Abstract
BACKGROUND: Neuroblastoma (NB) is the most common extracranial malignant tumor in children, presenting significant challenges in prognosis and treatment stratification. This study aims to identify novel biomarkers for NB and develop a prognostic model using comprehensive analytical methods, including Mendelian randomization (MR) analysis. METHODS: Utilizing bioinformatics and Mendelian randomization methods, we explored biomarkers associated with neuroblastoma at the mRNA level. We used chip expression data from the GEO database to screen for differentially expressed genes (DEGs) and conducted two-sample MR analysis using expression quantitative trait loci (eQTL) and neuroblastoma data from the IEU database to identify co-expressed genes through colocalization. A relevant prognostic model was constructed using lasso regression based on the co-expressed genes. Furthermore, we confirmed the correlation between high-risk and low-risk groups with the tumor microenvironment and immune cell infiltration. Subsequently, we evaluated the relationship between risk scores and sensitivity to immunotherapy and anti-tumor drugs. RESULTS: Differential analysis identified 485 downregulated and 349 upregulated genes that play important roles in NB. MR analysis identified 4 significant co-expressed genes associated with NB: CAV2, CTSK, LXN, and NDRG2. GO and KEGG enrichment analyses revealed that these genes are involved in crucial biological processes and pathways. A prognostic model based on these four genes was constructed, and its independence as a prognostic factor was confirmed. NB patients were divided into two different risk score groups, with survival analysis indicating that the high-risk group had poorer overall survival, lower immune infiltration, and poorer immune therapy response. In contrast, the low-risk group showed potential efficacy in immunotherapy and higher sensitivity to anti-tumor drugs. CONCLUSION: Our findings provide new insights into the molecular basis of NB, identifying four novel biomarkers and developing a risk scoring model based on four co-expressed genes. This model has the potential to become an effective tool for predicting prognosis and guiding treatment in NB patients.