Welcome, this website is intended for international healthcare professionals with an interest in the treatment of Advanced Prostate Cancer. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

Efficacy of Cabazitaxel in Castration-resistant Prostate Cancer Is Independent of the Presence of AR-V7 in Circulating Tumor Cells



Androgen receptor splice variant 7 (AR-V7) in circulating tumor cells (CTCs) from patients with metastatic castration-resistant prostate cancer (mCRPC) was recently demonstrated to be associated with resistance to abiraterone and enzalutamide. Cabazitaxel might, however, remain effective in AR-V7-positive patients.


To investigate the association between AR-V7 expression in CTCs and resistance to cabazitaxel.

Design, setting, and participants

We selected patients with mCRPC from the multicenter, randomized, phase 2, randomized, open-label, multicenter study in mCRPC on the pharmacodynamic effects of budesonide on cabazitaxel (Jevtana) (CABARESC). Before the start of the first and third cabazitaxel cycle, CTCs were enumerated using the CellSearch System. In patients with ≥10 CTCs in 7.5 ml blood at baseline, the expression of AR-V7 was assessed by quantitative polymerase chain reaction.

Outcome measures and statistical analysis

The primary end point was the association between the AR-V7 status and the CTC response rate (decrease to fewer than five CTCs in 7.5 ml blood during treatment). Secondary end points were the prostate-specific antigen (PSA) response rate (RR) and overall survival (OS). Analyses were performed using chi-square and log-rank tests.

Results and limitations

AR-V7 was detected in 16 of 29 patients (55%) with ≥10 CTCs and was more frequently found in abiraterone pretreated patients (5 of 5 [100%] treated vs 7 of 20 [35%] untreated; p = 0.009). We found no differences in CTC and PSA RRs. The presence of AR-V7 in CTCs was not associated with progression-free survival (hazard ratio [HR]: 0.8; 95% confidence interval [CI], 0.4–1.8) or overall survival (HR 1.6; 95% CI, 0.6–4.4).


The response to cabazitaxel seems to be independent of the AR-V7 status of CTCs from mCRPC patients. Consequently, cabazitaxel might be a valid treatment option for patients with AR-V7-positive CTCs.

Patient summary

Tools are needed to select specific treatments for specific patients at specific times. The presence of the gene AR-V7 in CTCs has been associated with resistance to anti-androgen receptor treatments. We investigated whether this holds true for cabazitaxel, but we found cabazitaxel to be effective independent of the presence of AR-V7.

Take Home Message

Characterization of AR-V7 expression is feasible in CellSearch-enriched circulating tumor cells and might guide optimal treatment sequencing in patients with metastatic castration-resistant prostate cancer. Presence of AR-V7 does not predict resistance to cabazitaxel and does not affect prognosis following cabazitaxel treatment.

Keywords: AR-V7, Androgen receptor, Cabazitaxel, Castration-resistant prostate cancer, CellSearch, Circulating tumor cells, Metastatic prostate cancer.

1. Introduction

Several new treatment options have become available for patients with metastatic castration-resistant prostate cancer (mCRPC). Abiraterone and enzalutamide, both acting on androgen receptor (AR) signaling, improve overall survival (OS) both in the pre- and post-docetaxel setting [1], [2], [3], [4], [5], and [6]. Cabazitaxel, the next-generation taxane, has been developed to overcome docetaxel resistance and improves OS in mCRPC patients pretreated with docetaxel [3] and [7]. With the arrival of these treatments, the question of how to optimally sequence treatment lines for mCRPC patients has arisen. Preclinical and clinical data indicate cross-resistance between abiraterone, enzalutamide, and docetaxel [8], [9], [10], [11], and [12]. However, patients pretreated with abiraterone, enzalutamide, and docetaxel still appear to benefit from cabazitaxel [7], [13], and [14]. Reliable predictive factors reflecting tumor characteristics in real-time are thus urgently needed to guide treatment selection.

A circulating tumor cell (CTC) count from peripheral blood before and during treatment is an independent prognostic factor for progression-free survival (PFS) and OS in mCRPC, and it outperforms prostate-specific antigen (PSA) measurements as an early treatment response marker [15], [16], [17], [18], and [19]. The presence of the AR splice variant 7 (AR-V7), coding for a truncated and constitutively active androgen receptor (AR), in CTCs has been found to be associated with resistance to enzalutamide and abiraterone but not to taxanes, mainly docetaxel [20] and [21]. We investigated the association of AR-V7 in CTCs with the response to cabazitaxel in docetaxel-pretreated mCRPC patients. We set up a highly specific reverse-transcriptase quantitative polymerase chain reaction (RT-qPCR) assay to measure messenger RNA (mRNA) expression levels of wild-type AR (AR-WT) and AR-V7 in CTCs enriched by the CellSearch System (Janssen Diagnostics LLC, Raritan, NJ, USA). Extensive and robust data are available concerning the clinical relevance of CTCs enumerated by this relatively widely available US Food and Drug Administration (FDA)-cleared technique. Next, we explored associations between the presence of AR-V7 in CTCs taken before start of cabazitaxel and the outcome to cabazitaxel.

