Trial characteristics specific to the assumption of noninferiority (NI)
| First author [ref.] | Rationale | Measure | Margin of NI | Choice of margin | Consistency of sample size estimation | Power estimation | Power % | Planned number per arm | E/O in control arm | Absolute decrease accepted |
|---|---|---|---|---|---|---|---|---|---|---|
| First-line setting | ||||||||||
| Negoro [13] | No | Hazard ratio | ⩽1.33 | NR | Yes | Superiority of experimental arm | 81 | 81 events | Median OS 35 weeks (E) | 7 weeks |
| Rosell [14] | No | Rates difference (E−C) | >−10% | NR | Yes | Equivalence | 80# | 260 | Response rate 30% (E) | 10% |
| Fossella [15] | No | Hazard ratio | ⩾0.85 | ERM (50%) | No¶ | NR | 85 | 360 | Median OS 8 months (E) | 1.2 months |
| Ohe [16] | No | Rates difference (E−C) | >−10% | NR | Yes | Superiority of experimental arm | 80 | 135 | 1-year survival 43% (E) | 10% |
| Novello [17] | No | Hazard ratio | ⩽1.33 | CM | Yes | Equivalence | 80# | 110 events | 1-year survival 33% (E) | 1.9 months |
| Park [18] | Yes | Rates difference (E−C) | >−15% | NR | Yes | NR | 80 | 109 | Median OS 6.2 months (E) | 15% |
| Scagliotti [19] | No | Hazard ratio | <1.176 | NR | Yes | Equivalence | 80 | 595 events | Median OS 10.3 months (O) | 1.5 months |
| Heymach [20] | No | Hazard ratio | NR | NR | Yes | NR | 75+ | NR | Median PFS 11.5 weeks (E) | NC |
| Mok [21] | No | Hazard ratio | ⩽1.20 | NR | Yes | Equivalence | 80 | 472 events | NR (NPH) | NA |
| Okamoto [22] | Yes | Hazard ratio | ⩽1.33 | NR | Yes | Equivalence | 85 | 280 | Median OS 14 months (E) | 3.5 months |
| Lee [23] | No | Rates difference (E−C) | >−15% | Other | Yes | Equivalence | 80 | 143 | Response rate 37% (E) | 15% |
| Second-line setting | ||||||||||
| Schuette [24] | Yes | Rates difference (E−C) | >−2%, <2%§ | NR | Unclear | NR | 80 | 102 | Unclear | 2% |
| Hanna [25] | Yes | Hazard ratio | ⩽1.11 | NR | Yes | Hazard ratio 0.83 | 81% | NR | Median OS: 7.9 months (O) | 0.7 months |
| ⩽1.21 | ERM (50%) | NA | NA | |||||||
| Ramlau [26] | No | Rates difference (E−C) | >−10% | NR | Yes | Equivalence | 81 | 380 | 1-year survival 37% | 10% |
| Krzakowski [27] | No | Hazard ratio | ⩾0.755 | NR | Yes | NR | 80 | 275 | Median PFS 2.3 months (O) | 0.6 months |
| Maruyama [28] | No | Hazard ratio | ⩽1.25 | NR | Yes | Superiority of experimental arm | 90 | 242 | Median OS 12 months (E) | 2.4 months |
| Kim [29] | Yes | Hazard ratio | ⩽1.154 | ERM (50%) | Yes | NR | NR | 585 events | Median OS: 8 months (O) | 1.1 months |
| Li [30] | Yes | Rates difference (E−C) | >−10% | NR | NR | NR | NR | NR | Response rate 4.9% (O) | 10% |
| Natale [31] | No | Hazard ratio | NR | ERM (50%) | No¶ | NR | NR | 555 events | Median PFS 2.2 months | 0.6 months |
| Shi [32] | No | Hazard ratio | <1.14 | ERM | Yes | NR | 80 | 200 | Median PFS 4.2 months | 0.5 months |