Background
The activity of enzalutamide after prior treatment with both abiraterone acetate (abiraterone) and docetaxel has been examined in several retrospective studies. However, limited data are available on the efficacy of enzalutamide following abiraterone in chemotherapy-naive patients with metastatic castration-resistant prostate cancer (mCRPC).
Objective
To compare the activity of enzalutamide after abiraterone in docetaxel-experienced and docetaxel-naive mCRPC patients.
Design, setting, and participants
The British Columbia Cancer Agency Cancer Registry was searched for mCRPC patients who received enzalutamide after prior abiraterone. Clinicopathologic characteristics, confirmed prostate-specific antigen (PSA) response rates (PSA decline ≥50% confirmed ≥3 wk later), and survival data were collected.
Outcome measurements and statistical analysis
Outcomes on enzalutamide were compared between docetaxel-experienced and docetaxel-naive patients using chi-square for PSA response and log-rank test for time to PSA progression and overall survival (OS). Univariate analysis was performed to identify variables associated with confirmed PSA response on enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables.
Results and limitations
A total of 115 patients received enzalutamide after abiraterone: 68 had received prior docetaxel and 47 were docetaxel naive. Median time on enzalutamide was 4.1 mo. Confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2) did not differ significantly between docetaxel-experienced and docetaxel-naive patients. No clinical variables (including prior response to abiraterone) were found to associate significantly with confirmed PSA response to enzalutamide.
Conclusions
Antitumour activity of enzalutamide following abiraterone was limited in mCRPC patients irrespective of prior docetaxel use. Identifying clinical and molecular factors predictive of response to enzalutamide remains a high priority for future research.
Patient summary
We looked at the effectiveness of enzalutamide after abiraterone acetate for treatment of advanced prostate cancer. We found that patients who had received docetaxel chemotherapy before abiraterone gained similar benefit from enzalutamide compared with patients who had not received docetaxel. These results suggest that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
The therapeutic armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has rapidly expanded in recent years with several novel agents demonstrating a benefit on overall survival (OS) [1] . Among these agents are the novel androgen receptor (AR)-targeted therapies abiraterone acetate (abiraterone) and enzalutamide that inhibit CYP17 and AR, respectively. Both abiraterone and enzalutamide are increasingly being used in chemotherapy-naive mCRPC patients on the basis of positive data from the COU-AA-302 (abiraterone) and PREVAIL (enzalutamide) phase 3 studies[2] and [3]. Despite their efficacy, however, many questions regarding the use of abiraterone and enzalutamide remain unanswered including the optimal sequencing of treatment. Several prior studies have reported on outcomes in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. All but one of these studies [9] were restricted to patients previously treated with docetaxel, and currently little is known about the activity of enzalutamide following abiraterone in docetaxel-naive mCRPC patients. The aim of this study was to examine the activity of enzalutamide after abiraterone in mCRPC patients not previously treated with docetaxel chemotherapy and to compare outcomes with docetaxel-experienced patients.
2.1. Patient population
The British Columbia Cancer Agency (BCCA) consists of six distinct centres located throughout British Columbia, Canada. The Cancer Registry at BCCA was reviewed for mCRPC patients treated with enzalutamide following abiraterone. Patient demographics, prior treatments including docetaxel and abiraterone, clinicopathologic characteristics, and outcomes on postabiraterone enzalutamide (prostate-specific antigen [PSA] response and survival parameters) were documented from medical records of each patient. Patients who had previously received a minimum of one cycle of docetaxel were classified as docetaxel experienced; all other patients were classified as docetaxel naive. Research ethics board approval was obtained prior to commencing this study.
2.2. Outcome measures
The primary objective of this study was to compare the efficacy of enzalutamide following abiraterone in docetaxel-experienced and docetaxel-naive patients. Data were collected for the following end points: confirmed PSA response rate (PSA decline ≥50% from baseline maintained ≥3 wk), time to progression, and OS (time from initiation of enzalutamide to death of any cause or censoring on March 31, 2014). Progression was defined as radiologic (Prostate Cancer Working Group 2 [PCWG2] criteria) or clinical (worsening disease-related symptoms requiring a change in antineoplastic therapy or a decrease in Eastern Cooperative Oncology Group performance status of two or more levels). Reasons for discontinuation of enzalutamide were recorded as follows: PSA progression (PCWG2 criteria), radiologic progression (PCWG2 criteria), clinical progression, and toxicity. Patients who stopped enzalutamide due to toxicity were censored for time to progression analysis.
