Artificial intelligence-assisted colonoscopy: excellent results, but will mucosal exposure devices push this boundary?

The quality of colonoscopy screening is frequently assessed using the adenoma detection rate (ADR); however, the number of subjects exhibiting high risk features, such as an advanced adenoma or five or more adenomas, at the index colonoscopy holds greater clinical significance, directly influencing surveillance recommendations [1]. In actual clinical settings, colonoscopy performance exhibits significant variation in the ADR and number of subjects with high risk features, potentially leading to inconsistencies in diagnostic accuracy and therapeutic effectiveness. Experience, and advancements in devices and technology can enhance the performance of endoscopists, with adequate bowel preparation serving as the foundation. Studies have indicated that computer-aided detection (CADe) systems and mucosal exposure devices, such as the Endocuff and balloon-based devices, can improve the ADR when the baseline ADR of the endoscopist is between 35% and 45% [2] [3] [4].

”Notably, the ADRs in this study were exceptionally high (almost 60%) compared with other reports of CRC screening performed by experienced endoscopists, including the authors' previous study in the same setting 5 years previously.”

Rocchetto et al. recently investigated the addition of a balloon-based device to colonoscopy, using CADe-assisted colonoscopy alone as the control group [5]. Their study population consisted of individuals with a positive fecal immunochemical test (FIT) at a cutoff value of 20 µg Hb/g feces. Standard parameters, including bowel preparation scores, withdrawal times, and endoscopist experience, were maintained at established levels. The endoscopists involved were highly experienced, with a median of 10 years in practice and a history of performing over 500 procedures annually for the preceding 3 years. Over 640 subjects were randomized into two arms to test the hypothesis of finding an absolute rate of 30% for high risk subjects in the study arm using the balloon device plus CADe. Secondary end points included the ADR, polyp detection rate (PDR), advanced adenoma detection rate (aADR), and other colonoscopy findings. The study revealed no significant difference in the number of subjects with high risk findings between the balloon-assisted CADe arm and the CADe-alone arm (19.3% vs. 23.1%; P = 0.24). Furthermore, the ADRs were remarkably high in both arms, at 59.1% and 55.4% in the control and study arms, respectively.

Notably, the ADRs in this study were exceptionally high (almost 60%) compared with other reports of CRC screening performed by experienced endoscopists, including the authors' previous study in the same setting 5 years previously [6]. In contrast to the high ADRs that were found, the anticipated synergistic effect of the two technologies (the balloon-based device and CADe) reaching a rate of 30% for high risk subjects was not achieved. This discrepancy further emphasizes the possibility of certain effects limiting the observed benefit of a combined approach [6]. As the authors suggested, this may reflect a "ceiling effect," where CADe elevates the ADR to its maximum potential, potentially underestimating the benefit of mucosal exposure devices. They also acknowledged the limitation of their trial in not including an arm using a balloon-based device alone or a standard white-light colonoscopy arm.

Ideally, mucosal exposure devices and CADe should complement each other. CADe assists endoscopists in detecting adenomas within their visual field that might be missed owing to human error. Conversely, mucosal exposure devices can unfold the colon, potentially revealing lesions behind folds that CADe may not detect [2]. The efficacy of different mucosal exposure devices does however vary. Many studies have struggled to demonstrate a clear advantage of one device over another, owing to limitations in study design and, critically, a lack of direct comparisons between mucosal exposure and full-spectrum exposure devices like the Endocuff, balloon-based endoscope, transparent caps, and the full-spectrum endoscopy (FUSE) system. To date, only FUSE has been confirmed as not being superior to standard white-light colonoscopy in terms of the ADR [7], while other devices have shown improved performance by increasing the ADR by 5%–10% compared with standard white-light endoscopy [2] [3] [4]. Unfortunately, direct head-to-head comparisons among these devices remain scarce. Notably, the majority of the additional adenomas detected were small-sized adenomas that may not have a clinical impact [2] [3] [4].

The balloon-based endoscope differs in design, as the endoscopist cannot directly visualize the device on the monitoring screen, and the balloon is usually placed a bit away from the lens when compared with other devices. Its ability to unfold the colon may be limited by its inability to maintain the unfolded position for an extended period, compared with devices like the Endocuff and transparent cap. In addition, the withdrawal time was also noted to be longer than standard colonoscopy [3]. In contrast, the transparent cap and Endocuff showed a shorter cecal intubation time [2] [4]; however, this point is debatable, and direct comparative studies evaluating the efficacy of different mucosal exposure devices are warranted. Moreover, judging the potential for ADR improvement among different mucosal exposure devices is challenging. Meta-analyses suggest that the Endocuff may only be beneficial when the baseline ADR is below 35% [8]. In the study of Rocchetto et al. [5], all endoscopists had already achieved a near-maximum ADR of almost 60% using white light and CADe. Whether the unfolding technique employed by these highly experienced endoscopists was so proficient that mucosal exposure devices offered little additional benefit remains unclear. A head-to-head study among balloon-based colonoscopy, CADe colonoscopy, and their combination, performed by endoscopists with average ADRs, is therefore recommended.

Publication History

Article published online:
08 August 2025

• References

• 1 Hassan C, Antonelli G, Dumonceau JM. et al. Post-polypectomy colonoscopy surveillance: European Society of Gastrointestinal Endoscopy (ESGE) Guideline – Update 2020. Endoscopy 2020; 52: 687-700
• 2 Aniwan S, Mekritthikrai K, Kerr SJ. et al. Computer-aided detection, mucosal exposure device, their combination, and standard colonoscopy for adenoma detection: a randomized controlled trial. Gastrointest Endosc 2023; 97: 507-516
• 3 Haghbin H, Zakirkhodjaev N, Beran A. et al. G-EYE improves polyp, adenoma, and serrated polyp detection rates in colonoscopy: a systematic review and meta-analysis. J Clin Gastroenterol 2024; 58: 668-673
• 4 Chirapongsathorn S, Suzuki S, Prasanwon P. et al. Comparison of colon adenoma detection rate using transparent cap-assisted conventional colonoscopy: result from an international trial in Asia. Dig Dis 2025; 43: 215-224
• 5 Rocchetto S, Rondonotti E, Alicante S. et al. Added value of a balloon-based mucosal exposure device in computer-aided detection-assisted colonoscopy: a randomized controlled trial (COMBAT study). Endoscopy 2025;
• 6 Rondonotti E, Di Paolo D, Rizzotto ER. et al. Efficacy of a computer-aided detection system in a fecal immunochemical test-based organized colorectal cancer screening program: a randomized controlled trial (AIFIT study). Endoscopy 2022; 54: 1171-1179
• 7 Hassan C, Senore C, Radaelli F. et al. Full-spectrum (FUSE) versus standard forward-viewing colonoscopy in an organised colorectal cancer screening programme. Gut 2017; 66: 1949-1955
• 8 Williet N, Tournier Q, Vernet C. et al. Effect of Endocuff-assisted colonoscopy on adenoma detection rate: meta-analysis of randomized controlled trials. Endoscopy 2018; 50: 846-860