Prognostic Value of a Ferroptosis-Related Gene Signature in Patients With Head and Neck Squamous Cell Carcinoma

Abstract

Background: Ferroptosis is an iron-dependent programmed cell death (PCD) form that plays a crucial role in tumorigenesis and might affect the antitumor effect of radiotherapy and immunotherapy. This study aimed to investigate distinct ferroptosis-related genes, their prognostic value and their relationship with immunotherapy in patients with head and neck squamous cell carcinoma (HNSCC).Methods: The differentially expressed ferroptosis-related genes in HNSCC were filtered based on multiple public databases. To avoid overfitting and improve clinical practicability, univariable, least absolute shrinkage and selection operator (LASSO) and multivariable Cox algorithms were performed to construct a prognostic risk model. Moreover, a nomogram was constructed to forecast individual prognosis. The differences in tumor mutational burden (TMB), immune infiltration and immune checkpoint genes in HNSCC patients with different prognoses were investigated. The correlation between drug sensitivity and the model was firstly analyzed by the Pearson method.Results: Ten genes related to ferroptosis were screened to construct the prognostic risk model. Kaplan-Meier (K-M) analysis showed that the prognosis of HNSCC patients in the high-risk group was significantly lower than that in the low-risk group (P < 0.001), and the area under the curve (AUC) of the 1-, 3- and 5-year receiver operating characteristic (ROC) curve increased year by year (0.665, 0.743, and 0.755). The internal and external validation further verified the accuracy of the model. Then, a nomogram was build based on the reliable model. The C-index of the nomogram was superior to a previous study (0.752 vs. 0.640), and the AUC (0.729 vs. 0.597 at 1 year, 0.828 vs. 0.706 at 3 years and 0.853 vs. 0.645 at 5 years), calibration plot and decision curve analysis (DCA) also shown the satisfactory predictive capacity. Furthermore, the TMB was revealed to be positively correlated with the risk score in HNSCC patients (R = 0.14; P < 0.01). The differences in immune infiltration and immune checkpoint genes were significant (P < 0.05). Pearson analysis showed that the relationship between the model and the sensitivity to antitumor drugs was significant (P < 0.05).Conclusion: Our findings identified potential novel therapeutic targets, providing further potential improvement in the individualized treatment of patients with HNSCC.

Clustering and machine learning-based integration identify cancer associated fibroblasts genes’ signature in head and neck squamous cell carcinoma

“…In addition, compared to the conventional tools such as age, gender, stage and HPV status for evaluating clinical outcomes, the risk score signature worked independently of these factors and had significantly superior efficiency in predicting prognosis in training and validation cohorts. We also reviewed previous published HNSCC-related risk models which including different genes’ combination ( Liu et al, 2020 ; Li et al, 2021b ; He et al, 2021 ; Wang et al, 2022a ; Huang et al, 2022b ; Wang et al, 2022b ; Chen et al, 2022 ; Chi et al, 2022 ; Du et al, 2022 ; Han et al, 2022 ; Peng et al, 2022 ), among these, none of them presented better AUC value performance than our model. Therefore, our risk score signature could be a promising surrogate for evaluating the prognosis of HNSCC in clinical practice.…”

Identification of Ferroptosis-Related lncRNA Pairs for Predicting the Prognosis of Head and Neck Squamous Cell Carcinoma

“…e comparison between the established models was utilized to assess the forecasting ability of the novel model [16,[24][25][26][27][28].…”

Identification of new head and neck squamous cell carcinoma subtypes and development of a novel score system (PGSscore) based on variations in pathway activity between tumor and adjacent non-tumor samples

“…Most previous studies have largely focused on one pathway, gene set, lncRNA set, or a single molecule that is considered to be associated with tumor development. For example, pyroptosis-related genes [11] , pyroptosis-related lncRNA [12] , ferroptosis-related genes [13] , m6A-modified lncRNA [14] , immune-related genes [15] , aging-related genes [16] , hypoxia-related genes [17] , autophagy-related genes [18] , inflammatory pathways [19] , fatty acid metabolism pathway [20] , Fascin Actin-bundling Protein 1 [21] , and dual-specificity phosphatase 1 [22] have been used as a basis to predict HNSC outcomes. Hence, these publications tend to choose single emerging or popular gene sets to establish a molecular signature or model to predict outcomes, while ignoring other gene sets.…”