Androgen deprivation therapy (ADT) is the longstanding treatment for advanced prostate cancer (PC) because androgen receptor (AR) is the key therapeutic vulnerability for this disease. Bipolar androgen therapy (BAT) - the rapid cycling of supraphysiologic androgen (SPA) and low serum testosterone levels - is an alternative concept, but not all patients respond and acquired resistance can occur. In this issue of the JCI, Sena et al. developed a gene signature indicative of high AR activity to predict patient response to BAT, including a decline in both serum prostate-specific antigen (PSA) and tumor volume. Preclinical models showed that AR-mediated suppression of MYC, known to drive PC, was associated with decreased cell growth following SPA treatment. Because BAT eventually leads to resistance, the authors tested cycling between SPA and AR antagonism in a patient-derived xenograft and observed a delay in tumor growth. These findings represent a major step toward the informed use of BAT for advanced PC.