Kamba Shunsuke, Tamai Naoto, Saitoh Iduru, Matsui Hiroaki, Horiuchi Hideka, Kobayashi Masakuni, Sakamoto Taku, Ego Mai, Fukuda Akihiro, Tonouchi Aya, Shimahara Yuki, Nishikawa Masako, Nishino Haruo, Saito Yutaka, Sumiyama Kazuki
Department of Endoscopy, The Jikei University School of Medicine, 3-25-8 Nishi-Shimbashi, Minato-ku, Tokyo, 105-8461, Japan.
Department of Endoscopy, The Jikei University School of Medicine Third Hospital, 4-11-1 Izumihoncho, Komae-shi, Tokyo, 201-8601, Japan.
J Gastroenterol. 2021 Aug;56(8):746-757. doi: 10.1007/s00535-021-01808-w. Epub 2021 Jul 3.
BACKGROUND: We have developed the computer-aided detection (CADe) system using an original deep learning algorithm based on a convolutional neural network for assisting endoscopists in detecting colorectal lesions during colonoscopy. The aim of this study was to clarify whether adenoma miss rate (AMR) could be reduced with CADe assistance during screening and surveillance colonoscopy. METHODS: This study was a multicenter randomized controlled trial. Patients aged 40 to 80 years who were referred for colorectal screening or surveillance at four sites in Japan were randomly assigned at a 1:1 ratio to either the "standard colonoscopy (SC)-first group" or the "CADe-first group" to undergo a back-to-back tandem procedure. Tandem colonoscopies were performed on the same day for each participant by the same endoscopist in a preassigned order. All polyps detected in each pass were histopathologically diagnosed after biopsy or resection. RESULTS: A total of 358 patients were enrolled and 179 patients were assigned to the SC-first group or CADe-first group. The AMR of the CADe-first group was significantly lower than that of the SC-first group (13.8% vs. 36.7%, P < 0.0001). Similar results were observed for the polyp miss rate (14.2% vs. 40.6%, P < 0.0001) and sessile serrated lesion miss rate (13.0% vs. 38.5%, P = 0.03). The adenoma detection rate of CADe-assisted colonoscopy was 64.5%, which was significantly higher than that of standard colonoscopy (53.6%; P = 0.036). CONCLUSION: Our study results first showed a reduction in the AMR when assisting with CADe based on deep learning in a multicenter randomized controlled trial.
背景：我们开发了一种计算机辅助检测（CADe）系统，该系统使用基于卷积神经网络的原创深度学习算法，以协助内镜医师在结肠镜检查期间检测大肠病变。本研究的目的是阐明在筛查和监测结肠镜检查中，CADe辅助是否可以降低腺瘤漏诊率（AMR）。 方法：本研究是一项多中心随机对照试验。在日本四个地点接受大肠筛查或监测的40至80岁患者按1:1的比例随机分配到“标准结肠镜检查（SC）优先组”或“CADe优先组”，接受背对背的串联检查程序。每位参与者的串联结肠镜检查由同一位内镜医师在同一天按预先指定的顺序进行。每次检查中检测到的所有息肉在活检或切除后进行组织病理学诊断。 结果：共纳入358例患者，179例患者被分配到SC优先组或CADe优先组。CADe优先组的AMR显著低于SC优先组（13.8%对36.7%，P<0.0001）。息肉漏诊率（14.2%对40.6%，P<0.0001）和无蒂锯齿状病变漏诊率（13.0%对38.5%，P=0.03）也观察到类似结果。CADe辅助结肠镜检查的腺瘤检出率为64.5%，显著高于标准结肠镜检查（53.6%；P=0.036）。 结论：我们的研究结果首次表明，在多中心随机对照试验中，基于深度学习的CADe辅助可降低AMR。
Gastroenterology. 2020-8
Lancet Gastroenterol Hepatol. 2020-1-22
Gastroenterology. 2022-7
Life (Basel). 2025-7-14
Ulster Med J. 2025-4
Therap Adv Gastroenterol. 2025-1-31
Zotero 插件