Efficacy and Safety Exposure–Response Relationships of Apalutamide in Patients with Nonmetastatic Castration-Resistant Prostate Cancer Free

This article is featured in Highlights of This Issue, p. 4427

Apalutamide is an orally administered, specific inhibitor of the androgen receptor (AR), which is approved in the United States and European Union for the treatment of patients with nonmetastatic castration-resistant prostate cancer (nmCRPC). Apalutamide acts via inhibition of AR nuclear translocation and of AR binding to androgen response elements (1). The efficacy and safety of apalutamide versus placebo in men with high-risk nmCRPC have been evaluated in a multicenter phase III study (ARN-509-003; SPARTAN; NCT01946204), as described previously (2). Addition of apalutamide 240 mg once daily to androgen deprivation therapy (ADT) resulted in a 72% decrease in the risk of developing metastatic disease or death [HR = 0.28, 95% confidence interval (CI), 0.23–0.35; P < 0.001], compared with placebo plus ADT (3). This study also confirmed the tolerability and acceptable safety profile of apalutamide observed in earlier studies.

An understanding of the relationship between dose, exposure, and response is important to assess the benefit–risk profile and individualize the appropriate dose (4). This exploratory exposure–response (ER) analysis was undertaken to explore the relationships between exposure to apalutamide and its active metabolite, N-desmethyl-apalutamide, and selected clinical efficacy and safety endpoints.

The ER analysis was performed with data collected from all patients enrolled in the phase 3 SPARTAN study (2, 3). Patients with nmCRPC were randomized in a 2:1 ratio to receive apalutamide 240 mg or matched placebo administered orally on a continuous daily dosing regimen. After the first dose on cycle 1 day 1 0.5–4 hours postdose, pharmacokinetic samples were collected while predose samples were available for cycles 2, 3, 6, 12, 18, and 24. All patients who not had bilateral orchiectomy had to continue medical ADT with a gonadotropin-releasing hormone (GnRH) agonist or antagonist. Two formulations of apalutamide were used, leading to dosing of 240 mg first as (8 × 30 mg) capsules then as (4 × 60 mg) tablets, with no previously identified clinically relevant exposure differences between these two formulations (5).

For patients who experienced treatment-emergent adverse events (TEAE), dose interruptions and/or reductions were permitted if study discontinuation criteria had not been met. When patients experienced an AE of grade 3 or higher, apalutamide was to be held until symptoms improved to grade 1 or lower. Reinitiation of apalutamide therapy was allowed at the same or lower dose [maximum of 2 dose level reductions (240 mg–180 mg; 180 mg–120 mg)].

Metastasis-free survival (MFS) was the primary endpoint of the SPARTAN trial, and was defined as the time from randomization to first evidence of detectable bone or soft-tissue distant metastasis as identified by independent central review or death due to any cause (whichever occurred earlier). The MFS data for patients without metastasis or death were censored on the date of the last tumor assessment (or, if no tumor assessment was performed after the baseline visit, at the date of randomization). Patients continued with study treatment until blinded independent central review confirmed the detection of distant metastatic disease, development of unacceptable toxicity, or withdrawal of consent.

SPARTAN was conducted in accordance with principles for human experimentation as defined in the Declaration of Helsinki and was approved by the Human Investigational Review Board of each study center and by the Competent Authority of each country. Written informed consent was obtained from each patient before enrollment in the study, after the patient was advised of the potential risks and benefits, as well as the investigational nature of the study.

The observed pharmacokinetic data was used as input in the population pharmacokinetic model to derive model-based exposure metrics (6). Individual steady-state exposure metrics [area under the concentration curve after 24 hours (AUC_0–24h,ss), predose concentration (C_min,ss), and maximum concentration (C_max,ss)] were computed at steady state for apalutamide and N-desmethyl-apalutamide. The active moiety was calculated as the sum of apalutamide and N-desmethyl-apalutamide exposures weighted by their relative potency, considering that N-desmethyl-apalutamide exhibits one-third of the activity of apalutamide in an in vitro transcriptional reporter assay (7). After 4 weeks of treatment, more than 95% of patients had reached steady-state exposure of apalutamide and N-desmethyl-apalutamide, well before the first visit scheduled for MFS assessment at 16 weeks after start of treatment. The exposure metrics were derived on the basis of the maximum a posteriori estimates of the individual pharmacokinetic model parameter for apalutamide and N-desmethyl-apalutamide, which were obtained from the plasma concentrations collected in the SPARTAN study and the reference population pharmacokinetic model, described in a separate article (6). Given the option of dose modifications for AEs, the average taken daily dose up to the event of interest (either MFS or first occurrence of AE) was used to compute the exposure metrics. The exposure metrics (AUC_0–24h,ss, C_min,ss, and C_max,ss), adjusted by the average daily dose, reflect the average exposure of a subject up to the time of the first event of interest, considering possible dose reductions, dose interruptions, or missed doses prior to first time of the event.

