基于肝胆胰的多模态数据集

该数据集主要面向肝胆胰相关疾病的医疗研究、智能会诊系统需求建设，其于医院真实诊疗场景中产生，主要记录了肝癌、胰腺癌等肝胆胰疾病患者的多模态信息，包括 CT 和 MRI 影像数据，涵盖初步诊断、手术情况及出院摘要的病历信息，以及血常规 + CRP、癌胚抗原测定、血清丙氨酸氨基转移酶等多项指标的生化检查数据。数据按疾病、患者 ID 分层构建目录，疾病文件夹下含对应患者的病历 Excel 文件及 CT、MRI 影像文件。本数据集需审批后共享，上传至医疗会诊平台后，可经智能体处理提取结构化数据，为多智能体协同会诊提供支持。

This dataset is targeted at medical research on hepatobiliary and pancreatic diseases and the development of intelligent consultation systems. It is collected from real clinical scenarios in hospitals, and mainly records multimodal data of patients with hepatobiliary and pancreatic diseases including liver cancer, pancreatic cancer and other related diseases. The data covers CT and MRI imaging data, medical records including preliminary diagnosis, surgical status and discharge summaries, as well as biochemical examination data of multiple indicators such as routine blood test + CRP, carcinoembryonic antigen (CEA) assay, serum alanine aminotransferase (ALT) and others. The dataset is structured into a hierarchical directory based on disease types and patient IDs, with each disease folder containing the Excel medical record files and CT/MRI imaging files of corresponding patients. Access to this dataset requires prior approval for sharing. After being uploaded to the medical consultation platform, it can be processed by AI Agents to extract structured data, providing support for multi-agent collaborative consultation.

搜集汇总

背景与挑战

背景概述

该数据集面向肝胆胰疾病医疗研究与智能会诊，包含来自真实诊疗场景的肝癌、胰腺癌等患者的多模态信息，具体有CT与MRI影像、病历文本及多项生化检查数据。数据按疾病与患者分层组织，经审批后可用于支撑多智能体协同会诊平台。