Table 3.

Average Sample Size, Proportion of Responders, and No. of Nonresponders for Fixed 1:1 and Adaptive Randomized Phase II Trial Design

Response Rates by Treatment Arm		Fixed 1:1 (maximum sample size = 190)			Adaptive Randomization Capped at 80% Assignment Probability (maximum sample size = 208)
Control Arm	Experimental Arm	Average sample size	P (responders) %	No. of Nonresponders	Average sample size	P (responders) %	No. of Nonresponders
0.2	0.2	177.9	20.3	142.3	194.3	20.2	155.5
0.2	0.3	135.2	25.9	101.4	147.6	26.3	109.7
0.2	0.4	78.4	31.4	54.8	83.7	32.1	57.3
0.2	0.5	43.3	36.6	28.2	45.3	37.1	28.8

NOTE. Adaptive randomization uses the method of Thall and Wathen¹². Trials are stopped early for superiority of the experimental treatment if P(E> C) > 0.984. One-sided type 1 error = 10%; power = 90% at 20% v 40% response rates; results based on 500,000 simulations. Characteristics of trial designs corresponding to the trial alternative hypothesis are in bold type. P (responders) % is the average proportions of responders given as a percentage.