We previously showed that pretreatment detection of androgen receptor splice variant-7 (AR-V7) in circulating tumor cells (CTCs) from men with castration-resistant prostate cancer is associated with resistance to abiraterone and enzalutamide, but not to taxane chemotherapies. Here, we conducted serial measurements of AR-V7 and evaluated patterns of longitudinal AR-V7 dynamics over the course of multiple sequential therapies.

Metastatic prostate cancer patients treated at Johns Hopkins with AR-directed therapies or taxane chemotherapies underwent serial liquid biopsies for CTC-based AR-V7 analysis at baseline, during therapy, and at progression. We used a CTC enrichment platform followed by multiplexed reverse-transcription polymerase chain reaction analysis to detect full-length androgen receptor and AR-V7 transcripts. Patients selected for inclusion in this report were those who provided ≥4 CTC samples, at least one of which was AR-V7 positive, over the course of ≥2 consecutive therapies.

We identified 14 patients who received a total of 37 therapies and contributed 70 CTC samples for AR-V7 analysis during a median follow-up period of 11 months. Three patients remained AR-V7 positive during the entire course of therapy. The remainder underwent transitions in AR-V7 status: there were eight instances of ‘conversions’ from AR-V7-negative to -positive status (during treatment with first-line androgen deprivation therapy, abiraterone, enzalutamide, and docetaxel), and six instances of ‘reversions’ from AR-V7-positive to -negative status (during treatment with docetaxel and cabazitaxel).

AR-V7 is a dynamic marker, and transitions in AR-V7 status may reflect selective pressures on the tumor exerted by therapeutic interventions. While ‘conversions’ to AR-V7-positive status were observed with both AR-directed therapies and taxane chemotherapies, ‘reversions’ to AR-V7-negative status only occurred during taxane therapies. Serial blood-based AR-V7 testing is feasible in routine clinical practice, and may provide insights into temporal changes in tumor evolution.