AI Technology Revolutionizes Colonoscopy
PHILADELPHIA—An analysis from the Minneapolis VA Medical Center indicates that colonoscopies enhanced with artificial intelligence (AI) technology lead to a greater prevalence of benign lesions being resected compared to traditional methods. Tessa Herman, MD, a chief resident at the University of Minnesota, explained that the data showed a 4% absolute increase in benign lesions resection due to AI assistance.
Impact of AI on Colonoscopy Procedures
The study highlights AI-assisted colonoscopy (AIAC), utilizing computer-aided detection (CADe) technology designed to enhance the quality of colonoscopy. AIAC has improved the adenoma detection rate and reduced the adenoma miss rate. With information drawn from 12 randomized controlled trials encompassing around 13,000 patients, the analysis confirmed that AI-led interventions notably increased the resection of non-neoplastic lesions.
Study Design and Findings
To evaluate the resection rates of benign lesions during colonoscopies aimed at colorectal cancer prevention, Dr. Herman and colleagues revisited data from a previous trial. They assessed colonoscopies performed for screening, surveillance, or positive fecal immunochemical tests, discounting those done for diagnostic purposes or other complications.
From 1,040 colonoscopies conducted, 599 utilized AI assistance while 441 were unassisted. The outcomes showed that the resection of benign non-adenomatous lesions was significantly higher at 30.9% in the AI group compared to 22.7% in the unassisted group.
Addressing Potential Challenges
Dr. Herman emphasized that while many benign lesions like colonic mucosa and lymphoid aggregates were successfully resected using AI, not all procedures may have been necessary, raising concerns regarding the potential increase in healthcare costs and procedural risks. In her presentation at ACG 2024, she highlighted that although AIAC improved benign lesion detection, it could lead to unnecessary interventions.
Future Implications
The research urges further exploration into refining AI technologies in colonoscopy. Vani J.A. Konda, MD, echoed Dr. Herman’s sentiments and remarked that current advancements call for a balanced approach toward clinical decision-making aided by AI, while ensuring the endoscopist’s judgment maintains paramount importance.
Conclusion
As AI continues to evolve in the healthcare sector, its integration into colonoscopy practices could significantly alter protocols aimed at colorectal cancer prevention. However, further innovations are necessary to enhance the precision of this technology and fine-tune its application in clinical settings.
- 0 Comments
- Benign Lesions
- Colonoscopy