LAUSR

We are pleased that our colleagues are reading our article [1] with great interest, and we appreciate their comments that raise valid questions. These questions are answered below one-by-one. First, head and neck squamous cell carcinoma (HNSCC) constitute a group of malignant tumors located in the oropharynx, larynx, hypopharynx, nasopharynx, and oral cavity. Oropharyngeal squamous cell carcinoma (OPSCC) is a major part of HNSCC. We did not draw a conclusion only based on OPSCC patients. Though we established the seven-gene panel solely based on advanced OPSCC patients, we further validated the prognostic value in other HNSCC patients and got strong evidence for the predictive power. Moreover, in Fig. 4a, five is lambda.min which is generally considered to be the best value. Seven is lambda.1se which represents a model with excellent performance and the least number of independent variables [2, 3]. To build a more accurate clinical model, we finally chose the sevengene panel. Second, there are too many variables in our research. To avoid overfitting the model, Lasso regression analysis is a good choice, which can filter the variables stably [4]. Furthermore, we were concerned about the whole sevengene panel, so we did not analyze the survival of each individual gene. And it is normal to have different results of your study. Moreover, the ROC curve analysis of GSE113282 data showed a favorable efficiency, so we next validated the prognostic value by survival analysis in GSE102995 validation set and got a good result. Furthermore, the treatments of patients in GSE113282 data and GSE102995 data were inconsistent in detail, but there were similarities because they were all radiotherapy or chemotherapy based on platinum. Third, progression-free survival (PFS) is an important index to evaluate the curative effect of cancer. There are a great number of authors use PFS to evaluate prognosis in Web of Science, PubMed and other databases. We did not think it was inappropriate to validate our panel. Finally, as we wrote in our article, the identification of the seven-gene panel gave us some useful clues to explore the mechanism of sensitivity and prognosis with HNSCC treated with platinum-based radio(chemo)therapy. Further basic experiments are necessary and we will confirm with molecular biological and cellular experiments in the future. We greatly appreciate the excellent points brought forth in this letter and again would like to thank Dr. Shi and Dr. Xie.