How AI is improving polyp detection

Artificial intelligence can improve the accuracy and sensitivity of adenoma detection, according to a poster presented at World Congress of Gastroenterology 2017, Oct. 13 to Oct. 18 in Orlando, Fla., Medscape report

William Karnes, MD, from UC Irvine, and colleagues found that AI can increase detection for lesions as small as 5 mm.

Out of 9,000 colonoscopy images, convolutional neural network technology was able to identify polyps with a 96 percent accuracy rate.

Dr. Karnes said the machine-learning system reads up to 170 images per second. He said the technology can "easily [be] applied to live video."

Dr. Karnes said, "Artificial intelligence for polyp detection has the potential to help all colonoscopists achieve detection rates closer to true prevalence, and to further reduce the risk of interval colorectal cancers."

In a separate presentation, Yuichi Mori, MD, of Shinagawa-ku, Japan-based Showa University, examined EndoBrain. EndoBrain uses AI to provide an automatic diagnosis at the push of a button.

The EndoBrain computer-aided endocytoscopy technology detected neoplastic polyps at a rate greater than 90 percent. In a subanalysis of rectosigmoid colon polyps smaller than 5 mm, Medscape reports the negative predictive value was 99 percent.

Dr. Mori presented a poster of the topic at United European Gastroenterology Week 2017, Oct. 28 to Nov. 1 in Barcelona, Spain.

Copyright © 2022 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.


Featured Webinars

Featured Whitepapers

Featured Podcast