2. Patients and methods

2.1. Patients

Patients with mCRPC were recruited from an ongoing, multicenter, randomized phase 2 trial investigating the effects of budesonide on cabazitaxel toxicity (CABARESC; Dutch Trial Registry no. NTR2991). All patients had progressive disease after docetaxel (three rising PSA measurements ≥2 wk apart, PSA increase ≥2.0 μg/l, or radiologic progression). Full inclusion criteria are listed in Supplement 1. All patients received 25 mg/m2 of cabazitaxel until progression, unacceptable toxicity, or the maximum of 10 cycles. The collection of CTC samples was a side study of the CABARESC trial. For this study, we selected patients who had been included between August 2012 and August 2014 with ≥10 CTCs in 7.5 ml blood before the start of cabazitaxel to ensure robust and CTC-specific downstream analysis. The Erasmus Medical Center and local institutional review boards approved the study (METC 11-324). All patients provided written informed consent.

2.2. Sample processing

Before the start of the first and the third cycle of cabazitaxel, CTCs were enumerated from 7.5 ml blood drawn in a CellSave tube using the CellSearch System (Janssen Diagnostics LLC, Raritan, NJ, USA). Characterization of CTCs was done before the first cycle of cabazitaxel from 7.5 ml ethylenediaminetetraacetic acid (EDTA) blood, which was processed using the CellSearch Profile Kit. After RNA isolation, cDNA generation, and preamplification, expression levels of AR-WT and AR-V7 were measured by RT-qPCR in an 11% aliquot of the original starting material using Taqman Gene Expression Assays (Applied Biosystems, Carlsbad, CA, USA) (Supplementary Fig. 1). Details on sample processing are available in Supplement 2.

The performance of the assays was tested through the analysis of 17 healthy blood donors (HBDs) and prostate (22RV1, LNCaP, PC3, and VCaP) and breast (CAMA1, MDA-MB-415, MDA-MB-453, MPE600, SUM185PE, and ZR75.1) cancer cell lines (Supplementary Fig. 1 and 2). A total of 100 cell-line cells were spiked in 7.5 ml HBD blood and CellSearch enriched to serve as negative and positive controls: 22RV1 (WThigh/V7high), CAMA1 (WTlow/V7neg), LNCaP (WThigh/V7low), MDA-MB-415 (WTlow/V7neg), MDA-MB-453 (WTlow/V7neg), MPE600 (WTlow/V7neg), PC3 (WTneg/V7neg), SUM185PE (WTlow/V7low), VCaP (WThigh/V7high), and ZR75.1 (WTlow/V7low). All samples were processed in a similar way to the patient blood samples.

2.3. Normalization and statistical analysis

Samples with an average cycle threshold for quantification (Cq) <26.5 for the three reference genes (glucoronidase, beta [GUSB]; hydroxymethylbilane [HMBS]; and hypoxanthine phosphoribosyltransferase 1 [HPRT1]) and an average Cq of the two epithelial genes <26.5 (epithelial cell adhesion molecule [EPCAM], keratin 19, type I [KRT19]) were considered evaluable. To correct for CTC count and epithelial tumor cell input, Cq values of AR-V7 and AR-WT were normalized to the average Cq value of the epithelial genes (Spearman's r [rs] with CTC count 0.7; p < 0.01) (Supplementary Fig. 3a). Final epithelial tumor cell input in the aliquot of RNA used was calculated using the equation derived from the regression line of the correlation between the epithelial genes and the CTC count, thereby taking into account that only 11% of the originally isolated RNA from all CTCs in the sample was used for the characterization of AR-V7 status (Supplementary Fig. 3a). A cut-off value for positivity for AR-V7 was determined based on the cell line and HBD experiments (Supplement 2).

