Combining multimodal imaging and treatment features improves machine learning‐based prognostic assessment in patients with glioblastoma multiforme

MRI-based features were the most relevant feature class for prognostic assessment. Combining clinical, pathological, and imaging information increased predictive power for OS and PFS. A further increase was achieved by adding treatment features.