What problem does it solve?
This Skill removes the friction of locating, filtering, and downloading large public cancer imaging datasets by automating metadata queries, license checks, and DICOM downloads from NCI Imaging Data Commons (IDC).
Core Features & Use Cases
- Dataset discovery by metadata: Query IDC for CT/MR/PET radiology series or digital pathology (slide microscopy) using
idc-index with the collection_id/patient/study/series hierarchy.
- Efficient downloads for selected cohorts: Download DICOM data by
collection_id, by specific SeriesInstanceUID lists, or in batches/organized directory templates.
- Browser visualization without special tooling: Generate viewer URLs for quick inspection of results before committing to downloads.
- License-aware usage: Retrieve and surface license information (via IDC metadata) so you can respect CC-BY vs CC-NC terms.
- Covers core modalities and derived objects: Supports indexing for segmentations (SEG), annotations (ANN), clinical tables, and geometry validation (volume_geometry_index) via on-demand index tables and references.
Quick Start
Use the imaging-data-commons skill to query and download a small subset of IDC CT CHEST series for a research experiment using the currently documented idc-index workflow and then open one downloaded series in the browser viewer to confirm correctness.