Prognosis prediction of head and neck squamous cell carcinoma through the basement membrane-related lncRNA risk model

Head and neck squamous cell carcinoma (HNSCC) ranks among the most widespread and significantly heterogeneous malignant tumors globally. Increasing evidence suggests that the basement membrane (BM) and associated long non-coding RNAs (lncRNA) are correlated with the onset of HNSCC and its prognosis. Our study aims to construct a basement membrane-associated lncRNAs (BMlncRNAs) marker to accurately predict the prognosis of HNSCC patients and find novel immunotherapy targets.

The Cancer Genome Atlas (TCGA) database was accessed to acquire the transcriptome expression matrices, somatic mutation data, and clinical follow-up data of HNSCC patients. Utilizing co-expression analysis, the BMlncRNAs were identified and the differentially expressed lncRNAs (DEBMlncRNA) were then filtered, The filtering thresholds are FDR<0.05 and |log2FC|≥1. Furthermore, univariate analysis, least absolute shrinkage and selection operator (LASSO), and multivariable Cox regression were utilized to develop the risk model. The model then underwent thorough evaluation across diverse perspectives, encompassing tumor immune infiltration, tumor mutation burden (TMB), functional enrichment, and chemotherapy sensitivity.

The risk assessment model consists of 14 BMlncRNA pairs. The acquired data is indicative of the reliability of the risk score in its capacity as a prognostic factor. Individuals at high risk exhibited a poorer prognosis, and a statistically significant variance was noted in TMB and tumor immune infiltration compared to the low-risk group. Additionally, heightened sensitivity to paclitaxel and docetaxel was evident in the patients at high risk.

We have established a BMLncRNA-based prognostic model that can provide clinical guidance for future laboratory and clinical studies of HNSCC.