Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Jun 28, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Can Small Open-Source Language Models With Retrieval-Augmented Generation Match GPT-4 Performance in Breast Cancer

Chanhee Park1, In Hae Park2, Minhyuk Kim1

  • 1Department of Computer Science and Engineering, Korea University, Seoul, Korea.

JCO Clinical Cancer Informatics
|June 26, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Data-Driven Quantum Simulation of Artificial Quantum Materials with Rydberg Atoms.

Materials (Basel, Switzerland)·2026
Same author

Streamlining Human-Robot Interaction: Integrating LLM-Based Planning into Modular Robotic Frameworks.

Sensors (Basel, Switzerland)·2026
Same author

Nanostructured branched Y-DNA promotes antitumor immunity through dual activation of cGAS/STING and TLR9.

Archives of pharmacal research·2026
Same author

Vacancy Cluster-Mediated Epitaxial Layer-by-Layer Growth of van der Waals Heterostructures.

ACS nano·2026
Same author

Bypassing Nonlocal Phenomena in Metals Using Phonon-Polaritons.

ACS nano·2025
Same author

A clean van der Waals interface between the high-<i>k</i> dielectric zirconium oxide and two-dimensional molybdenum disulfide.

Nature electronics·2025
Same journal

Effect of a Multidimensional Digital Health Intervention on Quality of Life in Breast Cancer Survivors: A Randomized Controlled Trial.

JCO clinical cancer informatics·2026
Same journal

Machine Learning Algorithm for the Detection of Tumor Microsatellite Instability Based on Multiomics Biomarkers.

JCO clinical cancer informatics·2026
Same journal

Foundation Model-Driven Regions of Interest Classification and Renaming in Cancer Radiotherapy: A Customizable, Retraining-Free Workflow Across Institutions.

JCO clinical cancer informatics·2026
Same journal

Announcing a New Article Type in <i>JCO Clinical Cancer Informatics</i>: The Resource Report.

JCO clinical cancer informatics·2026
Same journal

A Harmonized International Database of More Than 10,000 Pediatric Renal Tumor Patients From 30 Years of SIOP-RTSG Studies.

JCO clinical cancer informatics·2026
Same journal

Machine Learning for Monoclonal Gammopathy of Undetermined Significance Screening: Who, How, and Why?

JCO clinical cancer informatics·2026
See all related articles
This summary is machine-generated.

Small open-source large language models (LLMs) with retrieval-augmented generation (RAG) show promise for breast cancer clinical decision support. Optimized RAG approaches proprietary model performance, offering scalable and cost-effective solutions.

Area of Science:

  • Artificial Intelligence in Medicine
  • Clinical Decision Support Systems
  • Oncology Informatics

Background:

  • The dynamic nature of breast cancer treatment presents challenges for clinicians in synthesizing up-to-date information.
  • Proprietary large language models (LLMs) offer potential but face limitations in cost, privacy, and accessibility.
  • Open-source LLMs present an alternative for developing specialized clinical support tools.

Purpose of the Study:

  • To evaluate the performance of small, open-source LLMs augmented with retrieval-augmented generation (RAG) for breast cancer clinical guideline queries.
  • To compare the performance of RAG-enhanced open-source LLMs against state-of-the-art proprietary models.
  • To assess the feasibility of using these models for clinical decision support.

Main Methods:

Related Experiment Videos

Last Updated: Jun 28, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

  • A domain-specific RAG pipeline was developed using 1,356 ASCO breast cancer guideline documents.
  • Five LLMs (GPT-4-turbo, GPT-3.5-turbo, Qwen2.5-14B, LLaMA3-8B, OpenBioLLM-8B) were tested with and without RAG.
  • Performance was evaluated using expert-curated question-answer triplets and rubric-based scoring, with GPT-4-turbo as judge and human oncologist validation.

Main Results:

  • RAG-enhanced Qwen2.5-14B demonstrated performance comparable to GPT-4-turbo, with relative improvements in win rates of 16% to 46%.
  • Absolute gains in rubric scores were modest, but RAG consistently improved LLM performance.
  • Human expert validation confirmed the superiority of RAG but yielded more conservative scores than LLM judges.

Conclusions:

  • Optimized RAG with small open-source LLMs can achieve performance close to proprietary models for clinical decision support.
  • This approach offers a scalable, cost-effective, and privacy-preserving solution for clinical implementation.
  • Potential for deployment on single-GPU infrastructure under expert supervision exists.