Abstract
Background/Aim: No practical predictive biomarkers exist to date for the response to androgen receptor-axis targeted (ARAT) therapies in metastatic castration-resistant prostate cancer (mCRPC). This study investigated whether prostate-specific antigen (PSA) kinetics in primary androgen-deprivation therapy for advanced hormone-sensitive prostate cancer may be associated with the response to ARAT agents in mCRPC. Patients and Methods: This study assessed 102 patients with mCRPC treated with enzalutamide or abiraterone to evaluate the associations between clinical outcomes and PSA kinetics, including the ratio of initial to nadir PSA (I/N PSA) level in primary combined androgen blockade. The PSA response was defined as a ≥50% decrease at 3 months from baseline in patients with mCRPC. Results: In patients treated with enzalutamide, the optimal cut-off I/N PSA value for PSA response was 531 ng/ml (sensitivity=66.7%, specificity=88.2%, area under the curve=0.73, using a receiver operating characteristic curve). The PSA response was 83.3% and 25.0% in the high and low I/N PSA groups, respectively. The median overall survival and radiographic progression-free survival from enzalutamide initiation were longer for the high compared to the low I/N PSA group. Multivariate analysis revealed I/N PSA (hazard ratio=0.275, p=0.026) as an independent risk factor for overall survival in the patients treated with enzalutamide. In contrast, I/N PSA showed no predictive ability for PSA response in patients treated with abiraterone. Conclusion: In patients with mCRPC, I/N PSA can be a practical predictive biomarker for response to the ARAT agent enzalutamide.
- Prostate cancer
- enzalutamide
- initial prostate-specific antigen level
- nadir prostate-specific antigen level
- androgen-deprivation therapy
Prostate cancer (PC) is the second most common malignancy among men (1). An estimated 1.41 million men were diagnosed with PC in 2020 globally, representing 14.1% of all cancers diagnosed in men, with 375,000 deaths caused by this disease (2).
Androgen-deprivation therapy (ADT) has been the standard treatment for patients with metastatic hormone-sensitive PC. Although PC is highly androgen-dependent and sensitive to primary ADT and although PC patients benefit from ADT at the beginning of their treatment plan, patients with metastatic hormone-sensitive PC eventually (within 2-3 years) develop metastatic castration-resistant PC (mCRPC), a disease associated with a high mortality rate (1, 3).
The treatment landscape of advanced PC is changing rapidly. Multiple agents have been approved for advanced PC, including enzalutamide, abiraterone, apalutamide, darolutamide, docetaxel, cabazitaxel, immunotherapy, radium-223 and sipuleucel-T (4). Appropriate drug selection and sequencing remain crucial in this evolving landscape to derive maximum benefit for patients with mCRPC. However, the most effective use of these therapies – order of administration, duration of treatment, and efficacy of combinations – has not yet been defined because of the absence of direct comparative randomised data (4). Currently, there are no well-validated practical predictive markers of response to androgen receptor (AR)-axis targeted (ARAT) therapy or chemotherapy, including markers, such as AR splice variant 7 (4, 5). Therefore, the decision between ARAT therapy and chemotherapy after disease progression with ADT is based on clinical factors and practical considerations regarding drug availability.
In most patients with advanced PC treated with primary ADT, disease progression to castration-resistant PC (CRPC) is unavoidable, despite castrate levels of serum testosterone. Typically, CRPC is associated with increases in serum prostate-specific antigen (PSA) levels, suggesting that the disease continues to be driven by AR signaling. In vitro evidence suggests that AR overexpression is adequate to confer resistance to androgen-deprivation in PC cell lines (6, 7) and that intratumoral androgen levels are often increased in patients with progressive PC (8). These observations have formed an explicit basis for developing more effective methods, such as ARAT therapy, which work to treat PC by further suppressing AR signaling (9, 10). Indeed, large-scale phase III trials have demonstrated the clinical benefits of ARAT therapy in patients with mCRPC (4, 11, 12). Moreover, recent studies have shown the PSA response to be a surrogate or predictive biomarker for mCRPC treated with enzalutamide or abiraterone (13-15). Therefore, the hypothesis in the current study of patients with mCRPC was that PSA kinetics in primary ADT may be associated with the efficacy of ARAT therapy based on the PSA-dependent mechanisms of ARAT agents.