2.3. Statistics
Outcomes on enzalutamide were compared between docetaxel-naive and docetaxel-experienced patients using chi-square for PSA response and log-rank test for time to PSA progression and OS. Univariate analysis was performed to identify variables associated with confirmed PSA response to enzalutamide, using either chi-square for categorical variables or logistic regression for continuous variables. Progression-free survival (PFS) and OS were calculated using Kaplan-Meier estimates.
A total of 115 patients were treated with enzalutamide following abiraterone: 68 had received at least one cycle of docetaxel previously (“docetaxel experienced”); the remaining 47 patients were docetaxel naive. Table 1 lists the patient characteristics at the initiation of enzalutamide. Docetaxel-experienced and docetaxel-naive patients were matched for all clinicopathologic characteristics with the exception of age and serum lactate dehydrogenase (LDH).
Table 1 Baseline characteristics at initiation of enzalutamide in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All (n = 115) | Docetaxel experienced (n = 68) | Docetaxel naive (n = 47) | p | Test |
|---|---|---|---|---|---|
| Age | |||||
| Median (IQR) | 72 (67–81) | 70 (64–74) | 81 (72–86) | <0.001 | Mann-Whitney |
| Gleason score, n (%) | |||||
| 6–7 | 39 (34) | 21 (31) | 18 (38) | 0.5 | Chi-square |
| 8–10 | 65 (57) | 39 (57) | 26 (55) | ||
| Unknown | 11 (10) | 8 (12) | 3 (6) | ||
| Disease sites, n (%) | |||||
| Bone | 107 (93) | 64 (94) | 43 (91) | 0.6 | Chi-square |
| Lymph node | 40 (35) | 24 (35) | 16 (34) | 0.9 | Chi-square |
| Visceral | 17 (15) | 13 (19) | 4 (9) | 0.1 | Chi-square |
| Prior therapy | |||||
| Docetaxel, n (%) | 68 (59) | ||||
| No. of cycles, median (range) | 8 (1–12) | ||||
| No PSA decline, n (%) | 10 (15) | ||||
| PSA decline <50%, n (%) | 21 (31) | ||||
| PSA decline ≥50%, n (%) | 37 (54) | ||||
| Abiraterone, | |||||
| Treatment duration, mo, median (95% CI) | 7.56 (6.45–8.67) | 7.36 (6.00–8.72) | 7.89 (5.41–10.36) | 0.5 | Log-rank |
| No PSA decline, n (%) | 29 (25) | 20 (29) | 9 (19) | 0.3 | Chi-square |
| PSA decline <50%, n (%) | 30 (26) | 15 (22) | 15 (32) | ||
| PSA decline ≥50%, n (%) | 56 (49) | 33 (49) | 23 (49) | ||
| Other therapy * , n (%) | 26 (23) | 18 (26) | 8 (17) | 0.2 | Chi-square |
| Laboratory | |||||
| LDH, U/l, median (IQR) | 257 (204–351) | 246 (186–308) | 266 (233–354) | 0.046 | Mann-Whitney |
| Elevated, ≥250, n (%) | 62 (54) | 30 (44) | 32 (68) | ||
| Normal, n (%) | 49 (43) | 35 (51) | 14 (30) | ||
| Unknown, n (%) | 4 (3) | 3 (4) | 1 (2) | ||
| ALP, U/l, median (IQR) | 166 (107–316) | 178 (110–329) | 145 (92–263) | 0.3 | Mann-Whitney |
| Elevated (≥160), n (%) | 58 (50) | 37 (54) | 21 (45) | ||
| Normal, n (%) | 55 (48) | 29 (43) | 26 (55) | ||
| Unknown, n (%) | 2 (2) | 2 (3) | 0 (0) | ||
| Albumin, g/l, median (IQR) | 37 (34–40) | 37 (34–40) | 36 (33–40) | 0.7 | Mann-Whitney |
| Low, <40, n (%) | 83 (72) | 49 (72) | 34 (72) | ||
| Normal, n (%) | 24 (21) | 16 (24) | 8 (17) | ||
| Unknown, n (%) | 8 (7) | 3 (4) | 5 (11) | ||
| Haemoglobin, g/l, median (IQR) | 121 (108–129) | 121 (114–129) | 116 (106–129) | 0.2 | Mann-Whitney |
| Low (<100), n (%) | 18 (16) | 10 (15) | 8 (17) | ||
| Normal, n (%) | 97 (84) | 58 (85) | 39 (83) | ||
| Unknown, n (%) | 0 (0) | 0 (0) | 0 (0) | ||
* Systemic agents excluding androgen deprivation therapy (luteinising hormone-releasing hormone antagonist/agonist, antiandrogens).
