Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

EviAgent: Evidence-Driven Agent for Radiology Report Generation

Tuoshi Qi, Shenshen Bu, Yingfei Xiang, Zhiming Dai

Abstract

Automated radiology report generation holds immense potential to alleviate the heavy workload of radiologists. Despite the formidable vision-language capabilities of recent Multimodal Large Language Models (MLLMs), their clinical deployment is severely constrained by inherent limitations: their "black-box" decision-making renders the generated reports untraceable due to the lack of explicit visual evidence to support the diagnosis, and they struggle to access external domain knowledge. To address these challenges, we propose the Evidence-driven Radiology Report Generation Agent (EviAgent). Unlike opaque end-to-end paradigms, EviAgent coordinates a transparent reasoning trajectory by breaking down the complex generation process into granular operational units. We integrate multi-dimensional visual experts and retrieval mechanisms as external support modules, endowing the system with explicit visual evidence and high-quality clinical priors. Extensive experiments on MIMIC-CXR, CheXpert Plus, and IU-Xray datasets demonstrate that EviAgent outperforms both large-scale generalist models and specialized medical models, providing a robust and trustworthy solution for automated radiology report generation.

EviAgent: Evidence-Driven Agent for Radiology Report Generation

Abstract

Automated radiology report generation holds immense potential to alleviate the heavy workload of radiologists. Despite the formidable vision-language capabilities of recent Multimodal Large Language Models (MLLMs), their clinical deployment is severely constrained by inherent limitations: their "black-box" decision-making renders the generated reports untraceable due to the lack of explicit visual evidence to support the diagnosis, and they struggle to access external domain knowledge. To address these challenges, we propose the Evidence-driven Radiology Report Generation Agent (EviAgent). Unlike opaque end-to-end paradigms, EviAgent coordinates a transparent reasoning trajectory by breaking down the complex generation process into granular operational units. We integrate multi-dimensional visual experts and retrieval mechanisms as external support modules, endowing the system with explicit visual evidence and high-quality clinical priors. Extensive experiments on MIMIC-CXR, CheXpert Plus, and IU-Xray datasets demonstrate that EviAgent outperforms both large-scale generalist models and specialized medical models, providing a robust and trustworthy solution for automated radiology report generation.
Paper Structure (18 sections, 2 equations, 2 figures, 5 tables)

This paper contains 18 sections, 2 equations, 2 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. MLLMs in Radiology
  4. Medical Agents
  5. The EviAgent
  6. The Evidence-based Framework
  7. Experts & Knowledge Integration
  8. Discriminative Perception Experts
  9. Retrieval-Augmented Knowledge
  10. Extensible Customization via Configuration
  11. Experiments
  12. Experimental Setup
  13. Performance on Multiple Metrics
  14. Clinical Value Evaluation via LLM-as-a-Judge
  15. Ablation Study
  16. ...and 3 more sections

Figures (2)

  • Figure 1: Overview of the EviAgent. The agent operates via four sequential stages: (A) Planning, which decomposes the task into granular operational units; (B) Tool Use, which dynamically invokes tools using the ReAct loop; (C) Evidence Extraction, which consolidates discrete observations into traceable proofs; and (D) Output, which synthesizes the final report strictly based on the accumulated evidence.
  • Figure 2: Qualitative analysis. Top: A complex case (Study 56122911) where EviAgent achieves accurate diagnosis and localization. Bottom: A case demonstrating error traceability (Study 50239281). Text in green indicates consistency with the ground truth, while text in red denotes discrepancies.