{
  "abstract": "Commercial, regulatory-approved computer-aided detection (CADe) tools are widely used in colonoscopy; however, open-source multimodal large language models may present a promising alternative. This study compared the polyp detection performance of two such models—GPT-4o and Gemini 1.5 Pro—with ENDO-AID CADe, a commercial artificial intelligence (AI) system, using colonoscopy videos. Per-lesion sensitivities were 75% for GPT-4o, 50% for Gemini and 87% for ENDO-AID. While the large language models were less effective at detecting polyps ≤5 mm, they performed comparably for larger lesions, suggesting their potential as alternative AI tools in endoscopic practice.",
  "authors": [
    {
      "affiliations": [
        "Dipartimento di Elettronica, Informatica e Bioingegneria (DEIB), Politecnico di Milano, Milan, Italy"
      ],
      "name": "Luca Carlini"
    },
    {
      "affiliations": [
        "IRCCS Humanitas Research Hospital, Rozzano, Milan, Italy"
      ],
      "name": "Davide Massimi"
    },
    {
      "affiliations": [
        "Digestive Disease Center, Showa University Northern Yokohama Hospital, Yokohama, Kanagawa, Japan",
        "Clinical Effectiveness Research Group, University of Oslo, Oslo, Norway"
      ],
      "name": "Yuichi Mori"
    },
    {
      "affiliations": [
        "Gastroenterology and Digestive Endoscopy Unit, Ospedale Nuovo Regina Margherita, Roma, Italy"
      ],
      "name": "Giulio Antonelli"
    },
    {
      "affiliations": [
        "IRCCS Humanitas Research Hospital, Rozzano, Milan, Italy"
      ],
      "name": "Tommy Rizkala"
    },
    {
      "affiliations": [
        "IRCCS Humanitas Research Hospital, Rozzano, Milan, Italy"
      ],
      "name": "Marco Spadaccini"
    },
    {
      "affiliations": [
        "Dipartimento di Elettronica, Informatica e Bioingegneria (DEIB), Politecnico di Milano, Milan, Italy"
      ],
      "name": "Chiara Lena"
    },
    {
      "affiliations": [
        "Department of Gastroenterology, Swedish Medical Center, Seattle, Washington, USA"
      ],
      "name": "Sravanthi Parasa"
    },
    {
      "affiliations": [
        "Department of Gastroenterology and Hepatology, Uz Leuven, TARGID Kuleuven, Leuven, Belgium"
      ],
      "name": "Raf Bisschops"
    },
    {
      "affiliations": [
        "University of Montreal Hospital Centre, Montreal, Quebec, Canada",
        "Division of Gastroenterology, Montreal University Hospital Center (CHUM), Montreal, Quebec, Canada"
      ],
      "name": "Daniel von Renteln"
    },
    {
      "affiliations": [
        "Gastroenterology, St Vincent's Healthcare Group Ltd, Dublin, Ireland"
      ],
      "name": "Susanne O’Reilly"
    },
    {
      "affiliations": [
        "University of Kansas School of Medicine and VA Medical Center, Kansas City, Kansas, USA"
      ],
      "name": "Prateek Sharma"
    },
    {
      "affiliations": [
        "Department of Medicine, Indiana University School of Medicine, Indianapolis, Indiana, USA"
      ],
      "name": "Douglas K Rex"
    },
    {
      "affiliations": [
        "Clinical Effectiveness Research Group, University of Oslo, Oslo, Norway"
      ],
      "name": "Michael Bretthauer"
    },
    {
      "affiliations": [
        "IRCCS Humanitas Research Hospital, Rozzano, Milan, Italy",
        "Department of Biomedical Sciences, Humanitas University, Pieve Emanuele, Italy"
      ],
      "name": "Alessandro Repici"
    },
    {
      "affiliations": [
        "Dipartimento di Elettronica, Informatica e Bioingegneria (DEIB), Politecnico di Milano, Milan, Italy"
      ],
      "name": "Elena De Momi"
    },
    {
      "affiliations": [
        "IRCCS Humanitas Research Hospital, Rozzano, Milan, Italy",
        "Department of Biomedical Sciences, Humanitas University, Pieve Emanuele, Italy"
      ],
      "name": "Cesare Hassan"
    }
  ],
  "title": "Large language models for detecting colorectal polyps in endoscopic images",
  "uid": "739c2510-75a7-5b84-8134-e369fa2ccd07"
}