ALP = alkaline phosphatase; CI = confidence interval; IQR = interquartile range; LDH = lactate dehydrogenase; PSA = prostate-specific antigen.
Median duration of enzalutamide treatment was 4.1 mo and did not significantly differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Overall, 21% of patients (24 of 115) received enzalutamide ≥6 mo. Median time between cessation of abiraterone and initiation of enzalutamide was 21 d (interquartile range: 2–68). A total of 13% of patients (15 of 115) received another systemic agent between stopping abiraterone and starting enzalutamide. Seven of these 15 patients were docetaxel experienced and were rechallenged with docetaxel between abiraterone and enzalutamide. The remaining patients received BEZ235 (n = 2), PX-866 (n = 2), mitoxantrone (n = 1), and AZD2171 (n = 1); two patients were enrolled in the Study of Cabozantinib (XL184) versus Mitoxantrone Plus Prednisone in Men with Previously Treated Symptomatic Castration-resistant Prostate Cancer (COMET-2) trial (cabozantinib vs mitoxantrone).
Table 2 Outcomes with enzalutamide treatment in metastatic castration-resistant prostate cancer patients previously treated with abiraterone acetate
| Characteristic | All n = 115 | Docetaxel experienced n = 68 | Docetaxel naive n = 47 | Difference (95% CI) | HR (95% CI) | p | Test |
|---|---|---|---|---|---|---|---|
| PSA decline | |||||||
| ≥50%, n (%) | 27 (23.5) | 15 (22) | 12 (25.5) | 3.5 (−12 to 19) | NA | 0.8 | Chi-square |
| <50%, n (%) | 46 (40) | 29 (42.6) | 17 (36.2) | 6.4 (−12 to 23) | NA | ||
| None, n (%) | 42 (36.5) | 24 (35.3) | 18 (38.3) | 3 (−14 to 21) | NA | ||
| Time on treatment, mo, median | 4.1 | 4.1 | 4.6 | NA | 1.09 (0.69–1.74) | 0.7 | Log-rank |
| Time to progression, mo * , median | 5.3 | 4.6 | 6.6 | NA | 0.87 (0.50–1.51) | 0.6 | Log-rank |
| Overall survival, mo † , median | 10.6 | 10.6 | 8.6 | NA | 1.58 (0.82–3.02) | 0.2 | Log-rank |
* Radiologic and/or clinical progression.
† From start of enzalutamide.
CI = confidence interval, HR = hazard ratio; NA = not available; PSA = prostate-specific antigen.
A total of 61% of patients (70 of 115) were on corticosteroids at the start of enzalutamide and 28% (32 of 115) remained on corticosteroids at the cessation of enzalutamide. Reasons for enzalutamide discontinuation (more than one could apply) were clinical progression (43%), PSA progression (25%), radiologic progression (10%), and toxicity (9%); 28% of patients (32 of 115) remained on treatment as of March 31, 2014. Among patients stopping enzalutamide due to toxicity (n = 10), four were docetaxel experienced and six were docetaxel naive. Postenzalutamide systemic treatment was administered to 25% (17 of 68) and 16% (7 of 45) of docetaxel-experienced and docetaxel-naive patients, respectively (p = 0.3; chi-square).