Patients and Methods
Patients. This retrospective study was performed in accordance with the ethical standards of the Declaration of Helsinki. The study was approved by the Institutional Review Board of the University of Occupational and Environmental Health, Japan (authorization number: UOEHCRB21-074). An opt-out methodology was used to provide accessible information to all patients to facilitate informed consent without interfering with their medical consultation; the patients were informed of their inclusion in the study and were provided with study information on the institution’s website. This study enrolled consecutive male patients diagnosed with prostate adenocarcinoma at the University of Occupational and Environmental Health Hospital in Japan between August 2014 and December 2020. All study participants received primary ADT, following biochemically confirmed CRPC and castrate levels of testosterone (<50 ng/dl). These patients with CRPC received enzalutamide or abiraterone plus prednisolone 10 mg/day before or after docetaxel; patients who initiated ADT at other hospitals and were then referred to the study hospital were also included. Combined androgen blockade (CAB) consisted of medical or surgical castration plus a non-steroidal antiandrogen (bicalutamide 80 mg/day or flutamide 375 mg/day). Exclusion criteria were upfront combination therapy (use of docetaxel or ARAT agents prior to CRPC), non-CAB (medical or surgical castration alone) as ADT, history of definitive therapy for PC, non-metastatic hormone-sensitive PC and non-metastatic CRPC, all of which may affect nadir PSA and overall survival (OS). An additional exclusion criterion was second-line ARAT treatment following another ARAT agent, such as prior abiraterone use in enzalutamide-treated patients or prior enzalutamide use in abiraterone-treated patients, because of possible cross-resistance (16-18). Dose modification of either ARAT agent was permitted for patients with comorbidities or treatment-associated intolerable adverse events. Castration-resistant PC was defined according to the recommendations of the Prostate Cancer Clinical Trials Working Group 2 criteria (19), whereas radiologic surveillance was not routinely performed for all study participants. Serum PSA levels were measured every month for several months and imaging studies were conducted at the discretion of each physician after initiation of ARAT therapy.
Patient background assessments were as follows: age, Eastern Cooperative Oncology Group Performance Status, initial PSA (iPSA) levels, nadir PSA levels, initial to nadir PSA (I/N PSA) levels, time to nadir PSA, PSA doubling time, PSA levels at the time of ARAT initiation, primary tumour stage, International Society of Urological Pathology grade group, time to CRPC, extent of disease score, visceral metastases, and prior docetaxel use. I/N PSA was calculated by dividing the iPSA level by the nadir PSA level. Undetectable nadir PSA levels were categorized into high I/N PSA. Metastatic status was determined using computed tomography and bone scintigraphy before initiating CAB and ARAT therapy. OS was defined as the time between initiation of ARAT therapy and either the date of all-cause death or the date of last follow-up for surviving patients.
Treatment protocol. All patients with hormone-sensitive PC were initially treated with CAB. After CRPC diagnosis, patients received treatments comprising ARAT (enzalutamide and/or abiraterone) and/or systemic chemotherapy (docetaxel and/or cabazitaxel). Patients who chose not to receive chemotherapy or ARAT therapy were treated with alternative antiandrogen therapy, antiandrogen withdrawal therapy, estramustine, and/or low-dose oral steroid therapy. Changes of CRPC therapeutic agents occurred after PSA and/or radiologic progression or severe adverse events. Computed tomography and/or bone scan were performed when patients experienced any symptoms and/or PSA progression.
Statistical analysis. Patient characteristics were compared between the groups using the Mann–Whitney U-test or Fisher’s exact test for continuous or nominal variables, respectively. The optimal cut-off value for I/N PSA was defined as the closest point to the upper left corner of the receiver operating characteristic curve. Patients were divided into high or low I/N PSA groups based on the optimal cut-off value of I/N PSA. Patient characteristics, radiographic progression-free survival, and OS between the groups were compared. Survival rates were calculated using the Kaplan–Meier method and comparison was made using the log-rank test. The Cox proportional hazards model was used to determine statistical significance of predictors for 3-year OS in a multivariate setting. Variables with a value of p<0.05 were selected in the univariate models. All statistical analyses were performed using EZR (Easy R, Vienna, Austria), which is a graphical interface for R (The R Foundation for Statistical Computing). A value of p<0.05 was considered statistically significant.
Results
Patients. A total of 102 Japanese patients with mCRPC treated with ARAT therapy were identified during the study period. Among these patients, 57 were excluded based on the exclusion criteria or due to a lack of detailed clinical and follow-up data. The remaining 32 patients treated with enzalutamide and 13 patients treated with abiraterone were eligible for analysis.
The survival ROC curve analysis for PSA response revealed that the optimal cut-off value of I/N PSA was 531 ng/ml and the area under the curve (AUC) was 0.73 in the enzalutamide group (Figure 1). Of note, the abiraterone group showed no useful cut-off value of I/N PSA for PSA response (data not shown).
