Kings lead £2.5m AI trial for early detection of bowel cancer

Researchers at King’s College Hospital NHS Foundation Trust are testing an artificial intelligence (AI) tool designed to support the analysis of endoscopy findings by helping doctors identify abnormalities to speed up diagnosis .

The trial will explore how an AI platform can improve patient outcomes for colorectal cancer with faster diagnosis and quicker treatment.

Starting this summer, the test will run through the end of 2024. The platform will be used by endoscopists at King’s College Hospital to highlight pre-cancerous lesions with a visual marker in real time. The images will be processed using advanced algorithms that are able to identify and characterize abnormalities with the polyps.

The technology involved in the trial is GI Genius, an intelligent endoscopy module from Cosmo Pharmaceuticals and distributed by Medtronic. It uses computer-aided detection (CADE) to process colonoscopy images in milliseconds. As an invaluable ‘second pair of eyes’, GI Genius can identify even the subtlest possible lesion, highlighting it clearly on the endoscopy display.

Computer-aided diagnosis (CADX) is used to reliably predict the histopathology of identified polyps, helping to prevent unnecessary polypectomy if the platform detects a potential lesion.

This complex neural network was built using videos of 2,684 histologically confirmed polyps from 840 patients who underwent high-definition white-light colonoscopy in a previous randomized controlled study.

Dr Bu Hiyi, consultant gastroenterologist at King’s College Hospital and principle investigator of the study, said: “I am delighted to be working on this project to establish whether the use of artificial intelligence (AI) in colonoscopy can affect endoscopy performance. and improve outcomes. For patients. The need for innovation in the NHS has never been greater and this research may be able to shed light on the potential benefits this technology can provide.

“Our NAIAD study (National Study of Artificial Intelligence in Adenoma Detection for Colonoscopy) set out to explore the use of AI in a ‘real world’ setting, and how it might impact endoscopists in their day-to-day practice “

The study received set-up funding from NHS England, following a competition to identify and rapidly scale AI technology into clinical services. King’s College Hospital will lead the study, although it will eventually involve 20 NHS trusts across England.

Published last May, King’s College Hospital NHS Foundation Trust’s annual action plan included a key theme of being digitally enabled to support high quality care and use innovative digital tools.