The primary end point of this study was to compare the CTC response rate (CTC RR), defined as a decrease to fewer than five CTCs in 7.5 ml blood during treatment, between patients with AR-V7-positive and AR-V7-negative CTCs. Secondary objectives were the PSA RR (30% or 50% decline in PSA level from baseline to 12 wk or earlier in the case of treatment discontinuation), best PSA response during treatment, PFS (interval between registration and progression of disease or death), and OS (interval between registration and death). Associations between PFS or OS and the CTC response during treatment were analyzed after the second blood draw. Patients without events were censored at the last date recorded to be progression free and/or alive. Reported end points were based on the Prostate Cancer Working Group 2 guidelines [22].

The main hypothesis stated that there would be no difference in response to cabazitaxel by the presence or absence of AR-V7. Since limited data regarding the prevalence of AR-V7 were available at the time of the study design, no formal sample size calculations were performed. Therefore, our analyses were exploratory. Differences in the primary objective, CTC RR, and secondary objective, PSA RR, were analyzed using the chi-square or Fisher exact tests. Survival was analyzed using Cox regression models and visualized in Kaplan-Meier plots. Other applied tests were the Student t test, the Mann-Whitney U test, and Pearson or Spearman correlation, depending on the variable or distribution of a variable. All statistical tests were two-sided and performed using the SPSS 21.0 software package (IBM Corp., Armonk, NY, USA). A p value <0.05 was considered statistically significant.

3. Results

3.1. AR-WT and AR-V7 in CTCs

We first tested the sensitivity and specificity of our assays by comparing RNA fractions isolated from pure and spiked-in breast and prostate cancer cell-line cells before and after CellSearch enrichment (Supplementary Fig. 1 and 2). The AR-V7 status could be reliably determined in three or more spiked-in epithelial cells. The cut-off was confirmed in our clinical samples, in which two patients with RNA from three CTCs in the used aliquot were positive for AR-V7; none of the patients with fewer than three CTCs were positive for AR-V7 (Supplementary Fig. 2). The leukocyte background did not influence the outcomes of our analyses (Supplementary Fig. 3b and 3c). Of the 17 HBDs tested, 16 were negative for the expression of AR-WT and AR-V7 (Supplementary Fig. 2). One 67-yr-old male HBD had detectable AR-WT in his peripheral blood. Since this donor was anonymous, no follow-up or further diagnostics were done.

We next selected patients with ≥10 CTCs at baseline to limit stochastic variations between the CTC enumeration and isolation tubes and to assure epithelial input. Twenty-nine patients with sufficient RNA quality and quantity and sufficient epithelial cell input in the CTC samples were identified (Fig. 1). Table 1 shows all patient characteristics. Five patients had received abiraterone before enrollment. The expression of AR-WT in CTCs was detected in all patients, whereas AR-V7 was detected in 16 patients (55%). All 5 patients who had previously been treated with abiraterone expressed AR-V7 compared with 7 of the 20 patients (35%) who had not received abiraterone (p = 0.01). We found no significant correlation between the expression levels of AR-V7 and AR-WT in CTCs (rs = 0.3; p = 0.12) (Supplementary Fig. 4a) and no difference in AR-WT expression levels between patients with and without AR-V7 in the CTCs (p = 0.2) (Supplementary Fig. 4b).


Fig. 1 Study flow chart showing the selection of patients for the analyses.Cq = cycle threshold for quantification; CTC = circulating tumor cell.

Table 1 Patient characteristics

AR-V7 in CTCs at baseline
All patients Absent Present p value *
Patients 29 100% 13 100% 16 100%
Age at registration, yr, mean ± SD 70 ± 7 68 ± 9 71 ± 6 0.3
WHO performance score 0.4
0 11 38% 6 46% 5 31%
1 18 62% 7 54% 11 69%
Type of castration 0.4
Surgical 4 14% 1 8% 3 19%
LHRH agonist 25 86% 12 92% 13 81%
Prior chemotherapy lines, no. 0.9
1 (docetaxel) 27 93% 12 92% 15 94%
2 2 7% 1 8% 1 6%
Prior antiandrogens for mCRPC
Abiraterone 5 17% 0 0% 5 31% 0.009
Orteronel 3 10% 3 23% 0 0% 0.09
Baseline chemistry **
LDH, U/l, median (IQR) 453 (309–635) 431 (310–616) 456 (287–674) 0.9
ALP, U/l, median (IQR) 163 (106–375) 160 (96–358) 228 (107–384) 0.7
PSA, μg/l, median (IQR) 321 (76–649) 107 (68–439) 475 (78–885) 0.08
Baseline CTC count, median (IQR) 100 (50–243) 94 (38–260) 110 (52–254) 0.6

* Reported p values are from an independent sample Student t test (age), nonparametric Mann-Whitney U test (baseline chemistry and CTC count), and chi-square test (categorical variables).

