorithms; color channel separation; colorectal cancer; deep learning; narrow-band imaging; polyp discrimination
Authors: Raf Bisschops; James E East; Cesare Hassan; Yark Hazewinkel; Michał F Kamiński; Helmut Neumann; Maria Pellisé; Giulio Antonelli; Marco Bustamante Balen; Emmanuel Coron; Georges Cortas; Marietta Iacucci; Mori Yuichi; Gaius Longcroft-Wheaton; Serguei Mouzyka; Nastazja Pilonis; Ignasi Puig; Jeanin E van Hooft; Evelien Dekker Journal: Endoscopy Date: 2019-11-11 Impact factor: 10.093
Authors: Barham K Abu Dayyeh; Nirav Thosani; Vani Konda; Michael B Wallace; Douglas K Rex; Shailendra S Chauhan; Joo Ha Hwang; Sri Komanduri; Michael Manfredi; John T Maple; Faris M Murad; Uzma D Siddiqui; Subhas Banerjee Journal: Gastrointest Endosc Date: 2015-01-16 Impact factor: 9.427
Authors: Ann G Zauber; Sidney J Winawer; Michael J O'Brien; Iris Lansdorp-Vogelaar; Marjolein van Ballegooijen; Benjamin F Hankey; Weiji Shi; John H Bond; Melvin Schapiro; Joel F Panish; Edward T Stewart; Jerome D Waye Journal: N Engl J Med Date: 2012-02-23 Impact factor: 91.245
Authors: Michael F Byrne; Nicolas Chapados; Florian Soudan; Clemens Oertel; Milagros Linares Pérez; Raymond Kelly; Nadeem Iqbal; Florent Chandelier; Douglas K Rex Journal: Gut Date: 2017-10-24 Impact factor: 23.059
Authors: Stephanie Taha-Mehlitz; Silvio Däster; Laura Bach; Vincent Ochs; Markus von Flüe; Daniel Steinemann; Anas Taha Journal: J Clin Med Date: 2022-04-26 Impact factor: 4.964