{
  "abstract": "Introduction Conversational artificial intelligence(AI)models are increasingly used as clinical decision-support tools.This study aimed to compare ChatGPT(GPT-5) and Gemini(Google)in their ability to provide safe, justified,and guideline-compliant therapeutic strategies for esophageal achalasia and diverticula,with particular emphasis on the indication for Per Oral Endoscopic Myotomy(POEM).Methods Twenty-two simulated clinical cases were developed:18 achalasia cases(types I,II,III, atypical, post-radiotherapy, post-dilation, or post-POEM)and 4 esophageal diverticula cases (epiphrenic and Zenker, symptomatic and asymptomatic).Cases included complex clinical scenarios such as cardiac or neurological comorbidities,megaesophagus, pre-existing reflux, or post-treatment recurrence.Each case was independently submitted to both AI models using identical phrasing. The primary question was:‘Is POEM indicated for this patient?Justify your answer.’Responses were assessed by a clinical expert using five criteria:therapeutic relevance, clinical justification,safety,clarity, and adherence to guidelines.Scores ranged from 0–8 points. Qualitative analyses evaluated response structure,consistency,guideline adherence, and patient safety considerations.Results ChatGPT achieved a higher overall performance score than Gemini (96%vs89.8%).Both models demonstrated similar therapeutic relevance(90.9%) and clarity(100%).ChatGPT showed superior performance in safety (100%vs 95.5%) and explicit citation of international guidelines (86%vs 27%).Notable qualitative differences emerged.In elderly patients with megaesophagus and significant cardiac comorbidities,ChatGPT favored conservative or staged approaches (botulinum toxin injection or pneumatic dilation)before considering POEM, in line with guideline-based risk stratification.In contrast, Gemini more frequently proposed POEM directly,with limited reference to guideline-based precautions.In post-fundoplication achalasia recurrence,ChatGPT emphasized anatomical complexity and alternative therapies, whereas Gemini focused on the technical feasibility of POEM. ChatGPT consistently adhered to international guidelines and favored structured, risk-adapted, and less invasive strategies,whereas Gemini provided more detailed educational explanations but was less concise and occasionally favored more aggressive interventions. ChatGPT grounded its recommendations in explicit links between physiology,risks, and expected outcomes, while Gemini relied primarily on technical and pragmatic reasoning.Conclusions Both AI models generated clinically relevant and generally safe recommendations. However, ChatGPT demonstrated superior guideline adherence, structured reasoning, and risk-sensitive therapeutic decision-making.Gemini offered stronger pedagogical and technical explanations. Their complementary profiles suggest potential benefit in combined use for clinical decision support and medical education in esophageal motility disorders.",
  "authors": [
    {
      "affiliations": [
        "Department of Gastroenterology, Mohamed V Military Teaching Hospital, Rabat, Morocco"
      ],
      "name": "Docteur Meriem Amine"
    },
    {
      "affiliations": [
        "Department of Gastroenterology, Mohamed V Military Teaching Hospital, Rabat, Morocco"
      ],
      "name": "Samir Mrabti"
    },
    {
      "affiliations": [
        "Department of Gastroenterology, Mohamed V Military Teaching Hospital, Rabat, Morocco"
      ],
      "name": "Ahlame Benhamdane"
    },
    {
      "affiliations": [
        "Department of Gastroenterology, Mohamed V Military Teaching Hospital, Rabat, Morocco"
      ],
      "name": "Salma Azammam"
    },
    {
      "affiliations": [
        "Department of Gastroenterology, Mohamed V Military Teaching Hospital, Rabat, Morocco"
      ],
      "name": "Tarik Addajou"
    },
    {
      "affiliations": [
        "Department of Gastroenterology, Mohamed V Military Teaching Hospital, Rabat, Morocco"
      ],
      "name": "Ilham Elkoti"
    },
    {
      "affiliations": [
        "Department of Gastroenterology, Mohamed V Military Teaching Hospital, Rabat, Morocco"
      ],
      "name": "Fedoua Rouibaa"
    },
    {
      "affiliations": [
        "Department of Gastroenterology, Mohamed V Military Teaching Hospital, Rabat, Morocco"
      ],
      "name": "Hassan Seddik"
    }
  ],
  "title": "P416 Achalasia and esophageal diverticula: comparative performance of chatGPT-5 and Gemini in therapeutic decision-making",
  "uid": "c3ff308f-fdbc-5b07-9ce8-b6d327b3a815"
}