There is a need for response indicators that reliably reflect survival and occur early in patients with metastatic castration-resistant prostate cancer (mCRPC).

Recently, we demonstrated in real-life mCRPC patients that the measurement of circulating tumor cell (CTC) dynamics at 10-12 weeks during treatment with androgen-receptor signaling inhibitors provides independent prognostication over >50% PSA declines (De Laere, Prostate, 2018). Additionally, we demonstrated that absence of CTCs at 10-12 weeks, regardless if CTCs were detected at baseline, was associated with superior overall survival (OS).

Heller et al. demonstrated the high prognostic value of reaching undetectable CTC levels at week 13 after a nonzero baseline measurement (i.e. CTC0) in patients with mCRPC (Heller, J Clin Oncol, 2018) in a retrospective analysis of five prospective randomized phase III trials.

With the recent findings on the CTC0 endpoint, we revisited our data and performed a similar analysis.


We retrospectively analyzed data of a multicenter, prospective and non-interventional study in which blood samples for CTC enumeration were collected from patients with mCRPC starting treatment with abiraterone acetate or enzalutamide at baseline and at 10-12 weeks.


In our cohort, 63/102 (62%) patients with at least 10-12 weeks survival were CTC-positive at baseline and qualified for CTC0 response evaluation. In this group, 40/63 (63.5%) patients were chemotherapy-naive.

At 10-12 weeks, 28/63 (44%) of patients demonstrated a CTC0 response with a superior OS compared to CTC-positive patients (not reached versus 12.6 months, p<0.0001). In multivariable Cox regression analysis, not reaching a CTC0 response (HR 10.3, 95% CI 2.2-47.9, p<0.0001) was independently associated with shorter OS, whereas not reaching a ≥50% PSA decline was not.


The present analysis, although immature in OS events, demonstrates independent and confirmatory evidence of the potential clinical utility of CTC0 as measure of benefit in routine clinical practice. Interestingly, the evaluability rate in our study (62%) is lower than that in the study of Heller et al. (75%), probably because of the higher rate of chemotherapy-naive patients in our study. As the disease burden increases towards end-stage disease, the evaluability rates for CTC-driven response measure will increase accordingly.