Identification of ionizing-radiation-exposure-associated hub genes and a five-gene prognostic signature for pediatric medulloblastoma: An integrated bioinformatics study

This is the authors' abstract. We don't add key points for this paper.

Background Medulloblastoma is the most common pediatric malignant brain tumor, and radiotherapy is an important component of treatment. Molecular features associated with radiation-exposure responses and prognosis remain incompletely characterized. Methods We analyzed paired irradiated and unexposed human lymphoblastoid samples from GSE197236 (6 pairs) to identify ionizing-radiation-exposure-associated differentially expressed genes, then evaluated their prognostic associations in 103 pediatric medulloblastomas from GSE85217. Cox and LASSO regression were used to construct an internally evaluated prognostic signature, and PPI analysis was used for exploratory network-based candidate prioritization. Results A total of 947 differentially expressed genes were identified, predominantly associated with p53 signaling, cell-cycle regulation, DNA-damage repair, and apoptotic pathways. A five-gene prognostic model comprising CCNB1, CDKN1A, RRM2, TOP2A, and BIRC5 separated patients with different overall survival and showed internally estimated time-dependent AUCs of 0.834, 0.812, and 0.786 at 1, 3, and 5 years, respectively. Bootstrap 95% confidence intervals were 0.764-0.901, 0.730-0.879, and 0.681-0.862. The apparent C-index was 0.781 and the optimism-corrected C-index was 0.748. These values were generated in the same development cohort and are not independent validation estimates. Six network-prioritized hub-gene candidates were associated with molecular subgroup or survival in exploratory analyses. Conclusion Our study identified ionizing-radiation-exposure-related differentially expressed genes and developed a five-gene prognostic signature that robustly stratifies pediatric medulloblastoma patients by overall survival, with promising internal predictive performance. The signature, along with six network-prioritized hub-gene candidates, may inform future risk-adapted treatment strategies and therapeutic target discovery.