This Python package uses the official TCIA REST API to enable downloads from www.cancerimagingarchive.net from within Python scripts and Jupyter Notebooks.

The documentation can be found at https://moritzschwyzer.github.io/tciaclient/. This PyPI/Conda package is based on source code of the TCIA-API-SDK https://github.com/TCIA-Community/TCIA-API-SDK.

pip install tciaclient

Step 1: Import the TCIAClient from the tciaclient.core package.

from tciaclient.core import TCIAClient

Step 2: Create an instance of the TCIAClient.

tc = TCIAClient()

Step 3: Specify the collection you want to download (find the name on https://wiki.cancerimagingarchive.net/display/Public/Collections).

collection_name = "NSCLC-Radiomics"

Step 4: Get the series information of the chosen collection. In this example, we specify that we only want series that are CT scans.

series = tc.get_series(collection=collection_name, modality="CT")

Step 5: Download the dataset to the specified path.

download_path = "./tcia-downloads"
for i, s in enumerate(series):
    print(i)
    tc.get_image(seriesInstanceUid = s["SeriesInstanceUID"],
        downloadPath = download_path, zipFileName = str(i).zfill(3)+"-"+collection_name+".zip")

Information regarding data usage policies and restrictions can be found on https://wiki.cancerimagingarchive.net/display/Public/Data+Usage+Policies+and+Restrictions