SUMMARY Identifying patients with metastatic castration resistant prostate cancer (CRPC) that will no longer benefit from potent androgen receptor (AR)-directed therapies is an unmet need. One mechanism involves loss of AR signaling dependence and its early detection remains challenging. We have developed a genomic signature associated with AR-independence and the neuroendocrine phenotype that can be detected non-invasively using plasma samples from patients. In this proposal, we will develop a targeted assay applied to prospective cohorts to assess tumor dynamics and clinical impact of AR-independent genomic alterations in predicting response to subsequent AR-targeted therapies. We will also define the spectrum and pattern of circulating tumor clones compared to matched biopsies in patients with metastatic CRPC using a whole exome sequencing approach. This study would lead to further clinical development of a plasma genomic biomarker with several areas of potential impact including early detection of patients transforming towards AR- independence leading to early cessation of AR therapy and consideration of metastatic biopsy to look for neuroendocrine transformation and platinum-based therapies.