Mode
Text Size
Log in / Sign up

Bayesian network meta-analysis shows AI-assisted colonoscopy improves adenoma detection in adultsAI Boosts Colon Cancer Detection, But Not All Systems Are Equal

AI-generated summary of the cited source, checked by automated accuracy review. How we work

Key Takeaway
Consider AI-assisted colonoscopy for adenoma detection, noting small system differences and uncertain high-risk lesion benefits.

This Bayesian network meta-analysis examined the performance of artificial intelligence-assisted colonoscopy against standard colonoscopy in adults undergoing the procedure. The study included 34 106 participants and assessed multiple outcomes including adenoma detection rate, adenomas per colonoscopy, withdrawal time, and detection of advanced or sessile serrated lesions. The certainty of evidence was moderate.

The analysis demonstrated that artificial intelligence-assisted colonoscopy significantly improved adenoma detection rate compared with standard colonoscopy. EndoAngel achieved an odds ratio of 1.84 with a SUCRA of 0.9, while EndoAID showed an odds ratio of 1.64 and a SUCRA of 0.7. CAD-EYE and GI Genius also showed improvements with odds ratios of 1.46 and 1.45 respectively, both with a SUCRA of 0.5. EndoAID demonstrated the largest benefit for adenomas per colonoscopy with a mean difference of 0.62.

Regarding withdrawal time, EndoAngel modestly increased this metric with a mean difference of 1.14 minutes. However, no system significantly improved the detection of advanced or sessile serrated lesions. The authors note that differences between systems are small and benefits for high-risk lesions remain uncertain. Further head-to-head trials and cost-effectiveness studies are needed to clarify the clinical value of these technologies.

The Hidden Danger Inside Your Colon

You go in for a routine colonoscopy. You expect the doctor to find and remove any suspicious polyps. But here is the uncomfortable truth: even skilled doctors miss some polyps.

Colorectal cancer is the third most common cancer worldwide. Most cases start as adenomatous polyps. These are small growths that can turn into cancer over time. Finding them early saves lives.

That is why the idea of AI helping doctors see better is so exciting. But does it actually work?

What AI Brings to the Table

Think of a standard colonoscopy like searching a dark room with a flashlight. You see most things, but you might miss something in the corner.

AI acts like a second pair of eyes that never gets tired. It highlights suspicious spots on the screen in real time. The doctor still makes the final call. But the AI points out areas worth a closer look.

This doesn't mean AI replaces your doctor. It just gives them a better view.

The Numbers That Matter

Researchers looked at 48 clinical trials involving more than 34,000 people. They compared standard colonoscopy against four AI systems: EndoAngel, EndoAID, CAD-EYE, and GI Genius.

The results were clear. Every AI system helped doctors find more polyps than standard colonoscopy alone.

EndoAngel performed the best. It increased the chance of finding a polyp by 84 percent compared to standard exams. EndoAID came next at 64 percent. CAD-EYE and GI Genius both improved detection by about 45 percent.

But Here Is the Catch

The differences between AI systems were small. And none of them helped doctors find advanced polyps or sessile serrated lesions. These are the types most likely to turn into cancer.

That matters. Finding any polyp is good. But finding the dangerous ones is what really saves lives.

The best AI system, EndoAngel, also added about one extra minute to the procedure. That is not a big deal for most people. But it shows that AI is not a perfect solution.

Imagine you are looking at a photo of a crowded street. Your eyes naturally focus on the people and cars. But you might miss a small sign in the background.

AI works like a highlighter pen. It scans every pixel of the colon image. When it spots something that looks like a polyp, it draws a box around it. The doctor then decides if it is real.

This is called computer vision. It is the same technology that lets your phone recognize your face. But instead of faces, it learns to spot polyps by studying thousands of colon images.

If you are scheduled for a colonoscopy, you can ask if your doctor uses AI assistance. Many hospitals now offer it. But do not panic if yours does not.

Standard colonoscopy is still very good. AI just makes it a little better.

The bigger message is simple: get your screening. Colonoscopy remains the gold standard for preventing colorectal cancer. AI is a helpful upgrade, not a replacement.

The Limits of This Research

These results come from a network meta-analysis. That is a fancy way of saying researchers combined many studies to get a bigger picture.

But the studies were not all the same. Some used different AI systems. Some had different patient groups. And none of the studies directly compared one AI system against another.

The certainty of the evidence was moderate, not high. That means future research could change what we know.

What Happens Next

Researchers need to run head-to-head trials. They need to test EndoAngel against EndoAID directly. They also need to study whether AI actually prevents cancer deaths, not just finds more polyps.

Cost is another question. AI systems cost money. Hospitals need to decide if the small improvement is worth the price.

For now, the takeaway is clear. AI helps doctors find more polyps. But it is not a magic wand. The best tool is still a well-trained doctor using the best technology available.

Study Details

Study typeMeta analysis
Sample sizen = 106
EvidenceLevel 1
PublishedJun 2026
View Original Abstract ↓
BACKGROUND: Colorectal cancer is a major global health burden, with most cases arising from adenomatous polyps. Although colonoscopy is the gold standard for detection, its effectiveness is operator-dependent. Artificial intelligence-assisted systems have been developed to improve adenoma detection, but their comparative performance remains unclear. METHODS: We performed a systematic review and Bayesian network meta-analysis of randomized controlled trials comparing artificial intelligence-assisted with standard colonoscopy. PubMed, Scopus, and Google Scholar were searched up to 4 November 2025. Eligible studies included adults undergoing colonoscopy and reporting adenoma detection rate (ADR) and adenomas per colonoscopy (APC). Secondary outcomes included withdrawal time and detection of advanced and sessile serrated lesions. Risk of bias was assessed using Cochrane RoB 2.0, and certainty of evidence was evaluated with CINeMA. RESULTS: A total of 48 randomized controlled trials (34 106 participants) were included. Artificial intelligence-assisted colonoscopy significantly improved ADR compared with standard colonoscopy. EndoAngel showed the greatest effect [odds ratio (OR): 1.84, surface under the cumulative ranking curve (SUCRA): 0.9], followed by EndoAID (OR: 1.64, SUCRA: 0.7), CAD-EYE (OR: 1.46, SUCRA: 0.5), and GI Genius (OR: 1.45, SUCRA: 0.5). For APC, EndoAID demonstrated the largest benefit (mean difference: 0.62). EndoAngel modestly increased withdrawal time (mean difference: 1.14 minutes). No system significantly improved detection of advanced or sessile serrated lesions. Heterogeneity was low, and certainty of evidence was moderate. CONCLUSION: Artificial intelligence-assisted colonoscopy improves adenoma detection; however, differences between systems are small, and benefits for high-risk lesions remain uncertain. Further head-to-head trials and cost-effectiveness studies are needed.
Free Newsletter

Clinical research that matters. Delivered to your inbox.

Join thousands of clinicians and researchers. No spam, unsubscribe anytime.