Spatial heterogeneity in glioblastoma: linking single-cell and spatial multi-omics to vaccine immunotherapy

Abstract

Glioblastoma (GBM) remains the most lethal primary brain tumor, with survival still limited despite advances in multimodal treatment. The tumor’s profound heterogeneity--across patients, within individual tumors, and over time--has frustrated every class of therapeutic intervention. Recent developments in single-cell and spatial omics technologies have substantially advanced our understanding of this heterogeneity. No longer viewed as a uniform mass, GBM is now recognized as a spatially organized ecosystem of distinct malignant cell states, immune neighborhoods, and metabolic niches that interact across macroscopic distances. Spatial transcriptomics has revealed that malignant cells exist along a continuum of transcriptional programs--from neural progenitor-like to mesenchymal-like--that are not randomly distributed but organized into structured regions with a multi-layered architecture associated with hypoxia. Spatial proteomics and imaging mass cytometry have complemented these findings by mapping protein-level immune landscapes, uncovering cellular neighborhoods associated with survival and identifying spatially defined immunosuppressive niches. This review summarizes current knowledge from single-cell and spatial multi-omic studies of GBM, with an emphasis on the spatial organization of malignant cell states, the architecture of the tumor immune microenvironment, and the implications of these findings for therapeutic development, particularly vaccine immunotherapy. We argue that spatial technologies do not replace existing therapeutic pipelines but address a critical blind spot: the spatial dimension that bulk measurements cannot see.