Glioblastoma is the most frequent and aggressive type of primary brain tumor. Medical doctors are very interested in the development of prognostic metrics that can provide advance information on patient survival to chose therapies accordingly
In a paper just published in the journal European Radiology, a group of MOLAB researchers together with coworkers from seven hospitals involved in the TOG project developed simple metrics based on meaningful morphological MRI features based on mathematical principles that outperformed all predictive algorithms to date. Specially the simple metrics developed were superior to all previous 'blind' machine learning strategies.
MOLAB researchers are continuing the work in the framework of TOG project to design algorithms able to personalize current therapies and indentify potential responders to novel therapies.
Link to the paper:
Morphological MRI-based features provide pretreatment and post-surgery survival prediction in glioblastoma
J. Pérez-Beteta, D. Molina-García, A. Martínez-González, M. Amo, A. Henares-Molina, B. Luque, E. Arregui, M. Calvo, J.M.Borrás, J. Martino, C. Velasquez, B. Meléndez, A. R. de Lope, R. Moreno, J. A. Barcia, B. Asenjo, M. Benavides, I. Herruzo, P. C. Lara, R. Cabrera, D. Albillo, M. Navarro, L. A. Pérez-Romasanta, A. Revert, E. Arana, V. M. Pérez-García
European Radiology (2018)