{
  "abstract": "Background Colorectal Cancer (CRC) is one of the most common cancer and early detection by screening is critical for treatment. Previous studies indicate the effect of Computer Aided Detection (CADe) is significant in mixed settings, but its effect under screening settings remains unclear.Methods We systematically searched PubMed, Embase, Web of Science, Scopus, and the Cochrane Library for randomized controlled trials comparing AI-assisted colonoscopy with conventional colonoscopy in screening populations. Randomized Clinical Trails (RCTs) comparing the effect of CADe with conventional colonoscopy under screening settings or involving asymptomatic subjects and report Adenomda Dectection Rate (ADR) of those subjects were included in this meta-analysis. Risk of bias was assessed using a RoB2 analysis. Random-effects meta-analysis was performed to pool Risk Ratio (RR) with 95% confidence intervals.Results The initial screening included 495 articles from PubMed, Embase, Web of Science, Cochrane Library, and Scopus. Out of them, there were 267 unique articles, with 58 of them were recognized to undergo a full text review. A total of 8 studies involving 7598 subjects were included in this meta-analysis. CADe significantly improved ADR compared with conventional colonoscopy (RR = 1.23, 95% CI: 1.16–1.29) with low between-study heterogeneity (I 2 = 2.5%). By subgroup, the pooled results for pure screening setting RCTs show significant improvement in ADR (RR=1.25, 95% CI: 1.18-1.32). The pooled result for the screening population from mixed RCTs that include screening, diagnostic, and surveillance settings is insignificant (RR=1.08, 95% CI: 0.92-1.26). A funnel plot examination indicates that there is no recognized publication bias.Conclusions CADe significantly improves ADR under a screening setting. Such improvement may be caused by AI being trained with sufficient data of different types of adenoma, enabling CADe to perform better in screening settings compared to conventional coloscopy. This meta-analysis suggests that it could be effective to widely implant CADe in screening programs. This meta-analysis consists of a limited number of RCTs, and to reach a more generalizable result, further high-quality trials are needed to confirm these findings.",
  "authors": [
    {
      "affiliations": [
        "JC School of Public Health and Primary Care, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Junsheng Yang"
    },
    {
      "affiliations": [
        "JC School of Public Health and Primary Care, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Claire Chenwen Zhong"
    },
    {
      "affiliations": [
        "JC School of Public Health and Primary Care, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Martin CS Wong"
    },
    {
      "affiliations": [
        "JC School of Public Health and Primary Care, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Jiayu Yi"
    },
    {
      "affiliations": [
        "JC School of Public Health and Primary Care, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong"
      ],
      "name": "Junjie Huang"
    }
  ],
  "title": "IDDF2026-ABS-0313 Effectiveness of AI-assisted colonoscopy in colorectal cancer screening: a systematic review and meta analysis",
  "uid": "6ed6b717-ba2f-5404-8678-e70af5eee6d1"
}