TABLE 6.
Comparing sample sizes for cure-rate model (true) versus proportional time (PT) and piecewise exponential model (PEM), for one-sided hypothesis, r = 1
| Cure Rate Mixture CR model assumed true | Proportional Time PT model | Piecewise Exponential PE model | ||||
|---|---|---|---|---|---|---|
| Simulation Scenarios |
δ−1/γ | Power/NWLR; NSLR |
Calculation Parameters |
Power/Ntotal | Calculation Parameters |
Power/Ntotal |
| Exponential-logit mixture | 1.3/0.0 | 0.8/ | λ = 0.01 | 0.8/562 | mstd = 9 | 0.8/ NLR = 566, |
| 944; 1020 | σ = 1.5539 | mnew = 15 | NG = 460, NTW = 458 | |||
| 0.9/ | a = 12, f = 36 | 0.9/776 | a = 12 | 0.9/ NLR = 784, | ||
| 1306; 1412 | d = 0.874, ΔPT = 1.418 | f = 36 | NG = 638, NTW = 636 | |||
| π0 = 0.1 | 1.5/0.0 | 0.8/ | λ = 0.01 | 0.8/366 | mstd = 9 | 0.8/ NLR = 346, |
| λ0 = 0.1 | 380; 428 | σ = 1.5539 | mnew = 15 | NG = 274, NTW = 282 | ||
| a = 12 | 0.9/ | a = 12, f = 36 | 0.9/504 | a = 12 | 0.9/ NLR = 480, | |
| f = 36 | 526; 592 | d = 0.868, ΔPT = 1.545 | f = 36 | NG = 378, NTW = 388 | ||
| β = 1 | 1.8/0.0 | 0.8/ | λ = 0.01 | 0.8/190 | mstd = 9 | 0.8/ NLR = 188, |
| 174; 204 | σ = 1.5539 | mnew = 13 | NG = 142, NTW = 150 | |||
| 0.9/ | a = 12, f = 36 | 0.9/262 | a = 12 | 0.9/ NLR = 258, | ||
| 240; 282 | d = 0.859, ΔPT= 1.836 | f = 36 | NG = 198, NTW = 206 | |||
| 2/0.0 | 0.8/ | λ = 0.01 | 0.8/124 | mstd = 9 | 0.8/ NLR = 110, | |
| 122; 148 | σ = 1.5539 | mnew = 12 | NG = 84, NTW = 90 | |||
| 0.9/ | a = 12, f = 36 | 0.9/172 | a = 12 | 0.9/ NLR = 152, | ||
| 180; 204 | d = 0.847, ΔPT = 2.143 | f = 36 | NG = 116, NTW = 124 | |||
| Weibull-logit mixture | 1.3/0.0 | 0.8/ | λ = 0.2903 | 0.8/906 | mstd = 14 | 0.8/ NLR = 898, |
| 1030; 1096 | σ = 0.7705 | mnew = 14 | NG = 692, NTW = 732 | |||
| 0.9/ | a = 2, f = 6 | 0.9/1256 | a = 2 | 0.9/ NLR = 1244, | ||
| 1426; 1518 | d =, ΔPT= 1.145 | f = 6 | NG = 960, NTW = 1014 | |||
| π0 = 0.1 | 1.5/0.0 | 0.8/ | λ = 0.2903 | 0.8/372 | mstd = 14 | 0.8/ NLR = 344, |
| λ0 = 0.1 | 418; 460 | σ = 0.7705 | mnew = 14 | NG = 290, NTW = 292 | ||
| a = 12 | 0.9/ | a = 2, f = 6 | 0.9/516 | a = 2 | 0.9/ NLR = 476, | |
| f = 36 | 578; 636 | d = 0.875, ΔPT= 1.237 | f = 6 | NG = 400, NTW = 402 | ||
| β = 2 | 1.8/0.0 | 0.8/ | λ = 0.2903 | 0.8/192 | mstd = 14 | 0.8/ NLR = 188, |
| 192; 220 | σ = 0.7705 | mnew = 14 | NG = 138, NTW = 146 | |||
| 0.9/ | a = 2, f = 6 | 0.9/266 | a = 2 | 0.9/ NLR = 260, | ||
| 264; 304 | d = 0.870, ΔPT= 1.347 | f = 6 | NG = 190, NTW = 202 | |||
| 2.0/0.0 | 0.8/ | λ = 0.2903 | 0.8/134 | mstd = 14 | 0.8/ NLR = 134, | |
| 134; 157 | σ = 0.7705 | mnew = 13 | NG = 98, NTW = 104 | |||
| 0.9/ | a = 2, f = 6 | 0.9/184 | a = 2 | 0.9/ NLR = 184, | ||
| 186; 218 | d = 0.862, ΔPT = 1.432 | f = 6 | NG = 136, NTW = 146 | |||
See footnote below Table 5 for explanation of the notation used in this table. Additionally, δ−1 represents the hazard ratio comparing the standard treatment to the new treatment and γ = 0 represents no change in cure rate between the 2 treatment arms.