Scholarlink

Background <p>Recent advances in reasoning-enhanced large language models (LLMs) and multimodal LLMs (MLLMs) have significantly improved performance in complex tasks, yet medical AI models often overlook the structured reasoning processes inherent in clinical practice.</p> Methods <p>In this work, we propose a two-stage pipeline to train <b>ChestX-Reasoner</b>, a radiology diagnosis MLLM with process supervision mined directly from clinical reports, reflecting the step-by-step reasoning followed by radiologists. To facilitate and evaluate reasoning capabilities, we introduce <b>RadRBench-CXR</b>, a comprehensive benchmark featuring 59K visual question answering samples with 301K clinically validated reasoning steps, and <b>RadRScore</b>, a metric evaluating reasoning factuality, completeness, and effectiveness.</p> Results <p>Here we show that, ChestX-Reasoner achieves significant improvements of 16% and 8.5% in reasoning ability compared to leading medical and general-purpose models, respectively. Furthermore, diagnostic accuracy improves by 3.3% to 24% against baselines.</p> Conclusions <p>These results demonstrate that incorporating explicit reasoning steps improves diagnostic outcomes, and integrating process supervision enhances the reliability and transparency of medical diagnosis.</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

Advancing radiology foundation models with reasoning through step-by-step verification from daily reports

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月27日

查看全文
  • Ziqing Fan,
  • Cheng Liang,
  • Chaoyi Wu,
  • Ya Zhang,
  • Yanfeng Wang,
  • Weidi Xie

摘要

Background

Recent advances in reasoning-enhanced large language models (LLMs) and multimodal LLMs (MLLMs) have significantly improved performance in complex tasks, yet medical AI models often overlook the structured reasoning processes inherent in clinical practice.

Methods

In this work, we propose a two-stage pipeline to train ChestX-Reasoner, a radiology diagnosis MLLM with process supervision mined directly from clinical reports, reflecting the step-by-step reasoning followed by radiologists. To facilitate and evaluate reasoning capabilities, we introduce RadRBench-CXR, a comprehensive benchmark featuring 59K visual question answering samples with 301K clinically validated reasoning steps, and RadRScore, a metric evaluating reasoning factuality, completeness, and effectiveness.

Results

Here we show that, ChestX-Reasoner achieves significant improvements of 16% and 8.5% in reasoning ability compared to leading medical and general-purpose models, respectively. Furthermore, diagnostic accuracy improves by 3.3% to 24% against baselines.

Conclusions

These results demonstrate that incorporating explicit reasoning steps improves diagnostic outcomes, and integrating process supervision enhances the reliability and transparency of medical diagnosis.