Large Language Models with Vision on Diagnostic Radiology Board Exam Style Questions
Vision-capable large language models cannot effectively use images to increase performance on radiology board-style examination questions. When using textual data alone, Claude 3.5 Sonnet outperforms GPT-4V and Gemini 1.5 Pro, highlighting the advancements in the field and its potential for use in further research.
Authors
- Roozbeh Houshyar (ORCID: https://orcid.org/0000-0003-4263-5533)
- Shawn Sun (ORCID: https://orcid.org/0000-0002-6170-1949)
- Kasha Chen
- Mark D. Tran
- Gillean Cortes (ORCID: https://orcid.org/0000-0003-0387-1736)
- Vahid Yaghmai (ORCID: https://orcid.org/0000-0002-1440-3332)
- Kenneth N. Huynh (ORCID: https://orcid.org/0000-0002-3179-3424)
- Julia Tran (ORCID: https://orcid.org/0000-0002-9660-0080)
- Leslie Yeh
- Samuel Anavim
- Michael Phillipi (ORCID: https://orcid.org/0009-0007-2657-2989)
Institutions
- University of California, Irvine Medical Center (US)
Publication Details
- Journal
- Academic Radiology
- Published
- 2024-12-04
- DOI
- https://doi.org/10.1016/j.acra.2024.11.028
- Citations
- 20
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 2.41