Is it possible to download annotations for the images?

Yes, the skill provides access to the analysis_results_index and seg_index, which contain metadata for AI-generated and expert-curated segmentations and annotations.

How do I view the images without a local DICOM viewer?

The skill utilizes the get_viewer_URL method to generate direct links to web-based viewers, allowing you to inspect images in your browser.

What imaging modalities can I access with this skill?

You can access a variety of modalities including CT, MRI, PET, and Slide Microscopy (pathology) from various NCI-funded research collections.

Do I need an account to download data from IDC?

No, the Imaging Data Commons provides public access to its datasets via idc-index without requiring authentication or API keys.

Can I filter datasets by specific cancer types?

Yes, by joining the primary index with the collections_index, you can filter by cancer types, tumor locations, and specific research projects.

Imaging Data Commons Access

Name: Imaging Data Commons Access
Author: K-Dense-AI

byK-Dense-AI

•

7,774

•

Data Science & ML

Queries and downloads large-scale public cancer imaging datasets from the NCI Imaging Data Commons using Python and SQL.

This skill empowers researchers and AI developers to programmatically access the National Cancer Institute's Imaging Data Commons (IDC). By leveraging the idc-index Python package, users can perform complex SQL queries on rich metadata to filter massive datasets by modality (CT, MRI, PET), anatomical site, or cancer type. It streamlines the entire workflow from data discovery and license verification to bulk DICOM downloading and in-browser visualization, providing a critical bridge between raw medical archives and AI model training pipelines.

Key Features

01Seamless integration with pandas DataFrames for data exploration

02SQL-based metadata querying for precise dataset selection

03In-browser visualization links for quick medical image previews

04Programmatic bulk download of DICOM radiology and pathology images

05Unified access to AI-generated annotations and clinical metadata

067,774 GitHub stars

Use Cases

01Building diverse medical imaging datasets for training deep learning models

02Performing large-scale cohort discovery for oncology and radiology research

03Linking imaging series with associated clinical data and expert annotations

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add k-dense-ai/claude-scientific-skills imaging-data-commons

For use in Claude.ai and ChatGPT

Download Skill