零样本大模型自动提取肺部病理报告关键信息，省去标注成本。
Prompt, Plan, Extract: Zero-Shot Agentic LLMs Workflows for Lung Pathology Extraction from Clinical Narratives
- 设计零样本智能工作流，分三步：提示、规划、提取。
- 最佳模型零样本提取准确率达F1 0.893，接近监督模型的0.960。
- 无需训练即可处理复杂关系，适合医疗数据自动化场景。
从病理报告中提取信息对癌症分期和肿瘤登记至关重要，但关键数据常隐藏在自然语言中，手动提取耗时且易错。传统方法依赖全监督命名实体识别与关系抽取，需昂贵标注，且上游实体遗漏会引发连锁错误。本研究开发了一种零样本、代理式工作流，评估五种开源生成式大语言模型（LLMs），从肺癌切除病理报告中填充13个美国病理学家学院（CAP）结构化字段。采用新型注册表对齐评估框架，与最先进的监督模型GatorTron NER-RE基线对比：基线微调后达到微平均F1 0.960，而最佳零样本模型GPT-OSS-20B取得微平均F1 0.893（召回率0.949），可准确提取如病理分期等复杂关系，无需任务特定训练。结果表明，开源零样本代理式大模型在低成本提取肺部病理信息方面具有巨大潜力。
原文摘要 · Abstract (English)
Information extraction from pathology reports is essential for cancer staging, tumor registry population. Yet key data remains embedded in narrative reports, making manual extraction labor-intensive and error-prone. Traditional supervised Natural Language Processing pipelines address this through fully supervised Named Entity Recognition and Relation Extraction, but require expensive manual annotation and suffer cascading failures when upstream entities are missed. In this study, we developed a zero-shot, agentic workflow, and evaluated five open-source generative Large Language Models (LLMs) to populate 13 College of American Pathologists synoptic fields from lung resection pathology reports. We compared them against a state-of-the-art supervised GatorTron NER-RE baseline using a novel, registry-aligned evaluation framework. The baseline achieved Micro-F1of 0.960, while the best zero-shot model (GPT-OSS-20B) achieved Micro-F1 of 0.893 (recall: 0.949), accurately extracting complex relations like Pathologic Stage without task-specific training. These results suggest that open-source, zero-shot agentic LLMs show great potential as a low-cost solution for extracting lung pathology information.
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。