Researchers have made a set of 142 annotated CT scans from 101 patients with renal cell carcinoma (RCC) publicly available. The goal is to help other scientists develop and test artificial intelligence tools for detecting and classifying kidney cancer subtypes.
The scans come from patients with an average age of 56, including 26 women and 75 men. The dataset includes three common RCC subtypes: clear cell (95 scans), papillary (29 scans), and chromophobe (18 scans). All scans underwent quality control before release.
No safety concerns were reported, as this is a data-sharing effort, not a treatment study. However, the authors note that re-annotation may be needed because public annotations can be limited. This is an early resource, not a ready-to-use clinical tool.
For now, this open-source dataset is a step toward more reproducible AI research in kidney cancer. It does not change current diagnosis or treatment, but it may speed up future discoveries.