Appendix Table 1.

Sample sizes^* needed to achieve 90% power at a .025 one-sided significance level for a hazard ratio of 0.7 in the biomarker-positive subgroup (assuming 3 years of accrual and 2 years of follow-up)

	Biomarker positive, %
Design	100	80	60	50	40	30
Strategy	392	600 (613)	1008 (1089)	1440 (1568)	2220 (2450)	3880 (4356)
Enrichment or stratified	392	490	654	784	980	1307

^*Given in parentheses are the sample size estimates based on the approximate formula for the biomarker-strategy design (D_strategy).