Statistical methods

Statistical methods for primary and secondary outcomes {20a}

The experimental arm will be compared against the control arm for all primary analyses.
Cumulative proportions of time to event endpoints will be reported using Kaplan Meier method. the hazard ratio and the 95% confidence intervals for the locoregional recurrence rate will be computed using the Cox regression analysis. The upper bound of the 95% confidence intervals would be used to demonstrate non-inferiority for locoregional recurrence which would be concluded if the same exceeds the predefined threshold of 1.63. The Chi-square test with continuity correction will be used for comparing the binary / categorical variables, while the Wilcoxon test will be used for comparing continuous variables.
All tests will be two-sided and a significance level of 5% will be used for declaring statistical significance. The proportion of patients experiencing any Grade 3 - 4 AE will be reported in each arm and compared between the two arms using the Chi-square test or Fisher’s exact test. Similar comparisons will be done for each AE term between the two groups. Comparison of AE between the two arms will be done on full analysis set as well as the per-protocol set.
Multivariable analysis for the primary and secondary endpoints will be conducted using Cox regression analysis. For every model, the underlying assumptions for the model will be checked which would include checks for proportional hazards assumptions, outlier and overly influential detection as well as a check for the linearity assumptions. Additionally bootstrapping will be used to estimate the calibration and discrimination of the model. Further details can be found in the statistical analysis plan.

Interim analyses {21b}

An interim analysis for safety will be planned at 3 years when 16 events in total are expected to have occurred. As a result of this interim analysis, the sample size has been increased to 2100 patients. The critical Z value at the time of interim analysis is 3.47 for futility. The following table illustrates the key features of the design