胸部 CT 结节良恶性鉴别与病理对照高质量数据集

本数据集聚焦胸部 CT 结节良恶性鉴别与病理对照核心方向，原始数据均从合作三甲医院合规采集，经专业团队完成数据脱敏与文件校对全流程处理，最终形成高质量、标准化的专业医疗数据集，数据总量级达 1TB，可直接对接医疗算法研发、模型训练、临床研究等各类应用场景，为胸部 CT 结节相关人工智能技术开发、临床诊疗研究、医疗产品创新提供高价值、高可靠性的数据支撑。
数据集整体为多模态结构化医疗数据，单条数据围绕胸部 CT 结节良恶性鉴别与病理对照临床诊疗全维度构建，核心包含影像类数据与临床诊疗文本标注类数据两大类别。影像类数据为 dicom 格式原始影片，保留临床采集原始影像信息，格式标准化、信息完整度高；临床诊疗文本标注类数据涵盖住院病程、住院 / 出院记录、病理检验报告、放射及化验图文报告等核心医疗文件，字段信息完整覆盖胸部 CT 结节良恶性鉴别与病理对照临床诊疗的影像采集、病情诊断、检查分析、良恶性评估全流程，标注内容精准规范，与影像数据一一匹配，真实还原胸部 CT 结节良恶性鉴别与病理对照临床诊疗的实际场景与数据特征。

This dataset focuses on the core task of benign-malignant differentiation of pulmonary nodules in chest CT and pathological correlation. Its raw data are collected compliantly from cooperating tertiary hospitals, and have undergone full-process processing including data de-identification and document proofreading by a professional team. Eventually, a high-quality, standardized professional medical dataset is developed, with a total data volume of 1 TB. It can be directly integrated into various application scenarios such as medical algorithm research and development, model training, and clinical research, providing high-value and highly reliable data support for the development of AI technologies related to pulmonary nodules in chest CT, clinical diagnosis and treatment research, and medical product innovation. The dataset consists of multi-modal structured medical data. Each individual data entry is built across all dimensions of clinical diagnosis and treatment for benign-malignant differentiation of pulmonary nodules in chest CT and pathological correlation, and primarily includes two categories: imaging data and clinical diagnosis and treatment text annotation data. The imaging data are original films in DICOM format, which retain the original clinical imaging information, with standardized formats and high information completeness. The clinical diagnosis and treatment text annotation data cover core medical documents such as hospital course records, admission/discharge records, pathological examination reports, and radiological and laboratory graphic reports. The field information fully covers the entire clinical diagnosis and treatment process for benign-malignant differentiation of pulmonary nodules in chest CT and pathological correlation, including image acquisition, disease diagnosis, examination and analysis, and benign-malignant assessment. The annotation content is precise and standardized, with one-to-one correspondence to the imaging data, and truly restores the actual scenarios and data characteristics of clinical diagnosis and treatment for benign-malignant differentiation of pulmonary nodules in chest CT and pathological correlation.

搜集汇总

背景与挑战

背景概述

该数据集是一个专注于胸部CT结节良恶性鉴别与病理对照的高质量医疗数据集，数据来源于合作三甲医院，经过脱敏和标准化处理，总量达1TB。它包含dicom格式的影像数据和临床诊疗文本标注，覆盖诊疗全流程，适用于医疗AI开发、模型训练和临床研究。