Related Experiment Video
Updated: Aug 8, 2026
Establishing a Competing Risk Regression Nomogram Model for Survival Data
Published on: October 23, 2020
Development and validation of a nomogram for overall survival risk stratification in patients with brain metastases
Mengdie Zhao1, Yanqun Zhang1, Xiaoyu Zhong1
1The First Affiliated Hospital of Bengbu Medical University, Bengbu, Anhui, China.
Background:
Radiotherapy is a primary treatment modality for patients with brain metastases (BMs). However, individualized prediction of overall survival (OS) in this patient cohort remains a considerable clinical challenge. This study aimed to develop a prognostic model for estimating OS in patients with BMs receiving radiotherapy.
Methods:
A retrospective cohort study was conducted, enrolling 298 patients with radiologically confirmed BMs who received radiotherapy from January 2020 to December 2021. Least absolute shrinkage and selection operator (LASSO) Cox regression was used to identify key prognostic factors for OS. A nomogram was developed based on multivariate Cox regression models to predict 1-, 2-, and 3-year OS. The nomogram's performance was assessed using time-dependent receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA). A random survival forest (RSF) model was additionally used to evaluate the relative importance of each prognostic predictor.
Results:
LASSO regression identified six key prognostic variables: sex, primary tumor type, graded prognostic assessment (GPA) class, targeted/immunotherapy administration, primary tumor control status, and post-radiotherapy systemic therapy. All six variables were independently associated with OS. The nomogram showed moderate discriminative ability, with area under the curve (AUC) values of 0.773, 0.763, and 0.723 for 1-, 2-, and 3-year OS, respectively. Calibration plots showed reasonable agreement between predicted and observed OS outcomes. DCA suggested potential clinical net benefit across the assessed threshold probabilities. Using a total score cutoff of 224, patients were stratified into high-risk and low-risk groups, with significantly different OS between the two groups (P < 0.001). RSF analysis indicated that sex and primary tumor type were the most informative predictors of OS.
Conclusion:
The developed prognostic nomogram integrates readily available clinical and treatment-related characteristics to estimate OS in radiotherapy-treated patients with BMs. This model may assist individualized risk stratification and clinical decision-making, although further external validation is needed. An interactive dynamic nomogram is accessible online at: https://mengdie710.shinyapps.io/DynNomapp/.
