This article is more than 5 years old. Information may no longer be current.

AI System Meets ‘Resect and Discard’ Polyp Identification Standards

Issue: December 2018

Add topic to email alerts

Receive an email when new articles are posted on

Please provide your email address to receive an email when new articles are posted on .

Subscribe

Added to email alerts

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

Back to Healio

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Back to Healio

PHILADELPHIA — An artificial intelligence program helped achieve real-time pathology of diminutive polyps during colonoscopy screening, according to research presented at the American College of Gastroenterology Annual Scientific Meeting.

“Eighty-percent of polyps removed during colonoscopy are diminutive in nature, and these rarely contain any advanced pathology and cancer,” Robin Zachariah, MD, of the University of California Irvine Medical Center, said in his presentation. “Reliable optical diagnosis of diminutive polyps could allow for resect and discard and/or a leave along strategy. This translates to a potential savings of up to $1 billion annually in pathology-related costs in the United States.”

Zachariah and colleagues developed the program using a convolutional neural network and tested it by extracting more than 5,000 high-quality images of adenomas and serrated polyps of known locations, size and light source. Then, they calculated surveillance intervals and compared the optical pathology with true pathology of the polyps.

Among the polyps located throughout the colon, they found that the negative predictive value (NPV) for adenomas was 92% for white light endoscopy and 93% for narrow band imaging.

For diminutive polyps found throughout the colon, the NPV was 91% for white light and 92% for narrow band. When identifying diminutive polyps in the left colon, the NPV increased to 97% for white light and 95% for narrow band.

Zachariah said the optical pathology model meets the ASGE’s Preservation and Incorporation of Valuable endoscopic Innovations (PIVI) “resect and discard” guidelines that require an NPV of at least 90% for adenomas.

“To our knowledge we are the first deep learning, optical pathology technique challenged against surveillance intervals,” Zachariah said. “Potential applications of these results include point of care [adenoma detection rate] and point of care surveillance recommendations.” – by Alex Young

Reference: Zachariah R, et al. Abstract 41. Presented at: American College of Gastroenterology Annual Scientific Meeting; Oct. 5-10, 2018; Philadelphia.

Disclosure: Zachariah reports no relevant financial disclosures. Please see the study abstract for all other authors’ relevant financial disclosures.