Predicting and affecting response to anti-cancer therapy based on pathway-level biomarkers

The identification of robust, patient-specific, predictive biomarkers is a major obstacle in precision oncology. Their reproducibility is key requirement in the process of becoming clinically relevant. This crucial step is also needed to identify new personalized synergistic combination approaches. To address these problems and optimize patient-specific therapeutic strategies, we utilized pathway knowledge with drug sensitivity, RNAi, and CRISPR-Cas9 high-throughput screens of hundreds of cell lines from 9 tumor types. We found that pathway activity levels significantly predicted the essentiality of 15 genes that have the potential to act as therapeutic targets. Furthermore, we identified four signaling pathways that can act as strong and robust predictive biomarkers for BCL2-family, BRAF, and MEK inhibitors as well as microtubule inhibitors. Finally, this work demonstrates for the first time that pathway activity level modulation can sensitize NSCLC cells and human lung cancer tumors to microtubule inhibition therapy.