Table 4.

Summary of sensitivity analysis on the base case results

Sensitivity analysis	Sample size	Model coefficient, 95 % CI and p value	Model R2
HR for OS and PFS
Base case	75	0.18 (0.009 to 0.35); p = 0.04	0.31
Year of publication ≥2004	40	0.12 (−0.22 to 0.48); p = 0.48	0.31
First-line trials only	48	0.01 (−0.19 to 0.22); p = 0.90	0.30
Trials in 2nd line and beyond	27	0.40 (0.21 to 0.58); p < 0.001	0.55
Trials meeting the modern FDA definition of PFS	38	0.26 (0.05 to 0.42); p = 0.015	0.38
Trials with censored data	47	0.18 (−0.05 to 0.42); p = 0.12	0.35
Trials allowing cross over	20	0.24 (0.14 to 0.33); p < 0.001	0.49
Only globally conducted trials	34	0.031 (−0.27 to 0.33); p = 0.83	0.10
Δ in OS and Δ in PFS
Base case	79	0.79 (0.32 to 1.26); p = 0.001	0.44
Year of publication C2004	41	0.68 (0.08 to 1.28); p = 0.027	0.51
First-line trials only	51	0.77 (0.26 to 1.28); p = 0.004	0.47
Trials in 2nd line and beyond	28	1.1 (0.38 to 1.74); p = 0.004	0.49
Trials meeting the modern FDA definition of PFS	39	0.74 (0.10 to 1.37); p = 0.25	0.50
Trials with censored data	49	0.81 (0.25 to 1.36); p = 0.006	0.54
Trials allowing cross over	20	1.57 (0.66 to 2.48); p = 0.002	0.75
Globally conducted trials only	36	0.82 (0.14 to 1.51); p = 0.02	0.51

HR hazard ratio, OS overall survival, PFS progression-free survival, Δ change, FDA Food and Drug Administration