An Empirically Derived Seven-Gene Immune Prognostic Signature Enriched in T-Cell Exhaustion Mediators Stratifies Survival Outcomes in HPV-Positive and HPV-Negative Head and Neck Squamous Cell Carcinoma

Sharjeel Chaudhry; Zarmina Ehtesham; Fatmah Salem Alayidh

doi:10.31557/APJCC.2026.11.6.917

Authors

Sharjeel Chaudhry Department of Molecular Pathology, Dow Diagnostic Research and Reference Laboratory, Dow University of Health Sciences, Karachi, Pakistan. https://orcid.org/0000-0002-7070-1326
Zarmina Ehtesham Department of Molecular Pathology, Dow Diagnostic Research and Reference Laboratory, Dow University of Health Sciences, Karachi, Pakistan.
Fatmah Salem Alayidh Department of Medical Laboratory, King Fahad Military Medical Complex, Ministry of Defense Health Services, Dhahran, Kingdom of Saudi Arabia.

DOI:

https://doi.org/10.31557/APJCC.2026.11.6.917

Keywords:

head and neck squamous cell carcinoma, immune prognostic signature, T-cell exhaustion, HPV, LASSO regression, tumour microenvironment, checkpoint inhibitors

Abstract

Introduction: Head and neck squamous cell carcinoma (HNSCC) is a biologically heterogeneous malignancy with poor prognosis in advanced disease. Despite the introduction of immune checkpoint inhibitors, reliable biomarkers capable of translating tumour microenvironment (TME) complexity into clinically actionable risk stratification remain lacking. T-cell exhaustion, characterised by progressive effector dysfunction and inhibitory receptor upregulation, is a dominant immunological state in HNSCC that is differentially regulated by human papillomavirus (HPV) status. We aimed to develop and externally validate a T-cell exhaustion-based immune prognostic signature stratifying overall survival (OS) across HPV subgroups.

Materials and Methods: RNA-sequencing and clinical data from 502 TCGA-HNSC tumours were analysed. Immune-related genes (ImmPort v36) were assessed by differential expression analysis (limma/voom), univariate Cox regression, and least absolute shrinkage and selection operator (LASSO)-penalised regression to derive a parsimonious prognostic model, which was externally validated in GSE65858 (n = 270). Kaplan–Meier analysis, time-dependent receiver operating characteristic (ROC) curves, and multivariate Cox regression adjusting for clinicopathological covariates were used to evaluate performance.

Results: A seven-gene signature (PDCD1, HAVCR2, TOX, LAG3, TIGIT, CXCL13, ENTPD1) was identified. High-risk patients demonstrated significantly inferior OS in both the TCGA (median OS 24.1 versus 52.8 months; hazard ratio [HR] 2.63, 95% CI 2.01–3.44; p < 0.001) and GSE65858 cohorts (26.3 versus 50.6 months; HR 2.38, 95% CI 1.77–3.20; p < 0.001), with three-year area under the ROC curve (AUC) values of 0.74 and 0.71, respectively. Prognostic utility was retained in both HPV-positive and HPV-negative strata, with stronger discrimination in HPV-negative disease (AUC 0.76). The risk score remained an independent predictor of OS on multivariate analysis (HR 2.47, 95% CI 1.89–3.23; p < 0.001).

Conclusions: This seven-gene immune prognostic signature demonstrates robust, independent prognostic performance across HPV subgroups and cohorts, correlating with immunologically meaningful TME features.