Research

Paper

AI LLM March 10, 2026

Tracking Cancer Through Text: Longitudinal Extraction From Radiology Reports Using Open-Source Large Language Models

Authors

Luc Builtjes, Alessa Hering

Abstract

Radiology reports capture crucial longitudinal information on tumor burden, treatment response, and disease progression, yet their unstructured narrative format complicates automated analysis. While large language models (LLMs) have advanced clinical text processing, most state-of-the-art systems remain proprietary, limiting their applicability in privacy-sensitive healthcare environments. We present a fully open-source, locally deployable pipeline for longitudinal information extraction from radiology reports, implemented using the \texttt{llm\_extractinator} framework. The system applies the \texttt{qwen2.5-72b} model to extract and link target, non-target, and new lesion data across time points in accordance with RECIST criteria. Evaluation on 50 Dutch CT Thorax/Abdomen report pairs yielded high extraction performance, with attribute-level accuracies of 93.7\% for target lesions, 94.9\% for non-target lesions, and 94.0\% for new lesions. The approach demonstrates that open-source LLMs can achieve clinically meaningful performance in multi-timepoint oncology tasks while ensuring data privacy and reproducibility. These results highlight the potential of locally deployable LLMs for scalable extraction of structured longitudinal data from routine clinical text.

Metadata

arXiv ID: 2603.09638
Provider: ARXIV
Primary Category: cs.CL
Published: 2026-03-10
Fetched: 2026-03-11 06:02

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.09638v1</id>\n    <title>Tracking Cancer Through Text: Longitudinal Extraction From Radiology Reports Using Open-Source Large Language Models</title>\n    <updated>2026-03-10T13:13:43Z</updated>\n    <link href='https://arxiv.org/abs/2603.09638v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.09638v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Radiology reports capture crucial longitudinal information on tumor burden, treatment response, and disease progression, yet their unstructured narrative format complicates automated analysis. While large language models (LLMs) have advanced clinical text processing, most state-of-the-art systems remain proprietary, limiting their applicability in privacy-sensitive healthcare environments. We present a fully open-source, locally deployable pipeline for longitudinal information extraction from radiology reports, implemented using the \\texttt{llm\\_extractinator} framework. The system applies the \\texttt{qwen2.5-72b} model to extract and link target, non-target, and new lesion data across time points in accordance with RECIST criteria. Evaluation on 50 Dutch CT Thorax/Abdomen report pairs yielded high extraction performance, with attribute-level accuracies of 93.7\\% for target lesions, 94.9\\% for non-target lesions, and 94.0\\% for new lesions. The approach demonstrates that open-source LLMs can achieve clinically meaningful performance in multi-timepoint oncology tasks while ensuring data privacy and reproducibility. These results highlight the potential of locally deployable LLMs for scalable extraction of structured longitudinal data from routine clinical text.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <published>2026-03-10T13:13:43Z</published>\n    <arxiv:comment>6 pages, 2 figures</arxiv:comment>\n    <arxiv:primary_category term='cs.CL'/>\n    <author>\n      <name>Luc Builtjes</name>\n    </author>\n    <author>\n      <name>Alessa Hering</name>\n    </author>\n  </entry>"
}