Comparison of the oncological outcomes of abiraterone acetate and enzalutamide in metastatic castration-resistant prostate cancer: A multicenter real-life data


 Background: To compare enzalutamide (E) and abiraterone acetate (AA) in terms of efficacy, survival and prognostic factors affecting survival in metastatic castration-resistant prostate cancer (mCRPC) patients. Methods: A total of 250 patients treated with E or AA in 5 centers were included.Results: The number of patients with no prostate specific antigen (PSA) decline was higher in the AA group than that in the E group, and the proportion of patients with a PSA decline of ≥50% was higher in the E group (p = 0.020). The rate of progression in the AA group (82.2%) was significantly higher than that in the E group (p <0.001). Radiological progression free survival (rPFS) and overall survival (OS) were significantly longer in the E group when compared to that in the AA group (p <0.001 and p = 0.027, respectively). In the E group, rPFS was significantly longer than that in the AA group in both pre- and post-docetaxel settings (p=0.010 and p=0.003, respectively). OS was similar in the pre-docetaxel setting; but in the post-docetaxel setting, E group had a significantly longer OS than the AA group(p=0.021). In the multivariate analysis performed in the whole patient group, we found that good prognostic factors for rPFS were E treatment, being ≥75 years and a PSA decline of ≥50% while there was no factor affecting OS. Conclusion: With longer OS and PFS, E seems to be more suitable for mCRPC patients in the post-docetaxel setting than AA.


Introduction
Prostate cancer is the second most common cancer in men worldwide 1 . With the increasing use of screening tests, the majority of patients are at the local or locoregional stage at the time of diagnosis 2 .
Androgen deprivation therapy (ADT) alone or in combination with other options is the main treatment for metastatic prostate cancer 3 . The majority of patients with advanced disease eventually progress while on ADT; then the condition is called castration-resistant prostate cancer (CRPC). Other treatment options besides ADT for metastatic CRPC (mCRPC) patients are chemotherapy (CT) (docetaxel, cabazitaxel), androgen synthesis inhibitors, androgen receptor blockers, immunotherapy and Radium 223 radionuclide therapy 3 .
Abiraterone acetate (AA) and enzalutamide (E) are two main androgen receptor axis targeted agents used for the treatment of mCRPC 4,5 . Several pilot studies have shown that both drugs contribute signi cantly to overall survival (OS). After COU AA 301 and AFFIRM studies, AA and E were endorsed in the postdocetaxel setting. With the positive results obtained in COU AA 302 and PREVAIL, AA and E were approved in the pre-docetaxel setting.
The questions in front of us are the selection of patients we should use AA or E, and which one is advantageous in terms of e cacy and safety. Moreover, which one should be preferred before and after CT? Although the positive results of the Phase 3 studies were achieved, we need real-world data with the results of larger patient groups to answer these questions more clearly. There is no head-to-head comparative phase III study related to E and AA. In a study on the simultaneous use of E and AA, it was concluded that the combination therapy had a manageable safety pro le without signi cant drug-drug interaction; nevertheless it is not known whether the combination therapy is superior to the single agent therapy 6 . AQUARIUS, an observational, prospective study, which evaluated patient-reported outcomes in mCRPC patients who were treated with AA or E, suggested that AA was more advantageous than E in terms of fatigue and cognitive functions 7 .
Herein we aimed to compare E and AA in terms of baseline patient characteristics, e cacy and survival in mCRPC patients. Additionally, we analyzed prognostic factors affecting radiological progression free survival (rPFS) and OS in all patients.

Materials And Method Data Collection
A total of 250 patients diagnosed with mCRPC who were treated with E or AA between 2012 and 2020 in 5 centers were included in our study. The patients were treated with E at a dose of 160 mg daily or AA at a dose of 1000 mg daily with prednisolone 10 mg daily until disease progression, death, or unacceptable toxicity. All patients, except those who had bilateral orchiectomy, continued to use ADT with serum testosterone levels 50 ng / dL (≤ 2.0 nmol/L).

Clinical Assessment
Prostate Cancer Working Group 2 (PCWG-2) criteria, death, or unacceptable toxicity were used to de ne disease progression. Prostate speci c antigen (PSA) response was evaluated according to PCWG-2 criteria at the 12th week. The de nition of mCRPC was biochemical or radiological progression, in accordance with the criteria of the PCWG, in patients with blood testosterone levels < 50 ng/dl. Patients who did not have mCRPC or those who received both E and AA were excluded from the study. rPFS was de ned as the time from the date of initiation of E or AA until the date of radiological progression. OS was de ned as the time from the date of initiation of E or AA to the date of death from any cause. Increased or stable PSA levels at 12 weeks after E or AA initiation was de ned as No decline, and a decline in the PSA level was grouped as <50% PSA decline and ≥50% PSA decline, according to the decline rate. Radiological response rate (rRR) was evaluated in accordance with the Response Evaluation Criteria in Solid Tumors 1.1.