Confirmed PSA declines ≥90%, ≥50%, and ≥30% were seen in 4% (5 of 115), 23% (27 of 115), and 35% (40 of 115) of patients, respectively. Waterfall plots of maximal PSA decrease on enzalutamide are presented in Figure 1 A and stratified by prior treatment with docetaxel in Figure 1 B. Notably, confirmed PSA response rates (PSA decline ≥50%) did not differ between docetaxel-experienced and docetaxel-naive patients ( Table 2 ). Univariate analysis was also performed to identify any clinicopathologic factors predictive of response to enzalutamide. As shown in Table 3 , no factors significantly correlated with a confirmed PSA response to enzalutamide including response to prior abiraterone, duration of prior abiraterone, and corticosteroid treatment at commencement or cessation of enzalutamide. Treatment with another systemic agent between abiraterone and enzalutamide was also not significantly associated with PSA response on enzalutamide. Interestingly, however, no PSA responses to enzalutamide were observed in the seven docetaxel-experienced patients who were rechallenged with docetaxel after abiraterone.
Fig. 1 (A) Waterfall plot showing maximum prostate-specific antigen (PSA) decline in patients treated with enzalutamide following abiraterone; (B) maximal PSA decline after enzalutamide administration stratified by prior treatment with docetaxel.
Table 3 Univariate analysis examining association between baseline clinicopathologic factors and confirmed prostate-specific antigen response rate (decline ≥50% maintained ≥3 wk) in patients treated with enzalutamide after abiraterone acetate
| Variable | n | PSA RR, % | Difference (95% CI) | p |
|---|---|---|---|---|
| Docetaxel preabiraterone | ||||
| Yes | 68 | 22.0 | 3.5 (−12 to 19) | 0.7 |
| No | 47 | 25.5 | ||
| PSA decline ≥50% on docetaxel | ||||
| Yes | 37 | 27.0 | 10.9 (−9 to 29) | 0.3 |
| No | 31 | 16.1 | ||
| PSA decline ≥50% on abiraterone | ||||
| Yes | 56 | 28.6 | 10.0 (−6 to 25) | 0.2 |
| No | 59 | 18.6 | ||
| Gleason score | ||||
| ≤7 | 39 | 23 | 0 (−16 to 18) | 1.0 |
| 8–10 | 65 | 23 | ||
| ECOG performance status | ||||
| 0–1 | 67 | 25.4 | 4.1 (−12 to 19) | 0.6 |
| ≥2 | 47 | 21.3 | ||
| Lymph node metastases | ||||
| Absent | 75 | 26.7 | 9.2 (−8 to 23) | 0.3 |
| Present | 40 | 17.5 | ||
| Visceral metastases | ||||
| Absent | 98 | 24.5 | 6.9 (−18 to 22) | 0.5 |
| Present | 17 | 17.6 | ||
| On steroids at start of enzalutamide | ||||
| Yes | 70 | 20 | 8.9 (−7 to 25) | 0.3 |
| No | 45 | 28.9 | ||
| On steroids at end of enzalutamide | ||||
| Yes | 32 | 18.8 | 6.5 (−12 to 21) | 0.5 |
| No | 83 | 25.3 | ||
| Systemic treatment postabiraterone and pre-enzalutamide | ||||
| Yes | 15 | 13.3 | 11.7 (−14 to 25) | 0.3 |
| No | 100 | 25 | ||
| Duration of abiraterone (continuous) | 115 | NA | NA | 0.3 |
| Age at start of enzalutamide (continuous) | 115 | NA | NA | 0.9 |
| Time from start of ADT to start of enzalutamide (continuous) | 114 | NA | NA | 0.5 |
| Serum LDH (continuous) | 112 | NA | NA | 0.7 |
| Serum ALP (continuous) | 114 | NA | NA | 0.9 |
| Serum albumin (continuous) | 111 | NA | NA | 0.1 |
| Haemoglobin (continuous) | 115 | NA | NA | 0.4 |
ADT = androgen deprivation therapy; ALP = alkaline phosphatase; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; LDH = lactate dehydrogenase; NA = not available; PSA = prostate-specific antigen; RR = response rate.
