Prediction of radiotherapy response with a 5-microRNA signature-based nomogram in head and neck squamous cell carcinoma

Radiotherapy is unlikely to benefit all patients with head and neck squamous cell carcinoma (HNSCC). Therefore, novel method is warranted to predict the radiotherapy response. Our study aimed to construct a microRNA (miRNA)-based nomogram to predict clinical outcomes of patients with HNSCC receiving radiotherapy. We screened out 56 differential miRNAs by analyzing 44 paired tumor and adjacent normal samples miRNA expression profiles from The Cancer Genome Atlas (TCGA). A total of 307 patients with HNSCC receiving adjuvant radiotherapy were randomly divided into a training set (n = 154) and a validation set (n = 153). In the training set, we combined the differential miRNA profiles with clinical outcomes, and LASSO regression model was applied to establish a 5-miRNA signature. The prediction accuracy of the 5-miRNA signature was further validated. In addition, target genes of these miRNAs were predicted, and Gene Ontology (GO) analysis as well as KEGG pathway analysis was executed. A 5-miRNA signature including miR-99a, miR-31, miR-410, miR-424, and miR-495 was identified. With a cutoff value of 1.2201 from Youden's index, the training set was divided into high-risk and low-risk groups, and the 5-year overall survival was significantly different (30% vs. 73%, HR 3.65, CI 2.46–8.16; P < 0.0001). Furthermore, our 5-miRNA signature revealed that only low-risk group would benefit from radiotherapy. Then, a nomogram combining 5-miRNA signature with clinical variables to predict radiotherapy response was constructed. The analysis of 108 target genes of these miRNAs revealed some potential mechanisms in HNSCC radiotherapy response for future investigations. In conclusion, the 5-miRNA signature-based nomogram is useful in predicting radiotherapy response in HNSCC and might become a promising tool to optimize radiation strategies.