Ferroptosis-related Recurrence Risk Model Predicts Clinical Outcomes and Immune Infiltration in Glioblastoma

Abstract

Objective To construct a ferroptosis-related glioblastoma (GBM) recurrence risk model and evaluate the prognosis of patients. Methods Differentially expressed genes (DEGs) related to ferroptosis in recurrent GBM were screened by CGGA and FerrDb databases. Key genes were obtained by Lasso regression. Then, nomogram was constructed according to the key risk genes, and the prediction efficiency was verified using the TCGA database. GO, KEGG, and GSEA databases were used in exploring the mechanism of prognosis. ESTIMATE and TIMER were used in studying tumor immune infiltration and the expression of immune check points. Results WWTR1, PLIN2, and BID were important prognostic factors for GBM recurrence. The nomogram was constructed according to gender and age, and the observed values were in good agreement with the predicted values. The AUC values were 0.65 (1 year), 0.66 (3 years), and 0.63 (5 years) for CGGA and 0.68 (1 year), 0.76 (3 years), and 0.79 (5 years) for TCGA. Epithelial mesenchymal transition, KRAS pathway, and inflammatory response were significantly upregulated in the high-risk subtypes (P < 0.05). Immune cell infiltration was lower (P < 0.05). Risk score was positively correlated with the expression of immunosuppression check points. Conclusion Ferroptosis-related genes WWTR1, PLIN2, and BID can be used in constructing a nomogram with good predictive performance. These risk genes may affect prognosis through tumor-infiltrating immune cells and immune check points.

Keywords

• glioblastoma
• ferroptosis
• recurrence prognosis risk model
• immune infiltration