Scatterplots and linear regression analyses were performed between individual steady-state AUC_0–24h,ss, C_min,ss, and C_max,ss to select the exposure metrics for the ER analysis. If a variance inflation factor higher than 5 (r² > 0.80) was detected among the exposure metrics, the ER analysis was performed with the steady-state AUC_0–24h,ss only; if not, C_min,ss and C_max,ss were also considered. The assessment of the correlation between exposure metrics was performed separately for apalutamide and N-desmethyl-apalutamide.

MFS was the primary endpoint included in the ER analysis. The effects of the stratification factors prostate-specific antigen doubling time (PSADT; >6 months vs. ≤6 months), the presence of locoregional disease (N0 vs. N1), and the coadministration of a bone-sparing agent (yes vs. no) on MFS were considered in the efficacy analysis (3). Because age and Eastern Cooperative Oncology Group (ECOG) performance status were identified to have a statistically significant (P < 0.05) association with overall survival in SPARTAN, these prognostic factors were also included in the exposure–MFS analysis.

The exposure–safety analysis focused on TEAEs of clinical relevance (i.e., classified as TEAEs of special interest or related to the drug), with occurrence higher than 10%, at any grade. On the basis of these criteria, fatigue (30.4% apalutamide vs. 21.1% placebo), falls (15.6% apalutamide vs. 9.0% placebo), skin rash (23.8% apalutamide vs. 5.5% placebo), weight loss (16.1% apalutamide vs. 6.3% placebo), and arthralgia (15.9% apalutamide vs. 7.5% placebo) were included in the exposure safety analysis.

Pharmacokinetic analysis and the maximum a posteriori estimates of the individual pharmacokinetic model parameters for apalutamide and N-desmethyl-apalutamide were obtained using NONMEM software (Icon Development Solutions; ref. 8). The data management, diagnostic graphics, postprocessing of NONMEM analysis results, and statistical analysis were carried out using R Project for Statistical Computing, Version 3.4.1 or higher for Windows (Comprehensive R Network, http://cran.r-project.org/; ref. 9).

The relationships between apalutamide, N-desmethyl-apalutamide, and active moiety exposure and MFS were first evaluated using Kaplan–Meier curves of MFS by quartiles of exposure. Placebo was included as a separate group. In addition, a multivariate Cox regression analysis for apalutamide and N-desmethyl-apalutamide exposure was performed; covariates included were treatment arm (apalutamide vs. placebo), the prespecified stratification factors [PSADT (≤ 6 months vs. > 6 months), presence of locoregional disease (N1 vs. N0), and bone-sparing agent used (yes vs. no)], and prognostic factors including age and ECOG performance status (1 vs. 0).

Patients randomized to the placebo group had apalutamide and N-desmethyl-apalutamide exposure assigned to zero. The impact of apalutamide and N-desmethyl-apalutamide exposure on MFS, after adjusting for the covariates, was assessed by the HR and its 95% CI. The P values as well as the change in −2 loglikelihood (LL) after the inclusion of each exposure metric were used for model comparison.

The selected safety endpoints (fatigue, fall, skin rash, weight loss, and arthralgia) were dichotomized into presence or absence of any TEAEs regardless of grade. Patients with multiple occurrences of events were only counted once, when the first event was experienced. As half of fractures were preceded by falls, and causality was hard to determine, falls (regardless if this was followed by a fracture or not), but not fractures, were included in the exposure–safety analysis.

Univariate and multivariate logistic regression analysis were conducted to assess any possible association between treatment (apalutamide vs. placebo) and apalutamide and N-desmethyl-apalutamide exposure on the selected safety endpoints. The corresponding odds ratio (OR), 95% CI, χ², and P values were calculated. The P values as well as the change in −2LL were used for model comparison. In addition, a sensitivity analysis was conducted by including and excluding placebo patients in the exposure–safety analysis.

The ER analyses described above were exploratory and hypothesis-generating and were performed under the assumption that there was a sufficient number of events for a meaningful analysis. P values lower than 0.05 were considered statistically significant. No correction for multiple statistical testing was implemented.

