A central goal in single-cell cancer research is to identify phenotypically distinct cells that have prognostic significance for patients. This need is acute in brain tumor glioblastoma, where changes in gene sequence, copy number or transcript expression can define patient or cell subgroups, but these subgroups are not associated with large differences in patient clinical outcomes. I will present recent work using suspension mass cytometry and Imaging Mass Cytometry™ to study glioblastoma patient tissue samples, identify prognostic subpopulations of cancer and immune cells and locate these cells spatially, within both tumors and the brain. I will also discuss our ongoing work using mass cytometry to understand how tumor contact with a brain stem cell niche may alter the properties of cancer and immune cells in these tumors and impact patient survival.