AI-supported colorectal polyp detection (DAP78)

Colorectal cancer (CRC) also known as bowel cancer includes cancer of the colon or rectum (1).

It is the fourth most common cancer in the UK with over 44000 new cases each year and accounts for 10% of all cancer-related deaths in the UK.

It is the second most common cause of cancer-related death; the earlier it can be diagnosed the better the chances of survival (2-4).Colorectal polyps are small growths on the lining of the colon or rectum that are usually asymptomatic but may be detected on colonoscopy. Symptoms such as rectal bleeding or a change in bowel habits can occur in some cases.

Most colorectal polyps are harmless but some have the potential to turn into cancer over time if untreated (5). Colonoscopy is the gold standard for the detection of colorectal polyps. People may undergo colonoscopy through various pathways as summarised in Figure 1.

Overview of groups offered colonoscopy (reproduced from NICE DAP78 scope) During a colonoscopy concerning polyps can be removed. In the UK all removed polyps are sent for histopathological testing to classify them as neoplastic or non-neoplastic. Neoplastic polyps such as adenomas and serrated polyps have a higher risk of developing into cancer.

Some polyps may need to be removed during a separate procedure and endoscopists may choose not to remove certain polyps if they are confident they are not neoplastic.

A “resect and discard” strategy where an optical diagnosis by the endoscopist would suffice without histopathological testing is in its early stages of use within the Bowel Cancer Screening Programme (BCSP).In recent years artificial intelligence (AI)-supported colonoscopy technologies have been developed.

These include computer-aided detection (CADe) and computer-aided diagnosis (CADx).

CADe aims to reduce missed polyps by flagging areas of interest to for review while CADx involves characterising detected polyps to suggest whether they are likely neoplastic supporting decisions about resection.

This may reduce resections of non-neoplastic polyps and increase neoplastic polyp resection.This assessment will review the evidence for the clinical and cost-effectiveness of specific AI-supported colonoscopy technologies including CADe and CADx.

The systematic literature review (SLR) will aim to identify evidence related to clinical effectiveness diagnostic accuracy safety and other outcomes.

Previous economic evaluations will be reviewed and a new economic model may be developed to assess whether the technologies represent a good use of NHS resources where sufficient data are available.References:1.Cancer Research UK. About cancer - bowel cancer. Available from: https://www.cancerresearchuk.org/about-cancer/bowel-cancer.

Date accessed: Aug 24.2.Bowel Cancer UK. About bowel cancer - bowel cancer. Available from: https://www.bowelcanceruk.org.uk/about-bowel-cancer/bowel-cancer/. Date accessed: Aug 24.3.Cancer Research UK. Bowel cancer statistics.

Available from: https://www.cancerresearchuk.org/health-professional/cancer-statistics/statistics-by-cancer-type/bowel-cancer#heading-One. Date accessed: Aug 24.4.NHS England. Screening and earlier diagnosis. Available from: https://www.england.nhs.uk/cancer/early-diagnosis/screening-and-earlier-diagnosis/.

Date accessed: Aug 24.5.NHS. Bowel polyps 2023. Available from: https://www.nhs.uk/conditions/bowel-polyps/. Date accessed: Aug 24