Multimodal GPT model for assisting thyroid nodule diagnosis and management
Although using artificial intelligence (AI) to analyze ultrasound images is a promising approach to assessing thyroid nodule risks, traditional AI models lack transparency and interpretability. We developed a multimodal generative pre-trained transformer for thyroid nodules (ThyGPT), aiming to provide a transparent and interpretable AI copilot model for thyroid nodule risk assessment and management. Ultrasound data from 59,406 patients across nine hospitals were retrospectively collected to train and test the model. After training, ThyGPT was found to assist in reducing biopsy rates by more than 40% without increasing missed diagnoses. In addition, it detects errors in ultrasound reports 1,610 times faster than humans. With the assistance of ThyGPT, the area under the curve for radiologists in assessing thyroid nodule risks improved from 0.805 to 0.908 (p < 0.001). As an AI-generated content-enhanced computer-aided diagnosis (AIGC-CAD) model, ThyGPT has the potential to revolutionize how radiologists use such tools.
Authors
- Zhikai Lei
- Xiang Hao (ORCID: https://orcid.org/0000-0002-3931-6884)
- Fajin Dong (ORCID: https://orcid.org/0000-0002-4558-4885)
- Yahan Zhou (ORCID: https://orcid.org/0000-0001-5808-2801)
- Rongrong Ru (ORCID: https://orcid.org/0000-0002-4249-3152)
- Liping Wang (ORCID: https://orcid.org/0000-0003-2654-2337)
- Yidan Lu (ORCID: https://orcid.org/0000-0001-7476-2911)
- Jianhua Zhou (ORCID: https://orcid.org/0000-0003-2096-8126)
- Di Ou (ORCID: https://orcid.org/0000-0002-6987-7180)
- Shanshan Zhao (ORCID: https://orcid.org/0000-0002-9308-5562)
- Bojian Feng (ORCID: https://orcid.org/0000-0002-4559-0775)
- Yuan Gao (ORCID: https://orcid.org/0000-0002-0832-2982)
- Na Feng (ORCID: https://orcid.org/0000-0001-6205-1960)
- Dong Xu (ORCID: https://orcid.org/0000-0002-9451-1754)
- Yunpeng Wang (ORCID: https://orcid.org/0000-0002-0183-8345)
- Yang Chen (ORCID: https://orcid.org/0000-0003-0164-7079)
- Chen Chen (ORCID: https://orcid.org/0000-0003-2104-534X)
- Yanming Zhang (ORCID: https://orcid.org/0000-0002-1391-9572)
- Yaqing Chen (ORCID: https://orcid.org/0000-0002-0566-6130)
- Jincao Yao (ORCID: https://orcid.org/0000-0003-1543-6010)
- Ping Liang (ORCID: https://orcid.org/0000-0003-4292-5290)
- Xiaoxian Li (ORCID: https://orcid.org/0000-0002-0995-1721)
- Zhengping Wang (ORCID: https://orcid.org/0000-0001-5928-3816)
- Yuqi Yan (ORCID: https://orcid.org/0009-0002-9071-2441)
- Jiafei Shen
- Li‐Yu Chen (ORCID: https://orcid.org/0000-0002-5199-6237)
- Vicky Yang Wang
- Kai Wang (ORCID: https://orcid.org/0009-0004-6559-2585)
- Wei Li (ORCID: https://orcid.org/0009-0008-7802-1828)
Institutions
- Sun Yat-sen University (CN)
- Jinan University (CN)
- Chinese Academy of Sciences (CN)
- Wenzhou Medical University (CN)
- Zhejiang Provincial Hospital of TCM (CN)
- Zhejiang Cancer Hospital (CN)
- Affiliated Hospital of Hangzhou Normal University (CN)
- Hangzhou Academy of Agricultural Sciences (CN)
- Taizhou Central Hospital (CN)
- Sun Yat-sen University Cancer Center (CN)
- Dongyang People's Hospital (CN)
- Zhejiang Taizhou Hospital (CN)
- Artificial Intelligence in Medicine (Canada) (CA)
- Hangzhou Cancer Hospital (CN)
- Zhejiang University (CN)
Publication Details
- Journal
- npj Digital Medicine
- Published
- 2025-05-03
- DOI
- https://doi.org/10.1038/s41746-025-01652-9
- Citations
- 31
- Primary Topic
- Radiomics and Machine Learning in Medical Imaging
- Type
- article
- Field-Weighted Citation Impact
- 42.16
Funders
- National Natural Science Foundation of China
- National Key Research and Development Program of China