Fig. 3. Conventional group sequential design with non-binding two-sided hazard ratio bounds for stopping for efficacy or futility at each interim analysis.

The conventional group sequential design with non-binding two-sided hazard ratio bounds for stopping for efficacy or futility at each interim analysis. The red area in the figure is the futility boundary. If an observed HR lies within the red area, the study may be stopped for futility. The blue area in the figure is the efficacy boundary. Similarly, if an observed HR lies within the blue area, the study may be stopped for efficacy. The estimated event-free survival hazard ratio for lenalidomide treatment effect was 1.12, 0.92, 0.99, and 1.02 at interim analysis 1, 2, 3, and 4, respectively.