Accurate data on time to progression and OS were available in 108 and 115 patients (94% and 100%, respectively). Median time to progression and OS were 5.3 mo and 10.6 mo, respectively. Kaplan-Meier survival curves are shown in Figure 2 . Median time to progression and median OS on enzalutamide were not significantly different between docetaxel-experienced and docetaxel-naive patients ( Table 2 ).
In this retrospective study, we examined the antitumour activity of enzalutamide after abiraterone in mCRPC patients. We found that the efficacy of enzalutamide was comparable in docetaxel-experienced and docetaxel-naive patients in terms of confirmed PSA response rates (22% vs 26%;p = 0.8), median time to radiologic/clinical progression (4.6 mo vs 6.6 mo;p = 0.6), and median OS (10.6 mo vs 8.6 mo;p = 0.2). These findings are somewhat surprising because we had anticipated that enzalutamide would be less efficacious in more heavily pretreated patients, that is, those who had also received prior docetaxel. The differences in baseline age and LDH between the cohorts, which were biased against the docetaxel-naive group, may have had an impact on our results, although age and LDH have not been shown differentially to affect the benefit of enzalutamide[10] and [11].
Population-based studies have shown that only a third of mCRPC patients receive docetaxel [12] . However, results from the recently presented Androgen Ablation Therapy with or Without Chemotherapy in Treating Patients with Metastatic Prostate Cancer (CHAARTED) trial [13] are likely to lead to an increase in docetaxel use in patients with high-risk metastatic castration-sensitive disease. Our results, which indicate that earlier treatment with docetaxel does not have a large impact on the activity of enzalutamide after abiraterone, may therefore be of increasing relevance to the management of patients with metastatic prostate cancer.
The findings of this study point to a common mechanism of resistance to enzalutamide in docetaxel-experienced and -naive patients. Resistance mechanisms to enzalutamide include induction of AR splice variants [14] , an activating point mutation in the AR (F876L)[15] and [16], upregulation of the stress-activated chaperone protein clusterin [17] , enhanced protein kinase C/Twist1 signalling [18] , and activation of the glucocorticoid receptor [19] . These factors do not overlap with known mechanisms of docetaxel resistance [20] , although it has been postulated that cross-resistance may develop between docetaxel, enzalutamide, and abiraterone because all of these agents are known to modulate AR nuclear translocation [21] . The similar efficacy of enzalutamide regardless of prior docetaxel use in our study suggests that cross-resistance between abiraterone and enzalutamide emerges independently of docetaxel.
The PSA response rate to enzalutamide in our cohort was 23%, median time to progression was 5.3 mo, and median OS was 10.6 mo. These results are broadly in keeping with earlier studies that reported PSA response rates of 13–29%, median PFS of 2.9–4.0 mo, and median OS of 7.3–7.5 mo in mCRPC patients treated with enzalutamide after abiraterone[4], [5], [6], [7], [8], and [9]. In our chemotherapy-naive cohort, PSA declines ≥50% and ≥30% were seen in 23% and 35% of patients, respectively. These findings are also comparable with the only prior study to include chemotherapy-naive patients [9] , which reported PSA declines ≥30% in 40% of patients. Collectively, these data indicate that enzalutamide has relatively modest activity when used after abiraterone. Abiraterone is also reported to have limited efficacy after prior treatment with enzalutamide[22] and [23]. This lends weight to the possibility of a common mechanism of resistance between enzalutamide and abiraterone. In support of this, a recent report linked detection of the AR-V7 splice variant with primary and acquired resistance to both enzalutamide and abiraterone [24] .
