Gene Expression Subtype Predicts Nodal Metastasis and Survival in Human Papillomavirus–Negative Head and Neck Cancer

Objectives/Hypothesis: Gene expression analyses of head and neck cancer have revealed four molecular subtypes: basal (BA), mesenchymal (MS), atypical (AT), and classical (CL). We evaluate whether gene expression subtypes in oral cavity squamous cell carcinoma (OCSCC) and laryngeal squamous cell carcinoma (LSCC) can be used to predict nodal metastasis and prognosticate survival. Study Design: Retrospective cohort study and genomic analysis. Methods: OCSCC and LSCC cases were identified from the The Cancer Genome Atlas (TCGA) head and neck cancer cohort. RNA-seq by expected maximization (RSEM) was used to quantify gene expression levels from TCGA RNA-seq data and to assign each case to one of four subtypes. Descriptive statistics were used to describe patient, disease, and treatment characteristics in each subtype. Cox regression and Kaplan-Meier analyses were used to determine associations with survival. Results: OCSCC cases were comprised primarily of the MS and BA subtypes, whereas LSCC was comprised primarily of CL and AT subtypes. In OCSCC, the MS subtype was significantly associated with higher risk of nodal metastasis. In a subset analysis of clinically T1-2N0M0 OCSCC, we demonstrate that the MS subtype was predictive of occult nodal metastasis (relative risk = 3.38, 95% confidence interval [CI]: 1.08-10.69). In LSCC, the CL subtype was associated with significantly worse overall survival (hazard ratio = 4.32, 95% CI: 1.77-10.54, P =.001). Conclusions: Gene expression analysis reveals potential novel markers of nodal metastasis and survival in human papillomavirus–negative head and neck cancer. Future studies will continue to refine and validate these markers, with the goal of providing molecular risk assessments that guide therapy and improve patient outcomes. Level of Evidence: 2b Laryngoscope, 129:154–161, 2019.