Department of Medical Oncology, Netherlands Cancer Institute, Amsterdam, the Netherlands.
Corresponding author: Andries M. Bergman, MD, PhD, Department of Medical Oncology, Netherlands Cancer Institute/Antoni van Leeuwenhoek, Plesmanlaan 121, 1066CX Amsterdam, the Netherlands; Fax: (011) 31-20-512-2572; firstname.lastname@example.org
Enzalutamide (Enz) and abiraterone acetate (AA) are hormone treatments that have a proven survival advantage in patients with metastatic, castration-resistant prostate cancer who previously received docetaxel (Doc). Recently, limited activity of AA after Enz and of Enz after AA was demonstrated in small cohort studies. Here, the authors present the activity and tolerability of Enz in patients who previously received AA and Doc in the largest cohort to date.
The efficacy and tolerability of Enz were investigated in men with progressive, metastatic, castrate-resistant prostate cancer who previously received Doc and AA. Toxicity, progression-free survival, time to prostate-specific antigen (PSA) progression, and overall survival were retrospectively evaluated.
Sixty-one patients were included in the analysis. The median age was 69 years (interquartile range [IQR], 64-74 years), 57 patients (93%) had an Eastern Cooperative Oncology Group performance status from 0 to 2, 48 patients (79%) had bone metastases, 33 patients (54%) had lymph node metastases, and 13 patients (21%) had visceral metastases. The median duration of Enz treatment was 14.9 weeks (IQR, 11.1-20.0 weeks), and 13 patients (21%) had a maximum PSA decline ≥50%. The median progression-free survival was 12.0 weeks (95% confidence interval [CI], 11.1-16.0 weeks), the median time to PSA progression was 17.4 weeks (95% CI, >16.0 weeks), and the median overall survival was 31.6 weeks (95% CI, >28.7 weeks). Enz was well tolerated, and fatigue and musculoskeletal pain were the most frequent grade ≥2 adverse events. The PSA response to Doc and AA did not predict the PSA response to Enz.
Enz has modest clinical activity in patients with metastatic, castrate-resistant prostate cancer who previously received Doc and AA.
Prostate cancer is the most prevalent cancer among men in the western world and the second leading cause of male cancer death.[1, 2] After an initial response to medical or surgical castration, the disease will progress into castration-resistant prostate cancer (CRPC). However, CRPC is still driven by androgen receptor (AR) signaling, requiring lower than castration testosterone levels as a result of AR modulations.[4-6] Therefore, new drugs have been developed that more effectively inhibit AR signaling. Enzalutamide (Enz) (MDV3100; Xtandi; comarketed by Astellas Pharma US, Inc. [Northbrook, Ill] and Medivation, Inc. [San Francisco, Calif]), a novel nonsteroidal AR signaling inhibitor, has exhibited a survival advantage in patients with metastasized CRPC (mCRPC) who previously received docetaxel (Doc) in combination with prednisone. However, earlier, abiraterone acetate (AA) (Zytiga; Janssen Biotech, Inc., Horsham, Pa), an inhibitor of testosterone synthesis, demonstrated a comparable survival advantage in combination with prednisone in the same patient population. It is suggested that Enz and AA actions are nonoverlapping and, thus, are potentially synergistic. However, only limited activity of AA was described in 2 cohort studies of patients who previously received Enz, and limited activity of Enz was described in a single cohort of patients who previously received AA.[10-12] Knowledge of the clinical cross-resistance in both sequences of these antihormonal treatments is of value for future trial design.
Pending final registration of Enz in the Netherlands, Astellas Pharma Europe Ltd. established an Expanded Access Program (EAP) for patients with progressive disease who had no satisfactory alternative treatments available. Therefore, Enz treatment was positioned after Doc and AA treatment. Here, we report the tolerability and efficacy of Enz in the largest cohort to date of patients with mCRPC who previously received Doc and AA.
