data-code

BackgroundRetinoblastoma (RB) is the most common intraocular malignancy in children, with its oncogenesis and progression tightly linked to RB1 gene mutations. Despite advances in clinical management, the treatment of advanced and drug-resistant RB remains a formidable challenge. Ferroptosis, an iron-dependent form of regulated cell death, has emerged as a promising target for tumor suppression. However, its precise mechanism of action and associated biomarkers in RB have not been fully elucidated.MethodsIn this study, we integrated multiple RB transcriptome datasets (GSE24673, GSE97508, GSE208143, and GSE110811) retrieved from the Gene Expression Omnibus (GEO) database. Ferroptosis-related differentially expressed genes (DEGs) were screened by intersecting results from differential expression analysis, weighted gene co-expression network analysis (WGCNA), and a ferroptosis gene set. Two machine learning algorithms, Least Absolute Shrinkage and Selection Operator (LASSO) regression and Support Vector Machine-Recursive Feature Elimination (SVM-RFE), were jointly applied to identify core genes. The diagnostic efficacy of core genes was evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC) values. Immune cell infiltration patterns and their correlations with core gene expression were analyzed using CIBERSORT. Furthermore, potential targeted therapeutic agents were predicted based on the Drug Signature Database (DSigDB).ResultsA total of 15 ferroptosis-related differentially expressed genes were identified, including CDCA3, ENO3, FTL, HELLS, KIF20A, LINC00472, MAPK3, MAPKAP1, NOX4, NT5DC2, PPARA, PTGS2, RRM2, SQSTM1, and TFAP2A. Cross-validation via LASSO regression and SVM-RFE algorithms yielded three core genes: NT5DC2, CDCA3, and SQSTM1. Expression analysis revealed that NT5DC2 and CDCA3 were significantly upregulated in RB tissues, and NT5DC2 expression levels exhibited a strong positive correlation with GPX4 (r=0.86), a canonical regulator of ferroptosis. ROC analysis demonstrated that all three core genes possessed robust diagnostic performance, with AUC values consistently exceeding 0.7 in both training and validation cohorts. Immune infiltration analysis indicated that the expression of these core genes was significantly correlated with the abundance of specific immune cell subsets, including M0/M2 macrophages and CD8⁺ T cells. Drug prediction highlighted several ferroptosis inducers, such as gossypol and Ferroptocide, as potential therapeutic candidates for RB.ConclusionFor the first time, we systematically identified NT5DC2, CDCA3, and SQSTM1 as key ferroptosis-related biomarkers in RB through an integrated bioinformatics and machine learning approach. These genes not only exhibit excellent diagnostic potential but also are closely associated with the tumor immune microenvironment. Our findings provide novel insights into the molecular mechanisms underlying RB and lay a theoretical foundation for the development of ferroptosis-targeted diagnostic and therapeutic strategies for this devastating pediatric malignancy.