Privacy-ensuring Open-weights Large Language Models Are Competitive with Closed-weights GPT-4o in Extracting Chest Radiography Findings from Free-Text Reports
Authors:
- Sebastian NowakSNSebastian Nowak
- Yannik C. LayerYLYannik C. Layer
- Maike TheisMTMaike Theis
- Alexander IsaakAIAlexander Isaak
- Wolfgang BlockWBWolfgang Block
- Daniel KuettingDKDaniel Kuetting
- Claus C. PieperCPClaus C. Pieper
- Julian A. LuetkensJLJulian A. Luetkens
- Ulrike AttenbergerUAUlrike Attenberger
- Alois M. SprinkartASAlois M. Sprinkart
Top Cited Papers
- 1 January 2025
- journal article
- research article
- Published by Radiological Society of North America (RSNA) in Radiology
- Vol. 314 (1) , e240895
Abstract
Privacy-ensuring, open-weights large language models showed great potential in the extraction of structured content from free-text radiology reports, facilitating the secondary use of clinical databases for applications in data-driven medicine.Keywords
This publication has 13 references indexed in Scilit:
- Generative Pre-trained Transformer 4 makes cardiovascular magnetic resonance reports easy to understandJournal of Cardiovascular Magnetic Resonance, 2024
- Development of image-based decision support systems utilizing information extracted from radiological free-text report databases with text-based transformersEuropean Radiology, 2023
- The shaky foundations of large language models and foundation models for electronic health recordsnpj Digital Medicine, 2023
- Large language models in medicineNature Medicine, 2023
- A Context-based Chatbot Surpasses Radiologists and Generic ChatGPT in Following the ACR Appropriateness GuidelinesRadiology, 2023
- ChatGPT and the Future of Medical WritingRadiology, 2023
- ChatGPT and Other Large Language Models Are Double-edged SwordsRadiology, 2023
- Transformer-based structuring of free-text radiology report databasesEuropean Radiology, 2023
- ESR paper on structured reporting in radiologyInsights into Imaging, 2018
- Preparing a collection of radiology examinations for distribution and retrievalJournal of the American Medical Informatics Association, 2015