Grant: Collaborative Health Research Projects (NSERC partnered)

Breast cancer is the second leading cause of cancer deaths in Canadian women, with 23,800 new cases and an estimated death rate of 21% in Canada in 2013. The advent of personalized medicine, which consists of using molecular data (mutations in cancer-related genes for instance) to tailor treatment to the individual patient, holds the promise to significantly improve cancer management in our modern society. However, while thousands of anticancer drugs have been developed, few of them were translated into efficient therapeutic strategies in breast cancer. The primary reason for such a discrepancy is the inability to predict patients' response to a given therapeutic drug. There is therefore a dire need for developing new clinical tests that can help clinicians select the most beneficial therapy for each individual breast cancer patient. Predicting drug response is a challenging task due to the complex nature of breast cancer. New molecular features enabling prediction of therapy response, are usually discovered using either immortalized cancer cells or patients' tumor biopsies. More recently, patient-derived xenografts (PDX) where human tumors are implanted in mice, also enabled to test drugs in a model that recapitulate most of the features of real patients' tumors. These preclinical models have their own weaknesses and strengths. To date these data have never been combined to develop accurate predictors of drug response. We propose in our project to use our recently published network-based method called SNF to efficiently combine cancer cell lines, PDX and patients' tumors to build better predictors of drug response and test them in new patients' samples collected within ongoing clinical trials. Upon successfully validation, we will closely work with our collaborators to implement our new computational tool in clinical routine with the aim to improve matching of patients to the clinical trials testing the drugs from which they most benefit.