** Upper limit of normal: LDH, 247 U/l; ALP, 114 U/l; PSA, 6.4 μg/l.

ALP = alkaline phosphatase; AR-V7 = androgen receptor splice variant 7; CTC = circulating tumor cell; IQR = interquartile range; LDH = lactate dehydrogenase; LHRH = luteinizing hormone-releasing hormone; mCRPC = metastatic castration-resistant prostate cancer; PSA = prostate-specific antigen; SD = standard deviation; WHO = World Health Organization.

3.2. AR-V7 and response to cabazitaxel

The primary end point of this study was the CTC RR, defined as a decrease to fewer than five CTCs per 7.5 ml blood after two cabazitaxel cycles, determined by the presence or absence of AR-V7 in baseline CTCs. A secondary CTC sample was available from 25 patients. In three patients, the second blood draw was missed, and one patient died after the second cycle because of a non–disease-related event. The overall CTC RR to cabazitaxel was 5 of 25 patients (20%). Fifteen patients had AR-V7-positive and 10 had AR-V7-negative CTCs. The CTC RRs in both AR-V7-positive and AR-V7-negative patient groups were 20% (Table 2). Sequential PSA levels during cabazitaxel treatment for evaluation of the PSA RR were available from 26 patients. Five (17%) and three patients (10%) achieved a ≥30% and ≥50% PSA response, respectively, after 12 wk of treatment. At the end of treatment, the best PSA response was ≥30% in seven patients (24%) and ≥50% in three patients (10%). The 30% and 50% PSA RRs at 12 wk and at the end of treatment in patients with and without AR-V7 in CTCs were not statistically different (Table 2, Fig. 2). We found no statistical difference in CTC RR and PSA RR between patients who had or had not received abiraterone before cabazitaxel.

Table 2 Presence of AR-V7 in circulating tumor cells (CTCs) at baseline versus CTC response to cabazitaxel after two cycles, prostate-specific antigen (PSA) response after 12 weeks of treatment, and best PSA response at the end of treatment

AR-V7 CTC response PSA response at week 12 Best PSA response
No Yes No ≥30%* ≥50% No ≥30%* ≥50%
No 8 2 9 3 2 7 4 2
Yes 12 3 12 2 1 12 3 1
p value Fisher exact test, p = 1 χ2 test, p = 0.7 χ2 test, p = 0.6

* Numbers include patients with ≥50% PSA response. Sequential PSA values from three patients were missing (two AR-V7 positive and one negative). One AR-V7–positive patient discontinued treatment after two cycles of cabazitaxel and was not included in the analysis of PSA response after 12 wk. The AR-V7–negative patient was still undergoing treatment and thus was included only in the analysis for PSA response after 12 wk.

CTC = circulating tumor cells; PSA = prostate-specific antigen.


Fig. 2 Waterfall plots of prostate-specific antigen (PSA) responses to cabazitaxel treatment (A) after 12 wk and (B) at the end of treatment. The dashed lines represent 30% and 50% decreases in PSA level relative to the baseline level. No differences in PSA responses were observed between AR-V7–positive and AR-V7–negative patients. Sequential PSA values from three patients were missing (two AR-V7 positive and one negative). One AR-V7–positive patient discontinued treatment after two cycles of cabazitaxel and was not included in the analysis for PSA response after 12 wk. The AR-V7–negative patient was still undergoing treatment and thus was included only in the analysis for the PSA response after 12 wk.CTC = circulating tumor cell; PSA = prostate specific antigen.* Patients who were treated with abiraterone before cabazitaxel.

3.3. AR-V7 and survival

At the time of analysis, four patients were still receiving cabazitaxel treatment. The median follow-up time from the date of registration for the 12 patients still alive was 7 mo (range: 2–27 mo). The median OS in all 29 patients was 10 mo (95% CI, 5–14 mo); median PFS was 5 mo (95% CI, 2–8 mo). The five patients with a CTC response to cabazitaxel had a significantly longer OS than the 20 patients without a CTC response (hazard ratio [HR]: 0.1; 95% CI, 0.01–0.9; p = 0.04), but had a comparable PFS (HR: 0.7; 95% CI, 0.2–2.0; p = 0.5). The presence of AR-V7 in CTCs at baseline was not associated with PFS (HR: 0.8; 95% CI, 0.4–1.8; p = 0.6) or OS (HR: 1.6; 95% CI, 0.6–4.4; p = 0.4) (Fig. 3). Treatment with abiraterone before or after cabazitaxel had no influence on OS (Supplementary Fig. 5).


