In this study we assessed the clinical significance of an epithelial-mesenchymal transition (EMT) gene signature and explored its association with the tumor microenvironment related to immunotherapy in patients with head and neck squamous cell carcinoma (HNSCC). Genes were selected when mRNA levels were positively or negatively correlated with at least one well-known EMT marker. We developed an EMT gene signature consisting of 82 genes. The patients were classified into epithelial or mesenchymal subgroups according to EMT signature. The clinical significance of the EMT signature was validated in three independent cohorts and its association with several immunotherapy-related signatures was investigated. The mesenchymal subgroup showed worse prognosis than the epithelial subgroup, and significantly elevated PD-1, PD-L1, and CTLA-4 levels, and increased interferon-gamma, cytolytic, T cell infiltration, overall immune infiltration, and immune signature scores. The relationship between PD-L1 expression and EMT status in HNSCC after treatment with TGF-β was validated in vitro. In conclusion, the EMT gene signature was associated with prognosis in HNSCC. Additionally, our results suggest that EMT is related to immune activity of the tumor microenvironment with elevated immune checkpoint molecules.

SOX2 copy number and mRNA expression were analysed to examine the clinical significance of SOX2 activation in HNSCC. Gene expression signatures reflecting SOX2 activation were identified in an HNSCC cohort. Patients with HNSCC were classified into two subgroups according to the gene expression signature: SOX2-high and SOX2-low. The clinical significance of SOX2 activation was further validated in two independent cohorts. Moreover, clinical significance of SOX2 activation in response to radiotherapy was assessed in patients with HNSCC. The relationship between SOX2 activation and radiotherapy was validated in an in vitro experiment. Patients in the SOX2-high subgroup had a better prognosis than patients in the SOX2-low subgroup in all three patient cohorts. Results of multivariate regression analysis showed that SOX2 signature was an independent predictor of the overall survival of patients with HNSCC (hazard ratio, 1.45; 95% confidence interval, 1.09–1.92; P = 0.01). Interestingly, SOX2 activation was a predictor of therapy outcomes in patients receiving radiotherapy. Moreover, SOX2 overexpression enhanced the effect of radiotherapy in HNSCC cell lines. SOX2 activation is associated with improved prognosis of patients with HNSCC and might be used to predict which patients might benefit from radiotherapy.

Several studies have highlighted the diagnostic potential of salivary microRNA (miRNA) in head and neck squamous cell cancer (HNSCC). The purpose of this meta-analysis was to summarize published studies and evaluate the diagnostic accuracy of salivary miRNA in HNSCC detection. In this meta-analysis, we systematically searched PubMed, EMBASE, and Cochrane Library databases for studies on miRNA and HNSCC diagnosis. Pooled sensitivity, specificity, and diagnostic odds ratio (DOR) with a summary receiver-operating characteristic curve were calculated using a bivariate random-effect meta-analysis model. Furthermore, subgroup analyses were conducted to explore the main sources of heterogeneity. Seventeen studies from ten articles, including 23 miRNA and a total of 759 subjects, were included in this meta-analysis. The pooled sensitivity and specificity of salivary miRNA in the diagnosis of HNSCC were 0.697 (95% CI: 0.644–0.744) and 0.868 (95% CI: 0.811–0.910), respectively. The overall area under the curve was 0.803 with a DOR of 12.915 (95% CI: 9.512–17.534). Salivary miRNAs are a promising non-invasive diagnostic biomarker with moderate accuracy for HNSCC. These results must be verified by large-scale prospective studies.