Skip to main content


A pharmaceutical company developing a novel cancer drug was struggling to validate progression-free survival (PFS) or overall response rate (ORR) as surrogate endpoints for overall survival (OS) in non-small cell lung cancer (NSCLC). Two potential issues hampering endpoint validation were cross-over from the control to the treatment arm after progression and pseudo-progression—when a patient’s cancer looks like it is getting worse when he/she is actually improving.2-4

Certara Repurposing Data 1
Certara Repurposing Data 2


The Certara team used CODEx, an intuitive, interactive, web-based graphical interface, to unlock the value and richness of public and proprietary clinical outcome data. They performed an exploratory analysis on a large NSCLC database, including aggregate data from more than 1000 clinical trials published before the end of June 2021 to evaluate how the strength of surrogacy or uncertainty may be impacted by different model parameterizations and how surrogacy may vary across patient subgroups.

The Certara team used CODEx, an intuitive, interactive, web-based graphical interface, to unlock the value and richness of public and proprietary clinical outcome data.


The analysis revealed that surrogacy strength depends on model parameterization and how well nonlinearities and collinearities are captured. Surrogacy strength substantially improves when using a joint or bi-variate meta-analytic model5 (as recommended in the recently updated NICE methodological guideline6) for the surrogate and the final endpoint, which can be seen in Table 1: the adjusted R2 (correlation coefficient) of the joint model increased to 0.47 as compared to the one of the standard meta-regression model. In addition, while joint modelling does not impact mean predictions, it led to narrower prediction intervals compared to those based on the standard meta-regression model, see the last column of Table 1.

The study demonstrated the importance of considering these components when predicting a surrogate’s performance. The client was able to use these findings to gain valuable insights into the potential of developing an effective surrogate modeling strategy.

Certara Repurposing Data 3 2
Certara Repurposing Data 4


  1. Surrogate Endpoint Resources for Drug and Biologic Development. US Food & Drug Administration.
  2. Blumenthal GM, Karuri SW, Zhang H, et al. Overall response rate, progression-free survival, and overall survival with targeted and standard therapies in advanced non-small-cell lung cancer: US Food and Drug Administration trial-level and patient-level analyses. J Clin Oncol Off J Am Soc Clin Oncol. 2015;33(9):1008-1014. doi:10.1200/ JCO.2014.59.0489
  3. Clarke JM, Wang X, Ready NE. Surrogate clinical endpoints to predict overall survival in non-small cell lung cancer trials-are we in a new era? Transl Lung Cancer Res. 2015;4(6):804-808. doi:10.3978/j.issn.2218-6751.2015.05.03
  4. Ma Y, Wang Q, Dong Q, Zhan L, Zhang J. How to differentiate pseudoprogression from true progression in cancer patients treated with immunotherapy. Am J Cancer Res. 2019;9(8):1546-1553.
  5. Papanikos T, Thompson JR, Abrams KR, et al. Bayesian hierarchical meta‐analytic methods for modeling surrogate relationships that vary across treatment classes using aggregate data. Stat Med. 2020;39(8):1103-1124. doi:10.1002/sim.8465
  6. Welton NJ, Phillippo DM, Owen R, et al. CHTE2020 Sources and Synthesis of Evidence; Update to Evidence Synthesis Methods.; 2020.
  7. Wheaton L, Papanikos A, Thomas A, Bujkiewicz S. Using Bayesian Evidence Synthesis Methods to Incorporate Real World Evidence in Surrogate Endpoint Evaluation. Published online December 16, 2021. Accessed 14 December 2022.

How Can We Help?