Radiomics and radiogenomics

Brain tumor is the most frequently encountered pediatric tumor. The Herby project, lead by the Institut Gustave Roussy, aims at applying radiomic, and further genomic, analysis on pediatric glioma and invasive brain stem glioma.

The group is in charge of the servicing of three project cohorts which are composed of 117 patients with pediatric glioma (HERBY), but also 50 (PREBIOMEDE) and 38 (BIOMEDE) young patients with diffuse brain stem tumors.
For now, available data are composed of clinical tables and medical images (mostly T1Gd and FLAIR, but also T1, T2, diffusion, perfusion and asl).

Future prospects. The full pipeline we implemented complies with BIDS data reorganization, data pre-processing (based on state-of-the-art methods such as [1]), feature extraction and radiomics analysis. Each level of the pipeline is followed by a visual quality control step. Thus, we propose a solid base adaptable to other cohort, such as metastases or primary central nervous system lymphomas. The next step will be to study the prognostic performance of classifiers

[1] Shinohara, Russell T., et al. “Statistical normalization techniques for magnetic resonance imaging.” NeuroImage: Clinical 6 (2014): 9-19.