Multimodal Integration of Transcriptomics, Proteomics and Radiomics Improves Prediction of Recurrence in Patients with IDH-Mutant Glioma

Isocitrate dehydrogenase mutant gliomas remain lethal brain cancers which impair quality of life in young adults. These tumors are molecularly and cellularly heterogeneous and have a wide range of survival prognoses. Consequently, the identification of patients at risk of early recurrence remains an unmet need. Here, we analyzed imaging, transcriptomic, and proteomic profiles using machine learning to 1) describe the biological characterization of subtypes of IDH-mutant gliomas categorized by PET and histology, 2) reinforce the integration of PET metrics in the classification of IDH-mutant gliomas, and 3) improve patient stratification with novel signatures of patient risk of recurrence based gene expression, protein level, and imaging. Our integrative analysis provides a better stratification of IDH-mutant gliomas patients and their risk of recurrence, which will lead to a better monitoring of the clinical evolution of the disease.