Search papers, labs, and topics across Lattice.
EviDx introduces an evidence-aware active diagnosis framework that enhances clinical decision-making by integrating patient-specific environments with a structured scaffold for LLM agents. This approach allows for dynamic evidence acquisition and hypothesis updating, addressing the limitations of static diagnosis systems. Experimental results demonstrate that EviDx not only improves diagnostic performance and stability but also clarifies the operational boundaries of LLM capabilities in clinical settings.
EviDx reveals that structured scaffolding and dynamic evidence acquisition can significantly enhance the reliability and effectiveness of AI-driven clinical diagnosis.
Clinical diagnosis is an active evidence-seeking process in which clinicians acquire evidence, update competing hypotheses, and decide when the available evidence is sufficient for diagnosis. Yet many medical diagnosis systems built around large language models (LLMs) still formulate diagnosis as static case-to-answer prediction, with limited support for evidence acquisition. Agentic LLMs offer a dynamic alternative through tool use and intermediate diagnostic trajectories, but existing systems often under-specify how patient evidence should be exposed, scaffolded, and controlled at runtime. We introduce EviDx, an evidence-aware active diagnosis framework that pairs patient-specific diagnostic environments with a clinical diagnostic scaffold and an observer-guided runtime harness. In EviDx, $\mathcal{E}$-Synthesis constructs interactive environments from raw clinical cases; the scaffold organizes role-specialized agents, evidence tools, and evolving evidence states; and the harness regulates diagnostic termination by tracking uncertainty and evidence coverage. A 3-level evaluation pyramid assesses execution robustness, reasoning dynamics, and diagnostic outcomes. Experiments show that EviDx improves diagnostic performance and process stability while revealing model-dependent capability boundaries.
