First, why not apply the multivariate analysis to further identify prognostic values of the 13 metabolic genes after utilizing univariate Cox and LASSO regression analyses. The paper lacks mention of statistical method involved the multivariate analysis of gene signature, and thus is not rigorous. On the basis of univariate Cox regression analysis, we wanted to further define the similarities and differences in the microcosmic characterization of the metabolic genes.

Then, we applied a least absolute shrinkage and selection operator (LASSO) regression to identify the final elimination of potential predictors with nonzero coefficients [1], which can avoid model overfitting, to select the optimal metabolic signature. Finally, multivariate Cox regression analysis was conducted for the risk signature. Afterwards, multivariate Cox regression was carried out to determine the risk signature was an independent prognostic factor for LC in part of result.

Second, thank you for your positive and constructive comments on our paper. The purpose of this study is to establish a metabolic gene model to evaluate the prognosis of laryngeal cancer. Our study confirmed that the risk score is not only effective for stratification of laryngeal cancer survival, but also effective for recurrence patients. If we only verify it in the TCGA recurrence cohort without external verification, our study is not convincing. Thank you again for your comments.