4.6. Statistical Analyses of Prospective Clinical Trial Data and Real-World Data

The clinical features of interests were summarized using descriptive statistics including median and interquartile range for continuous variables and proportions and frequencies for categorical variables. In general, Kruskal–Wallis tests for continuous variables and chi-squared tests for categorical variables were conducted to compare the differences among the multiple groups. For those categorial variables in which some levels were less than 5 counts, the Fisher’s exact test was applied.

For survival analyses of the prospective clinical trial, progression-free survival (PFS) time was defined as the time between the date of study enrollment and the date of progressive disease or death whichever happened first or otherwise censored at the last date of known alive. Overall survival (OS) time was defined as the time between the date of study enrollment and the date of death or censored at the last date known alive. For survival analyses of the RWD, PFS time was not available. The OS was defined as the time between the date of first dose of given treatments and the date of death or censored at the last date known alive. The propensity score matching [25] was applied to match our prospective clinical trial and the RWD based on the prognostic factors which are significantly different across the three cohorts: cetuximab and nivolumab combination, cetuximab monotherapy (RWD-cetuximab), and RWD-nivolumab/pembrolizumab CPI monotherapy (RWD-CPI). For subset analyses of the RWD, propensity score weighing [25] was used for given 2-group comparisons. The Kaplan–Meier method was used for the PFS and OS analyses, and log-rank tests were adopted to compare survival differences between two groups. Univariate Cox proportional hazards (PH) model was conducted to evaluate the association of OS with the sequential treatment or the individual clinical feature. All statistical analyses were performed using SAS (version 9.4, SAS Institute Inc., Cary, NC, USA) and the R 3.6.0 software (https://www.R-project.org, accessed on 24 November 2020).