Artificial Intelligence Enhances Detection of High-Risk Colon Polyps

Colorectal cancer remains one of Europe's most prevalent cancers, but early detection through screening significantly reduces its impact. Colonoscopy is considered the gold standard for screening, allowing healthcare providers to identify and remove potentially dangerous colon polyps before they develop into cancer. Nevertheless, accurately differentiating benign from potentially malignant polyps demands substantial expertise among endoscopists.

This process, known as "optical diagnosis," has traditionally been performed by experienced specialists, making it challenging for less seasoned doctors to make confident assessments.

Recent research from Lower Austria offers promising advancements. A study conducted at St. Pölten University Hospital examined whether young physicians could reliably distinguish between harmless and dangerous polyps with the support of artificial intelligence (AI). The team utilized the GI Genius AI system to assist endoscopy trainees, and the results were compelling: junior doctors supported by AI achieved diagnostic accuracy comparable to seasoned specialists.

This pioneering study, published in the American Journal of Gastroenterology, was led by Dr. Andreas Maieron. The findings suggest that integrating AI into colonoscopy procedures could make colorectal cancer screening safer, more efficient, and more cost-effective. Moreover, it has potential to improve medical training and expand access to high-quality preventive care.

AI acts as an invaluable partner during procedures. The GI Genius system analyzes real-time images captured during colonoscopy, offering immediate guidance on polyp assessment. In the study, 225 participants—comprising junior doctors supported by AI—performed evaluations that closely matched the histological analysis of excised polyps and the judgments of experienced physicians.

Remarkably, for small rectal polyps (≤5 mm), junior doctors correctly identified benign polyps over 90% of the time, aligning with expert-Level accuracy. The AI system alone achieved more than 93% accuracy in polyp characterization.

This advancement holds significant promise for preventive healthcare. AI-assisted colonoscopies can empower less experienced clinicians to deliver high-quality care, reducing unnecessary polyp removals, minimizing patient risk, and decreasing healthcare costs. Ultimately, this technology enhances patient safety and offers more effective long-term protection against colorectal cancer.

Source: https://medicalxpress.com/news/2025-06-ai-dangerous-colon-polyps.html