RadOncRAG: A Novel Retrieval-Augmented Generation Framework Improves Large Language Model Benchmark Performance in Radiation Oncology
Radiation-oncology-specific retrieval-augmented generation pipeline enhances nonreasoning LLM performance in radiation oncology by integrating domain-specific evidence, whereas it does not improve performance of reasoning models. These findings demonstrate that RAG can elevate clinical decision support by enabling simpler, cost-effective nonreasoning models to tackle complex tasks through retrieval capabilities-an efficient alternative to extensive model training that also yields citable, evidence-based explanations.
Authors
- Ajay Choudhri
- Gautam Thaker
- Vivek Subbiah (ORCID: https://orcid.org/0000-0002-6064-6837)
- Adam P. Dicker (ORCID: https://orcid.org/0000-0003-0733-3337)
- Navid Redjal (ORCID: https://orcid.org/0000-0003-0951-2195)
- Trevor J. Royce (ORCID: https://orcid.org/0000-0001-9879-1290)
- Arturo Loaiza‐Bonilla (ORCID: https://orcid.org/0000-0002-3655-9602)
- Nikhil G. Thaker (ORCID: https://orcid.org/0000-0003-3608-0025)
- Neena Shetty
- Jonathan R Gabriel
- Vikash Deendyal
Institutions
- Lee College (US)
- Atrium Health Wake Forest Baptist (US)
- St. Luke's University Health Network (US)
- University of Arizona (US)
- Sarah Cannon (US)
- Capital Health (US)
Publication Details
- Journal
- JCO Clinical Cancer Informatics
- Published
- 2025-11-01
- DOI
- https://doi.org/10.1200/cci-25-00220
- Primary Topic
- Topic Modeling
- Type
- article
- Field-Weighted Citation Impact
- 0.00