A novel nomogram and risk stratification system for cancer-specific survival of elderly patients with glioblastoma:a SEER-based study

WU Yujuan; MEI Li; HUANG Guanyou

doi:10.3639/j.issn.1671-2897.2025.04.010

Abstract：Objective To develop a nomogram and risk stratification system to predict the cancer-specific survival(CSS)of elderly patients with glioblastoma multiforme(GBM).Methods A total of 5 634 elderly GBM patients diagnosed between 2004 and 2021 were identified from the SEER(Surveillance,Epidemiology,and End Results)database and randomly divided into a training set(3 946 cases)and a validation set(1 688 cases)in a 7:3 ratio.Independent prognostic variables were identified through univariate and multivariate Cox regression analyses,and these variables were used to construct a nomogram to predict CSS at 1-,2-,and 3-years.The predictive accuracy of the nomogram was assessed using the area under the time-dependent receiver operating characteristic curve(AUC)and calibration curves.Additionally,a risk stratification system was developed based on each patient's total nomogram score.Results Six independent prognostic factors were identified through univariate and multivariate Cox regression analyses for constructing the nomogram,including age,tumor site,tumor size,chemotherapy,radiotherapy and surgery.Based on these factors,the nomogram achieved AUC values of 0.766,0.714 and 0.669 for predicting 1-,2-,and 3-year CSS in elderly GBM patients.Calibration curves indicated strong agreement between observed and nomogram-predicted CSS values.Furthermore,a risk stratification system was developed based on the quartiles of the total risk score,categorizing patients into four risk groups:low risk,low-medium risk,high-medium risk,and high risk.Significant differences in survival rates were observed among the risk groups in both the training and validation sets(P<0.0001),with survival rates as follows:low-risk group>low-medium risk group>high-medium risk group>high-risk group.Conclusion The developed nomogram and corresponding risk stratification system demonstrates high predictive accuracy for CSS in elderly GBM patients,provides clinicians with a valuable tool for predicting survival probabilities and informs treatment decisions through effective risk stratification.