Predict progression free survival and overall survival using objective response rate for anti - PD1 / PDL1 therapy development

Abstract Background: In anti - PD1 (programmed death-1) / PDL1 (programmed death ligand 1) therapy development for solid tumors, the objective response rate (ORR) is a commonly used clinical endpoint for early phase study decision making, while progression free survival (PFS) and overall survival (OS) are widely used for late phase study decision making. Developing predictive models for median PFS (mPFS) and median OS (mOS) based on early phase clinical outcome ORR could inform late phase study design optimization and probability of success (POS) evaluation. In the existing literature, there are ORR / mPFS / mOS associations and surrogacy investigations with limited number of included clinical trials. In this paper, without establishing surrogacy, we attempt to predict mPFS and mOS based on early efficacy in ORR and to optimize late phase trial design for anti - PD1 / PDL1 therapy development. Methods: In order to include an adequate number of eligible clinical trials, we built a comprehensive clinical trial quantitative landscape database (QLD) by combining information from different sources, such as clinicaltrial.gov, publications, company press releases for relevant indications and therapies. We developed a generalizable algorithm to systematically extract structured data and manual curation for scientific accuracy and completeness. More than 150 late phase clinical trials were identified for ORR / mPFS (ORR / mOS) predictive model development, while existing literature included at most 40 trials. A tree-based machine learning regression model was derived to account for ORR / mPFS (ORR / mOS) relationship heterogeneity across tumor type, stage, line of therapy, treatment class and borrow strength simultaneously when homogeneity persists. Results: The proposed method ensures that the predictive model is robust and has explicit structure for clinical interpretation. Over 1000 times cross validation, the average predictive mean square error of the proposed model is competitive to random forest and extreme gradient boosting methods and outperforms commonly used additive or interaction linear regression models. Conclusions: An example application of the proposed ORR / mPFS (ORR / mOS) predictive model on late phase trial POS evaluation for anti - PD1 / PDL1 combination therapy was illustrated.