Graphical schematic of study design: Severity of 22 MDASI-BT symptoms reported by 1128 primary brain tumor (PBT) patients was assessed for redundancy, consolidating 2 pairs of redundant symptoms. The resulting severity dataset was used to construct symptom networks for the all-patient cohort, and patient subgroups identified by unsupervised clustering of symptom co-severities. The unsupervised clustering step presents an optional branching in the workflow; if taken this branch divides the cohort into groups, which are then submitted to the network construction branch and for assessment of clinical/demographic differences. Symptom clusters were discovered for each network, and symptoms were prioritized based on their importance in the network architecture.