发表机构
Durham University（杜伦大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出Sparse Multi-Stage Expert-Agent Routing框架，通过自适应激活稀疏专家智能体实现高效临床推理，引入ClinFEScore评估协议，在200例真实MDT病例上达91.5%诊断准确率，激活专家数从17.0降至3.0。
AI 中文摘要
复杂临床推理要求模型在新证据出现时更新诊断假设，并在有限会诊资源下协调不同医学专科。现有基于大语言模型（LLM）的临床推理系统通常执行单次预测或依赖固定多智能体工作流，导致专家参与要么静态要么不必要地详尽。本文提出Sparse Multi-Stage Expert-Agent Routing，一种基于语言的临床推理框架，将诊断建模为分阶段路由过程。给定从多模态逐步获取的临床证据，该框架维护不断演变的病例状态，并自适应激活稀疏的医学专家智能体集合，各阶段由专家专属记忆支持。为评估自由文本诊断结论而非表面相似性，本文进一步引入ClinFEScore，一种针对临床推理输出的事实感知语义评估协议。在从MAC和AgentClinic-NEJM重构的多阶段病例上，本文框架将激活专家的平均数量从17.0降至3.0，同时保持较强的事实级诊断质量。在200个真实世界医院多学科诊疗（MDT）病例上，ClinFEScore与临床医生判断具有强相关性（Spearman's ρ=0.81；Pearson's r=0.87），而本文方法实现了91.5%经临床医生验证的诊断准确率，每个病例仅需约5次专家智能体/LLM调用。这些结果表明，稀疏分阶段协调是一种高效且符合临床需求的基于LLM的临床推理方法。
英文摘要
Complex clinical reasoning requires models to update diagnostic hypotheses as new evidence emerges and to coordinate different medical specialities under limited consultation resources. Existing LLM-based clinical reasoning systems typically perform single-pass prediction or rely on fixed multi-agent workflows, making expert participation either static or unnecessarily exhaustive. We propose Sparse Multi-Stage Expert-Agent Routing, a language-based clinical reasoning framework that models diagnosis as a stage-wise routing process. Given progressively available clinical evidence derived from multiple modalities, the framework maintains an evolving case state and adaptively activates a sparse set of medical expert agents, supported by expert-specific memory across stages. To evaluate free-text diagnostic conclusions beyond surface similarity, we further introduce ClinFEScore, a fact-aware semantic evaluation protocol for clinical reasoning outputs. On reconstructed multi-stage cases from MAC and AgentClinic-NEJM, our framework reduces the average number of activated experts from 17.0 to 3.0 whilst maintaining strong fact-level diagnostic quality. On 200 real-world hospital MDT cases, ClinFEScore correlates strongly with clinician judgements (Spearman's $ρ=0.81$; Pearson's $r=0.87$), whilst our method achieves 91.5\% clinician-verified diagnostic accuracy with approximately five expert-agent/LLM calls per case. These results support sparse stage-wise coordination as an efficient and clinically relevant approach to LLM-based clinical reasoning.
