Comparison of a Natural Language Processing Model and Large Language Models for Extracting Incidental Lung Nodule Data
A task-specific natural language processing model demonstrated higher accuracy and reliability when compared with seven general-purpose large language models for extracting incidental lung nodule data from unstructured chest CT radiology reports.
Authors
- Marco Aurélio Veiga Conrado (ORCID: https://orcid.org/0000-0003-2341-978X)
- João Martins da Fonseca (ORCID: https://orcid.org/0009-0003-3978-3727)
- Bruno Hochhegger
- Rodrigo Basilio
- Thiago Machuca
- Rosana Souza Rodrigues
- Alysson Roncally Carvalho
Institutions
- D’Or Institute for Research and Education (BR)
- Florida College (US)
Publication Details
- Journal
- Radiology Cardiothoracic Imaging
- Published
- 2026-02-01
- DOI
- https://doi.org/10.1148/ryct.250210
- Primary Topic
- Lung Cancer Diagnosis and Treatment
- Type
- article
- Field-Weighted Citation Impact
- 0.00