The Landscape of Medical Agents: A Survey

Abstract

Medical Agents are an emerging class of agentic systems deployed in clinical settings that operate over multimodal, longitudinal data, maintain internal state, plan and adapt sequences of actions, and interact with clinical information systems under governance constraints. They extend traditional medical artificial intelligence (MedAI) beyond narrow diagnostic and predictive models toward workflow-centric architectures that address persistent challenges such as administrative burden, fragmented workflows, and workforce strain. In this paper, we (i) propose a functional definition and three-level developmental roadmap for Medical Agents, linking architectural capabilities (planning, memory, tool use, long-horizon control) to degrees of workflow integration and autonomy; (ii) map representative deployments across hospital departments and tasks, including domain-specific agents and multi-agent hospital simulations; and (iii) synthesize cross-cutting challenges in safety, robustness, fairness, evaluation, and governance, outlining research directions for advancing capabilities under clinical constraints and achieving system-level impact. We argue that Medical Agents should be treated as emerging infrastructure for learning health systems, whose value will be measured less by benchmark accuracy than by reliable restructuring of clinical workflows. CodeRepository: https://github.com/NUS-Project/Landmark-of-medical-agent

Ähnliche Arbeiten