Data from all 1,207 randomized patients (806 in the apalutamide group and 401 in the placebo group) were included in the exposure–safety analysis while 1 patient from the apalutamide treatment group was excluded in the exposure–efficacy analysis because of missing values in a significant prognostic factor (ECOG performance status).

Consistent with the long half-life of apalutamide, pharmacokinetic data showed that 240 mg once-daily dosing led to a relatively constant apalutamide plasma concentration during the dosing interval at steady state, with limited peak-to-trough fluctuation ranging from a mean C_max,ss value of 5.81 μg/mL (range: 1.46–13.9) to a mean C_min,ss value of 4.24 μg/mL (range: 0.60–11.1). The peak-to-trough fluctuation of N-desmethyl-apalutamide plasma concentration at steady state is negligible, with relatively constant N-desmethyl-apalutamide plasma concentrations over the dosing interval at a plasma concentration average value of 6.3 μg/mL (range: 1.8–12.5). The apalutamide and N-desmethyl-apalutamide exposures, measured as AUC_0–24h,ss, were 115 (range: 19.8–291) and 152 (range: 44.1–299) μg·hour/mL, respectively. The interindividual variability in apalutamide and N-desmethyl apalutamide exposure, expressed as coefficient of variation of AUC_0–24h,ss was considered low to moderate (≤27%). In addition, following once-daily oral apalutamide administration of 240 mg, apalutamide and N-desmethyl-apalutamide systemic exposure showed a 5.3- and 85.2-fold accumulation (calculated, based on AUC) in plasma, respectively.

The correlations between the individual exposure metrics at steady state (AUC_0–24h,ss, C_min,ss, C_max,ss) for apalutamide and N-desmethyl-apalutamide were high (r² > 0.95) between all parameters. Because of this high correlation, AUC_0–24h,ss was selected as the unique exposure metric to conduct the ER analysis for efficacy and safety endpoints. In contrast, the correlation between AUC_0–24h,ss of apalutamide and N-desmethyl-apalutamide was estimated to be below 0.80 (r² = 0.646, P < 0.001), and therefore both apalutamide and N-desmethyl-apalutamide AUC_0–24h,ss were evaluated in the ER analysis, and the collinearity due to these two variables was further monitored.

Given the dose modification rate in the phase III SPARTAN study (21% dose reductions in the apalutamide group and 15% in the placebo group), the average daily dose (instead of the start dose of 240 mg) up to the event of interest (either MFS or first occurrence of TEAEs) was used to compute the exposure metrics. An average daily dose of 209 mg (apalutamide group) versus 222 mg (placebo group) in patients with one or more dose reductions was observed (Table 1); however, the relative difference between the derived AUC_0–24h,ss from the average daily dose and the dose-normalized 240 mg (assuming 100% treatment adherence) was relatively small (<13%). In addition, predicted dose-normalized AUC_0–24h,ss was similar in patients with or without dose reductions.

A summary of the prognostic factor distribution across treatment groups and quartiles of apalutamide and N-desmethyl-apalutamide exposure is presented in Table 2. These results suggest that the relevant prognostic factors were balanced among the treatment and exposure quartiles evaluated, and the potential imbalances that may exist should be controlled by the multivariate analysis. The Kaplan–Meier curves of MFS stratified by the quartiles of apalutamide, N-desmethyl-apalutamide, and the active moiety AUC_0–24h,ss are shown in Fig. 1. In all pairwise comparisons of the exposure subgroups in the apalutamide group versus the placebo group, a statistically significant (P < 0.0001) increase in MFS was observed for the apalutamide, N-desmethyl-apalutamide, and active moiety exposure subgroups.

However, after excluding the patients randomized to placebo, none of the other pairwise comparisons among the exposure subgroups for apalutamide, N-desmethyl-apalutamide, and active moiety were identified as statistically significant. These findings suggest that all exposure levels in patients treated with apalutamide experienced benefit. Because the results of the active moiety are comparable with the results of apalutamide and N-desmethyl-apalutamide, only apalutamide and N-desmethyl-apalutamide were carried forward to the multivariate Cox regression analysis.

The multivariate Cox regression analysis found a statistically significant association between apalutamide treatment, PSADT, presence of locoregional disease, and ECOG performance status on MFS. The effect of apalutamide on MFS (HR = 0.30; 95% CI, 0.24–0.36) was found to be similar after adjusting for the prognostic factors described previously (ref. 3; HR = 0.28, 95% CI, 0.23–0.35), suggesting that inclusion of the prognostic factors in the model did not produce any treatment-effect modification, and consequently the prognostic factors can be considered independently associated with MFS.

