Table 1.

Second column contains the approximate number of events required for 80% power with 5% two-sided log-rank test for comparing arms of targeted design shown in Figure 3. Only marker positive patients are randomized. Treatment hazard ratio for marker positive patients is shown in first column. Time-to-event distributions are exponential and all patients are followed to failure. Last three columns show the approximate number of events required for comparing the arms of a traditional design in which unclassified marker positive and marker negative patients are randomized. Treatment hazard ratio for marker negative patients is assumed to be 1.

Treatment Hazard Ratio for Marker Positive Patients	Number of Events for Targeted Design	Number of Events for Traditional Design
		Percent of Patients Marker Positive
		20%	33%	50%
0.5	74	2040	720	316
0.67	200	5200	1878	820