Developing a named entity framework for thyroid cancer staging and risk level classification using large language models.

Matrix MH Fung ; Eric HM Tang ; Tingting Wu ; Yan Luk ; Ivan CH Au ; Xiaodong Liu ; Victor HF Lee ; Chun Ka Wong ; Zhili Wei ; Wing Yiu Cheng ; +8 more... Isaac CY Tai ; Joshua WK Ho ; Jason WH Wong ; Brian HH Lang ; Kathy SM Leung ; Zoie SY Wong ; Joseph T Wu ; Carlos KH Wong ORCID logo ; (2025) Developing a named entity framework for thyroid cancer staging and risk level classification using large language models. NPJ digital medicine, 8 (1). 134-. ISSN 2398-6352 DOI: 10.1038/s41746-025-01528-y
Copy

We developed a named entity (NE) framework for information extraction from semi-structured clinical notes retrieved from The Cancer Genome Atlas-Thyroid Cancer (TCGA-THCA) database and examined Large Language Models (LLMs) strategies to classify the 8th edition of American Joint Committee on Cancer (AJCC) staging and American Thyroid Association (ATA) risk category for patients with well-differentiated thyroid cancer. The NE framework consisted of annotation guidelines development, ground truth labelling, prompting approaches, and evaluation codes. Four LLMs (Mistral-7B-Instruct, Llama-3.1-8B-Instruct, Gemma-2-9B-Instruct, and Qwen2.5-7B-Instruct) were offline utilised for information extraction, comparing with expert-curated ground truth. Our framework was developed using 50 TCGA-THCA pathology notes. 289 TCGA-THCA notes and 35 pseudo-clinical cases were used for validation. Taking an ensemble-like majority-vote strategy achieved satisfactory performance for AJCC and ATA in both development and validation sets. Our framework and ensemble classifier optimised efficiency and accuracy of classifying stage and risk category in thyroid cancer patients.

picture_as_pdf

picture_as_pdf
Fung-etal-2025-Developing-a-named-entity-framework-for-thyroid-cancer-staging-and-risk-level-classification-using-large-language-models.pdf
subject
Published Version
Available under Creative Commons: Attribution-NonCommercial-No Derivative Works 4.0

View Download

Atom BibTeX OpenURL ContextObject in Span Multiline CSV OpenURL ContextObject Dublin Core Dublin Core MPEG-21 DIDL Data Cite XML EndNote HTML Citation JSON MARC (ASCII) MARC (ISO 2709) METS MODS RDF+N3 RDF+N-Triples RDF+XML RIOXX2 XML Reference Manager Refer Simple Metadata ASCII Citation EP3 XML
Export

Downloads