Machine-Learning-Based Survival Prediction in Glioblastoma Using Graph-Theoretical Analysis of Structural Network Alterations
Recent Publication
A recent study in Cancers shows that glioblastomas should not only be viewed as a single tumour, but as a disease that disrupts central networks in the brain. The researchers led by Prof. Dr Andreas Stadlbauer analysed MRI data from 871 patients and were able to show that changes in important nerve connections are closely linked to survival. Disruptions in network nodes of the temporal lobe and in other central brain regions were particularly significant. Based on this network data, the models were able to predict with high accuracy whether patients would survive longer than one year. The results thus open up new perspectives for prognosis, treatment planning and a more networked view of this aggressive brain tumour disease. The work was funded by the Research Impulse programme of the state of Lower Austria and the KL Open Access Publication Fund.
Stadlbauer, A., Oberndorfer, S., Heinz, G., Marhold, F., Kinfe, T. M., Dorostkar, M., Schnell, O., Meyer-Bäse, U., & Meyer-Bäse, A. (2026). Machine-Learning-Based Survival Prediction in Glioblastoma Using Graph-Theoretical Analysis of Structural Network Alterations. Cancers, 18(7), 1161. https://doi.org/10.3390/cancers18071161
Prof. Dr. Andreas Stadlbauer
Institute of Diagnostic and Interventional Radiology (University Hospital St. Pölten)
Prim. Univ.-Prof. PD Dr. Stefan Oberndorfer FEAN
Division of Neurology (University Hospital St. Pölten)
Prim. Univ.-Prof. Dr. Gertraud Heinz MBA
Institute of Diagnostic and Interventional Radiology (University Hospital St. Pölten)
OA Dr. Mario Dorostkar PhD
Institute of Clinical Pathology and Molecular Pathology of the Lower Austria Central Region (University Hospital St. Pölten)