MicroRNA expression is associated with human papillomavirus status and prognosis in mucosal head and neck squamous cell carcinomas.

Abstract

The major cause of mucosal squamous cell carcinomas of the head and neck (HNSCCs) has been attributed to human papillomavirus (HPV) infection. Here we investigate if microRNA expression in HNSCC can be used as a prognostic tool with or without HPV status. We performed a discovery miRNA microarray (miRBase v.21) profiling of 52 tonsillar SCCs with TaqMan real-time PCR validation of 228 HNSCCs. Patients had a histologically confirmed primary SCC of the oropharynx, oral cavity, hypopharynx or larynx. Logistic regression models were used to estimate the magnitude of the effect of association with clinical factors and miRNAs associated with HPV status. For recurrence and survival analysis, we used unadjusted and multivariable adjusted Cox proportional hazard regression models. Seventeen miRNAs were significantly associated with better prognosis in the discovery phase and were validated in the extended dataset. The best fitting model (AUC=0.92) for HPV status included age, smoking, and miRNAs: miR-15b, miR-20b, miR-29a, miR-29c, miR-142, miR-146a and miR-205. Using Cox regression model for recurrence, miR-29a was associated with 49% increased risk of recurrence while miR-30e and miR-342 were associated with decreased risk of recurrence with HRs 0.92 (95% CI 0.85-0.99) and 0.84 (95% CI 0.73-0.98), respectively. Our best fitting model for survival included age, gender, alcohol consumption, N stage, recurrence, HPV status, together with miRNAs-20b, 29a, and 342. miRNAs show potential to serve as usual biomarkers to predict the clinical course of patients with mucosal HNSCC.