301医院肝胆胰患者病历、影像和会诊语音的数据集

数据来源于中国301医院，基于真实患者的临床数据构建，并在数据采集过程中通过去隐私化手段对患者的个人信息进行处理，包括病历资料、影像数据、语音数据在内的3000 份肝胆胰科室患者的真实病例。具体内容如下：包含病历资料、影像数据、语音数据等。病历数据包含肝胆胰科室患者的真实病例共4610份（基本信息、患者在院病历、检查结果、检验结果）；影像数据包含肝癌MRI1229份，肝内胆管细胞癌MRI32份，胰腺癌MRI903份，胰腺癌PET-CT873份，共计影像数据3037份；语音数据包含真实会诊场景中的原始音频、音频转成的word版、内容整理版。

This dataset is built on real clinical patient data obtained from the 301st Hospital of China. All patients' personal identifiable information was processed using de-identification measures during data acquisition. It comprises 3000 real patient cases from the Department of Hepatobiliary and Pancreatic Surgery, encompassing three main data types: medical records, imaging data, and voice data.

Specific details are listed below:
1. Medical records: A total of 4610 real cases of patients from the hepatobiliary and pancreatic department, including basic patient information, in-hospital medical records, examination results, and laboratory test results.
2. Imaging data: A total of 3037 scans, consisting of 1229 MRI scans of liver cancer, 32 MRI scans of intrahepatic cholangiocarcinoma, 903 MRI scans of pancreatic cancer, and 873 PET-CT scans of pancreatic cancer.
3. Voice data: Raw audio recordings from real clinical consultation scenarios, Word transcripts converted from the audio, and organized content versions of the recordings.

搜集汇总

背景与挑战

背景概述

该数据集是一个多模态医疗数据集，源自中国301医院的真实肝胆胰患者临床数据，包含病历资料、影像数据（如MRI和PET-CT）和会诊语音数据，总计覆盖3000多份病例，数据量达356.96GB。数据集经过去隐私化处理，适用于医疗AI研究，特别是人机融合会诊技术开发，并关联国家重点研发计划项目，具有较高的科研和应用价值。