Related Experiment Video
Updated: Jul 3, 2026
06:19
Constructing and Visualizing Models using Mime-based Machine-learning Framework
Published on: July 22, 2025
Improved prognostic survival models for pediatric medulloblastoma using high dimensional gene expression data
Elizabeth B Amona1, Mst Sharmin Akter Sumy2, Tyler Jones1
1Biostatistics Core, Brown Cancer Center, University of Louisville, Louisville, KY, USA.
BMC Medical Genomics
|July 2, 2026
Summary
This study identifies key genes to improve medulloblastoma survival predictions using advanced statistical models. The findings offer a robust and interpretable framework for personalized cancer treatment strategies.
Area of Science:
- Oncology
- Genomics
- Biostatistics
Background:
- Medulloblastoma (MB) is stratified into WNT, SHH, Group 3, and Group 4 subgroups, with recent proposals for a combined Non-WNT/Non-SHH category.
- Accurate prognostic markers are crucial for tailoring treatment in pediatric and young adult MB patients.
Purpose of the Study:
- To identify genes that enhance prognostic accuracy for medulloblastoma survival.
- To develop and evaluate a multi-stage framework for prognostic gene identification and survival modeling.
Main Methods:
- Gene screening using Benjamini-Hochberg adjusted Cox regression at varying False Discovery Rate (FDR) thresholds (1%-6%).
- Evaluation of multiple survival models (LASSO, Elastic Net, Ridge, SCAD, MCP, PCA-Cox, Random Survival Forests) via ten-fold cross-validation.
- Utilized Integrated Brier Score for calibration and concordance index for discrimination.
Main Results:
- The 6% FDR Elastic Net model balanced predictive accuracy and model sparsity, reducing gene set from 146 to 49.
- Identified genes associated with poorer survival (e.g., FKBP4, GATA3) and improved survival (e.g., ZNF774, COX10).
- Demonstrated a clear prognostic gradient based on gene-level effects.
Conclusions:
- Combining FDR-based screening with Elastic Net-penalized Cox modeling provides a robust, parsimonious, and interpretable prognostic framework for medulloblastoma.
- This approach achieves strong predictive performance in high-dimensional genomic settings for medulloblastoma.