Frontiers
Browse

Image_1_Development and validation of a GRGPI model for predicting the prognostic and treatment outcomes in head and neck squamous cell carcinoma.jpeg

Download (865.11 kB)
figure
posted on 2023-01-12, 05:09 authored by Fei Han, Hong-Zhi Wang, Min-Jing Chang, Yu-Ting Hu, Li-Zhong Liang, Shuai Li, Feng Liu, Pei-Feng He, Xiao-Tang Yang, Feng Li
Background

Head and neck squamous cell carcinoma (HNSCC) is among the most lethal and most prevalent malignant tumors. Glycolysis affects tumor growth, invasion, chemotherapy resistance, and the tumor microenvironment. Therefore, we aimed at identifying a glycolysis-related prognostic model for HNSCC and to analyze its relationship with tumor immune cell infiltrations.

Methods

The mRNA and clinical data were obtained from The Cancer Genome Atlas (TCGA), while glycolysis-related genes were obtained from the Molecular Signature Database (MSigDB). Bioinformatics analysis included Univariate cox and least absolute shrinkage and selection operator (LASSO) analyses to select optimal prognosis-related genes for constructing glycolysis-related gene prognostic index(GRGPI), as well as a nomogram for overall survival (OS) evaluation. GRGPI was validated using the Gene Expression Omnibus (GEO) database. A predictive nomogram was established based on the stepwise multivariate regression model. The immune status of GRGPI-defined subgroups was analyzed, and high and low immune groups were characterized. Prognostic effects of immune checkpoint inhibitor (ICI) treatment and chemotherapy were investigated by Tumor Immune Dysfunction and Exclusion (TIDE) scores and half inhibitory concentration (IC50) value. Reverse transcription-quantitative PCR (RT-qPCR) was utilized to validate the model by analyzing the mRNA expression levels of the prognostic glycolysis-related genes in HNSCC tissues and adjacent non-tumorous tissues.

Results

Five glycolysis-related genes were used to construct GRGPI. The GRGPI and the nomogram model exhibited robust validity in prognostic prediction. Clinical correlation analysis revealed positive correlations between the risk score used to construct the GRGPI model and the clinical stage. Immune checkpoint analysis revealed that the risk model was associated with immune checkpoint-related biomarkers. Immune microenvironment and immune status analysis exhibited a strong correlation between risk score and infiltrating immune cells. Gene set enrichment analysis (GSEA) pathway enrichment analysis showed typical immune pathways. Furthermore, the GRGPIdel showed excellent predictive performance in ICI treatment and drug sensitivity analysis. RT-qPCR showed that compared with adjacent non-tumorous tissues, the expressions of five genes were significantly up-regulated in HNSCC tissues.

Conclusion

The model we constructed can not only be used as an important indicator for predicting the prognosis of patients but also had an important guiding role for clinical treatment.

History