Patients were divided into two groups based on the type of surgery: primary TL group and salvage TL group. Each patient’s clinical characteristics (gender, age, tumor site, T- and N-stage, etc.) of these two groups of patients were compared using Chi-square tests.
We analyzed the impact on OS, CSS and RFS of the type of surgery (primary vs. salvage TL) and of the following factors: age (< vs. >70 years), gender (male vs. female), ASA score (< vs. ≥3), primary tumor site (larynx vs. hypopharynx), preoperative tracheostomy, surgical margins (negative vs. positive), T-stage (3 vs. 4) and N-stage (< vs. ≥2A). Univariate analyses were performed using Log Rank tests. For multivariate analysis (conducted only when more than one factor was significant in univariate analyses), all variables associated with p < 0.05 in univariate analysis were included in Cox regression models with forward stepwise selection. The predictive ability of each Cox regression model was determined using the area under the receiver operating characteristic (ROC) curve (AUC) with a clinical endpoint fixed at 5 years.
All statistical analyses were performed at 5% alpha risk or 95% confidence interval by the biostatistician using R.3.0.1 software on Windows.
We analyzed the impact on OS, CSS and RFS of the type of surgery (primary vs. salvage TL) and of the following factors: age (< vs. >70 years), gender (male vs. female), ASA score (< vs. ≥3), primary tumor site (larynx vs. hypopharynx), preoperative tracheostomy, surgical margins (negative vs. positive), T-stage (3 vs. 4) and N-stage (< vs. ≥2A). Univariate analyses were performed using Log Rank tests. For multivariate analysis (conducted only when more than one factor was significant in univariate analyses), all variables associated with p < 0.05 in univariate analysis were included in Cox regression models with forward stepwise selection. The predictive ability of each Cox regression model was determined using the area under the receiver operating characteristic (ROC) curve (AUC) with a clinical endpoint fixed at 5 years.
All statistical analyses were performed at 5% alpha risk or 95% confidence interval by the biostatistician using R.3.0.1 software on Windows.
Free full text: Click here
