EUVIMEDEUVIMED entdecken
Open Access

Fine‐tuning open‐source large language models to improve their performance on radiation oncology tasks: A feasibility study to investigate their potential clinical applications in radiation oncology

Peilong Wang, Zhengliang Liu, Yiwei Li, Jason Holmes, Peng Shu, Lian Zhang, Xiang Li, Quanzheng Li, Brady S Laughlin, Diego Santos Toesca, Carlos E Vargas, Sujay A Vora und 4 weitere

Medical Physics · 2025-07 · 52 (7)

DOI: 10.1002/mp.17985 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Medical Physics, online erschienen am 15.07.2025. Verfasst von Peilong Wang und 15 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiation Oncology, Mayo Clinic Arizona, Phoenix, Arizona, USA. Hauptthemen laut MEDLINE: Radiation Oncology, Natural Language Processing. Die Arbeit wurde bisher 17-mal zitiert (Crossref) und nennt 35 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NCI NIH HHS, National Cancer Institute.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

The radiation oncology clinical practice involves many steps relying on the dynamic interplay of abundant text data. Large language models have displayed remarkable capabilities in processing complex text information. But their direct applications in specific fields like radiation oncology remain underexplored.

Purpose

This study aims to investigate whether fine‐tuning LLMs with domain knowledge can improve the performance on Task (1) treatment regimen generation, Task (2) treatment modality selection (photon, proton, electron, or brachytherapy), and Task (3) ICD‐10 code prediction in radiation oncology.

Methods

Data for 15 724 patient cases were extracted. Cases where patients had a single diagnostic record, and a clearly identifiable primary treatment plan were selected for preprocessing and manual annotation to have 7903 cases of the patient diagnosis, treatment plan, treatment modality, and ICD‐10 code. Each case was used to construct a pair consisting of patient diagnostics details and an answer (treatment regimen, treatment modality, or ICD‐10 code, respectively) for the supervised fine‐tuning of these three tasks. Open source LLaMA2‐7B and Mistral‐7B models were utilized for the fine‐tuning with the Low‐Rank Approximations method. Accuracy and ROUGE‐1 score were reported for the fine‐tuned models and original models. Clinical evaluation was performed on Task (1) by radiation oncologists, while precision, recall, and F‐1 score were evaluated for Task (2) and (3). One‐sided Wilcoxon signed‐rank tests were used to statistically analyze the results.

Results

Fine‐tuned LLMs outperformed original LLMs across all tasks with p value ≤ 0.001. Clinical evaluation demonstrated that over 60% of the fine‐tuned LLMs‐generated treatment regimens were clinically acceptable. Precision, recall, and F1‐score showed improved performance of fine‐tuned LLMs.

Conclusion

Fine‐tuned LLMs demonstrated statistically significant improvements over original LLMs upon three clinically important tasks in radiation oncology. This study explored the feasibility of applying fine‐tuned LLMs in radiation oncology, inspiring further development of utilizing LLMs to assist with radiation oncology tasks.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Radiation OncologyNatural Language ProcessingFeasibility StudiesHumansLarge Language Models

Stichwörter der Autor:innen

clinical tasksfine‐tuninglarge language modelsradiation oncology

Autor:innen und Institutionen

  • Department of Radiation Oncology, Mayo Clinic Arizona, Phoenix, Arizona, USA.
  • School of Computing, University of Georgia, Athens, Georgia, USA.
  • Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA.

Förderung

  • NCI NIH HHS
  • National Cancer Institute

Zitieren

APA

Wang, P., Liu, Z., Li, Y., Holmes, J., Shu, P., Zhang, L., Li, X., Li, Q., Laughlin, B. S., Toesca, D. S., Vargas, C. E., Vora, S. A., Patel, S. H., Sio, T. T., Liu, T., & Liu, W. (2025). Fine‐tuning open‐source large language models to improve their performance on radiation oncology tasks: A feasibility study to investigate their potential clinical applications in radiation oncology. Medical Physics, 52(7). https://doi.org/10.1002/mp.17985

Vancouver

Wang P, Liu Z, Li Y, Holmes J, Shu P, Zhang L, et al. Fine‐tuning open‐source large language models to improve their performance on radiation oncology tasks: A feasibility study to investigate their potential clinical applications in radiation oncology. Medical Physics. 2025;52(7). doi:10.1002/mp.17985

BibTeX
@article{Wang2025,
  author = {Wang, Peilong and Liu, Zhengliang and Li, Yiwei and Holmes, Jason and Shu, Peng and Zhang, Lian and Li, Xiang and Li, Quanzheng and Laughlin, Brady S and Toesca, Diego Santos and Vargas, Carlos E and Vora, Sujay A and Patel, Samir H and Sio, Terence T and Liu, Tianming and Liu, Wei},
  title = {Fine‐tuning open‐source large language models to improve their performance on radiation oncology tasks: A feasibility study to investigate their potential clinical applications in radiation oncology},
  journal = {Medical Physics},
  year = {2025},
  volume = {52},
  number = {7},
  doi = {10.1002/mp.17985}
}
Kostenlos starten →