AI-assisted endoscopy for gastrointestinal cancer
Topic Status Complete
Artificial Intelligence (AI)-assisted endoscopy in the detection of gastrointestinal cancer and pre-cancerous lesions
Outcome of the appraisal
The evidence supports the routine adoption of computer aided detection (CADe) colonoscopy for the detection of lower gastrointestinal cancer and pre-cancerous lesions.
Compared with standard colonoscopy, CADe is associated with improved detection of adenomas, polyps, and sessile serrated lesions, without considerable increases to withdrawal time.
Economic modelling suggests that CADe is cost effective compared with standard colonoscopy with an incremental cost-effectiveness ratio (ICER) of £4,197 per quality-adjusted life year (QALY) gained.
HTW recommends the collection of data on the real-world implementation and effectiveness of CADe.
Why was this topic appraised?
Lower gastrointestinal (GI) cancer is the third most common cancer globally. This type of cancer can develop from the slow progression of asymptomatic pre-cancerous polyps, such as adenomas or sessile serrated lesions. Identification and removal of pre-cancerous polyps reduces the risk of future cancers. Polyps are identified using colonoscopy, during which a flexible lighted tube is inserted into the lower GI tract and an endoscopist reviews video images in real-time. Colonoscopy relies on the experience of the endoscopist to recognise areas of concern. Artificial intelligence (AI) assisted lower GI endoscopy, known as computer aided detection (CADe) colonoscopy, uses AI technology to assist endoscopists in the identification of polyps and cancers. Computer aided diagnosis (CADx) systems can also assist in the characterisation of any identified polyps. It has been suggested that CADe colonoscopy can improve important outcomes such as adenoma detection rate, which, in turn, should reduce the incidence of cancer.
This topic was submitted by an endoscopist working in the NHS in Wales.
Plain language summary
While there is no official definition of Artificial Intelligence (AI), it is defined by the Oxford Dictionary as “the theory and development of computer systems able to perform tasks normally requiring human intelligence”. AI currently has many roles in healthcare, such as domestic, administrative, diagnosis, testing and treatment. Different terms can be used to describe the AI dependant on its role.
Lower-gastrointestinal (lower-GI) cancer is the uncontrolled growth of abnormal cells (cancer) in the lower digestive tract. This includes the colon, rectum, and anus.
An Endoscopy is a test to look inside your body. A long, thin tube with a small camera inside, called an endoscope, is passed into your body through a natural opening such as your mouth. For people with possible lower gastrointestinal cancer, the endoscopist looks at the organs involved for signs of cancer. The endoscopist decides if tissue needs to be sampled and may take some for testing.
AI assisted endoscopy uses a from of AI to assist the endoscopist in identifying possible cancer. This type of AI is known as “computer aided detection (CADe)”. Unlike other forms of AI, CADe does not ‘learn’ or acquire more knowledge than it is programmed with. The CADe is programmed to identify the various features of lower-GI cancer. During the endoscopy, areas of concern are flagged by the CADe system to the endoscopist by drawing a box, typically with a blue or green border, around areas of concern. The endoscopist then makes a final decision on whether to test the identified area. It is proposed that the use of AI in this context will help identify possible lower-GI cancer earlier.
Health Technology Wales looked for evidence on the use of Artificial intelligence assisted endoscopy in the detection of lower gastrointestinal cancer and pre-cancerous lesions. The evidence supports the routine adoption of computer aided detection
Topic Exploration Report
TER463 05.2023
Evidence Appraisal Review
EAR055 04.2024
Guidance
GUI055 04.2024
Health Technology Wales Partners
