Table 3.
Statistical analysis based on permutation testing
| Study pool A | Study pool B | Differential median pooled PFS (A-B) | Estimated P-value (based on 20,000 random permutations) |
|---|---|---|---|
| Any line of therapy | |||
| Chemotherapy | EGFR TKI | −4.7 | 0.000 |
| Chemotherapy | Gefitinib | −3.8 | 0.000 |
| Chemotherapy | Erlotinib | −6.8 | 0.000 |
| Predominantly first line* | |||
| Chemotherapy | EGFR TKI | −4.7 | 0.000 |
| Chemotherapy | Gefitinib | −4.0 | 0.000 |
| Chemotherapy | Erlotinib | −6.2 | 0.000 |
| Lines of therapy other than first | |||
| Chemotherapy | EGFR TKI | −6.1 | 0.0028 |
| Chemotherapy | Gefitinib | −5.1 | 0.0039 |
| Chemotherapy | Erlotinib | −8.8 | 0.0022 |
| Chemotherapy | |||
| Predominantly first line* | Lines of therapy other than first | 1.7 | 0.012 |
| Erlotinib | |||
| Predominantly first line* | Lines of therapy other than first | −0.9 | 0.678 |
| Gefitinib | |||
| Predominantly first line* | Lines of therapy other than first | 0.6 | 0.283 |
*
Predominantly first line is ≥90% of patients treated in the first-line setting. This comparative test is statistically valid, but only refers to the given study pool (conditional test) and cannot be readily extrapolated to the total patient population.