• Joeky T Senders, Patrick Staples, Alireza Mehrtash, David J Cote, TaphoornMartin J BMJBDepartment of Neurology, Haaglanden Medical Center, The Hague, The Netherlands., David A Reardon, William B Gormley, Timothy R Smith, Marike L Broekman, and Omar Arnaout.
• Computational Neuroscience Outcomes Center (CNOC), Department of Neurosurgery, Brigham and Women's Hospital, School of Medicine, Harvard University, Boston, Massachusetts.
• Neurosurgery. 2020 Feb 1; 86 (2): E184E192E184-E192.

BackgroundAlthough survival statistics in patients with glioblastoma multiforme (GBM) are well-defined at the group level, predicting individual patient survival remains challenging because of significant variation within strata.ObjectiveTo compare statistical and machine learning algorithms in their ability to predict survival in GBM patients and deploy the best performing model as an online survival calculator.MethodsPatients undergoing an operation for a histopathologically confirmed GBM were extracted from the Surveillance Epidemiology and End Results (SEER) database (2005-2015) and split into a training and hold-out test set in an 80/20 ratio. Fifteen statistical and machine learning algorithms were trained based on 13 demographic, socioeconomic, clinical, and radiographic features to predict overall survival, 1-yr survival status, and compute personalized survival curves.ResultsIn total, 20 821 patients met our inclusion criteria. The accelerated failure time model demonstrated superior performance in terms of discrimination (concordance index = 0.70), calibration, interpretability, predictive applicability, and computational efficiency compared to Cox proportional hazards regression and other machine learning algorithms. This model was deployed through a free, publicly available software interface (https://cnoc-bwh.shinyapps.io/gbmsurvivalpredictor/).ConclusionThe development and deployment of survival prediction tools require a multimodal assessment rather than a single metric comparison. This study provides a framework for the development of prediction tools in cancer patients, as well as an online survival calculator for patients with GBM. Future efforts should improve the interpretability, predictive applicability, and computational efficiency of existing machine learning algorithms, increase the granularity of population-based registries, and externally validate the proposed prediction tool.Copyright © 2019 by the Congress of Neurological Surgeons.

23 keywords...

hide…