Despite its limited overall activity, enzalutamide did confer longer-term benefit for a subset of patients in our study with 21% of patients remaining on treatment for at least 6 mo. However, predicting which patients will benefit from treatment with enzalutamide remains a major challenge. We could not find any clinicopathologic factors that predicted a PSA decline ≥50% on enzalutamide. Although prior PSA decline ≥50% on abiraterone has been linked to subsequent response to enzalutamide[8] and [9], we and other investigators[4] and [5]have not corroborated these findings. Gleason score has also been implicated as a predictor for enzalutamide response following abiraterone [4] ; however, we did not find this to be the case. Similarly, although coadministration of corticosteroids was found to be associated with inferior survival parameters in a post hoc analysis of the Safety and Efficacy Study of MDV3100 in the Patients with Castration-resistant Prostate Cancer Who Have Been Previously Treated with Docetaxel-based Chemotherapy (AFFIRM) trial [25] , we did not observe lower PSA response rates in patients receiving corticosteroids at the initiation of enzalutamide. Participation in clinical trials has also been linked to improved outcomes with systemic therapy in mCRPC patients [26] . However, this was not relevant to our study because no patients received enzalutamide in the context of a clinical trial.
The absence of predictive clinicopathologic factors for enzalutamide efficacy illustrates the need for further translational studies to identify novel predictive biomarkers. Pretreatment measurement of serum androgens [27] and enumeration of circulating tumour cells (CTCs)[28], [29], and [30]are avenues of interest, having been identified as prognostic factors in mCRPC patients receiving systemic therapy. Correlating outcomes on enzalutamide with genomic aberrations (eg,ARamplification,ARmutations, andTMPRSS2-ERGrearrangement) detectable in tumour biopsies, CTCs, and/or circulating tumour DNA would also be of significant interest.
We acknowledge that this study has several limitations. The limited cohort size means our data must be interpreted with due caution and require validation in larger, prospective series. Other limitations of this study include its retrospective design, its restriction to a single province in Canada, and the relatively short follow-up period. We also could not assess radiologic response to treatment because imaging was not consistently performed, and postenzalutamide treatment may have confounded assessment of OS.
We found that the antitumour efficacy of enzalutamide following abiraterone is similar in docetaxel-experienced and docetaxel-naive mCRPC patients. Our data suggest that therapeutic resistance to enzalutamide is not linked to prior treatment with docetaxel. Notably, enzalutamide had limited overall activity after abiraterone with or without prior docetaxel. This illustrates the importance of conducting clinical trials examining novel therapeutic approaches in patients progressing on abiraterone. Nevertheless, a subset of patients did attain relatively durable benefit from treatment with enzalutamide following abiraterone, and developing robust predictive biomarkers to identify these patients is paramount. In turn, this would optimise the use of enzalutamide and potentially lead to improved clinical outcomes for patients with mCRPC.
Author contributions: Kim N. Chi 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: Chi, Eigl, Azad.
Acquisition of data: Chi, Eigl, Azad, Kollmannsberger, Murray.
Analysis and interpretation of data: Kollmannsberger, Chi, Eigl, Azad.
Drafting of the manuscript: Chi, Eigl, Azad, Kollmannsberger, Murray.
Critical revision of the manuscript for important intellectual content: Chi, Eigl, Azad, Kollmannsberger, Murray.
Statistical analysis: Azad.
Obtaining funding: None.
Administrative, technical, or material support: Chi, Eigl, Azad.
Supervision: Chi, Eigl.
Other(specify): None.
Financial disclosures: Kim N. Chi 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: Kim N. Chi receives research funding from OncoGenex Technologies Inc., Astellas, Janssen, and Novartis, and is a consultant for Janssen, Astellas, Amgen, Bayer, Millennium, Novartis, and Sanofi. Bernhard J. Eigl has received an unrestricted educational grant from Janssen. Arun A. Azad receives research funding from Astellas.
Funding/Support and role of the sponsor: None.
Earlier docetaxel did not affect the response to enzalutamide after abiraterone and PSA responses are seen in one quarter of men. The small population (n=115) prohibits solid conclusions: although not significant OS was 2 months longer for men that earlier received docetaxel suggesting differences in populations based on criteria for choosing docetaxel as earlier treatment.