Lopez-Gonzalez Ruth, Sanchez-Cordero Sergi, Pujol-Gebellí Jordi, Castellvi Jordi
General and Digestive Surgery, Moises Broggi University Hospital, C Oriol Martorell 12, 08970, Barcelona, Spain.
Obes Surg. 2025 Jan;35(1):19-24. doi: 10.1007/s11695-024-07279-1. Epub 2024 May 17.
With the growing interest in artificial intelligence (AI) applications in medicine, this study explores ChatGPT's potential to influence surgical technique selection in metabolic and bariatric surgery (MBS), contrasting AI recommendations with established clinical guidelines and expert consensus.
Conducting a single-center retrospective analysis, the study involved 161 patients who underwent MBS between January 2022 and December 2023. ChatGPT4 was used to analyze patient data, including demographics, pathological history, and BMI, to recommend the most suitable surgical technique. These AI recommendations were then compared with the hospital's algorithm-based decisions.
ChatGPT recommended Roux-en-Y gastric bypass in over half of the cases. However, a significant difference was observed between AI suggestions and actual surgical techniques applied, with only a 34.16% match rate. Further analysis revealed any significant correlation between ChatGPT recommendations and the established surgical algorithm.
Despite ChatGPT's ability to process and analyze large datasets, its recommendations for MBS techniques do not align closely with those determined by expert surgical teams using a high success rate algorithm. Consequently, the study concludes that ChatGPT4 should not replace expert consultation in selecting MBS techniques.
随着医学领域对人工智能（AI）应用的兴趣日益浓厚，本研究探讨了ChatGPT在代谢和减重手术（MBS）中影响手术技术选择的潜力，并将人工智能的建议与既定的临床指南和专家共识进行对比。
本研究采用单中心回顾性分析，纳入了2022年1月至2023年12月期间接受MBS的161例患者。使用ChatGPT4分析患者数据，包括人口统计学、病史和体重指数，以推荐最合适的手术技术。然后将这些人工智能的建议与医院基于算法的决策进行比较。
ChatGPT在超过半数的病例中推荐了Roux-en-Y胃旁路术。然而，人工智能的建议与实际应用的手术技术之间存在显著差异，匹配率仅为34.16%。进一步分析显示，ChatGPT的建议与既定的手术算法之间没有显著相关性。
尽管ChatGPT有处理和分析大量数据集的能力，但其对MBS技术的建议与使用高成功率算法的专家手术团队所确定的建议并不紧密一致。因此，该研究得出结论，在选择MBS技术时，ChatGPT4不应取代专家咨询。