MATERIALS AND METHODS
Patients and Treatment
In this multicenter, observational study, we included patients with progressive mCRPC enrolled in the Dutch EAP for ENZ who had previously received Doc and AA. All patients either progressed on or did not tolerate AA treatment and consented to join the program. Inclusion criteria for the EAP included: effective surgical or medical castration, progressive disease, an Eastern Cooperative Oncology Group (ECOG) performance from 0 to 2, and no satisfactory alternative treatment at the physician's discretion. The exclusion criteria of the EAP included: earlier treatment with or participation in a clinical trial with ENZ; severe concurrent disease; inadequate bone marrow, liver, vascular, heart, and kidney functions; and prior chemotherapy, biologic therapy, or radiation therapy within 3 weeks before treatment and radionucleotide treatment 8 weeks before treatment.
Patients received Enz at a once-daily dose of 160 mg. All patients received at least 1 dose of Enz. Treatment was continued until clinical deterioration, disease progression, and/or unacceptable adverse effects, all at the physician's discretion or until death.
Study Procedures and Data Collection
Patient baseline characteristics were documented, including age, ECOG performance status, disease characteristics (including Gleason score, involved metastatic sites, number of metastatic sites), blood test results (including hemoglobin concentration, liver chemistry tests, creatinine, testosterone, and prostate-specific antigen [PSA]), clinical signs of disease progression, previous antihormonal and chemotherapy treatments (including duration of Doc and AA treatment, best PSA response to Doc and AA treatment, and reason for AA discontinuation), and use of concomitant medication with known interaction with Enz (macrolide antibiotics, benzodiazepines, immune modulators, antiepileptics, coumarins, colchicine, digoxin).
Clinical and biochemical activity and toxicity were assessed at Enz treatment initiation and every month according to the EAP protocol, including ECOG performance status and toxicity, which was recorded using Common Toxicity Criteria (CTC), version 4.0. Imaging studies were performed at the discretion of the physician. Patients were followed for prostate cancer progression and survival until April 2013.
PSA response was evaluated using the Prostate Cancer Clinical Trials Working Group 2 (PCWG2) recommendations. Progression-free survival (PFS) was calculated from the time treatment started to the first date of confirmed progression or the date of last follow-up. Progression was defined as PSA progression, and/or radiographic progression, and/or clinical progression. PSA response, as recommended by the PCWG2, was defined as a decline ≥50% from baseline, and PSA progression was defined as an increase of 25% and a minimum level of 2 ng/mL confirmed by a second PSA reading a minimum of 3 weeks later. When no decline from baseline was documented, PSA progression was defined as an increase of 25% from baseline value along with an increase in absolute PSA value of at least 2 ng/mL after 12 weeks of treatment. PSA declines <30%, <50%, and <90% from baseline after 12 weeks with or without conformation also were evaluated. Objective responses were measured according to Response Evaluation Criteria in Solid Tumors. Bone progression based on bone scans was assessed according to PCWG2 criteria. Progression after initial response and no response as reasons for AA discontinuation were defined as PSA progression (PCWG2), radiographic or clinical progression after an initial PSA response (PCWG2), or no PSA response, respectively. Overall survival (OS) was calculated from the date Enz treatment started to the date of death or last follow-up. Informed consent was obtained from all patients before enrolment into the EAP, and patients' outcomes were analyzed with ethics committee approval.
In line with PCWG2 criteria, waterfall plots with maximum PSA decline from baseline and PSA after 12 weeks of treatment were constructed. Survival and progression were evaluated using Kaplan-Meier estimates. Patients who did not achieve a 50% decline in PSA on Doc or AA were designated Doc nonsensitive or AA nonsensitive, respectively; and patients who had a ≥50% decline in PSA on Doc or AA treatment were designated Doc sensitive or AA sensitive, respectively. The best PSA responses on Enz were compared between subpopulations of patients according to Doc and AA sensitivities using a 2-sample t test. The predictive power of the best Doc and AA responses and their interaction for maximum PSA decline on Enz were evaluated by means of linear regression using a log link. Statistical analyses were conducted using Statistical Analysis System (SAS) statistical software (SAS Institute Inc., Chicago, Ill) and R (R Foundation for Statistical Computing, Vienna, Austria).
