A Clinically Curated and Provenance-Aware TCGA-BRCA Cohort of Triple-Negative Breast Cancer and Age-Matched Luminal A Controls with Linked Whole-Slide Images

This dataset provides a clinically curated and fully reproducible subset of The Cancer Genome Atlas Breast Invasive Carcinoma (TCGA-BRCA) cohort designed for computational pathology and breast cancer classification research. The resource contains: 153 clinically confirmed Triple-Negative Breast Cancer (TNBC) patients. 153 age-matched Luminal A control patients. Linked diagnostic whole-slide image (WSI) metadata and Genomic Data Commons (GDC) download manifests. Complete cohort provenance through a documented 19-step clinical filtering pipeline. Full exclusion audit logs enabling reconstruction of cohort selection decisions. Reproducible Python scripts for cohort generation and validation. The dataset was derived from the TCGA-BRCA cohort consisting of 1,107 patients and 145 clinical variables obtained through cBioPortal and linked to publicly available diagnostic whole-slide images hosted by the National Cancer Institute Genomic Data Commons (GDC). Released files include clinical metadata tables, WSI linkage files, exclusion logs, download manifests, and reproducible curation scripts. License: CC BY 4.0