Follow-up and statistical analysis
The survival stage and prognosis of all patients were followed up by means of telephone return visits, outpatient reexaminations, and other methods. Follow-up period was once every 3 months in the first and second years and once every 6 months in the third to fifth years. The re-examination included physical examination, laryngoscopy and imaging examination (cervicothoracic enhanced CT). Patients with clinically suspected tumour recurrence were examined by PET–CT. Survival rate was analysed via Kaplan–Meier analysis. The difference between groups was tested through log-rank method, and independent prognostic factors were analysed by Cox regression model in multivariate survival analysis. The statistical software SPSS 25 was used to analyse the data. P<0. 05 was considered statistically significant.