Starting in June 2012, an EAP for Enz was established in the Netherlands. At the time the program was closed in March 2013, 61 patients in 9 hospitals who were previously treated with both Doc and AA were evaluable for treatment outcome and tolerability assessments. Patient characteristics at the time of Enz treatment initiation are summarized in Table 1. The median patient age before starting Enz treatment was 69 years (interquartile range [IQR], 64-74 years); the ECOG performance was 0 or 1 in 57% of patients and 2 in 36% of patients; and 79% of patients had bone metastases, 54% had lymph node metastases, and the majority (93%) had more than 1 metastatic site. With respect to laboratory results, the median hemoglobin concentration was 11.0 g/dL (IQR, 9.9-12.5 g/dL), the median alkaline phosphatase level was 191 U/L (IQR, 100-288 U/L), and the median lactate dehydrogenase level was 241 U/L (IQR, 191-385 U/L) at the time of Enz treatment initiation. Thirty-eight patients (70%) had increased alkaline phosphatase levels, and 25 patients (48%) had increased lactate dehydrogenase levels at baseline. Disease progression at the time of Enz treatment initiation presented in 95% of patients as PSA progression and in 87% of patients as clinical progression and could be confirmed in 57% of patients by a bone scan and in 30% of patients as progression of measurable lesions.
Previous antihormonal treatment (other than abiraterone)
Concomitant medication with known interaction with enzalutamide
Ninety percent of patients received 1 course of Doc treatment, whereas 10% received 2 or more courses. The median number of cycles of Doc in all courses combined was 8 (IQR, 6-10 cycles). Thirty percent of patients received at least 1 course of cabazitaxel in combination with prednisone. The median duration of AA treatment was 26 weeks (IQR, 13-37 weeks), and the reason for AA discontinuation was disease progression after initial response in 54% and no initial response in 38% of patients. All patients received a luteinizing hormone-releasing hormone antagonist/agonist or underwent orchidectomy, whereas 18% received steroid drugs as monotherapy at the time of Enz treatment initiation. Twenty percent of patients received drugs that had a known interaction with Enz.
Patients started Enz treatment a median of 60.7 weeks (IQR, 36.6-78.2 weeks) after discontinuing Doc and 8.9 weeks (IQR, 4.3-28.9 weeks) after discontinuing AA (Table 2). The median duration of Enz treatment was 14.9 weeks (IQR, 11.1-20.0 weeks), and the median follow-up was 16.3 weeks (IQR, 13.7-21.1 weeks). A maximum PSA decline ≥30% was observed in 28 patients (46%), a PSA decline ≥50% was observed in 13 patients (21%), and a PSA decline ≥90% was observed in 2 patients (3%) (Fig. 1). The maximum PSA decline was reached after a median of 5.0 weeks (IQR, 4.0-8.6 weeks). Eighteen patients (30%) had no PSA response at any time. For each patient, we collected the PSA measurement at the time point closest to 12 weeks after Enz treatment initiation (median, 12 weeks; IQR, 10.7-13.6 weeks). A PSA decline ≥30% was observed in 17 patients (28%), a PSA decline ≥50% was observed in 9 patients (15%), and a PSA decline ≥90% was observed in 1 patient (2%). Reasons for Enz treatment discontinuation included no initial activity in 16 patients (26%), progressive disease in 22 patients (36%), and death in 2 patients (3%). One patient died of a hemorrhagic stroke, and 1 patient died of disease progression while receiving Enz treatment. One patient (2%) was intolerant to Enz and experienced severe nausea and fatigue (Table 2). At the time of analysis, 19 patients (31%) were still receiving Enz treatment.
Table 2. Outcomes of Enzalutamide Treatment After Abiraterone Treatment
PFS is depicted in the top portion of Figure 2. The Kaplan-Meier estimate for median PFS was 12.0 weeks (95% confidence interval [CI], 11.1-16.0 weeks) (Table 2), the estimated median time to PSA progression was 17.4 weeks (95% CI, >16.0 weeks) (Table 2), and the median OS was 31.6 weeks (95% CI, >28.7 weeks) (Table 2; Fig. 2, bottom). Of the 11 patients (18%) who were receiving steroid therapy at the time Enz was initiated, 4 (36%) had a maximum PSA response ≥30%, 2 (18%) had a maximum PSA response ≥50%, and 1 (9%) had a PSA response ≥90%. The median time to maximum PSA response was 6.9 weeks (IQR, 3.6-10.5 weeks) in these patients.
