Table 5.

Hypothetical Example 2. Stopping probability at each analysis and conjunctive power when the FS procedure is used.

Correlation	Boundary function combination	Stopping probability (%) at each analysis (month)				Conjunctive power (%)
		24 (l =1)	36 (l =2)	48 (l =3)	60 (l =4)
0.3	(a) OBF-OBF	4.52	36.62	19.92	10.53	71.59
	(b) OBF-POC	20.31	19.89	9.95	15.13	65.29
	(c) POC-OBF	5.60	35.69	19.23	8.97	69.49
	(d) POC-POC	25.90	19.21	9.44	9.25	63.79
0.5	(a) OBF-OBF	5.06	37.39	20.18	10.10	72.73
	(b) OBF-POC	21.88	20.14	10.10	14.27	66.39
	(c) POC-OBF	6.01	36.65	19.49	8.68	70.82
	(d) POC-POC	27.20	19.54	9.58	8.83	65.15
0.8	(a) OBF-OBF	6.12	38.92	20.98	9.22	75.25
	(b) OBF-POC	25.39	21.05	10.56	11.65	68.65
	(c) POC-OBF	6.68	38.65	20.41	8.12	73.87
	(d) POC-POC	29.95	20.61	10.07	7.47	68.11

The accrual duration is τ_A =24 months and the follow-up duration is ${\tau }_{f_1}=12$ months for PFS and ${\tau }_{f_2}=36$ months for OS. The HR is ψ₁ =0.65 for PFS and ψ₂ =0.70 for OS. The proportion of survivors after the follow-up duration is $S_1^{(1)}(36)=6.0\text{\%}$ for PFS and $S_2^{(1)}(60)=7.0\text{\%}$ for OS. Two analyses at 24 and 36 months for PFS and up to four analyses at 24, 36, 48 or 60 months for OS are planned. A correlation is assumed to be ρ =0.3, 0.5 and 0.8. The conjunctive probability at each anal and conjunctive power are calculated under a sample size of 304, which provides at least 80% power to detect 30% reduction in the HR for OS at the 2.5% significance level by a one-sided logrank test in the fixed-sample design.