Sarcopenia-driven gene model as a clinically actionable prognostic signature for head and neck squamous cell carcinoma.

[BACKGROUND] Sarcopenia is increasingly recognized as a critical factor that can diminish treatment tolerance, exacerbate therapy-related toxicity, and ultimately impair clinical outcomes in cancer patients. This highlights the need to move beyond prognostic modeling and focus on its impact on treatment decisions and outcomes. Regarding head and neck squamous cell carcinoma (HNSCC), the association between sarcopenia and clinical outcomes is still not well characterized. Our study is therefore designed to bridge this gap and further explore the underlying multi-omics mechanisms, with the ultimate goal of informing treatment personalization and intervention strategies to improve therapeutic efficacy and patient survival in HNSCC.

[METHODS] A total of 167 HNSCC patients from The Cancer Genome Atlas (TCGA) with paired computed tomography (CT) imaging and genomic data were analyzed. Sarcopenia assessment was based on the measurements of skeletal muscle area at the level of the third cervical vertebra (C3) from baseline CT scans. Overall survival (OS) was compared using Cox regression. Multi-omics differences (gene expression, transcriptomics, and immune microenvironment) were examined. A Genetic Risk Score (GRS) derived via least absolute shrinkage and selection operator (LASSO) regression was validated in the Gene Expression Omnibus (GEO) database under accession number GSE65858.

[RESULTS] Patients with sarcopenia demonstrated significantly worse OS [hazard ratio (HR) =2.36; 95% confidence interval (CI): 1.33-4.19; P=0.003]. Transcriptomic profiling revealed distinct sarcopenia-associated gene signatures in HNSCC, with pathway enrichment analysis implicating inflammatory responses and metabolic dysregulation in disease pathogenesis. A robust four-gene prognostic signature comprising fibrinogen beta chain (), kinesin family member 1A (), hepatic leukemia factor (), and yippee-like 1 () was derived from LASSO regression and demonstrated strong predictive value for survival (HR =2.86; 95% CI: 1.73-4.73; P<0.001), with subsequent confirmation in an independent validation cohort. Characteristic depletion of CD8 and CD4 T cell infiltration was observed in sarcopenic patients.

[CONCLUSIONS] Sarcopenia is significantly associated with poorer prognosis in HNSCC. Sarcopenic patients exhibit distinct gene expression, transcriptomic profiles related to lipid metabolism, and immune infiltration levels compared to non-sarcopenic patients. A GRS model constructed with four genes demonstrated favorable performance in predicting prognosis in HNSCC.