What happened
Health Technology Wales (HTW) appraised AI-assisted endoscopy for lower gastrointestinal cancer and published its guidance in April 2024. In computer-aided detection (CADe), the software draws a box around areas of concern on the live colonoscopy video, and the endoscopist decides what to remove. HTW's researchers pooled 36 randomized trials with 27,870 participants. They estimated that the rate of adenoma detection was 23% higher with CADe than with standard colonoscopy, at 44.3% against 37.1%. Detection of polyps and sessile serrated lesions also improved, and differences in withdrawal time were small. HTW found no evidence that CADe detected more advanced adenomas or cancers, and no trial measured cancer rates or survival over the long term. In its own economic model, HTW estimated that CADe costs an extra £4,197 for each quality-adjusted life year gained. The appraisal panel concluded that the evidence supports routine adoption, and HTW asked for data on use in practice. The main source of bias was that endoscopists were not blinded, and HTW reports high heterogeneity between the trials.
Sources
-
Assessment report, by authors with no tie to the system.
Health Technology Wales. Artificial Intelligence (AI)-assisted endoscopy in the detection of gastrointestinal cancer and pre-cancerous lesions. Guidance GUI055, April 2024 (topic page, 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.
-
Assessment report, by authors with no tie to the system.
It was estimated from HTW’s meta-analysis that ADR was 23% higher in the CADe group than the standard colonoscopy group (Table 2).
Details
AI system
- Workflow
-
CADe drew a box around areas of concern on the live colonoscopy video. Endoscopists decided whether to remove or sample each flagged area.
Follow-up
- Conditions for replication
-
Patients resemble the trial populations, which excluded people with inflammatory bowel disease, colorectal cancer or earlier colorectal surgery. Services collect data on how CADe performs in routine use, as HTW recommends.
Entry details
- Sector
-
Healthcare.
- Purpose
-
Medtronic, Olympus, Fujifilm, Wision AI, Wuhan EndoAngel and other makers built CADe systems to flag possible polyps and cancers to endoscopists during colonoscopy.
Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). Wales backs CADe colonoscopy after it found more adenomas in trials (AIBD-0742). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0742.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)