In this study, we developed a long noncoding RNA (lncRNA)-based prognostic signature for stratification of patients with head a nd neck squamous cell carcinoma (HNSCC). In total, 493 HNSCC samples obtained from the Cancer Genome Atlas database were divided into training and testing cohorts (3:2 ratio). We identified 3913 immune-related lncRNAs in the HNSCC training cohort by Pearson correlation analysis; only seven were independently associated with overall survival and were used to develop an immune-related lncRNA prognostic signature (IRLPS) grouping of HNSCC patients into high- and low-IRLPS subgroups. Univariate and multivariate Cox analyses revealed that low-IRLPS patients had a better prognosis in all the cohorts, which was retained after stratification by sex, grade, and HPV status. Although the TNM stage was also an independent prognostic factor, the IRLPS had a better discriminability with higher AUC at the 3- and 5-year follow-ups in all cohorts. Low-IRLPS samples had more immune cell infiltration and were enriched in immune-related pathways, while high- IRLPS samples were enriched in metabolic pathways. A nomogram constructed including age, TNM stage, and IRLPS showed good calibration. Thus, IRLPS improves the prognostic prediction and also distinguishes different tumor microenvironment (TME) in HNSCC patients.
Keywords: head and neck squamous cell carcinoma; immune; lncRNA; tumor microenvironment.
© 2021 The Authors. Cancer Medicine published by John Wiley & Sons Ltd.