Developing and Evaluation of Computational Phenotypes of Metastatic Breast Cancer Using All of Us Data

Breast cancer is the most frequent cancer among women worldwide,1the metastatic breast cancer is the main cause of mortality for breast cancer patients. Computational phenotyping, an informatics approach, that extracts phenotypes from real-world data such as electronic health records (EHRs), has the potential to advance medicine’s understanding of who is most at risk for metastatic breast cancer. All of US (AoU) is a research program funded by the National Institutes of Health that aims to recruit one million highly diverse patients across the United States.2 The program makes data from participants accessible to both participants and approved researchers through the AoU “workbench.” In this study, we attempt to implement computational phenotypes from multiple sources to describe the prevalence of metastatic breast cancer in the AoU patient cohort. We used the AoU research workbench to implement the metastatic breast cancer phenotypes and describe cases within this rapidly growing a real clinical research database.