Validating ORR and PFS as surrogate endpoints in phase II and III clinical trials for NSCLC patients: difference exists in the strength of surrogacy in various trial settings

Table 4 Trial-level associations of all included studies (Simple linear regression)

Treatment effect of surrogate endpoint	Treatment effect of true endpoint	No. of trials	Slope estimate	95% CI	R²	Adjusted R²	STE
ΔORR	HR of OS	112	0.011	0.007, 0.014	0.261	0.254	——^b
ΔORR	HR of PFS^a	114	0.023	0.019, 0.027	0.578	0.574	−26.30%
HR of PFS^a	HR of OS^a	138	0.403	0.313, 0.494	0.360	0.355	——^b

Simple linear regression model form
log (HR of OS) = −0.062 + 0.011 * ΔORR;
log (HR of PFS) = − 0.103 + 0.023 * ΔORR;
log (HR of OS) = − 0.053 + 0.403 * log (HR of PFS)
^aNatural log hazard ratio in the analysis
^bThe STE will be given only when the upper limit of r_s (in Table 3) is more than 0.7

ISSN: 1471-2407