Statistical Analysis
Statistical analyses were performed using the IBM Statistical Package for the Social Science Statistics for Windows, version 22.0 (IBM Corp., Armonk, MY, USA). The variables were investigated using the Kolmogorov-Smirnov test to determine whether or not they were normally distributed. The continuous variables were expressed as mean ± standard deviation (for normally distributed variables) or median and interquartile range (IQR) (for not normally distributed variables). The Chi-square test or the Fisher's exact test was used to compare the proportions in two groups. The Mann-Whitney U test was used to compare median PSA and median follow up time and the Student's t-test was used to compare mean age. The Kaplan-Meier method was used to estimate survival. The log-rank test was used to identify the univariate effects of treatments and other factors on rPFS and OS of mCRPC patients. Possible factors associated with survival outcomes (p≤0.250) in univariate analysis were selected for testing in multivariate models. The independent predictors of survival were determined with multivariate Cox regression models. A 5% type-I error level was used to infer statistical signi cance.

Ethics approval and consent to participate
The study protocol was approved by the ethics committee of Sakarya University Medical Faculty and was conducted in accordance with the principles of the Declaration of Helsinki (05.03.2020-71522473/050.01.04/43). All participants provided written informed consent.

Results
A total of 250 patients diagnosed with mCRPC were analyzed. The median follow-up time was 13 months (IQR: 6-21, E: 12 months, AA: 13 months, p=0.169). The baseline characteristics of patients are summarized in Table 1. In the AA group, the rate of patients with metastatic disease at the time of diagnosis were signi cantly higher than that in the E group (p=0.016). The number of patients with no PSA decline was higher in the AA group than that in the E group, and the proportion of patients with a PSA decline of ≥50% was higher in the E group (p=0.020) ( Fig. 1). At the end of the 12th week, progressive disease rate was higher and stable disease rate was lower in the AA group compared to the E group. The rate of progression in the AA group (82.2%) was signi cantly higher than the E group (p<0.001). Radiological PFS and OS analysis were in favor of E group at a signi cant level (p<0.001 and =0.027, respectively) (Fig. 2).
The subgroup analysis results of the pre-docetaxel and post-docetaxel settings are demonstrated in Table  2. In the E group, rPFS was signi cantly longer than that in the AA group in both pre-docetaxel and postdocetaxel settings. OS was similar in the pre-docetaxel setting; but in the post-docetaxel setting, E group had a signi cantly longer OS than the AA group (Fig. 3).
The univariate analysis results for the factors affecting rPFS and OS are summarized in Table 3. In the univariate analysis, the treatment agent (E and AA) signi cantly predicted both rPFS and OS. The other factors that signi cantly predicted rPFS were age, pre-or post-docetaxel setting, PSA decline rate and the line of therapy. Besides, the factors that signi cantly affected OS were pre-or post-docetaxel setting and PSA decline rate. Multivariate Cox regression analysis was performed for parameters that had a signi cant or near-signi cant effect (p < 0.250) on rPFS and OS (Table 4). Multivariate analysis results showed that age, treatment agent, PSA decline rate and metastatic sites were independently associated with rPFS. No factor was detected as an independent predictor of OS.

