Data from NCI Imaging Data Commons
Cancer imaging data + annotations and analysis results
Overview
Imaging Data Commons (IDC)
IDC hosts a growing number of imaging collections that are contributed by either funded US National Cancer Institute (NCI)
Image data hosted by IDC is stored in DICOM
: these include DICOM metadata attributes extracted from the DICOM data into BigQuery tables, which are further enriched by including collection-level metadata that is not available in DICOM.
: these files are available in Storage buckets.
This dataset is rather large (~40TB), and is updated monthly, which makes is challenging to download all of the files. Instead, users should utilize the BigQuery tables to search and identify files of interest, which then can be downloaded selectively from Cloud Storage. Please see the download instructions page in Imaging Data Commons documentation: https://learn.canceridc.dev/data/downloading-data
See further details about data organization in IDC documentation
Additional details
Data Healthcare & Life Sciences, Science & research NCI Imaging Data Commons BigQuery Monthly
Samples
Get the list of GCS URLs for all of the DICOM files that have the Modality value of “MR” and BodyPartExamined value resembling “PROSTATE”.
This query examines the IDC metadata in BigQuery and returns a list of images that meet the criteria in the WHERE statement
Run this query
What are the distinct values of ROIName available in RTSTRUCT series, and the counts of the series containing those values?
This query examines the IDC metadata in BigQuery to determine the number of distinct ROI names in the dataset and the number of series each ROI name has.
Run this query
Explore the dashboard
You can check out the IDC Data Studio dashboard
Terms of Service
This dataset consists of multiple collections, each governed by its own license/attribution terms. Each file has an `instance_uuid` corresponding to the prefix of the file name in the `auxiliary_metadata_table` BigQuery table. The `source_doi` column corresponds to the Digital Object Identifier that should be referenced to learn about attribution requirements. `license_*` columns contain the details of the license. The data are provided "AS IS" without any warranty, express or implied, from Google.
