Summary

This data container describes both computer-aided and manually-corrected segmentation labels for the pre-operative multi-institutional scans of The Cancer Genome Atlas (TCGA) Glioblastoma Multiforme (GBM) collection, publicly available in The Cancer Imaging Archive (TCIA), coupled with a rich panel of radiomic features along with their corresponding skull-stripped and co-registered multimodal (i.e. T1, T1-Gd, T2, T2-FLAIR) magnetic resonance imaging (MRI) volumes in NIfTI format.

Pre-operative multimodal MRI scans were identified in the TCGA-GBM collection via radiological assessment. These scans were initially skull-stripped and co-registered, before their tumor segmentation labels were produced by an automated hybrid generative-discriminative method, ranked first during the International multimodal BRAin Tumor Segmentation challenge (BRATS 2015). These segmentation labels were revised and any label misclassifications were manually corrected by an expert board-certified neuroradiologist. The final labels were used to extract a rich panel of imaging features, including intensity, volumetric, morphologic, histogram-based and textural parameters, as well as spatial information and diffusion properties extracted from glioma growth models. The generated computer-aided and manually-revised labels enable quantitative computational and clinical studies without the need to repeat manual annotations whilst allowing for comparison across studies. They can also serve as a set of manually-annotated gold standard labels for performance evaluation in computational challenges. The provided panel of radiomic features may facilitate research integrative of the molecular characterization offered by TCGA, and hence allow associations with molecular markers, clinical outcomes, treatment responses and other endpoints, by researchers without sufficient computational background to extract such features.