We consider the multiscale model for glioma growth introduced in (Math. Biosci. Eng. 71: 443–460, 2016) to accommodate tumor heterogeneity by relying on the go-or-grow dichotomy and extend it to account for therapy effects. Thereby, three treatment strategies involving surgical resection, radio-, and chemotherapy are compared for their efficiency. The chemotherapy relies on inhibiting the binding of cell surface receptors to the surrounding tissue, which impairs both migration and proliferation. The multiscale features of our model allow to connect subcellular level information to individual cell dynamics and—upon scaling—carry over such information to the population level on which a tumor is clinically observed. This makes it particularly appropriate for investigating the effects of therapy, as both ionizing radiation and chemotherapeutic agents act on the subcellular level, but their outcome is assessed on the macroscopic scale. The model includes patient-specific brain structure available in the form of DTI data and the numerical simulations are performed relying on these.