DRAP: a toolbox for drug response analysis and visualization tailored for preclinical drug testing on patient-derived xenograft models

Abstract

Abstract Background One of the key reasons for the high failure rate of new agents and low therapeutic benefit of approved treatments is the lack of preclinical models that mirror the biology of human tumors. At present, the optimal cancer model for drug response study to date is patient-derived xenograft (PDX) models. PDX recaptures both inter- and intra-tumor heterogeneity inherent in human cancer, which represent a valuable platform for preclinical drug testing and personalized medicine applications. Building efficient drug response analysis tools is critical but far from adequate for the PDX platform. Results In this work, we first classified the emerging PDX preclinical trial designs into four patterns based on the number of tumors, arms, and animal repeats in every arm. Then we developed an R package, DRAP, which implements Drug Response Analyses on PDX platform separately for the four patterns, involving data visualization, data analysis and conclusion presentation. The data analysis module offers statistical analysis methods to assess difference of tumor volume between arms, tumor growth inhibition (TGI) rate calculation to quantify drug response, and drug response level analysis to label the drug response at animal level. In the end, we applied DRAP in two case studies through which the functions and usage of DRAP were illustrated. Conclusion DRAP is the first integrated toolbox for drug response analysis and visualization tailored for PDX platform. It would greatly promote the application of PDXs in drug development and personalized cancer treatments.

Keywords

• Patient-derived xenograft
• Tumor model
• Drug response analysis
• Preclinical drug testing
• Personalized drug selecting