Skip to main content

Table 5 Trial-level association of all included studies (Multiple linear regression, adjusted for line, phase and maskingb)

From: 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

Treatment effect of surrogate endpoint

Treatment effect of true endpoint

No. of trials

Slope estimate

95% CI

R2

Adjusted R2

ΔORR

HR of OSa

103

0.012

0.008, 0.015

0.313

0.299

ΔORR

HR of PFSa

104

0.023

0.119, 0.027

0.571

0.562

HR of PFSa

HR of OSa

125

0.418

0.312, 0.524

0.393

0.380

  1. Multivariable linear regression model form:
  2. Log (HR of OS) = − 0.003 + 0.012 * ΔORR − 0.092 * phase;
  3. Log (HR of PFS) = − 0.066 + 0.023 * ΔORR – 0.093 * masking;
  4. Log (HR of OS) = 0.009 + 0.418 * log (HR of PFS) - 0.072 * phase
  5. aNatural log hazard ratio in the analysis
  6. bCode for line, phase and masking in the regression:
  7. Line: First-line = 0, ≥2nd-line = 1
  8. Phase: Phase II = 0, Phase III = 1
  9. Masking: Open-label = 0, Double-blind = 1
Back to article page

BMC Cancer

ISSN: 1471-2407

Contact us