Nature Scientific Data Publications Underscore the Growing Features and Usage of NCI's Cancer Imaging Archive

The Cancer Imaging Archive (TCIA), a service provided by the National Cancer Institute to the research community, is an archive of cancer-specific medical images and supporting data. TCIA de-identifies, collects, and curates radiology, radiation therapy, and pathology images to address data availability gaps that impede specific research objectives. The resulting datasets are made freely available to the public. A review paper describing many of the available collections was recently published in Nature Scientific Data (Prior, 2017).

In 2017 TCIA completed a multi-year collection activity associated with NCI/NIH's The Cancer Genome Atlas (TCGA) project. Images from nearly 2,000 subjects were archived in TCIA with corresponding genomic and clinical data available through the NCI Genomic Data Commons. This effort created an unprecedented opportunity to stimulate radiogenomic and precision medicine research. Over 50 manuscripts have since been published based on TCGA imaging data on TCIA.

One of those manuscripts, also published in Nature Scientific Data (Bakas, 2017), describes the release of analysis data that includes tumor labels and radiomic features derived from TCGA glioma images. Gliomas contain sub-regions that have variable histologic, genomic, and radiographic phenotypes, for which the medical images are a source of information that could contribute to the development of clinical biomarkers. The use of radiomic features for research may require knowledge of the patient's specific tumor location, and TCIA has begun accepting analysis data, such as tumor labels and annotations. Bakas, et al, describe the process by which glioma subjects were selected from TCIA, as well as how sub-region segmentation labels were applied, and radiomic features were determined. The generated data describe collections of multi-institutional, pre-operative magnetic resonance imaging (MRI) scans for glioblastoma (n=135) and low-grade-glioma (n=108). The segmentation labels that were shared enabled the authors' own research aims, but their availability on TCIA will also allow future quantitative computational and clinical studies to quickly develop new image-based predictive, prognostic, and diagnostic assessments, with potential to evaluate disease via non-invasive approaches.

References:
Prior F, Smith K, Sharma A, Kirby J, Tarbox L, Clark K, Bennett W, Nolan T, Freymann J. The public cancer radiology imaging collections of The Cancer Imaging Archive. Sci Data. 2017 Sep 19;4:170124.

Bakas S, Akbari H, Sotiras A, Bilello M, Rozycki M, Kirby JS, Freymann JB, Farahani K, Davatzikos C. Advancing The Cancer Genome Atlas glioma MRI collections with expert segmentation labels and radiomic features. Sci Data. 2017 Sep 5;4:170117.