## Attribution
This dataset contains 224x224 patches from whole-slide images originally downloaded from The Cancer Genome Atlas (TCGA) that are available in the NCI Genomic Data Commons (GDC) Open Access tier. We mirror and repackage a commonly used ~12k WSI subset in LitData format for ease of training. We exclude patches that did not pass HSV thresholding, following the procedure in Kaiko.AI's Midnight paper. Patches were randomly sampled across magnification levels.
**Primary source**
- The Cancer Genome Atlas Program (TCGA), National Cancer Institute and National Human Genome Research Institute
- NCI Genomic Data Commons (GDC) Data Portal
**Required acknowledgment**
> The results here are based upon data generated by the TCGA Research Network (https://www.cancer.gov/tcga).
**Use and redistribution**
- These WSIs originate from the GDC Open Access tier. Redistribution of Open Access TCGA content is permitted.
- Do not attempt to re-identify participants. Follow NIH Gen…
