Endocuff With Or Without Artificial Intelligence-Assisted Colonoscopy In Detection Of Colorectal Adenoma: A Randomized Colonoscopy Trial
Thomas Ka-Luen Lui,Carla Pui-Mei Lam,Elvis Wai-Pan To,Michael Kwan-Lung Ko,Vivien Wai Man Tsui,Kevin Sze-Hang Liu,Cynthia Ka-Yin Hui,Michael Ka-Shing Cheung,Loey Lung-Yi Mak,Rex Wan-Hin Hiu,Siu-Yin Wong,Wai Kay Seto,Wai K Leung
DOI: https://doi.org/10.14309/ajg.0000000000002684
2024-02-03
The American Journal of Gastroenterology
Abstract:INTRODUCTION: Both artificial intelligence (AI) and distal attachment devices have been shown to improve adenoma detection rate and reduce miss rate during colonoscopy. We studied the combined effect of Endocuff and AI on enhancing detection rates of various colonic lesions. METHODS: This was a three-arm prospective randomized colonoscopy study involving patients aged 40 years or above. Participants were randomly assigned in a 1:1:1 ratio to undergo Endocuff with AI, AI alone, or standard high-definition (HD) colonoscopy. The primary outcome was adenoma detection rate (ADR) between the Endocuff-AI and the AI group, while secondary outcomes included detection rates of polyp (PDR), sessile serrated lesion (SDR), and advanced adenoma (AADR) between the two groups. RESULTS: A total of 682 patients were included (mean age 65.4 years, 52.3% male), with 53.7% undergoing diagnostic colonoscopy. The ADR for the Endocuff-AI, AI, and HD group was 58.7%, 53.8%, and 46.3%, respectively; while the corresponding polyp detection rate (PDR) was 77.0%, 74.0%, and 61.2%. A significant increase in ADR, PDR and SDR was observed between the Endocuff-AI and the AI group (ADR difference: 4.9%, 95% CI: 1.4% to 8.2%, p=0.03; PDR difference: 3.0%, 95% CI: 0.4% to 5.8%, p=0.04; SDR difference: 6.4%, 95% CI: 3.4% to 9.7%, p<0.01). Both Endocuff-AI and AI groups had higher ADR ,PDR, SDR and AADR than the HD group (all p<0.01). Conclusions: Endocuff in combination with AI further improves various colonic lesion detection rates when compared to AI alone.
gastroenterology & hepatology