Fig. 3 (A) Progression-free survival and (B) overall survival as a function of AR-V7 in circulating tumor cells at baseline. The reported p value is from a log-rank test.AR-V7 =androgen receptor splice variant 7; CI = confidence interval; HR = hazard ratio.

4. Discussion

The presence of AR-V7 in CTCs of mCRPC patients is associated with resistance to enzalutamide or abiraterone but not to taxanes [20] and [21]. In these studies, CTCs were enriched using an mRNA-based method; limited data exist about the method's clinical relevance in mCRPC. We explored the feasibility of the characterization of the presence of AR-V7 in CTCs captured by the CellSearch System, which obtained FDA clearance for clinical use of the CTC count. We set up a robust RT-qPCR assay that reliably detects AR-V7 in three or more CTCs and investigated the association between the AR-V7 status of CTCs and outcome to cabazitaxel. In contrast to docetaxel, no cross-resistance seemed to emerge among cabazitaxel, abiraterone, and enzalutamide [8] and [14]. Consequently, we hypothesized that patients with AR-V7-positive CTCs would still benefit from cabazitaxel.

The prevalence AR-V7 in our cohort of 29 docetaxel-pretreated mCRPC patients with ≥10 CTCs per 7.5 ml blood was 55%, which seems higher than the 29% in the previously reported enzalutamide/abiraterone cohort [20], but comparable to the 46% in the prior taxane cohort [21]. We confirmed the higher prevalence of AR-V7 in abiraterone-resistant patients. In line with our hypothesis, we found indications that the presence of AR-V7 in CTCs taken prior to treatment might not be associated with the outcomes of cabazitaxel treatment in terms of CTC RR, PFS, and OS. The CTC RR, defined as a decrease to fewer than five CTCs in 7.5 ml blood during treatment, has been shown to be a robust surrogate end point for PFS and OS in several prior studies [15], [16], [17], [18], [19], and [23].

The lack of an association between AR-V7 in CTCs and outcome is in agreement with the findings in 37 patients starting treatment with docetaxel (n = 30) or cabazitaxel (n = 7) [21], and contrasts with the results in patients treated with enzalutamide or abiraterone [20]. However, comparisons have to be made with caution because of the differences in methodology and patient selection. In the previous studies, CTCs were detected using the AdnaTest (AdnaGen, Langenhagen, Germany), whereas we used the CellSearch System. Both methods immunomagnetically enrich CTCs based on the expression of the EpCAM, but there are important differences to consider; for example, the AdnaTest also enriches CTCs expressing the human epidermal growth factor receptor 2 (HER2). Although frequently detected on breast cancer CTCs [24], the expression and clinical relevance to prostate cancer CTCs is unclear. After enrichment, the CellSearch system identifies a CTC as an intact, nucleated cell with expression of cytokeratin as assessed by immunofluorescence. In the AdnaTest, all morphologic information is lost after lysis of the enriched cells, so CTC enumeration is not possible. The presence of CTCs is assumed by the presence of the epithelial gene transcripts, thereby disregarding other characteristics, such as the presence of a nucleus or intact cell membrane. Considering the differences in methodology, the AdnaTest and the CellSearch System might not detect comparable cell populations. Therefore, we have started a clinical trial to investigate the predictive value of the presence of AR-V7 in CellSearch-enriched CTCs for outcome to cabazitaxel as well as to AR-targeted treatments.

The limitations of our study concern the CellSearch System's dependency on EpCAM expression on CTCs of the CellSearch System. In breast cancer, EpCAM-negative CTCs have been detected and have even been reported to be more strongly predictive of treatment resistance [25] and [26]. Whether this applies to prostate cancer CTCs remains to be investigated. Since only an aliquot of the total of isolated RNA could be used for the current study, patients with ≥10 CTCs—and thus a poor prognosis based on the high baseline CTC count—were selected. Patients with insufficient quality and quantity of mRNA were excluded from the analyses. Although potentially introducing a selection bias, this assured sufficient epithelial input to reliably measure the AR-V7 status. In addition, patients were recruited from a phase 2 study investigating cabazitaxel toxicity. As survival was not an end point of the main study and PFS was not defined in the study protocol, our PFS analyses were a composite of PSA, radiographic, and clinical progression, which were assessed at the discretion of the treating physician. This might explain the lack of prognostic value of CTCs for PFS. Last, our analyses were exploratory, because no formal power calculations were possible at the time of study design and only a small group of patients was included. We plan to validate our findings by extending the patient cohort, thereby including patients with <10 CTCs. Ultimately, prospective, randomized trials, taking into account all other baseline characteristics that might affect outcome, should offer insights into the exact role of cabazitaxel in the treatment of AR-V7-positive patients.

