Abstract 2991: Integrative analysis of molecular and drug response data from clinical samples and PDTXs to identify pharmacogenomic associations in breast cancer

Breast cancer is the most common cancer among women showing high clinical and molecular heterogeneity. Current clinical management causes patients overtreatment with implications on both patients’ quality of life and healthcare costs. Moreover, intrinsic or acquired tumor resistance to treatment leads to incurable metastatic progression in a significant proportion of patients. Consequently, there is an urgent need for better predictive biomarkers and a better understanding of the mechanisms driving response to treatment. As part of the METABRIC initiative, we fully molecularly characterized 2000 breast primary tumors, measuring gene expression, copy number aberration, somatic mutations and methylation. In addition, a biobank of breast cancer patient-derived tumor xenograft (PDTX) models (n=92) has been generated in our lab and a comprehensive molecular characterization was also obtained. We recently demonstrated that breast cancer PDTXs maintain originating cancers intra-tumor heterogeneity, hence representing a more relevant preclinical model than cell lines. An ex-vivo drug screening was performed in these models generating response data (IC50 and AUC) for 100 different drugs, including "best in class" PI3K, PARP and CDK4/6 inhibitors, novel biological and chemical inhibitors of HER2, ER, IGF1R and HER3, as well as standard of care agents. Here we derived signatures of pathway activation/disruption by integrating different data types in the METABRIC cohort. Their association with previous breast cancer classifications, as well as their prognostic significance was studied. The predictive power of these signatures was investigated in the PDTX cohort to identify novel pharmacogenomics associations. They were tested independently as well as in combination, to derive molecular predictors of response to treatment. We found known and novel associations between genomic/transcriptomic features and drug response. For example, our results confirmed the known association between estrogen receptor-related genes and response to tamoxifen treatment. We also identified markers of response to inhibitors of the PI3K/AKT/mTOR pathway. Selected findings were validated in clinical cohorts as well as in independent PDTX models. In conclusion, by integrating molecular data from large cohorts of clinical samples and PDTX we have generated a computational framework for the systematic identification of pharmacogenomics associations in breast cancer and to generate hypothesis for rational drug-drug combinations. Citation Format: Maurizio Callari, Rajbir N. Batra, Ankita Sati Batra, Wendy Greenwood, Suet-Feung Chin, Alejandra Bruna, Oscar M. Rueda, Carlos Caldas. Integrative analysis of molecular and drug response data from clinical samples and PDTXs to identify pharmacogenomic associations in breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 2991. doi:10.1158/1538-7445.AM2017-2991