About Dataset

Summary

The data is a preprocessed subset of the TCIA Study named Soft Tissue Sarcoma. The data have been converted from DICOM folders of varying resolution and data types to 3D HDF5 arrays with isotropic voxel size. This should make it easier to get started and test out various approaches (NN, RF, CRF, etc) to improve segmentations.

TCIA Summary

This collection contains FDG-PET/CT and anatomical MR (T1-weighted, T2-weighted with fat-suppression) imaging data from 51 patients with histologically proven soft-tissue sarcomas (STSs) of the extremities. All patients had pre-treatment FDG-PET/CT and MRI scans between November 2004 and November 2011. (Note: date in the TCIA images have been changed in the interest of de-identification; the same change was applied across all images, preserving the time intervals between serial scans). During the follow-up period, 19 patients developed lung metastases. Imaging data and lung metastases development status were used in the following study:

Vallières, M. et al. (2015). A radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremities. Physics in Medicine and Biology, 60(14), 5471-5496. doi:10.1088/0031-9155/60/14/5471.

Imaging data, tumor contours (RTstruct DICOM objects), clinical data and source code is available for this study. See the DOI below for more details and links to access the whole dataset. Please contact Martin Vallières (mart.vallieres@gmail.com) of the Medical Physics Unit of McGill University for any scientific inquiries about this dataset.

Acknowledgements

We would like to acknowledge the individuals and institutions that have provided data for this collection:
McGill University, Montreal, Canada - Special thanks to Martin Vallières of the Medical Physics Unit

License

This collection is freely available to browse, download, and use for commercial, scientific and educational purposes as outlined in the Creative Commons Attribution 3.0 Unported License. See TCIA's Data Usage Policies and Restrictions for additional details. Questions may be directed to help@cancerimagingarchive.net.

Citation

Data Citation
Vallières, Martin, Freeman, Carolyn R., Skamene, Sonia R., & El Naqa, Issam. (2015). A radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremities. The Cancer Imaging Archive. http://doi.org/10.7937/K9/TCIA.2015.7GO2GSKS

Publication Citation

Vallières, M., Freeman, C. R., Skamene, S. R., & Naqa, I. El. (2015, June 29). A radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremities. Physics in Medicine and Biology. IOP Publishing. http://doi.org/10.1088/0031-9155/60/14/5471

TCIA Citation

Clark K, Vendt B, Smith K, Freymann J, Kirby J, Koppel P, Moore S, Phillips S, Maffitt D, Pringle M, Tarbox L, Prior F. The Cancer Imaging Archive (TCIA): Maintaining and Operating a Public Information Repository, Journal of Digital Imaging, Volume 26, Number 6, December, 2013, pp 1045-1057. (paper)