Chest X-ray (CXR) plays a pivotal role in clinical diagnosis, and various task-specific and foundation models have been developed for automatic CXR interpretation. However, these models often struggle to adapt to new diagnostic tasks and complex reasoning scenarios. Recently, LLM-based agent models have emerged as a promising paradigm for CXR analysis, enhancing models' capability through tool coordination, multi-step reasoning, and team collaboration.
Existing agents often rely on a single diagnostic pipeline and lack mechanisms for assessing tools' reliability, limiting their adaptability and credibility. To address this, we propose CXRAgent, a director-orchestrated, multi-stage agent for CXR interpretation. The central director coordinates the following stages:
-
Tool Invocation: The agent strategically orchestrates a set of CXR-analysis tools, with outputs normalized and verified by the Evidence-driven Validator (EDV), grounding diagnostic outputs with visual evidence to support reliable downstream diagnosis.
-
Diagnostic Planning: Guided by task requirements and intermediate findings, the agent formulates a targeted diagnostic plan, assembling an expert team accordingly, defining member roles and coordinating their interactions for adaptive and collaborative reasoning.
-
Collaborative Decision-making: The agent integrates insights from the expert team with accumulated contextual memories, synthesizing them into an evidence-backed diagnostic conclusion.
Experiments on various CXR interpretation tasks show that CXRAgent delivers strong performance, providing visual evidence and generalizes well to clinical tasks of different complexity. Code and data are available at this link.
Blogger's Review: CXRAgent significantly enhances chest X-ray interpretation through its multi-stage collaborative reasoning approach. Its innovative tool invocation and validation mechanisms provide higher reliability and adaptability for clinical applications, marking a technological advancement in medical imaging analysis.