5. Conclusions

We demonstrated the feasibility of measuring the AR-V7 status of mCRPC patients with ≥10 CTCs after CellSearch enrichment. We showed that the outcome of cabazitaxel treatment in these patients is not associated with the presence of this particular splice variant. Our results add important information to the existing evidence that CTCs are an invaluable tool for personalizing cancer treatments and improving the prognosis of mCRPC patients by allowing optimal treatment sequencing.

Author contributions: Wendy Onstenk had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: De Laere, Jenster, Foeken, Martens, Dirix, Lolkema, Mathijssen, de Wit, Sleijfer, Onstenk, van Weerden.

Acquisition of data: Nieuweboer, Sieuwerts, Kraan, Van, van Soest, Onstenk.

Analysis and interpretation of data: Sieuwerts, Onstenk.

Drafting of the manuscript: Sieuwerts, Onstenk.

Critical revision of the manuscript for important intellectual content: Sieuwerts, Jenster, Kraan, Foekens, Martens, Lolkema, Hamberg, de Wit, Mathijssen, Sleijfer, van Weerden.

Statistical analysis: Sieuwerts, Onstenk.

Obtaining funding: Lolkema, Mathijssen, de Wit, Sleijfer.

Administrative, technical, or material support: Nieuweboer, De Laere, Kraan, Van, van Soest.

Supervision: Foekens, Martens, de Wit, Sleijfer.

Other (specify): Patient inclusion and collection of blood samples: Nieuweboer, Meulenbeld, Hamberg, van Soest, de Wit, Sleijfer.

Financial disclosures: Wendy Onstenk certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: A. Sieuwerts is supported by Cancer Genomics Netherlands and a grant from the Dutch Organization of Scientific Research (NWO). J. Martens, J. Foekens, and S. Sleijfer received research funding from Janssen Diagnostics (formerly Veridex LLC). M. Lolkema received research funding from Astellas. R. Mathijssen received funding from Sanofi and Astellas. R. de Wit received research funding and consultancy fees from Sanofi, Janssen, Astellas, and Millenium. W. van Weerden has received research funding from Janssen, Sanofi, Millenium, and AstraZeneca.

Funding/Support and role of the sponsor: Sanofi supported this study and reviewed and approved the manuscript.

Acknowledgments: The authors thank Michelle Daane for her technical support and the all patients for their willingness to participate in this study. We also thank all the medical oncologists, research nurses, and laboratory technicians involved in the study for their assistance with patient inclusion, sample processing, and data analysis.

Appendix A. Supplementary data

Supplementary Fig. 1 – Details of the assays used in the RT-qPCRs.

Supplementary Fig. 2 – RT-qPCR data of all tested samples. Five healthy blood donors and 11 breast and prostate cancer cell lines – pure and after spiking-in healthy blood donor blood and CellSearch enrichment – served as negative and positive controls. In total, 44 patient samples were tested, of which 29 were of sufficient quality to be entered in the analyses. After CellSearch enrichment, an aliquot of 11% was used for the assessment of AR-V7 status. The calculated final number of CTCs used for the PCR analysis was derived from the equation of the regression line of the correlation between the CTC count from 7.5 mL blood – as measured by CellSearch enumeration from the CellSave tube – and the average Cq of EPCAM and KRT19 (Supplementary Fig. 3a), thereby taking into account that only 11% of the original sample was used.

Supplementary Fig. 3 – Sensitivity and specificity of the assays. (a) Material from 40 mCRPC patients was used to evaluate the correlation between CTC count after CellSearch enrichment and the average Cq value of EPCAM and KRT19 in corresponding RNA samples (CTC count = 260466e-0.333*Cqaverage(EPCAM+KRT19)). (b) Data from 62 individual experiments with RNA input from 4–580 VCAP cells in the RT-qPCR were used to evaluate the linear correlation between the Cq value of AR-V7 and the average Cq value of EPCAM +KRT19. Samples with an average Cq value of EPCAM + KRT19 below 26.5 Cq were considered to contain a sufficient epithelial signal to allow measurement of AR-V7 in these cells. Circles: unspiked VCAP cells; squares: VCAP cells spiked in HBD blood. (c) Sensitivity and specificity measuring AR-WT and AR-V7 by RT-qPCR in VCAP cells before and after spiking in HBD, before and after CellSearch enrichment, and before and after pre-amplification. Data are expressed relative to the average expression of EPCAM + KRT19 (ΔCq) measured in these preparations. Within a window of ± 1.1 Cq, both transcripts can be reproducibly measured in material from as little as two VCAP cells. Dark blue circles: ΔCqAR-WT; squares; light blue circles: ΔCqAR-V7.

