Development and validation of a novel palmitoylation-related prognostic signature in head and neck squamous cell carcinoma.

Background: Palmitoylation, a reversible lipid modification, plays a critical role in protein trafficking and signaling and has been implicated in various cancers. However, its function in head and neck squamous cell carcinoma (HNSCC)-a highly aggressive and heterogeneous malignancy-remains largely unexplored. This study aimed to establish a prognostic model based on palmitoylation to improve risk stratification and support clinical decision-making in HNSCC patients. Methods: The HNSCC training and validation cohorts were obtained from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO, GSE41613), respectively. A total of 4,052 palmitoylation-related genes (PRGs) were curated from the Molecular Signatures Database (MsigDB), and 282 differentially expressed PRGs were identified. Univariate Cox regression analysis yielded 49 prognostic genes, of which 16 were further selected by least absolute shrinkage and selection operator (LASSO) regression. Subsequent multivariate Cox analysis led to the construction of an 8-gene PRG-based risk model. Clinical variables, including age, sex, stage, and risk score, were integrated into a nomogram. Kaplan-Meier survival curves, time-dependent receiver operating characteristics (ROC) analyses, immune infiltration profiling, gene set enrichment analysis (GSEA), and drug sensitivity analysis were performed to assess the model's clinical utility. Results: The model effectively stratified patients into high- and low-risk groups with significantly different overall survival (OS) outcomes. Validation in an independent GEO dataset confirmed its robustness. A prognostic nomogram integrating the PRG-based risk score with clinical variables further enhanced predictive performance, achieving an area under curve (AUC) of 0.709, which was superior to stage (0.578), age (0.568), T stage (0.539), N stage (0.512), and sex (0.469). Immune analysis revealed that the low-risk group had higher ImmuneScores and ESTIMATEScores, along with reduced tumor purity, suggesting stronger immune cell infiltration. GSEA indicated enrichment of cancer-related pathways in the high-risk group. Drug sensitivity analysis based on the Genomics of Drug Sensitivity in Cancer (GDSC) database showed that low-risk patients had lower half-maximal inhibitory concentration (IC50) values for several targeted agents. Conclusions: Our findings highlight the prognostic significance of PRGs in HNSCC and suggest their potential utility in guiding personalized treatment strategies. Specifically, the PRG-based risk model demonstrated superior predictive performance compared to conventional clinical factors, supporting its value in improving patient stratification. Furthermore, the associations with immune infiltration patterns and drug sensitivity indicate that PRGs may not only serve as prognostic biomarkers but also help identify patients who could benefit from targeted or immunotherapeutic approaches.