EN
TÜRKÇE
Poster Display - 11
The challenges and possibilities of the use of ChatGPT-4o in pediatric surgery and urology
Jan Trachta, Lucie Poš
Department of Pediatric Surgery, 2nd Faculty of Medicine, Charles University and Motol University Hospital, Prague, Czech Republic
Introduction
The advent of multimodal large language models (LLMs) such as ChatGPT-4o presents novel opportunities for the integration of artificial intelligence (AI) into pediatric surgical practice. ChatGPT-4o can generate human-like text and interact in real time, offering potential support across clinical, educational, and administrative domains.
Methodology
This state-of-the-art presentation is based on a targeted literature search and personal experience with ChatGPT-4o Plus. The primary research tool was the PubMed database using the Boolean operator: ("ChatGPT" OR "GPT-4" OR "GPT-4o" OR "large language model" OR "generative AI") AND ("pediatric surgery" OR "pediatric surgery" OR "pediatric urology" OR "pediatric urology"). This strategy identified 43 articles published between 2023 and 2025; 12 were excluded after abstract review due to insufficient relevance. The remaining 31 articles were analyzed in detail. Key references within these articles led to the inclusion of 14 additional publications.
Results
Clinically, ChatGPT-4o may support patient triage, differential diagnosis brainstorming, and treatment planning. It can generate structured summaries from symptoms, lab results, and imaging reports or draft operation notes from brief physician input. To address GDPR concerns and prevent data sharing with private companies (such as OpenAI), offline open-source LLMs (e.g. LLaMA3) may be used for integration into electronic health records. In communication, ChatGPT-4o may help explain complex information to children and families in age-appropriate, understandable language. It can assist in informed consent and preoperative counselling reducing caregiver anxiety. In education, it offers personalized tools for students and trainees, including quizzes, simulated interviews, and adaptive explanations of surgical knowledge and procedures. ChatGPT-4o's use is limited by factual inaccuracies, inability to interpret imaging, lack of GDPR compliance, and ethical concerns including bias, transparency, and accountability.
Conclusion
ChatGPT-4o can support pediatric surgical practice and clinical reasoning, though full responsibility remains with the surgeon.