Supplementary Fig. 4 – (a) Transcription levels of AR-WT versus AR-V7. No correlation was observed (Spearman's r=0.3, P=0.1). (b) Transcription levels of AR-WT in patients without AR-V7 (left panel) and with AR-V7 (right panel) in CTCs. Horizontal lines represent the medians. There was no difference in transcription levels (Mann Whitney U, P=0.2).

Supplementary Fig. 5 – (a) Overall survival in patients who did or did not receive abiraterone before cabazitaxel, and (b) patients who did or did not receive abiraterone or enzalutamide after cabazitaxel treatment. Clinical data concerning pre-treatments of five patients were still missing, as the clinical trial was ongoing and recruiting at the time of data analysis. From one patient, no follow-up data was yet available.


  • [1] H.I. Scher, K. Fizazi, F. Saad, et al. Increased survival with enzalutamide in prostate cancer after chemotherapy. N Engl J Med. 2012;367:1187-1197
  • [2] K. Fizazi, H.I. Scher, A. Molina, et al. Abiraterone acetate for treatment of metastatic castration-resistant prostate cancer: final overall survival analysis of the COU-AA-301 randomised, double-blind, placebo-controlled phase 3 study. Lancet Oncol. 2012;13:983-992 Crossref
  • [3] J.S. De Bono, S. Oudard, M. Ozguroglu, et al. Prednisone plus cabazitaxel or mitoxantrone for metastatic castration-resistant prostate cancer progressing after docetaxel treatment: a randomised open-label trial. Lancet. 2010;376:1147-1154 Crossref
  • [4] E. Basch, K. Autio, C.J. Ryan, et al. Abiraterone acetate plus prednisone versus prednisone alone in chemotherapy-naive men with metastatic castration-resistant prostate cancer: patient-reported outcome results of a randomised phase 3 trial. Lancet Oncol. 2013;14:1193-1199 Crossref
  • [5] C.J. Ryan, M.R. Smith, J.S. de Bono, et al. Abiraterone in metastatic prostate cancer without previous chemotherapy. N Engl J Med. 2013;368:138-148 Crossref
  • [6] T.M. Beer, A.J. Armstrong, D.E. Rathkopf, et al. Enzalutamide in metastatic prostate cancer before chemotherapy. N Engl J Med. 2014;371:424-433 Crossref
  • [7] R. De Leeuw, L.D. Berman-Booty, M.J. Schiewer, et al. Novel actions of next-generation taxanes benefit advanced stages of prostate cancer. Clin Cancer Res. 2015;21:795-807 Crossref
  • [8] R.J. Van Soest, M.E. van Royen, E.S. de Morree, et al. Cross-resistance between taxanes and new hormonal agents abiraterone and enzalutamide may affect drug sequence choices in metastatic castration-resistant prostate cancer. Eur J Cancer. 2013;49:3821-3830 Crossref
  • [9] R. Nadal, Z. Zhang, H. Rahman, et al. Clinical activity of enzalutamide in docetaxel-naive and docetaxel-pretreated patients with metastatic castration-resistant prostate cancer. Prostate. 2014;74:1560-1568 Crossref
  • [10] S. Badrising, V. van der Noort, I.M. van Oort, et al. Clinical activity and tolerability of enzalutamide (MDV3100) in patients with metastatic, castration-resistant prostate cancer who progress after docetaxel and abiraterone treatment. Cancer. 2014;120:968-975 Crossref
  • [11] D. Bianchini, D. Lorente, A. Rodriguez-Vida, et al. Antitumour activity of enzalutamide (MDV3100) in patients with metastatic castration-resistant prostate cancer (CRPC) pre-treated with docetaxel and abiraterone. Eur J Cancer. 2014;50:78-84 Crossref
  • [12] M.T. Schweizer, X.C. Zhou, H. Wang, et al. The influence of prior abiraterone treatment on the clinical activity of docetaxel in men with metastatic castration-resistant prostate cancer. Eur Urol. 2014;66:646-652 Crossref
  • [13] C.J. Pezaro, A.G. Omlin, A. Altavilla, et al. Activity of cabazitaxel in castration-resistant prostate cancer progressing after docetaxel and next-generation endocrine agents. Eur Urol. 2014;66:459-465 Crossref