Survival receiver operating characteristics curve analysis of the ratio of initial to nadir prostate-specific antigen (PSA) levels for predicting PSA response in patients with metastatic castration-resistant prostate cancer. The area under the curve (AUC), sensitivity and specificity are shown. The black circle indicates the optimal point.
The baseline characteristics of the enzalutamide group categorized by I/N PSA is shown in Table I. The median follow-up time from enzalutamide initiation was 18 months (interquartile range=11-27 months). The I/N PSA was high in 12 (37.5%) patients and low in 20 (62.5%) patients (median=2113.4 ng/ml and 66.2 ng/ml, respectively). The high I/N PSA group had statistically lower nadir PSA levels and prior use of docetaxel, compared with the low I/N PSA group. Intriguingly, patients with high I/N PSA showed a statistically higher rate of PSA response (83.3%) to enzalutamide compared with those with low I/N PSA (25.0%) (p=0.003). No significant differences were observed for the other variables.
Patient characteristics based on I/N PSA status in the enzalutamide group.
Kaplan–Meier analyses for patients treated with enzalutamide. Heterogeneity in the PSA response was investigated using a cut-off value for I/N PSA of 531 ng/ml. The median OS from enzalutamide initiation was significantly longer in patients with high versus low I/N PSA (49.0 vs. 18.0 months, HR=0.27, 95% CI=0.10-0.73; p=0.010) (Figure 2A). The 3-year OS rates for patients with high and low PSA were 52.4% and 5.7%, respectively. The OS from CRPC diagnosis and radiographic progression-free survival from enzalutamide initiation also showed similar patterns to the OS from enzalutamide initiation (Figure 2B and C). Patients with high I/N PSA had significantly longer OS from CRPC diagnosis (median of 38 vs. 22 months, HR=0.36, 95% CI=0.14-0.92, p=0.033) and radiographic progression-free survival from enzalutamide initiation (median of 58 vs. 9 months, HR=0.18, 95% CI=0.04-0.83, p=0.028) than patients with low I/N PSA.
(A) Kaplan–Meier curves for overall survival (OS) from the start of enzalutamide. (B) OS from the diagnosis of castration-resistant prostate cancer (CRPC) and (C) radiographic progression-free survival (rPFS) from enzalutamide initiation, in patients with metastatic CRPC treated with enzalutamide (ENZ). The OS and rPFS were stratified by the ratio of initial to nadir prostate-specific antigen levels (I/N PSA). p-Values were calculated using the log-rank test.
Analyses of prognostic factors in mCRPC patients treated with enzalutamide Table II shows the results of the univariable and multivariable survival analyses using a Cox regression model for OS from enzalutamide initiation. Univariate analysis revealed that I/N PSA ≥531 ng/ml, time to CRPC ≥10 months, and PSA levels at enzalutamide initiation <16.57 ng/ml were significantly associated with long-term survival. Multivariate analysis revealed that I/N PSA ≥531 ng/ml (HR=0.28, 95% CI=0.09-0.86, p=0.026), and time to CRPC ≥10 months (HR=0.24, 95% CI=0.08-0.74, p=0.013) were significant prognostic factors.
Univariate and multivariate analyses to predict overall survival in the enzalutamide group.
Discussion
The study findings demonstrate that I/N PSA is associated with PSA response to enzalutamide and independently predicts OS in patients with mCRPC treated with enzalutamide. First, the optimal cut-off value of I/N PSA in the enzalutamide group was analyzed. Next, the results showed that I/N PSA was significantly associated with PSA response, radiographic progression-free survival, and OS, after initiation of enzalutamide. Finally, Cox hazard model analysis revealed that I/N PSA was an independent prognostic factor for OS in patients treated with enzalutamide. Of note, both of the independent prognostic factors for OS, namely I/N PSA and time to CRPC, during CAB may reflect hormone sensitivity of CRPC as well as that of hormone-sensitive PC. Interestingly, I/N PSA showed no predictive ability for PSA response using the ROC curve in patients treated with abiraterone. Therefore, I/N PSA may be a useful predictive biomarker for enzalutamide in practice, which allows clinicians to choose a suitable ARAT agent for patients with mCRPC.
