Programmatic access to NCI Imaging Data Commons - the largest public collection of cancer imaging data
NCI Imaging Data Commons (IDC) is a cloud-based platform providing researchers with free access to a large and growing collection of cancer imaging data. This includes radiology images (CT, MRI, PET), digital pathology slides, and more - all in standard DICOM format with rich clinical and research metadata.
idc-index is the official Python package for querying IDC metadata and
downloading imaging data - no cloud credentials or complex setup required.
- Query metadata with SQL - Search across ~100TB of data using DuckDB-powered SQL queries
- High-speed downloads - Parallel downloads from AWS and Google Cloud public buckets via s5cmd
- Browse hierarchically - Navigate collections → patients → studies → series programmatically
- Generate viewer URLs - Create links to view images in OHIF (radiology) or Slim (pathology) web viewers
- Command line interface - Download data directly from the terminal with
idccommands - No authentication required - All data is publicly accessible
pip install idc-indexRequires Python 3.10+. Downloads are powered by the bundled s5cmd tool.
The package version is updated with each new IDC data release. Upgrade regularly to access the latest collections and data:
pip install --upgrade idc-indexfrom idc_index import IDCClient
client = IDCClient.client()
# List all available collections
collections = client.get_collections()
print(f"IDC has {len(collections)} collections")
# Download a small collection (10.5 GB)
client.download_from_selection(collection_id="rider_pilot", downloadDir="./data")Find CT scans of the chest and download them:
from idc_index import IDCClient
client = IDCClient.client()
query = """
SELECT
collection_id,
PatientID,
SeriesInstanceUID,
SeriesDescription,
series_size_MB
FROM index
WHERE Modality = 'CT'
AND BodyPartExamined = 'CHEST'
LIMIT 10
"""
results = client.sql_query(query)
print(results)
# Download the matching series
client.download_dicom_series(
seriesInstanceUID=results["SeriesInstanceUID"].tolist(), downloadDir="./chest_ct"
)Navigate from collection to viewable images:
from idc_index import IDCClient
client = IDCClient.client()
# Get patients in a collection
patients = client.get_patients("tcga_luad", outputFormat="list")
print(f"Found {len(patients)} patients")
# Get studies for a patient
studies = client.get_dicom_studies(patients[0])
# Get series in that study
series = client.get_dicom_series(studies[0]["StudyInstanceUID"])
# Generate a viewer URL
viewer_url = client.get_viewer_URL(seriesInstanceUID=series[0]["SeriesInstanceUID"])
print(f"View in browser: {viewer_url}")Download data directly from the terminal using idc download, which
auto-detects the input type:
# Download a collection
idc download rider_pilot
# Download a specific series by UID
idc download 1.3.6.1.4.1.14519.5.2.1.6279.6001.100225287222365663678666836860
# Download from a manifest file
idc download manifest.s5cmd
# Specify output directory
idc download rider_pilot --download-dir ./data
# See all options
idc --help- Full Documentation - API reference and guides
- Tutorial Notebook - Interactive introduction to idc-index
- IDC Portal - Browse IDC data in your web browser
- IDC Forum - Community discussions and support
- idc-claude-skill - Claude AI skill for querying IDC with natural language
- SlicerIDCBrowser - 3D Slicer extension using idc-index
- s5cmd - The high-performance S3 client powering downloads
If idc-index helps your research, please cite:
Fedorov, A., Longabaugh, W. J. R., Pot, D., Clunie, D. A., Pieper, S. D., Gibbs, D. L., Bridge, C., Herrmann, M. D., Homeyer, A., Lewis, R., Aerts, H. J. W., Krishnaswamy, D., Thiriveedhi, V. K., Ciausu, C., Schacherer, D. P., Bontempi, D., Pihl, T., Wagner, U., Farahani, K., Kim, E. & Kikinis, R. National Cancer Institute Imaging Data Commons: Toward Transparency, Reproducibility, and Scalability in Imaging Artificial Intelligence. RadioGraphics (2023). https://doi.org/10.1148/rg.230180
This software is maintained by the IDC team, which has been funded in whole or in part with Federal funds from the NCI, NIH, under task order no. HHSN26110071 under contract no. HHSN261201500003I.