No unexpected toxicity of Enz was reported. Of the 382 adverse events (AEs) collected, the majority (247 AEs; 65%) were grade 1 (Table 2). Grade 2 AEs (n = 101; 26%) and grade 3 AEs (n = 34; 9%) were less frequent. Hot flushes were all grade 1. Fatigue was the most frequent grade 2 and 3 AE and was observed in 60 patients (59%) and 16 patients (47%), respectively, followed by musculoskeletal pain in 27 patients (27%) and 7 patients (20%), respectively.
Relation Between Doc and AA Response and Response to Enz
Table 3 summarizes characteristics of the patients who reached a PSA response <50% on Doc or AA treatment, patients who reached a PSA response ≥50% on Doc or AA treatment, and the entire population. The 24 patients who had a PSA response <50% on Doc treatment received fewer cycles of Doc and had a lower median best PSA response to Doc treatment compared with the 29 patients who had a PSA response ≥50% on Doc and compared with the entire population (6 cycles, 10 cycles, and 8 cycles, respectively; −20.1%, −78.3%, and −59.7%, respectively) (Table 3). The duration of AA treatment was not different between patients who had a PSA response <50% on Doc treatment, those who had a PSA response ≥50% on Doc treatment, and the entire population (26.1 weeks, 21.9 weeks, and 25.9 weeks, respectively). Although the difference was not statistically significant, the median maximum response on AA treatment was better in patients who had a PSA response <50% on Doc treatment compared with the population that had a PSA response ≥50% on Doc (−27.7% and −9.2%, respectively). Five patients (21%) who had a PSA response <50% on Doc and 5 patients (18%) who had a PSA response ≥50% on Doc had a PSA response ≥50% on AA treatment. However, there was no difference in the maximum PSA response to Enz treatment (median: −28.4%, −22.8%, and −25.2%, respectively) (Table 3) and the duration of Enz treatment (median: 14.6 weeks, 14.9 weeks, and 14.9 weeks, respectively) between patients who had a PSA response <50%, those who had a PSA response ≥50% on Doc, and the entire population. The 43 patients who had a PSA response <50% on AA received AA treatment for a shorter period than patients who had a PSA response ≥50% on AA (21.6 weeks and 39.1 weeks, respectively). There was no difference in the median maximum PSA response to Enz treatment (−31% and −20%, respectively) (Table 3) and the duration of Enz treatment (median: 15 weeks and 17.4 weeks, respectively) between patients who had a PSA response <50% and those who had a PSA response ≥50% on AA treatment. In linear regression analyses with and without an interaction term, the best responses on AA and Doc were not predictive of the response to Enz. By using 30% and 80% PSA declines as cutoff points, similar correlations between responses to Doc, AA, and Enz were observed.
Table 3. Response on Abiraterone Acetate Plus Prednisone Treatment and Enzalutamide Treatment in Relation to Docetaxel Plus Prednisone Sensitivity and Response on Enzalutamide Treatment in Relation to Abiraterone Acetate Plus Prednisone Sensitivitya
The response rates to Enz among men with mCRPC who received previous treatment with Doc and AA and their survival outcomes suggest modest activity in this cohort of patients. Forty-six percent of patients had a PSA decline ≥30% on Enz; however, only 21% had a PSA decline ≥50%. These PSA response rates are comparable to the PSA response rates reported in a smaller cohort of 35 mCRPC patients who received Enz after previous AA treatment. The median time to PSA progression on Enz, as estimated by the Kaplan-Meier method, was 17.4 weeks, and the median OS was 31.6 weeks. These outcomes compare unfavorably with the outcomes of the AFFIRM study of Enz in the postdocetaxel setting, in which 54% of patients reportedly had a PSA response ≥50%, the median time to PSA progression was 36.1 weeks, and the median OS was 80 weeks. However, this comparison is hampered by the more advanced disease stage among patients in the current analysis than among patients in the AFFIRM study, which is reflected by a poorer performance score, lower hemoglobin concentration, and a higher PSA levels at baseline. A post hoc analysis of the AFFIRM data suggested a worse outcome for patients who received corticosteroids during Enz treatment. Eighteen percent of patients in the current cohort were receiving corticosteroids; however, the maximum PSA response and the time to maximum PSA response did not appear to differ from those in the total population, although these conclusions are hampered by the small sample size. Enz treatment was well tolerated in these extensively pretreated patients with advanced disease, and no unexpected AEs were reported. Like in the AFFIRM study, fatigue was the most frequent AE, and all hot flushes were grade 1. However, in contrast, diarrhea was an infrequent AE in the current study, and musculoskeletal pain was more common than in the AFFIRM study. Our small retrospective study has several limitations, including selection bias, variable assessment intervals, incomplete data on the efficacy and tolerability of Enz, and a relatively short follow-up.