To the best knowledge of the authors, this study is the first to show a practical predictive biomarker of response to enzalutamide in mCRPC. The efficacy of I/N PSA was assessed for the following reasons. Firstly, PSA kinetics during primary ADT/CAB may reflect hormone sensitivity of CRPC as well as hormone-sensitive PC because the significant antitumor effects of ARAT agents confirms that mCRPC remains hormonally driven and dependent on AR signaling (4, 11, 12). The kinetics of PSA, such as nadir PSA, time to nadir PSA, and 3-month PSA, during primary ADT/CAB are recognized as prognostic factors in patients with metastatic hormone-sensitive PC (20-22). High nadir PSA and short time to nadir during primary ADT are also poor prognostic factors in patients with mCRPC (23). Importantly, PSA is recognized as a surrogate biomarker for mCRPC treated with enzalutamide or abiraterone (13, 14).
Secondly, although high iPSA level is recognized as a poor prognostic factor in metastatic hormone-sensitive PC (24), recent studies have demonstrated that high iPSA does not always suggest unfavorable prognosis in CRPC. A pre-ARAT-era Japanese study revealed that high iPSA was a favorable prognostic factor in mCRPC (25). A pre- and post-ARAT-era study also revealed that high iPSA was a favorable prognostic factor for OS after CRPC (23). Another study using genomic data showed that patients with low iPSA and high-grade PC had a higher risk for PC-specific mortality, poor response to ADT, and neuroendocrine genomic features (26). The results allow for a hypothesis that metastatic hormone-sensitive PC with high iPSA may include a high rate of castration/hormone-sensitive clones. On the other hand, metastatic hormone-sensitive PC with low iPSA may include a high rate of castration-resistant clones with or without neuroendocrine features, which may lead to clinical CRPC with short ‘time to CRPC’. Because the ratio of iPSA and nadir PSA consists of I/N PSA, the latter is a potential predictive biomarker of ARAT agents in mCRPC that responds to further suppressing AR signaling. The results of the current study successfully demonstrated that I/N PSA was a predictive biomarker for enzalutamide and an independent prognostic factor for OS in patients treated with enzalutamide.
Notably, this study demonstrated that I/N PSA predicted the efficacy of enzalutamide but not abiraterone. This finding may be explained by two possible reasons. First, the result may be attributed to the different mechanisms between enzalutamide and abiraterone. Enzalutamide (MDV3100) is an AR blocker, which binds to AR with greater relative affinity than bicalutamide, reduces the efficiency of its nuclear translocation, and impairs both DNA binding to androgen response elements and recruitment of coactivators (27). In contrast, abiraterone is a selective and irreversible cytochrome P450 17 inhibitor, which blocks androgen biosynthesis (28). Second, I/N PSA may reflect tumor heterogeneity from the aspect of AR activation. A study by Nelson demonstrated that four discrete states of PC can be defined based on the sources of androgens/androgenic ligands and the activity of AR (29). According to the Nelson states, because enzalutamide is a potent AR blocker and abiraterone is a direct inhibitor of androgen biosynthesis, ‘State 2: intracrine androgen dependent and AR dependent’ clones respond to both enzalutamide and abiraterone, whereas ‘State 3: androgen (ligand) independent and AR dependent’ clones still respond to enzalutamide, but not abiraterone. The Nelson states and results of the current study allow for a hypothesis that metastatic hormone-sensitive PC with higher I/N PSA (higher initial-lower nadir) during primary CAB may develop a higher rate of State 2-3 clones, whereas hormone-sensitive PC with low I/N PSA (higher initial-higher nadir, lower initial-lower nadir, and lower initial-higher nadir) may have homogeneous clones, including State 2-4 CRPC.
The current study has several limitations. The study design was retrospective, the sample size was relatively small, and the follow-up was relatively short. The results only showed the value of I/N PSA as a predictive biomarker for enzalutamide in patients with hormone-sensitive PC treated with CAB, but not with ADT alone. Further large-scale, prospective studies are needed to validate these results.
Conclusion
In conclusion, I/N PSA may be a practical biomarker to predict response to enzalutamide in patients with mCRPC, which leads to optimizing the therapeutic strategies for this population. In patients with mCRPC treated with enzalutamide, I/N PSA may be a more useful prognostic biomarker combined with other variables, such as time to CRPC. Our results also suggest that patients with low I/N PSA, especially low initial-high nadir PSA, are suitable for chemotherapy because they may develop a higher rate of AR-pathway independent clones. Further large-scale and prospective studies are warranted.
Footnotes
Authors’ Contributions
YN: project development, data collection, data analysis, manuscript writing and editing. TM: data collection, data analysis. IT: critical review and supervision. NF: project development, critical review, and supervision. All Authors discussed, verified, and approved the final version of the article.
Conflicts of Interest
The Authors declare that there are no financial disclosures or conflicts of interest regarding this article.
- Received October 27, 2022.
- Revision received November 5, 2022.
- Accepted November 8, 2022.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
