Abstract 5410: Predictive biomarker discovery method to bridge the gap between preclinical disease model dose-response and clinical trials
David Futorian, Oren Fischman, Gali Arad, Nitzan Simchi, Omri Erez, Eran Seger, Rozanne Groen, Kirill Pevzner · Cancer Research · 2023
Abstract The translational gap of drug and biomarker discovery remains one of the biggest challenges in the pharmaceutical industry. On one hand high throughput screening and omics methods facilitate the generation of in-vitro and in-vivo data of dose-response for a particular drug or combinations. On the other hand, clinical omics techniques allow for the creation of large scale treatment-naive clinical datasets, with the TCGA and CPTAC as prominent examples. However, the question of which disease model best represents the response/resistance mechanisms of an oncology patient remains challenging. In this study we present a machine learning (ML) based computational technique for integrating omics from preclinical dose-response studies with clinical treatment-naive samples to create putative predictive biomarkers, and exemplify its application 2 case studies - PARP or AKT inhibitors. We utilize the Cancer Cell Line Encyclopedia (CCLE) and Genomics of Drug Sensitivity in Cancer (GDSC) as a resource for in-vitro dose response data for PARPi and AKTi, coupled with multi-omic molecular data. We utilize the treatment naive molecular characterization from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) data of Breast cancer patients. Multi-omic analysis of these datasets derived 2 putative predictive biomarkers for PARPi and AKTi. The multi-omic analysis resembles the data availability at the critical drug development stage of transition from preclinical response models to a clinical trial. These biomarkers are tested on a validation dataset obtained from the iSPY adaptive clinical trial and achieve superior results compared to the original enrolment biomarkers and even the retrospectively derived biomarkers. This study presents a novel computational approach to bridge the gap between preclinical dose-response and clinical datasets and suggests an efficient way to discover predictive biomarkers based on data accessible in a preclinical setting. Citation Format: David Futorian, Oren Fischman, Gali Arad, Nitzan Simchi, Omri Erez, Eran Seger, Rozanne Groen, Kirill Pevzner. Predictive biomarker discovery method to bridge the gap between preclinical disease model dose-response and clinical trials. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5410.