Recently, Loriot et al and Noonan et al reported modest response rates and survival outcomes of AA treatment in cohorts of 38 patients and 30 patients, respectively, who progressed on Enz treatment.[10, 11] In those studies, PSA response rates ≥50% were reported in 8% and 4% of patients, respectively, which was lower than the 21% of patients in the current cohort. However, the median PFS reported by Loriot et al and Noonan et al was in the same range as the median PFS for the current cohort (11.7 weeks, 15.4 weeks, and 12.0 weeks, respectively) as was the reported median OS for the 3 studies (31.3 weeks, 50.1 weeks, and 31.6 weeks, respectively).
Enz inhibits AR signaling by competitively inhibiting the binding of androgens, inhibiting the translocation of ligand-bound receptor to the nucleus, and binding to its response elements in the DNA. In contrast, AA inhibits CYP17A, which is crucial for testosterone synthesis, with potent suppression of extragonadal androgen production as a result. Although by different means, both drugs target persistent AR signaling. Despite the differences in patient populations between the AFFIRM study and the studies into the sequence of AA and Enz, data from those studies suggest the possibility of cross-resistance. However, in the current study, PSA response on AA treatment did not predict for PSA response on Enz treatment, which is in line with the nonstatistically significant correlation between PSA response on Enz and PSA response on AA in the study by Loriot et al. However, Schrader et al suggested a correlation between PSA response on AA and subsequent Enz treatment in their cohort of 35 patients.
There are limited data on the mechanism of resistance against AA and Enz. Because AA and Enz both target AR signaling, the mechanism of cross-resistance may be at the AR level, including AR modifications, complex interactions of AR with coactivators or correpressors, and mutual regulation of microRNA and AR. In xenografts of human mCRPC that were treated with AA, induction of AR expression and AR splice variants was demonstrated[20, 21]; whereas, in another study, AR splice variants were identified as key mediators of persistent AR signaling and resistance to Enz. Increased steroidogenesis activation also may mediate cross-resistance between AA and Enz. Testosterone levels in blood and bone marrow from patients who received AA treatment were undetectable on treatment discontinuation; however, Enz treatment increased testosterone levels in the bone marrow of patients and decreased nuclear AR expression. In a mouse model of human prostate cancer, the oncogenic AKt pathway was activated when the AR was inhibited. This reciprocal activation of oncogenic pathways upon AR inhibition may represent a mechanism of resistance to AR antagonists.
Combination treatment of Enz and AA may reverse some mechanisms of drug resistance. A study in human prostate cancer cell lines provided evidence that the glucocorticoid drugs administered with AA to prevent side effects and mineralocorticoid receptor antagonists could activate mutant AR, which was inhibited by Enz. These findings provide a rationale for combination treatment. Currently, a phase 2 trial into the safety and tolerability of Enz in combination with AA is enrolling (National Clinical Trials identifier NCT01650194).
In conclusion, patients in the current study who previously progressed on AA had a modest PSA response rate and limited survival on subsequent Enz treatment. PSA response on AA treatment did not seem to predict for Enz treatment outcome. PFS and OS were in the same range as those described in 2 cohorts of men with mCRPC who received AA after progressing on Enz. Data from these 3 studies suggest limited activity and no preference for either sequence of treatments. However, a recommendation on the sequencing of Enz and AA in the postdocetaxel setting cannot be made, because these data are only hypothesis generating. Further studies are needed to establish the sequence of treatments with these new drugs.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
Drs. van Oort, van den Berg, and Bergman report personal fees from Astellas Pharma and Jansen Pharma outside the submitted work. Drs. Hamberg and de Jong report personal fees from Jansen Pharma outside the submitted work. Dr. van den Eertwegh reports personal fees from Astellas Pharma outside the submitted work.