Fig. 1.

Power of the four response-based end-points for six different scenarios in the one-arm trial setting. Measurement error variance is set to 0. Mean tumour shrinkage is described further in the supplementary material – negative values indicate average shrinkage in the tumour size. The failure intercept parameter in scenario 3 is the parameter that determines the probability of progressing for a non-growth reason. More highly negative values mean lower probabilities of progressing for non-growth reasons.