The addition of N-desmethyl-apalutamide AUC_0–24h,ss in the model on top of apalutamide AUC_0–24h,ss led to a substantial change of the treatment effect and its uncertainty (HR = 0.205, 95% CI, 0.097–0.435), which may be explained by the collinearity between prognostic factors, apalutamide treatment, and the exposure variables. A multivariate Cox regression analysis evaluated in the 806 patients treated with apalutamide (i.e., excluding placebo patients), showed a decrease in the collinearity and confirmed the results of the univariate analysis, that neither apalutamide AUC_0–24h,ss nor N-desmethyl-apalutamide AUC_0–24h,ss were statistically associated with MFS.

A summary of the incidence of the treatment-emergent events for placebo, apalutamide, and the quartiles of apalutamide and N-desmethyl-apalutamide exposure is presented in Table 3.

The univariate logistic regression analysis for the treatment effect showed that the probability of experiencing one of the described TEAEs was statistically significantly higher in the apalutamide group than the placebo group. Moreover, the probability of experiencing one of the described TEAEs significantly increases as apalutamide AUC_0–24h,ss increases (Fig. 2). The same effect was observed for N-desmethyl-apalutamide AUC_0–24h,ss (Fig. 3).

The treatment effect was included together with apalutamide or N-desmethyl-apalutamide exposure in the multivariate regression analysis. On the basis of the 95% CI of the OR, the contribution of apalutamide or N-desmethyl-apalutamide AUC_0–24h,ss, when adjusted by treatment effect differences, was significant for the probability of experiencing skin rash and weight loss, but not for the probability of experiencing fatigue, fall, or arthralgia at any grade.

The results of the multivariate logistic regression analysis were further confirmed in a univariate exposure–safety analysis in patients treated with apalutamide, which showed a statistically significant association between apalutamide treatment and skin rash and weight loss.

In the phase III SPARTAN study, the addition of apalutamide to ADT resulted in significant improvement over placebo plus ADT in the primary endpoint, MFS, as well as in other secondary endpoints, including time to symptomatic progression. However, apalutamide plus ADT was also associated with higher rates of skin rash, fatigue, arthralgia, weight loss, falls, and fracture (3). Some of these adverse events led to dose adjustments, which resulted in decreased drug exposure, and which provided the opportunity to undertake exploratory analyses to elucidate the relationship (if any) between apalutamide and N-desmethyl-apalutamide exposure and efficacy (MFS) as well as adverse drug reactions of special interest associated with the use of apalutamide.

The observed high correlation between the three exposure parameters selected (C_max,ss, C_min,ss, and AUC_0–24h,ss) can be expected considering that the apalutamide and N-desmethyl-apalutamide terminal half-life is substantially longer than the dosing interval; therefore, fluctuations in drug concentration during the dosing interval are low and drug concentrations collected at steady state are highly correlated with AUC. It is useful to correlate long-term drug effects with steady-state AUC_0–24h (as was done in this analysis), because AUC_0–24h,ss represents the average drug concentration over a longer time period following multiple dosing (10–12).

The 240 mg dose of apalutamide used in SPARTAN was selected on the basis of phase I data as well as on the inhibition of uptake of fluoro-5α dihydrotestosterone (13), a pharmacodynamic biomarker for AR inhibition. This dose level also ensured that exposures were within the range previously identified required for maximum tumor regression in murine CRPC models. The apalutamide exposure–efficacy analysis presented here revealed that in the relatively narrow observed exposure range (for both apalutamide and its metabolite), no statistically significant differences in MFS between exposure quartiles were determined, further confirming that 240 mg once-daily dose of apalutamide provides efficacious exposure in patients with nmCRPC, regardless of dose adjustments. Furthermore, the exposure was similar in patients with and without dose reductions and interruptions. Therefore, dose reductions/interruptions for the management of TEAEs are not expected to reduce the efficacy of apalutamide.

These results are in line with those obtained with another AR antagonist used for the treatment of metastatic CRPC, enzalutamide, which was evaluated in the AFFRIM trial (14). In this analysis, no apparent ER relationship was found for overall survival with an oral dose of 160 mg/day. The analysis was conducted using C_trough plasma values as the exposure metric instead of AUC_0–24h,ss, but the exposure–efficacy analysis was also conducted classifying the C_trough plasma values into quartiles, as in this analysis, and no significant differences in overall survival were seen regarding the quartiles of exposure.

