混合云本地架构让医学大模型在保护隐私前提下精准解读病历。
MedOrchestra: A Hybrid Cloud-Local LLM Approach for Clinical Data Interpretation
- 云端分解任务并生成提示，本地模型安全执行，全程不接触真实数据。
- 胰腺癌分期任务中，自由文本准确率达70.21%，显著优于本地基线和医生。
- 适合需高隐私保护的临床场景，尤其适用于医疗数据敏感的机构使用。
将大语言模型（LLM）部署于临床环境面临重大权衡：云端模型参数多、性能优，但存在敏感临床数据泄露风险；本地模型虽保障隐私，却难以胜任复杂临床解读任务。本文提出MedOrchestra，一种混合框架——云端模型基于临床指南与合成测试案例，将复杂任务拆解为可管理的子任务并生成、验证提示；本地模型在不访问原始数据的前提下，本地执行子任务并整合输出。在遵循NCCN指南的100份放射科报告上评估胰腺癌分期任务：自由文本场景下，准确率达70.21%，显著优于本地基线（无指南：48.94%；有指南：56.59%），也高于胃肠科医生（59.57%）、外科医生（65.96%）、放射科医生（55.32%）。结构化报告场景下准确率高达85.42%，各项指标均表现优异。
原文摘要 · Abstract (English)
Deploying large language models (LLMs) in clinical settings faces critical trade-offs: cloud LLMs, with their extensive parameters and superior performance, pose risks to sensitive clinical data privacy, while local LLMs preserve privacy but often fail at complex clinical interpretation tasks. We propose MedOrchestra, a hybrid framework where a cloud LLM decomposes complex clinical tasks into manageable subtasks and prompt generation, while a local LLM executes these subtasks in a privacy-preserving manner. Without accessing clinical data, the cloud LLM generates and validates subtask prompts using clinical guidelines and synthetic test cases. The local LLM executes subtasks locally and synthesizes outputs generated by the cloud LLM. We evaluate MedOrchestra on pancreatic cancer staging using 100 radiology reports under NCCN guidelines. On free-text reports, MedOrchestra achieves 70.21% accuracy, outperforming local model baselines (without guideline: 48.94%, with guideline: 56.59%) and board-certified clinicians (gastroenterologists: 59.57%, surgeons: 65.96%, radiologists: 55.32%). On structured reports, MedOrchestra reaches 85.42% accuracy, showing clear superiority across all settings.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。