Habeeb Amir, Lim Kim Hui, Kochilas Xenofon, Bhat Nazir, Amen Furrat, Chan Samuel
Academic Clinical Fellow Association, Queen Mary University of London, London, UK.
Ear, Nose and Throat Surgery Department, Peterborough City Hospital, Peterborough, UK.
Clin Otolaryngol. 2025 Jul;50(4):769-774. doi: 10.1111/coa.14305. Epub 2025 Mar 20.
ChatGPT is one of the most publicly available artificial intelligence (AI) softwares. Ear, nose and throat (ENT) services are often stretched due to the increasing incidence of thyroid malignancies. This study aims to investigate whether there is a role for AI software in providing accurate thyroid multidisciplinary team (MDT) outcomes.
A retrospective study looking at unique thyroid MDT outcomes between October 2023 and May 2024. ChatGPT-4TM was used to generate outcomes based on the British Thyroid Association (BTA) Guidelines for Management of Thyroid Cancer. Concordance levels were collected and analysed.
Thirty thyroid cases with a mean age of 58 (n = 24 female, n = 6 male) were discussed. The MDT's outcome had a 100% concordance with BTA Guidelines. When comparing ChatGPT-4TM and our MDT the highest level of concordance Y1 was seen in 67% of case while 13% of cases were completely discordant.
CONCLUSIONS/SIGNIFICANCE: AI is cheap, easy to use can optimise complex thyroid MDT decision making. This could free some clinicians allowing them to meet other demands of the ENT service. Some key issues are the inability to completely rely on the AI software for outcomes without being counterchecked by a clinician.
ChatGPT是最容易获取的人工智能(AI)软件之一。由于甲状腺恶性肿瘤发病率不断上升,耳鼻喉科(ENT)服务常常不堪重负。本研究旨在调查AI软件在提供准确的甲状腺多学科团队(MDT)结果方面是否能发挥作用。
一项回顾性研究,观察2023年10月至2024年5月期间独特的甲状腺MDT结果。使用ChatGPT-4TM根据英国甲状腺协会(BTA)甲状腺癌管理指南生成结果。收集并分析一致性水平。
讨论了30例甲状腺病例,平均年龄58岁(n = 24名女性,n = 6名男性)。MDT的结果与BTA指南的一致性为100%。将ChatGPT-4TM与我们的MDT进行比较时,67%的病例一致性水平最高为Y1,而13%的病例完全不一致。
结论/意义:AI价格低廉、易于使用,可以优化复杂的甲状腺MDT决策。这可以使一些临床医生从相关工作中解脱出来,去满足耳鼻喉科服务的其他需求。一些关键问题是,在未经临床医生核对的情况下,无法完全依赖AI软件得出的结果。