Data extraction from free-text stroke CT reports using GPT-4o and Llama-3.3-70B: the impact of annotation guidelines
LLMs have utility in data extraction from radiology reports, but the role of annotation guidelines remains underexplored. Data extraction accuracy from stroke CT reports by GPT-4o and Llama-3.3-70B improved when well-defined annotation guidelines were incorporated into the model prompt. Well-defined annotation guidelines can improve the accuracy of LLMs in extracting imaging findings from radiological reports.
Authors
- Jan S. Kirschke (ORCID: https://orcid.org/0000-0002-7557-0003)
- Lisa C. Adams (ORCID: https://orcid.org/0000-0001-5836-4542)
- Keno K. Bressem (ORCID: https://orcid.org/0000-0001-9249-8624)
- Francisco Rodrigues Pinto (ORCID: https://orcid.org/0000-0002-4217-0054)
- Sebastian Ziegelmayer (ORCID: https://orcid.org/0000-0001-8724-4718)
- Severin Schramm (ORCID: https://orcid.org/0000-0002-3299-5861)
- Enrike Rosenkranz
- Dennis M. Hedderich (ORCID: https://orcid.org/0000-0001-8994-5593)
- Su Hwan Kim (ORCID: https://orcid.org/0000-0002-5383-2041)
- Piotr Woźnicki (ORCID: https://orcid.org/0000-0002-9437-3785)
- Cornelius Berberich
- Michael Griessmair (ORCID: https://orcid.org/0009-0007-3518-0792)
- Jonas Wihl
- Claus Zimmer
- Benedikt Wiestler
Institutions
- University of Bern (CH)
- University Hospital of Bern (CH)
- Deutsches Herzzentrum München (DE)
- Technical University of Munich (DE)
Publication Details
- Journal
- European Radiology Experimental
- Published
- 2025-06-19
- DOI
- https://doi.org/10.1186/s41747-025-00600-2
- Citations
- 7
- Primary Topic
- Radiomics and Machine Learning in Medical Imaging
- Type
- article
- Field-Weighted Citation Impact
- 9.52