Early introduction of Video Capsule with AI for Upper GI bleeding
Research type
Research Study
Full title
An evaluation of AI- and clinician-read video capsule enteroscopy for the investigation of non-haematemesis upper GI bleeding
IRAS ID
342699
Contact name
Prof Bu'Hussain Hayee
Contact email
Sponsor organisation
King's College Hospital
Duration of Study in the UK
0 years, 10 months, 5 days
Research summary
This research proposal aims to investigate whether we can use artificial intelligence (AI) to identify the source of bleeding earlier, in patients arriving at UK hospitals with Non-Hematemesis acute upper gastrointestinal (GI) bleeding, using video capsule endoscopy (VCE-AI). Presently, only half of patients with acute upper GI bleeding (AUGIB) undergo gastroscopy within the first 24 hours, contrary to NICE recommendations. With mortality from AUGIB affecting over 5000 patients annually in the UK, early identification of the bleeding source and subsequent intervention could potentially reduce this mortality and associated morbidity.
The use of VCE in AUGIB is not new, having been established by randomised trials. Uptake is limited, however, given the time currently required for video to be analysed by an expert physician. We believe that using VCE-AI, which 'highlights' abnormalities for quicker validation by a clinician, will lead to a faster diagnosis.
The study will focus on patients admitted to a UK hospital with non-haematemesis AUGIB, with eligibility criteria including:
• Non-haematemesis UGIB (Melaena, Coffee ground vomit, IDA)
• Haemodynamically stable
• 18 years and over
• Glasgow Blatchford Score (GBS) ≥1 (Patient already planned for admission)The research will be conducted at two sites within the same trust over a period of Nine months. The study aims to recruit 78 patients. It is important to note that this is not a therapy study, and there will be no alteration to the current pathway for managing AUGIB patients as part of the study. It is an observational study to see if the introduction of VCE-AI (which is already in the pathway of management of UGIB), earlier in the diagnosis algorithm, can lead to earlier identification of the source of bleeding.
The findings of this research endeavour have the potential to enhance the pathway for patients presenting to the emergency department with AUGIB, ultimately improving patient outcomes and healthcare delivery.
REC name
West of Scotland REC 5
REC reference
25/WS/0037
Date of REC Opinion
11 Mar 2025
REC opinion
Favourable Opinion