NCI Imaging Data Commons
Queries and downloads cancer radiology and pathology datasets from NCI Imaging Data Commons using DuckDB SQL and fast S3/GCS streaming.
The Imaging Data Commons skill enables Claude Code to search, filter, and stream medical imaging cohorts from the National Cancer Institute's Imaging Data Commons (IDC). Leveraging the idc-index Python client and local DuckDB execution against an embedded Parquet catalog, it allows zero-latency SQL cohort selection without requiring cloud credentials, BigQuery billing, or external API keys. It streamlines building reproducible cancer machine learning datasets across 175+ DICOM collections, orchestrating high-speed downloads via bundled s5cmd, generating interactive OHIF and Slim viewer URLs, and prepping medical imaging data for downstream deep learning workflows.