To Evaluate the Efficacy of Zero-Shot Prompting Using Large Language Models in the Extraction of Microbial Keratitis Descriptors
Both GPT-4o and GPT-4o mini showed good agreement with human annotations in extracting MK descriptors. Detection of MK descriptors was influenced by limitations in the quality and consistency of electronic health record documentation. The trade-off between efficiency and performance among LLM architectures remains a consideration for implementation of LLM for large-scale MK data analysis.
Authors
- Nambi Nallasamy (ORCID: https://orcid.org/0000-0001-7501-7198)
- Tushar Mungle (ORCID: https://orcid.org/0000-0002-2109-3248)
- Tos T. J. M. Berendschot (ORCID: https://orcid.org/0000-0002-8101-939X)
- Maria A. Woodward (ORCID: https://orcid.org/0000-0001-7060-9004)
- Lokeshwari Aruljyothi (ORCID: https://orcid.org/0000-0002-4882-6288)
- Rudy M.M.A. Nuijts (ORCID: https://orcid.org/0000-0003-2218-9660)
- Namperumalsamy Venkatesh Prajna (ORCID: https://orcid.org/0000-0003-2019-4044)
Institutions
- University of Michigan (US)
- Aravind Eye Hospital (IN)
- Maastricht University (NL)
- Stanford University (US)
Publication Details
- Journal
- Cornea
- Published
- 2025-12-09
- DOI
- https://doi.org/10.1097/ico.0000000000004049
- Citations
- 1
- Primary Topic
- Data-Driven Disease Surveillance
- Type
- article
- Field-Weighted Citation Impact
- 1.35