文献速递消化内镜消化道肿瘤与早癌

首款美国FDA批准的结肠镜计算机辅助质量评估工具的前瞻性验证

Prospective Validation of the First US FDA-Approved Computer-Aided Quality Assessment Tool for Colonoscopy: An Initial Clinical Experience.
PubMed 文献 · Am J Gastroenterol · 2025-11-19

该前瞻性先导研究在110例门诊结肠镜检查中,将GI Genius ColonPro 4.0(首款获美国FDA批准的结肠镜计算机辅助质量平台)测量的退镜时间、盲肠插管和波士顿肠道准备评分(BBPS)与独立视频复核进行比较。结果显示,计算机辅助质量评估与医师在BBPS上一致性差(ICC=0.40,P<0.001),但在退镜时间(ICC=0.89,P<0.001)和盲肠插管(κ=0.79,P<0.001)上一致性强。结论认为该平台可可靠测量退镜时间和盲肠插管,但基于人工智能的BBPS评分在整合入现有结肠镜质量范式前需进一步验证。

要点

  • ColonPro 4.0与医师退镜时间一致性高(ICC=0.89)。
  • 盲肠插管一致性也强(κ=0.79)。
  • BBPS评分一致性差(ICC=0.40),需进一步验证。

核心结论

  • 计算机辅助质量评估可可靠测量退镜时间和盲肠插管,但人工智能BBPS评分需进一步验证后才能整合入现有结肠镜质量范式。

研究设计与样本

  • 前瞻性先导研究,在110例门诊结肠镜检查中比较ColonPro 4.0测量值与独立视频复核。

主要发现

  • BBPS的组内相关系数为0.40(P<0.001),一致性差。
  • 退镜时间的组内相关系数为0.89(P<0.001),一致性强。
  • 盲肠插管的κ值为0.79(P<0.001),一致性强。

临床怎么用

  • 可将该工具用于退镜时间和盲肠插管的客观质量监测,BBPS评分暂不宜直接用于临床质量评价。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

INTRODUCTION: We report a prospective pilot of GI Genius ColonPro 4.0 (Medtronic, Minneapolis, MN), the first US Food and Drug Administration-approved computer-aided quality (CAQ) platform for colonoscopy. METHODS: ColonPro measurements of withdrawal time, cecal intubation, and Boston Bowel Preparation Score (BBPS) were compared with independent video review in 110 outpatient colonoscopies. RESULTS: Agreement between the CAQ and physicians was poor for BBPS (intraclass correlation coefficient = 0.40, P < 0.001) but strong for withdrawal time (intraclass correlation coefficient = 0.89, P < 0.001) and cecal intubation (κ = 0.79, P < 0.001). DISCUSSION: CAQ reliably measures withdrawal time and cecal intubation, but artificial intelligence-based BBPS scoring requires further validation before integration into current paradigms of colonoscopy quality.

来源:Am J Gastroenterol,查看原文。版权归原作者所有。

引用

  1. Brenner TA, Labaki C, Feuerstein JD, et al. Prospective Validation of the First US FDA-Approved Computer-Aided Quality Assessment Tool for Colonoscopy: An Initial Clinical Experience. Am J Gastroenterol. 2025. PMID: 41257531. doi:10.14309/ajg.0000000000003855

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。