Scholarlink

<p>Large language models (LLMs) show promise for radiology decision support, yet conventional retrieval-augmented generation (RAG) relies on single-step retrieval and struggles with complex reasoning. We introduce radiology Retrieval and Reasoning (RaR), a multi-step retrieval framework that iteratively summarizes clinical questions, retrieves evidence, and synthesizes answers. We evaluated 25 LLMs spanning general-purpose, reasoning-optimized, and clinically fine-tuned models (0.5B → 670B parameters) on 104 expert-curated radiology questions and an independent set of 65 real radiology board-exam questions. RaR significantly improved mean diagnostic accuracy versus zero-shot prompting (75% vs. 67%; <i>P</i> = 1.1 × 10<sup>−7</sup>) and conventional online RAG (75% vs. 69%; <i>P</i> = 1.9 × 10<sup>−6</sup>). Gains were largest in mid-sized and small models (e.g., Mistral Large: 72% → 81%), while very large models showed minimal change. RaR reduced hallucinations and provided clinically relevant evidence in 46% of cases, improving factual grounding. These results show that multi-step retrieval enhances diagnostic reliability, especially in deployable mid-sized LLMs. Code, datasets, and RaR are publicly available.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-step retrieval and reasoning improves radiology question answering with large language models

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月22日

查看全文
  • Sebastian Wind,
  • Jeta Sopa,
  • Daniel Truhn,
  • Mahshad Lotfinia,
  • Tri-Thien Nguyen,
  • Keno Bressem,
  • Lisa Adams,
  • Mirabela Rusu,
  • Harald Köstler,
  • Gerhard Wellein,
  • Andreas Maier,
  • Soroosh Tayebi Arasteh

摘要

Large language models (LLMs) show promise for radiology decision support, yet conventional retrieval-augmented generation (RAG) relies on single-step retrieval and struggles with complex reasoning. We introduce radiology Retrieval and Reasoning (RaR), a multi-step retrieval framework that iteratively summarizes clinical questions, retrieves evidence, and synthesizes answers. We evaluated 25 LLMs spanning general-purpose, reasoning-optimized, and clinically fine-tuned models (0.5B → 670B parameters) on 104 expert-curated radiology questions and an independent set of 65 real radiology board-exam questions. RaR significantly improved mean diagnostic accuracy versus zero-shot prompting (75% vs. 67%; P = 1.1 × 10−7) and conventional online RAG (75% vs. 69%; P = 1.9 × 10−6). Gains were largest in mid-sized and small models (e.g., Mistral Large: 72% → 81%), while very large models showed minimal change. RaR reduced hallucinations and provided clinically relevant evidence in 46% of cases, improving factual grounding. These results show that multi-step retrieval enhances diagnostic reliability, especially in deployable mid-sized LLMs. Code, datasets, and RaR are publicly available.