肝癌分期病理报告

肝癌临床诊疗多源文本数据集”是由福建大数据一级开发有限公司依托省内三医（医疗、医保、医药）数据资源中心构建的高质量医疗专识数据集。数据集总规模为56.1MB，共包含13655条高质量问答对数据，数据以UTF-8编码的JSON/CSV格式扁平化存储。在数据结构上，输入端（Problem）深度整合了非结构化的多源临床综合诊断信息，涵盖腹部CT/MRI影像描述、超声检查结果以及详细的肉眼与组织学病理报告。输出端（Answer）为严格依据《原发性肝癌诊疗指南（2024年版）》标注的标准化CNLC临床分期标签（如CNLC IA期至IV期）。在合规性方面，本数据集经过严格脱敏与审查，在生命周期内坚守“原始数据不出域、数据可用不可见”的安全红线。

The Multi-source Text Dataset for Clinical Diagnosis and Treatment of Liver Cancer is a high-quality medical expert knowledge dataset constructed by Fujian Big Data First-level Development Co., Ltd. based on the provincial three-medical (medical care, medical insurance, pharmaceutical administration) data resource center. With a total size of 56.1 MB, this dataset contains 13,655 high-quality question-answer pairs, which are stored in a flat JSON/CSV format encoded with UTF-8. In terms of data structure, the input end (Problem) deeply integrates unstructured multi-source comprehensive clinical diagnostic information, including abdominal CT/MRI image descriptions, ultrasound examination results, and detailed macroscopic and histopathological reports. The output end (Answer) consists of standardized CNLC clinical staging labels (e.g., CNLC Stage IA to Stage IV) annotated strictly in accordance with the *Guidelines for the Diagnosis and Treatment of Primary Liver Cancer (2024 Edition)*. In terms of compliance, this dataset has undergone strict de-identification and review, and adheres to the security bottom line of "raw data shall not leave the domain, and data is accessible but not visible" throughout its lifecycle.

搜集汇总

背景与挑战

背景概述

该数据集是一个肝癌临床诊疗多源文本数据集，包含13655条高质量问答对，数据以JSON/CSV格式存储。输入端整合了腹部CT/MRI影像描述、超声检查结果及病理报告等非结构化临床信息，输出端为基于《原发性肝癌诊疗指南（2024年版）》标注的标准化CNLC临床分期标签，适用于医疗大模型的微调训练与评估。