EAGLE Trial

  • Research type

    Research Study

  • Full title

    Evaluation of AI for detection of Gastrointestinal Lesions in Endoscopy (EAGLE)

  • IRAS ID

    319223

  • Contact name

    Manish Chand

  • Contact email

    m.chand@ucl.ac.uk

  • Sponsor organisation

    ODIN MEDICAL Ltd

  • Clinicaltrials.gov Identifier

    NCT05730192

  • Clinicaltrials.gov Identifier

    N/A, N/A

  • Duration of Study in the UK

    0 years, 9 months, 1 days

  • Research summary

    Summary of Research

    Abnormal growths in the colon and rectum are referred to as polyps. Polyps can be harmless (such as hyperplastic polyps in the rectum) or harmful (for example adenomatous polyps or sessile serrated lesions (SSLs)). Colonoscopy is considered the ‘gold standard’ for detecting polyps and cancer in the large intestine. Yet a major challenge in colonoscopy is the significant miss rates for these polyps.

    Using ‘artificial intelligence’ (AI), Odin Vision has developed a device called CADDIE which can help the endoscopist to detect polyps, with an aim to reduce the amount of “missed” polyps.

    EAGLE is a multicentre study to evaluate the clinical performance and safety of CADDIE in patients referred for screening or surveillance colonoscopy. CADDIE comprises a computer attached to the output of the endoscopy stack and a screen which shows normal colon video output in real-time. CADDIE processes information contained in digital images in real-time during endoscopy procedures. It informs the endoscopist where there might be a polyp by highlighting areas of interest on the screen with a bounding box and an optional sound alarm. CADDIE does not involve any personal contact with the patient. All decisions are still made by the endoscopist.

    The target population are patients undergoing a screening or surveillance colonoscopy. Consenting patients will be registered and randomised to intervention (=routine colonoscopy with CADDIE), or control group (=routine colonoscopy without CADDIE).

    To answer the main research question, we will evaluate whether more polyps, particularly precancerous adenomatous polyps, are identified when using CADDIE compared to without as well as whether using CADDIE leads the endoscopist to incorrectly identifying polyps.

    We will recruit 946 patients to allow for an expected dropout rate of 10% leaving 860 patients in the study. There are no study-related follow up visits.

    Summary of Results

    Using CADDIE (a computer-aided detection tool) during colonoscopy led to the detection and removal of more clinically important bowel lesions than standard colonoscopy, in screening and surveillance patients. The study therefore met its main aim, showing that the device can help clinicians find more pre-cancerous growths (adenomas) and early cancers during routine colonoscopy.

    Improvements were also seen across several related outcomes. With the device, clinicians found more polyps overall, more neoplastic polyps (polyps more likely to progress to cancer), and more serrated polyps (a pathway that can also lead to bowel cancer). The benefits were seen across different patient and procedure subgroups, suggesting the results are applicable to a broad range of people undergoing screening or surveillance colonoscopy. Importantly, the device-assisted group detected more larger lesions, which are generally considered more clinically significant because they have a higher likelihood of progressing to cancer if left untreated.

    The safety results did not suggest that use of the device increased unnecessary biopsies or removals. The additional tissue removed was mainly clinically relevant polyps rather than normal bowel lining. The time taken for examinations where no polyps were found was similar in both groups, indicating the device did not prolong routine negative procedures. Total procedure time was slightly longer in the device-assisted group, consistent with the increased number of polyps removed. No serious harms were identified that were judged to be caused by the device during the study.

  • REC name

    London - Dulwich Research Ethics Committee

  • REC reference

    23/LO/0106

  • Date of REC Opinion

    3 Mar 2023

  • REC opinion

    Further Information Favourable Opinion