Comparative study of endocuff assisted colonoscopy and artificial intelligence in colonoscopy in adenoma and polyp detection: a randomized, back-to-back study

An endocuff is attached at the tip of colonoscope to improve mucosal visualization. An artificial intelligence (AI) assisted colonoscopy is a computer system designed to enhance detection and differentiation polyps. The aim of this study was to evaluate the endocuff assisted colonoscopy (E) and the AI assisted colonoscopy (A) in the detection of adenomas and polyps, in a randomized, back-to-back study. This study was done in a single tertiary center. Patients were randomly assigned to either the EA group, which underwent endocuff-assisted colonoscopy (E) followed by AI-assisted colonoscopy (A), or the AE group, which underwent AI-assisted colonoscopy followed by endocuff-assisted colonoscopy. All polyps were removed as appropriate during the colonoscopy in which they were detected. Three hundred sixty patients were randomly assigned to either the EA group or the AE group, with 180 patients in each group. The adenoma detection rates (ADR) of the first pass were 64.44% in the EA group and 65% in the AE group, p = 0.912. The median number of adenomas per patient was 2 and the most common size of an adenoma was 1–4 mm in both groups. In the second pass, the adenoma missed rates (ADR of the second pass) were 34.44% in the EA group and 42.78% in the AE group, p = 0.104. The median number of adenomas per patient was 1 and the most common size of an adenoma was 1–4 mm in both groups. There was no complication during the study. There was no significant difference between endocuff assisted colonoscopy and AI assisted colonoscopy in the detection of adenomas. Both systems demonstrated high ADRs but were also associated with high adenoma missed rates. As a high ADRs in this study, the post hoc analysis indicated that the larger sample size may detect a significant difference of ADR between two systems. This study was registered with Thai Clinical Trial Registry; identification number TCTR 20,240,904,003. (https://www.thaiclinicaltrials.org/show/TCTR20240904003).