AI-Assisted Colonoscopy: New Research and Guidelines for Clinical Use

Author(s):

Rachel Martin

Key Takeaways

AI-assisted colonoscopies improve adenoma detection rates, offering a "second set of eyes" for gastroenterologists.
A meta-analysis of 44 studies indicates CADe systems enhance adenoma detection, though advanced neoplasia detection improvement is minimal.
The American Gastroenterological Association is drafting guidelines for AI use in colonoscopies, with final publication anticipated in March 2025.
Clinicians should remain informed on AI advancements and guidelines to optimize colorectal cancer screening and patient outcomes.

Computer-aided detection during colonoscopies with AI may help increase adenoma detection rates.

Gyanprakash Ketwaroo, MD

Colorectal cancer is the third most common cancer and a leading cause of cancer death in the United States. Colonoscopy, recommended for all adults aged 45 or older, is a crucial screening tool where gastroenterologists remove polyps to prevent them from becoming cancerous. Recognizing the importance of detecting these polyps, the FDA authorized the first AI-assisted device for colonoscopies in 2021.

Dr. Gyanprakash Ketwaroo at the West Haven VA emphasizes the value of AI in providing a "second set of eyes" during colonoscopies. This technology, known as computer-aided detection (CADe), has shown promise in increasing adenoma detection rates, according to a meta-analysis led by Dr. Dennis Shung. The analysis, which reviewed 44 studies, found that CADe systems improved adenoma detection compared to standard colonoscopies, although the increase in advanced colorectal neoplasia detection was slight.

Shung's meta-analysis suggests further research is needed to assess AI's impact on interval colorectal cancer rates and recommends that future studies should randomize colonoscopists to understand better where CADe can be most effective. Yale has implemented this AI technology at the West Haven VA campus, providing additional support for endoscopists and trainees during procedures.

The American Gastroenterological Association (AGA) is also preparing new clinical guidelines for using AI in colonoscopies, for which Dr. Shung is the co-first author. These draft guidelines conditionally recommend CADe for polyp detection in adults, considering the technology's safety and efficiency. The AGA is gathering feedback from both providers and patients on AI use in screening, aiming to publish the final guideline in March 2025, with regular updates as the technology evolves.

Overall, AI-assisted colonoscopies represent a significant advancement in colorectal cancer screening, with the potential to enhance detection rates and outcomes for patients. Clinicians are advised to stay informed on evolving guidelines and technological developments to offer the best care possible.