Abstract BACKGROUND Glioblastoma (GBM) is the most aggressive and common brain tumor. Despite multimodal therapy which consists of complete tumor removal, chemo- and radiation therapy in most patients, the tumor recurs, and the overall survival of only 14 months remains very poor. To identify tumor-induced immune evasion and treatment resistance mechanisms induced during GBM recurrence, we aimed at characterizing the immunological/molecular changes in patient-matched treatment-naïve and recurrent GBM samples. MATERIAL AND METHODS Proteins and total RNA were extracted from 16 patient-matched treatment-naïve (primary) and recurrent GBM fresh frozen tumor samples. The expression of genes involved in brain immune responses were compared using Nanostring panels encompassing Cancer and Neuroinflammation genes. Proteomic was conducted using 6 patient-matched primary and recurrent tumor samples. Immunostaining was performed on a tissue microarray containing matched primary and recurrent tumor samples to identify tumor cells and microglia in order to assess spatial transcriptome of these specific cell types in situ using GeoMX technology. Moreover, correlation analysis of differentially genes and proteins with time to relapse was performed. RESULTS Genes significantly differentially expressed between primary and recurrent tumors revealed genes overexpressed in recurrent tumors belonging to “FCGR-mediated Phagocytosis” and “synapse pruning” pathways. Proteomic confirmed the synaptic signalling considering overexpressed proteins. Correlation studies with the time to relapse indicated that higher expression of FCG receptors, complement molecules and synaptic components correlate with shorter time to relapse, sharing similarities with neurodegenerative/cognitive disorders such as schizophrenia and Alzheimer disease. The source for these molecules is under evaluation with the spatial transcriptomic analysis. CONCLUSION Upregulation of specific proteins such as FCG receptors, complement and synaptic components may result in more favourable conditions for tumor regrowth and may therefore be identified as negative predictors patient outcome. Interfering with these molecules may represent novel therapeutic options after initial GBM surgery.