Skip to main content
Fig. 5 | Biology Direct

Fig. 5

From: Large-scale bulk and single-cell RNA sequencing combined with machine learning reveals glioblastoma-associated neutrophil heterogeneity and establishes a VEGFA+ neutrophil prognostic model

Fig. 5

Development and evaluation of VNRS prediction model. (A) The figure shows the univariate analysis results of differentially expressed genes in the VEGFA+GBMAN subgroup. The right figure shows the expression of each gene in the TCGA-GBM cohort. (B) Heatmap summarizing the construction of 117 predictive models and their corresponding concordance index (C-index) across three large datasets. The C-index quantifies the predictive accuracy of each model, with higher values indicating better performance in discriminating between outcomes. (C) Table displaying the results of univariate regression analysis for the predictive model in three independent cohorts. (D) Kaplan-Meier curves comparing OS between high- and low-risk groups as defined by the predictive model. The curves are shown for both the training set (used to develop the model) and validation sets (used to test its generalizability), demonstrating the model’s ability to stratify patients based on risk. (E) ROC curves evaluating the predictive performance of the model at 1-, 2-, and 3-year time points in the three large cohorts. The AUC values are reported, reflecting the model’s accuracy in predicting survival outcomes at different time intervals

Back to article page

Biology Direct

ISSN: 1745-6150

Contact us