  • [14] R.J. Van Soest, E.S. de Morree, C.F. Kweldam, et al. Targeting the androgen receptor confers in vivo cross-resistance between enzalutamide and docetaxel, but not cabazitaxel, in castration-resistant prostate cancer. Eur Urol. 2015;67:981-985 Crossref
  • [15] J.S. De Bono, H.I. Scher, R.B. Montgomery, et al. Circulating tumor cells predict survival benefit from treatment in metastatic castration-resistant prostate cancer. Clin Cancer Res. 2008;14:6302-6309 Crossref
  • [16] H.I. Scher, X. Jia, J.S. de Bono, et al. Circulating tumour cells as prognostic markers in progressive, castration-resistant prostate cancer: a reanalysis of IMMC38 trial data. Lancet Oncol. 2009;10:233-239 Crossref
  • [17] D. Olmos, H.T. Arkenau, J.E. Ang, et al. Circulating tumour cell (CTC) counts as intermediate end points in castration-resistant prostate cancer (CRPC): a single-centre experience. Ann Oncol. 2009;20:27-33
  • [18] M.H. Strijbos, J.W. Gratama, P.I. Schmitz, et al. Circulating endothelial cells, circulating tumour cells, tissue factor, endothelin-1 and overall survival in prostate cancer patients treated with docetaxel. Eur J Cancer. 2010;46:2027-2035 Crossref
  • [19] A. Goldkorn, B. Ely, D.I. Quinn, et al. Circulating tumor cell counts are prognostic of overall survival in SWOG S0421: a phase III trial of docetaxel with or without atrasentan for metastatic castration-resistant prostate cancer. J Clin Oncol. 2014;32:1136-1142 Crossref
  • [20] E.S. Antonarakis, C. Lu, H. Wang, et al. AR-V7 and resistance to enzalutamide and abiraterone in prostate cancer. N Engl J Med. 2014;371:1028-1038 Crossref
  • [21] E.S. Antonarakis, C. Lu, B. Luber, et al. Androgen receptor splice variant 7 and efficacy of taxane chemotherapy in patients with metastatic castration-resistant prostate cancer. JAMA Oncol. 2015;1:582-591 Crossref
  • [22] H.I. Scher, S. Halabi, I. Tannock, et al. Design and end points of clinical trials for patients with progressive prostate cancer and castrate levels of testosterone: recommendations of the Prostate Cancer Clinical Trials Working Group. J Clin Oncol. 2008;26:1148-1159 Crossref
  • [23] H.I. Scher, G. Heller, A. Molina, et al. Circulating tumor cell biomarker panel as an individual-level surrogate for survival in metastatic castration-resistant prostate cancer. J Clin Oncol. 2015;33:1348-1355 Crossref
  • [24] W. Onstenk, J.W. Gratama, J.A. Foekens, S. Sleijfer. Towards a personalized breast cancer treatment approach guided by circulating tumor cell (CTC) characteristics. Cancer Treat Rev. 2013;39:691-700 Crossref
  • [25] A. Sieuwerts, J. Kraan, J. Bolt, et al. Anti-epithelial cell adhesion molecule antibodies and the detection of circulating normal-like breast tumor cells. J Natl Cancer Inst. 2009;101:61-66 Crossref
  • [26] M. Yu, A. Bardia, B.S. Wittner, et al. Circulating breast tumor cells exhibit dynamic changes in epithelial and mesenchymal composition. Science. 2013;339:580-584 Crossref


a Erasmus Medical Center Cancer Institute and Cancer Genomics Netherlands, Department of Medical Oncology, Rotterdam, The Netherlands

b Sint Franciscus Gasthuis, Department of Internal Medicine, Rotterdam, The Netherlands

c Gelre Ziekenhuizen, Department of Internal Medicine, Zutphen, The Netherlands

d GZA Hospitals Saint Augustinus, Wilrijk, Belgium

e Center for Oncological Research, University of Antwerp, Antwerp, Belgium

f Erasmus MC, Department of Urology, Rotterdam, The Netherlands

Corresponding author. Erasmus MC Cancer Institute, Department of Medical Oncology and Cancer Genomics Netherlands, PO Box 2040, 3000 CA, Rotterdam, The Netherlands. Tel. +31 10 704 13 24; Fax: +31 10 704 10 05.