An Online Calculator for the Prediction of Survival in Glioblastoma Patients Using Classical Statistics and Machine Learning
Problem
Group-level glioblastoma survival statistics are well defined, but predicting individual patient survival is difficult due to variation within strata.
Why it matters
Individual survival estimates inform treatment decisions for patients with an aggressive brain cancer.
What the AI did
Fifteen statistical and machine learning algorithms were trained on SEER data (20,821 patients, 13 features) to predict overall survival and 1-year status, with the best model deployed as an online calculator.
Results
20,821 patients met inclusion criteria; model performance metrics truncated in abstract.
Impact
Evidence supports a deployed online survival calculator; clinical impact not stated.
Record
| Venue | Neurosurgery |
|---|---|
| Published | 2019 |
| Work type | article |
| Authors | Joeky T. Senders, Patrick C. Staples, Alireza Mehrtash, David J. Coté, MARTIN J.B. TAPHOORN, David A. Reardon, William Brian Gormley, Timothy Richard Smith et al. |
| Citations | 127 (OpenAlex, at import time) |
| References | 32 |
| Model or method | 15 statistical and ML algorithms compared; accelerated failure time model (details truncated) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1093/neuros/nyz403 |
| Open access | hybrid |
| Keywords | glioblastoma, survival prediction, machine learning, SEER, prognostic calculator |
Importance rating
real but narrower advance
Individual survival estimates inform treatment decisions for aggressive brain cancer; abstract gives no affected-population scale, only dataset size (20,821 SEER patients).