Fatigue, fall, fracture, skin rash, weight loss, and arthralgia were adverse drug reactions that occurred at an incidence greater than 10% in SPARTAN patients receiving apalutamide. The exposure–safety analysis demonstrated that within the observed exposure range in apalutamide-treated patients, the exposure–TEAE relationship was only statistically significant for skin rash and weight loss. Simulations based on the developed exposure–safety relationship demonstrate that dose reductions will likely improve apalutamide tolerability in patients who develop toxicity after starting apalutamide treatment at 240 mg once daily.

Although these results suggest that reductions and interruptions are appropriate to prevent patients from discontinuing apalutamide, this is a relatively small dataset, and further study is warranted in real world use regarding the effect of drug holidays and extended use of lower doses.

In summary, ER analyses demonstrated that the MFS benefit in patients with nmCRPC was similar across the range of apalutamide and N-desmethyl-apalutamide exposure following the recommended apalutamide dose of 240 mg once daily. A significant association with apalutamide exposure was observed for the incidence of skin rash and weight loss TEAEs, at any grade. The ER relationship with these AEs suggests that patients who experience AEs may benefit from reducing the dose of apalutamide without compromising efficacy.

C. Perez-Ruixo reports personal fees from Janssen RD (employee) during as well as outside the submitted work. O. Ackaert reports personal fees from Janssen RD (employee) during as well as outside the submitted work and reports other from Johnson & Johnson (stock ownership) outside the submitted work. D. Ouellet reports other from Janssen (former employee) during the conduct of the study and other from Pfizer (current employee) outside the submitted work. C. Chien reports other from Janssen R&D (holds JNJ stocks) outside of the submitted work. D. Olmos reports grants, personal fees, non-financial support, and other from AstraZeneca [clinical trial steering committee (unpaid), travel support to scientific meetings]; grants and personal fees from Astellas; grants, personal fees, non-financial support, and other from Bayer Healthcare [clinical trial steering committee (unpaid), travel support to scientific meetings] and Janssen (clinical trial steering committees, travel support to scientific meetings); personal fees from Clovis Oncology and Daiichi-Sankyo; non-financial support and other from F-Hoffman-La Roche [clinical trial steering committee (unpaid), travel support to scientific meetings]; non-financial support from Genentech [clinical trial steering committee (unpaid)]; other from Ipsen (travel support to scientific meetings); and grants from Sanofi outside the submitted work. P. Mainwaring reports personal fees from Janssen (lectures, translational research) during the conduct of the study. M.K. Yu reports other from Janssen R& D (employee) during as well as outside the submitted work; in addition, M.K. Yu is listed as a co-inventor on a patent regarding a method of use of apalutamide for the treatment of patients with nmCRPC owned by Janssen. J.J. Perez-Ruixo reports personal fees from Janssen RD (employee) and other from Johnson & Johnson (stock owner) during the conduct of the study. M.R. Smith reports other from Janssen (clinical research support to institution) during the conduct of the study, and personal fees from Bayer, Astellas, and Janssen (consultant) outside the submitted work. E.J. Small reports personal fees from Janssen (advisory board, speakers fees) outside the submitted work. No potential conflicts of interest were disclosed by the other authors.

C. Perez-Ruixo: Formal analysis, methodology, writing-original draft, project administration, writing-review and editing. O. Ackaert: Formal analysis, methodology, writing-original draft, project administration, writing-review and editing. D. Ouellet: Formal analysis, methodology, writing-review and editing. C. Chien: Conceptualization, formal analysis, writing-review and editing. H. Uemura: Investigation, writing-review and editing. D. Olmos: Investigation, writing-review and editing. P. Mainwaring: Investigation, writing-review and editing. J.Y. Lee: Investigation, writing-review and editing. M.K. Yu: Conceptualization, writing-review and editing. J.J. Perez-Ruixo: Formal analysis, methodology, writing-original draft, writing-review and editing. M.R. Smith: Investigation, writing-review and editing. E.J. Small: Investigation, writing-review and editing.

The authors would like to thank the patients, investigators, and their medical, nursing, and laboratory staff who participated in the clinical studies included in the present work. The authors acknowledge Jonás Samuel Pérez-Blanco for his support during the exposure–response analysis. This study was funded by Janssen Research and Development.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.