COLO-DETECT
Research type
Research Study
Full title
COLO-DETECT: A Randomised Controlled Trial of Lesion Detection at Colonoscopy Using the GI Genius™ Artificial Intelligence Platform
IRAS ID
286426
Contact name
Colin Rees
Contact email
Sponsor organisation
South Tyneside and Sunderland NHS Foundation Trust
Clinicaltrials.gov Identifier
Clinicaltrials.gov Identifier
ClinicalTrials.gov, Pending
Duration of Study in the UK
1 years, 6 months, 0 days
Research summary
Summary of Research
Bowel cancer (also known as colorectal cancer) is common in the UK, affecting around 1 in 15 men and 1 in 18 women at some point in their lifetime. Many colorectal cancers develop from polyps (small growths of the bowel lining). Some polyps can develop into cancer over a period of time, the most common type of polyp to do this is called an adenoma. Polyps can be detected and removed during a camera test (called a colonoscopy) before they become cancerous.
GI Genius™ is an Artificial Intelligence technology device designed to assist the person performing the colonoscopy in identifying polyps so that they may be removed, thus reducing the risk of future bowel cancer. The GI Genius™ is a box that integrates with existing colonoscopy equipment; it analyses the video feed from the colonoscope in real-time and highlights any area of the bowel lining that may be abnormal within a green box. This enables the person performing the colonoscopy to inspect the area more closely and if a polyp is identified, it can then be removed if appropriate. The GI Genius™ does not affect the size or function of the colonoscope in any other way.
The COLO-DETECT study is a clinical randomised controlled trial investigating the ability of colonoscopy assisted by GI Genius™ to detect polyps (and in particular, adenomas), when compared to standard colonoscopy.
Summary of Results
COLO-DETECT studied the effect of the GI Genius artificial intelligence device on polyp detection during colonoscopy for bowel cancer screening, for bowel symptoms, or for surveillance after previous polyps or bowel cancer. We compared it to normal (standard) colonoscopy.
We recruited a total of 2032 participants across 10 NHS Trusts in England (some in the North East, North West, Midlands, and South of England).
60% of those taking part were having a bowel cancer screening colonoscopy, 40% were having a colonoscopy for bowel symptoms or surveillance.
Participants were split evenly between the GI Genius group and the standard colonoscopy group.
The results are really exciting and make this trial a really important addition to the published literature on the subject of AI in colonoscopy.
MAP – mean adenomas per procedure (average number of adenomas in each colonoscopy)
• When using GI Genius colonoscopists detected (on average) an extra 0.36 adenomas per colonoscopy.
o Adenomas are the main polyp type that cause bowel cancer.
o This is a big difference compared to other things that have been introduced to improve polyp detection.
ADR – adenoma detection rate (proportion of colonoscopies with ≥1 adenoma)
• When using GI Genius colonoscopists found at least one adenoma in an extra 8 people out of every 100 people having a colonoscopy (56·6% vs 48·4%).
o Like MAP, this is a big difference when compared to other things that try to improve polyp detection.
o It could mean a large reduction in the number of bowel cancers that occur in people who have had a colonoscopy.
SSLDR – Sessile Serrated Lesion Detection Rate (proportion of colonoscopies with ≥1 SSL)
• When using GI Genius colonoscopists found at least one SSL in an extra 3 people out of every 100 people having a colonoscopy (11.6% vs 8.3%).
o SSLs cause about 20-30% of bowel cancers.
o Finding more SSLs seems to have an even larger impact on reducing bowel cancers after colonoscopy than finding more adenomas does.
Other findings:
• Colonoscopists found more smaller polyps, and more flat polyps when they used the GI Genius than when they didn’t.
• They did not find more ‘advanced’ adenomas or more bowel cancers.
• There was no difference in which parts of the bowel the polyps were located in.
• Colonoscopies took (on average) 1.5 minutes longer with GI Genius than without it
• Using GI Genius did not increase the risk of complications occurring after colonoscopy.REC name
West of Scotland REC 4
REC reference
21/WS/0003
Date of REC Opinion
25 Jan 2021
REC opinion
Further Information Favourable Opinion