We present a method for automatic segmentation of high-grade gliomas and their subregions from multi-channel MR images. Be-sides segmenting the gross tumor, we also differentiate between active cells, necrotic core, and edema. Our discriminative approach is based on decision forests using context-aware spatial features, and integrates a generative model of tissue appearance, by using the probabilities ob-tained by tissue-specific Gaussian mixture models as additional input for the forest. Our method classifies the individual tissue types simulta-neously, which has the potential to simplify the classification task. The approach is computationally efficient and of low model complexity. The validation is performed on a labeled database of 40 multi-channel MR images, including DTI. We assess the effects of using DTI, and vary-ing the amount of training data. Our segmentation results are highly accurate, and compare favorably to the state of the art.