Influence of prior probability information on large language model performance in radiological diagnosis
Providing information that may influence prior probabilities significantly affects the diagnostic performance of the LLM in radiological cases. This suggests that LLMs may incorporate Bayesian-like principles and adjust the weighting of their diagnostic responses based on prior information, highlighting the potential for optimizing LLM's performance in clinical settings by providing relevant contextual information.
Authors
- Ryo Kurokawa (ORCID: https://orcid.org/0000-0002-1186-8900)
- Akifumi Hagiwara (ORCID: https://orcid.org/0000-0001-5277-3249)
- Osamu Abe (ORCID: https://orcid.org/0000-0002-1180-2629)
- Yusuke Asari
- Mariko Kurokawa (ORCID: https://orcid.org/0000-0002-3907-9188)
- Wataru Gonoi (ORCID: https://orcid.org/0000-0002-5018-4683)
- Jun Kanzawa (ORCID: https://orcid.org/0009-0008-8203-8665)
- Yuki Sonoda (ORCID: https://orcid.org/0009-0005-8622-2433)
- Takahiro Fukushima
Institutions
- The University of Tokyo (JP)
Publication Details
- Journal
- Japanese Journal of Radiology
- Published
- 2025-02-05
- DOI
- https://doi.org/10.1007/s11604-025-01743-3
- Citations
- 4
- Primary Topic
- Topic Modeling
- Type
- article
- Field-Weighted Citation Impact
- 9.92
Funders
- University of Tokyo