Development of a novel prognostic signature derived from essential genes in head and neck squamous cell carcinoma.

Abstract

Substantial heterogeneity in head and neck squamous cell carcinoma (HNSCC) compromise accurate patient stratification and personalized treatment planning. Current molecular classification is largely based on genes with highly variable expression without considering their functional roles. Here, we sought to identify HNSCC essential genes for patient stratification and prognostication. Essential genes for HNSCC were screened from genome-wide CRISPR knockout datasets. Candidates were further identified through univariate Cox regression. The least absolute shrinkage and selection operator was utilized to develop the prognostic signature. Candidate essential genes were exploited to classify patients into subgroups by consensus clustering. Survival outcomes, genomic alterations, signaling activities, and therapeutic vulnerabilities were compared between patient subgroups. Sixty-eight genes were identified as candidates and utilized to develop an 8-gene prognostic signature. Patients were segregated into two clusters with distinct survival rates across multiple cohorts based on upregulated essential genes. Cluster 2 exhibited higher TP53, CDKN2A, and NOTCH1 mutations, higher stromal activities, worse prognosis as well as and sensitivities to cell cycle inhibitors. Cluster 1 was characterized by a better prognosis and susceptibility to PI3K/AKT and MAPK inhibitors. Our study developed a novel and robust prognostic signature and classification derived from essential genes for HNSCC, which sheds new light on HNSCC precision oncology.