Discussion
In this study, we mainly aimed to examine whether there is a difference between rPFS and OS of mCRPC patients treated with E and AA. In addition, we evaluated prognostic factors affecting rPFS and OS in this patient group. Although E was statistically signi cantly superior to AA in terms of rPFS and OS, it didn't provide a signi cant reduction in death risk compared to AA. In all patients, being <75 years of age, PSA decline of <50% at 12 weeks of treatment were found to be poor risk factors for rPFS. In our study, rPFS and OS were 12 months and 20 months in the entire cohort, 15 months and 29 months in the E group, and 7 months and 16 months in the AA group, respectively (E vs. AA; p<0.001 for rPFS, p=0.027 for OS).
In the COU AA-301 study, AA treatment was compared with placebo in mCRPC patients in post-docetaxel setting. Median rPFS was 8.5 months at a follow-up period of 20.2 months, and median OS was 15.8 months at the nal analysis in the AA group; both superior to placebo (HR: 0.65; p<0.001) 4 . In the COU AA-302 study (pre-docetaxel setting trial of AA) at the end of a follow-up period of 49.2 months, the median rPFS was 16.3 months in patients given AA, and median OS was 34.7 months and 30.3 months, in the AA and placebo groups, respectively (HR: 0.81, p=0.003) 8 . In our study, AA treatment given in post-docetaxel and pre-docetaxel setting revealed an rPFS of 5 months and 12 months, and an OS of 13 months and 24 months, respectively. The reason for the shorter rPFS and OS compared to the COU AA301 and 302 studies may be the shorter follow-up time in our study. Unlike the COU-AA-302 study, our patient population in the AA group included patients with visceral metastasis. E was approved in August 2012, based on the phase III, randomized trial AFFIRM, including 1199 patients who were randomized to receive E or placebo prior CT. Median OS was 18.4 months in patients who received E and 13.6 months in patients who received placebo (HR, 0.63, p<0.001). The time to PSA progression was 8.3 vs. 3.0 months (HR, 0.25; P<0.001) and rPFS was 8.3 vs. 2.9 months in the E and placebo groups, respectively (HR, 0.40; P<0.001) 9 . E was approved to be used in the pre-docetaxel setting with the PREVAIL study, which was terminated early, as pre-docetaxel E treatment demonstrated a clear superiority in rPFS (20 months vs. 5.4 months) over placebo. In the placebo arm, 167 patients had crossed over to receive E. At a median follow-up of 69 months, median OS was 35.3 and 31.3 months in the treatment arm and placebo arm, respectively 10 . In our study, E given in post-docetaxel and predocetaxel settings revealed a rPFS of 11 months and 17 months, an OS of 26 months and 29 months, respectively. Although the median follow-up time was shorter when compared to the studies mentioned above, the results were consistent with the literature. Even, our post-docetaxel rPFS and OS were longer than the AFFIRM study.  13 . Another real-world study evaluated 110 patients with mCRPC who were treated with AA. Of the patients, 58 and 52 received AA in prechemotherapy (preCT) and postchemotherapy (postCT) settings, respectively. Median PFS was 15.5 and 6.4 months, and OS was 18.1 and 6.7 months for preCT and postCT groups, respectively. Similar to our study, the factor affecting PFS and OS was a decline of> 50% in PSA levels in the rst 3 months. Survival was signi cantly lower in patients with visceral metastasis 14 .
Nadal et al. examined 107 patients who were treated with E. Of the patients, 60 were pretreated with docetaxel and 47 were docetaxel-naive. Median PFS was superior in the docetaxel naive group (p<0.0001). They claimed that E activity was lower in patients who had previously received docetaxel CT and thought that there might be cross resistance between docetaxel and E. The follow-up period in the study of Nadal et al. was shorter than that in our study 15 . In a Japanese retrospective study about the treatment e cacy, safety pro le, and prognostic factors of E, 184 patients with non-mCRPC and mCRPC were analyzed; 44 (23.9%) non-mCRPC patients, 89 (48.4%) docetaxel-naive mCRPC patients, and 51 mCRPC patients pretreated with docetaxel (27.7%) mCRPC patients underwent E therapy. The median PSA PFS was 16.5 and 7.0 months, and overall survival was 59.8 and 30.4 months for docetaxel-naive and for docetaxel-pretreated mCRPC patients, respectively. Multivariate analysis identi ed that the predictive factor for a shorter OS was 4-week PSA decline <50%. This study had a relatively longer observation period with a median follow-up of 41.3 months, than the other retrospective studies and our study 16 .
The other retrospective studies comparing E and AA in a design similar to our study were reviewed. Al-Ali et al. analyzed 457 patients with CRPC who received AA and/or E in preCT and postCT settings. The median OS of the entire cohort was 21 months, 15 months for the AA group, 24 months for the E group, 26 months for the sequence group, and 10 months for the sequence group after switching. Median OS in the pre-CT setting was 25 months (mean: 21.5 ± 1.1 months) in the entire cohort, 18 months in AA group (mean: 18.9 ± 1.5 months) and 17 months in E treatment group (mean: 18.2 ± 1.9 months). In the post-CT setting, the median OS was 14 months in the AA group (mean: 15.8 ± 0.9 months), 19 months in the E group (mean: 17.2 ± 1.4 months) and 25 months in the sequence group (mean: 22.7 ± 0.8 months) 17 . In the study of Al-Ali et al., OS was shorter than our study and the other pilot studies in mCRPC patients treated with AA and E. Miyake et al compared the e cacy of AA and E in mCRPC patients in pre-CT setting. The study included 280 mCRPC patients, of the patients 113 and 167 were receiving AA and E, respectively. In the E group, PSA response rate and PSA PFS were signi cantly higher than that in the AA group. Duration of ADT treatment and ECOG PS for the AA group, age and ECOG-PS for the E group, and ECOG-PS for the overall patients were identi ed as the independent predictors of PSA PFS. The rate of patients with grade ≥ 3 side effects in the E group (11.4%) was signi cantly higher than that in the AA group (4.4%) 18  Due to the shorter follow-up time compared to other studies, rPFS and OS were relatively shorter in our study when compared to the results of other E and AA studies. Our study was retrospective, but as the patient groups had similar clinicopathological features, the results can be used to compare the e cacy of the two drugs in mCRPC patients. There is currently no prospective study in the literature similar to the design of this study. We suggest that the reasons why E was found to be signi cantly superior to AA in terms of rPFS and OS, were that the rate of metastasis at the time of diagnosis was signi cantly higher in the AA group (E vs. AA, 55.6 vs. 71.1 p=0.016) and the PSA decline rates were lower in the AA group.
When the PSA decrease rates were examined, it was found that only ≥50% PSA decrease in the AA group had a signi cant effect on both rPFS and OS. In the real-world studies performed with E, the most noticeable side effect was fatigue, which was higher than that found in phase 3 prospective trials. Since our study lacked side effect data, we could not give any results on this topic.
With the recent use of E and AA in hormone sensitive patients, the question is whether our real-world data will be compatible with the results of randomized prospective studies in the literature. Therefore, real-life parameters affecting rPFS and OS in our study and other retrospective studies will guide us to discover new indications for these drugs.