{
  "abstract": "Introduction The detection and resection of adenomatous polyps during colonoscopy decrease the incidence of colorectal cancer and associated mortality, with adenoma detection rate (ADR) being a key quality performance indicator. While higher ADR is associated with lower post-colonoscopy colorectal cancer (PCCRC) rates, the latter varies significantly across the UK, and the true ‘baseline’ ADR is unknown. Meeting the rising demand for high quality colonoscopy is a significant challenge.The use of computer assisted detection (CADe) is associated with a significant improvement in ADR across multiple clinical trials, but real-world performance data is lacking.We sought to evaluate effect on adenoma detection of implementing CADe in a ‘real-world’ setting, in the largest such study to date.Method This was randomised; stepped-wedge, open-label trial evaluated the site-level effect of the GI Genius™ (GIG) computer-aided detection (CADe) system on adenoma detection. Both District General Hospitals and Teaching Hospitals participated, with procedures performed by expert (≥2,000 lifetime colonoscopies; PDR > 40%) and non-expert (<2,000 lifetime colonoscopies) endoscopists.All patients undergoing routine colonoscopy were eligible, excluding those attending for polyp or IBD surveillance. The study reflected real-world practice, with no procedural modifications other than CADe activation.Data were collected over three consecutive 3-month phases: Phase 1 (baseline; CADe inactive), Phase 2 (intervention; CADe active), and Phase 3 (post-intervention; CADe withdrawn).The primary outcome, adenoma detection rate (ADR), was analysed using generalised linear mixed-effects models, with study phase, patient age, endoscopist gender and expertise, bowel preparation quality, hospital type, and withdrawal time as fixed effects, and site and endoscopist as random effects to account for clustering.Results Between November 2023 and February 2025, 4,514 colonoscopies from 25 hospitals (England, Scotland, Wales) were analysed. ADR increased from 30.0% in Phase 1 to 34.8% during the GI Genius intervention (Phase 2) and declined to 27.6% post-intervention (p<0.001). After adjustment, Phase 2 was associated with a 17% higher ADR (aOR 1.17, 95% CI 1.01–1.35, p = 0.040). Older age, endoscopist expertise, male gender, and withdrawal time independently predicted higher ADR.Conclusion In this large, pragmatic trial, CADe use during colonoscopy was associated with a modest but clinically relevant improvement in adenoma detection, driven primarily by increased detection of small adenomas. The absence of sustained improvement following AI withdrawal highlights the importance of positioning CADe as a supportive technology embedded within high-quality colonoscopy practice rather than as a substitute for training or technique.Funding Funded by DHSC/NIHR with logistical support from Medtronic",
  "authors": [
    {
      "affiliations": [
        "King’s College Hospital, London, United Kingdom"
      ],
      "name": "Olaolu Olabintan"
    },
    {
      "affiliations": [
        "Community Health Sciences (Biostatistics) University of Calgary, Canada"
      ],
      "name": "Olayinka Arimoro"
    },
    {
      "affiliations": [
        "King’s College Hospital, London, United Kingdom"
      ],
      "name": "Alena Marynina"
    },
    {
      "affiliations": [
        "King’s College Hospital, London, United Kingdom"
      ],
      "name": "Mehul Patel"
    },
    {
      "affiliations": [
        "King’s College Hospital, London, United Kingdom"
      ],
      "name": "Andrew Emmanuel"
    },
    {
      "affiliations": [
        "King’s College Hospital, London, United Kingdom"
      ],
      "name": "Amyn Haji"
    },
    {
      "affiliations": [
        "King’s College Hospital, London, United Kingdom"
      ],
      "name": "Bu’Hussain Hayee"
    }
  ],
  "title": "P3 A Nationwide study of artificial intelligence in adenoma detection during colonoscopy - final result from the NAIAD trial",
  "uid": "d2bf2169-2faa-553c-a079-01660446d2ee"
}