Clinical Evaluation of Computer-Aided Colorectal Neoplasia Detection Using a Novel Endoscopic Artificial Intelligence: A Single-Center Randomized Controlled Trial

Abstract

<b><i>Introduction:</i></b> Computer-aided diagnostic systems are emerging in the field of gastrointestinal endoscopy. In this study, we assessed the clinical performance of the computer-aided detection (CADe) of colonic adenomas using a new endoscopic artificial intelligence system. <b><i>Methods:</i></b> This was a single-center prospective randomized study including 415 participants allocated into the CADe group (<i>n</i> = 207) and control group (<i>n</i> = 208). All endoscopic examinations were performed by experienced endoscopists. The performance of the CADe was assessed based on the adenoma detection rate (ADR). Additionally, we compared the adenoma miss rate for the rectosigmoid colon (AMRrs) between the groups. <b><i>Results:</i></b> The basic demographic and procedural characteristics of the CADe and control groups were as follows: mean age, 54.9 and 55.9 years; male sex, 73.9% and 69.7% of participants; and mean withdrawal time, 411.8 and 399.0 s, respectively. The ADR was 59.4% in the CADe group and 47.6% in the control group (<i>p</i> = 0.018). The AMRrs was 11.9% in the CADe group and 26.0% in the control group (<i>p</i> = 0.037). <b><i>Conclusion:</i></b> The colonoscopy with the CADe system yielded an 11.8% higher ADR than that performed by experienced endoscopists alone. Moreover, there was no need to extend the examination time or request the assistance of additional medical staff to achieve this improved effectiveness. We believe that the novel CADe system can lead to considerable advances in colorectal cancer diagnosis.

Advancing Colorectal Cancer Screening: A Comprehensive Systematic Review of Artificial Intelligence (AI)-Assisted Versus Routine Colonoscopy

“…The trials were published from 2019 to 2023. Among the 13 studies, eight were conducted in China, two in Japan, one in France, one in Italy, and one in the USA; three out of 13 trials were non-blinded [8,14,15], eight trials were single-blinded RCTs [16,17,19,20,[22][23][24][25], and two were double-blinded RCTs [18,21]. Table 3 provides a detailed description of each study.…”

“… Studies - Wang et al, 2019 [ 14 ], Liu et al, 2020 [ 15 ], Repici et al, 2020 [ 8 ], Gong et al, 2020 [ 16 ], Su et al, 2020 [ 17 ], Wang et al, 2020 [ 18 ], Wang et al, 2020 [ 19 ], Kamba et al, 2021 [ 20 ], Yao et al, 2022 [ 21 ], Glissen Brown et al, 2022 [ 22 ], Xu et al, 2023 [ 23 ], Nakashima et al, 2023 [ 24 ], Karsenti et al, 2023 [ 25 ] …”

“…In 2023, Nakashima et al [ 24 ] also conducted a single-center, single-blind prospective randomized study in Japan to assess the clinical performance of CADe to detect colonic adenomas compared to the control group. A total of 415 patients underwent colonoscopies, randomly assigned to either the CADe group or the control group.…”

The Role of Artificial Intelligence in Colorectal Cancer Screening: Lesion Detection and Lesion Characterization

“…Since then, there has been a vast amount of research published on real-time CADe systems, with strong support for their efficacy in polyp detection. Of the 15 RCTs reviewed here, 10 demonstrated a statistically significant increase in adenoma detection, although baseline and CADe adenoma detection rates (ADRs) are highly varied because of differing populations and study designs (Table 2) [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24]. Although overall lesion detection is generally improved, many of these systems have been criticised for a lack of impact on the detection of advanced adenomas of heightened clinical significance.…”

Computer‐aided detection in colonoscopic practice: Ready for prime time?

“…To my knowledge, in Japan, there is only one randomized controlled trial (RCT) using the CAD system developed by LPIXEL and Jikei University to date 3 . As for the CAD‐EYE system, two RCTs 4,5 have been published in the past 2 years, both successfully demonstrating a significantly higher adenoma detection rate (ADR) as the primary outcome. In contrast, Yamaguchi et al 6 .…”