Performance of Large Language Models in Diagnosing Rare Hematologic Diseases and the Impact of Their Diagnostic Outputs on Physicians: Combined Retrospective and Prospective Study
Without fine-tuning, new-generation commercial LLMs, particularly those with chain-of-thought reasoning, can identify diagnoses of rare hematologic diseases with high accuracy and significantly enhance the diagnostic performance of less-experienced physicians. Nevertheless, biased LLM outputs may mislead clinicians, highlighting the need for critical appraisal and cautious clinical integration with appropriate safeguard systems.
Authors
- Chenlu Yang (ORCID: https://orcid.org/0000-0003-3527-1874)
- Kai Shen (ORCID: https://orcid.org/0000-0002-7772-8011)
- Wenjiao Tang (ORCID: https://orcid.org/0000-0001-5576-1788)
- Qinyu Liu (ORCID: https://orcid.org/0000-0001-9194-8320)
- Hongbin Yu (ORCID: https://orcid.org/0000-0001-9511-9044)
- Xushu Zhong (ORCID: https://orcid.org/0000-0001-5194-1394)
- Chiyi Wang
- Xinmei Yu (ORCID: https://orcid.org/0000-0003-3044-4086)
- Yu Wu (ORCID: https://orcid.org/0000-0001-8708-9711)
- Huanling Zhu (ORCID: https://orcid.org/0000-0002-5226-3828)
- He Li (ORCID: https://orcid.org/0000-0002-9470-7167)
- Tian Chen (ORCID: https://orcid.org/0000-0002-5544-5125)
- Yunfan Yang (ORCID: https://orcid.org/0000-0002-2924-7002)
- Xin Zhang (ORCID: https://orcid.org/0000-0001-7975-9965)
- Xiao Shuai (ORCID: https://orcid.org/0000-0003-0896-2506)
- Y. P. Liao (ORCID: https://orcid.org/0000-0003-4665-4870)
Institutions
- Sichuan University (CN)
- West China Hospital of Sichuan University (CN)
Publication Details
- Journal
- Journal of Medical Internet Research
- Published
- 2025-10-09
- DOI
- https://doi.org/10.2196/77334
- Citations
- 2
- Primary Topic
- Genomics and Rare Diseases
- Type
- article
- Field-Weighted Citation Impact
- 3.63