OpenTME
收藏资源简介:
OpenTME是由Aignostics联合多家研究机构开发的开放数据集,包含3634张源自癌症基因组图谱(TCGA)的H&E染色全切片图像,涵盖膀胱癌、乳腺癌、结直肠癌、肝癌和肺癌五种类型。数据集通过Atlas H&E-TME人工智能应用生成,提供单细胞分辨率的组织质量控制、组织分割、细胞检测与分类等4500余项定量指标。数据来源于TCGA公共队列,经严格质量控制筛选,旨在降低空间生物学研究门槛,支持生物标志物发现、生存建模和计算病理学方法开发等应用。
OpenTME is an open dataset developed by Aignostics in collaboration with multiple research institutions. It contains 3,634 H&E-stained whole-slide images sourced from The Cancer Genome Atlas (TCGA), covering five cancer types: bladder cancer, breast cancer, colorectal cancer, liver cancer, and lung cancer. Generated via the Atlas H&E-TME AI application, the dataset provides over 4,500 quantitative metrics including single-cell resolution tissue quality control, tissue segmentation, cell detection and classification, among others. Derived from public TCGA cohorts and screened through rigorous quality control procedures, OpenTME aims to lower the barrier to spatial biology research, and supports applications including biomarker discovery, survival modeling, and computational pathology method development.

- 1OpenTME: An Open Dataset of AI-powered H&E Tumor Microenvironment Profiles from TCGAAignostics; 柏林夏里特医学院·病理学研究所; 梅奥诊所·实验室医学与病理学系; 柏林工业大学·机器学习组; 柏林学习与数据基础研究所; 韩国大学·人工智能系; 马克斯·普朗克信息学研究所; 德国癌症研究中心; 慕尼黑大学·病理学研究所; 巴伐利亚癌症研究中心 · 2026年



