This collection contains clinical data and computed tomography (CT) from 22 non-small cell lung cancer (NSCLC) radiotherapy patients. For 21 of these patients with pre-treatment CT scans, repeated blinded manual delineations by five different radiation oncologists of the 3D volume of the gross tumor volume on CT and clinical outcome data are available. The above was repeated with the same set of five radiation oncologists, using an in-house autosegmentation tool for initial delineation followed by manual adjustment of the primary gross tumor volume outline. For one patient, clinical data and CT was available but the tumor delineations were not extracted. This patient was included in this collection for the sake of completeness.
This dataset refers to the "Multiple delineation" dataset of the study published in Nature Communications (http://doi.org/10.1038/ncomms5006). In short, the publication used a radiomics approach to computed tomography data of 1,019 patients with lung or head-and-neck cancer. Radiomics refers to the comprehensive quantification of tumor phenotypes by applying a large number of quantitative image features. In the published analysis, 440 features quantifying tumor image intensity, shape, and texture were extracted. We found that a large number of radiomic features have prognostic power in independent data sets, many of which were not identified as significant before. Radiogenomics analysis revealed that a prognostic radiomic signature, capturing intra-tumor heterogeneity, was associated with underlying gene-expression patterns. These data suggest that radiomics identifies a general prognostic phenotype existing in both lung and head-and-neck cancer. This may have a clinical impact as imaging is routinely used in clinical practice, providing an unprecedented opportunity to improve decision-support in cancer treatment at low cost.
The delineations are provided in two formats; DICOM RTSTRUCT contains slice by slice contour points of the external outline of the primary tumour. DICOM SEGMENTATION contains binary masks of the same primary tumour. The nomenclature of the structures are as follows:
Side note : Radiation oncologists denoted “1” and “3” were trainee radiation oncologists at the time of this experiment. Radiation oncologists “2”, “4” and “5” were extensively experienced at the time of this experiment.
This dataset is intended to be open access to support repeatability and reproducibility of research in the radiomics domain. This dataset has been referenced in Medical Physics Dataset Article addressing FAIR radiomics practices to support transparency, harmonization and collaboration on radiomics (https://doi.org/10.1002/mp.14322).
Other data sets in the Cancer Imaging Archive that were used in the same study published in Nature Communications: NSCLC-Radiomics, NSCLC-Radiomics-Genomics, Head-Neck-Radiomics-HN1, RIDER Lung CT Segmentation Labels from: Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach (RIDER-LungCT-Seg).
For scientific or other inquiries about this dataset, please contact the TCIA Helpdesk.
We would like to acknowledge the individuals and institutions that have provided data for this collection: