Does AI-assisted Colonoscopy Improve Adenoma Detection in Screening Colonoscopy? A Multi-center Randomized Controlled

There are several studies showing that AI-assisted colonoscopy can help in identifying and characterizing polyps found on colonoscopy.

• Byrne et al demonstrated that their AI model for real-time assessment of endoscopic video images of colorectal polyp can differentiate between hyperplastic diminutive polyps vs adenomatous polyps with sensitivity of 98% and specificity of 83% (Byrne et al. GUT 2019)
• Urban et al designed and trained deep CNNs to detect polyps in archived video with a ROC curve of 0.991 and accuracy of 96.4%. The total number of polyps identified is significantly higher but mainly in the small (1-3mm and 4-6mm polyps) (Urban et al. Gastroenterol 2018)
• Wang et al conducted an open, non-blinded trial consecutive patients (n=1058) prospectively randomized to undergo diagnostic colonoscopy with or without AI assistance. They found that AI system increased ADR from 20.3% to 29.1% and the mean number of adenomas per patients from 0.31 to 0.53. This was due to a higher number of diminutive polyps found while there was no statistic difference in larger adenoma. (Wang et al. GUT 2019). In this study, they excluded patients with IBD, CRC and colorectal surgery. The patients presented with symptoms to hospital for investigation.

To date, there is a lack of large-scale randomized controlled study using AI assistance in the detection of polyps/adenoma in a screening population. The correlation of fecal occult blood test (FIT or FOBT) and the advantage of AI-assisted colonoscopy has not been investigated. There is also a lack of information of the benefit of AI-assisted colonoscopy in experienced colonoscopist versus trainee/resident.