Overall Survival After Systemic Treatment in High-volume Versus Low-volume Metastatic Hormone-sensitive Prostate Cancer: Systematic Review and Network Meta-analysis

Mike Wenzel a,b,*, Christoph Wu¨rnschimmel b,c, Luigi Nocera b,d, Claudia Colla` Ruvolo b,e, Zhe Tian b, Shahrokh F. Shariat f,g,h,i,j,k, Fred Saad b, Alberto Briganti d, Derya Tilki c,l, Markus Graefen c, Luis A. Kluth a, Frederik C. Roos a, Philipp Mandel a, Felix K.H. Chun a,
Pierre I. Karakiewicz b
a Department of Urology, University Hospital Frankfurt, Frankfurt am Main, Germany; b Cancer Prognostics and Health Outcomes Unit, Division of Urology, University of Montreal Health Center, Montreal, QC, Canada; c Martini-Klinik Prostate Cancer Center, University Hospital Hamburg-Eppendorf, Hamburg, Germany; d Department of Urology and Division of Experimental Oncology, Urological Research Institute, IRCCS San Raffaele Scientific Institute, Milan, Italy; e Department of Neurosciences, Reproductive Sciences and Odontostomatology, University of Naples Federico II, Naples, Italy; f Department of Urology, Comprehensive Cancer Center, Medical University of Vienna, Vienna, Austria; g Department of Urology, Weill Cornell Medical College, New York, NY, USA; h Department of Urology, University of Texas Southwestern, Dallas, TX, USA; i Department of Urology, Second Faculty of Medicine, Charles University, Prague, Czech Republic; j Institute for Urology and Reproductive Health, I.M. Sechenov First Moscow State Medical University, Moscow, Russia; k Division of Urology, Department of Special Surgery, Jordan University Hospital, The University of Jordan, Amman, Jordan; l Department of Urology, University Hospital Hamburg- Eppendorf, Hamburg, Germany

Article info

Article history:
Accepted April 4, 2021

Associate Editor: Malte Rieken

Keywords:
High volume Low volume
Androgen deprivation therapy Docetaxel
Abiraterone Apalutamide Enzalutamide Metastatic burden

Abstract

Context: Novel prospective randomized controlled observations addressing combina- tion therapy in metastatic hormone-sensitive prostate cancer (mHSPC) have demon- strated promising overall survival (OS) outcomes.
Objective: To compare these novel findings and systematically review and address them within formal network meta-analyses (NMAs) that include observations from other prospective randomized controlled trials (RCTs).
Evidence acquisition: First, we focused on abiraterone, enzalutamide, apalutamide, and docetaxel effects on OS in mHSPC using the PRISMA methodology. PubMed and abstracts identified prospective RCTs in first-line mHSPC. Second, we focused on mature studies that reached median OS and tested OS between abiraterone and docetaxel with tumor burden stratification.
Evidence synthesis: The first part included seven studies (n = 6639) and the second part, five studies (n = 4462). In the first part, abiraterone ranked first for high-volume mHSPC. Conversely, enzalutamide ranked first for low-volume mHSPC. In the second part, abiraterone treatment in high-volume mHSPC resulted in median OS of 50.1 mo and exceeded that with docetaxel (45.9 mo) and ADT alone (34.0 mo). Docetaxel treatment in low volume mHSPC resulted in median OS of 69.5 mo versus 67.7 mo with ADT alone. Conclusions: In conventional NMA that relied on conventional hazard ratios, differences were identified with respect to the relative efficacy of the combination therapies examined; abiraterone dominated the alternatives in high-volume mHSPC. In part

* Corresponding author. Department of Urology, University Hospital Frankfurt, Theodor-Stern Kai 7, 60590 Frankfurt, Germany. Tel. +49 69 630183147; Fax: +49 69 630183140.
E-mail address: [email protected] (M. Wenzel).

https://doi.org/10.1016/j.euf.2021.04.003

2405-4569/© 2021 European Association of Urology. Published by Elsevier B.V. All rights reserved.

2 E U R O P E A N U R O L O G Y F O C U S X X X ( 2 0 2 1 ) X X X – X X X

two, which relied on trials for which median OS is available, comparison of abir- aterone versus docetaxel revealed a 4-mo difference in OS in high-volume mHSPC. Conventional NMA may have overestimated the importance of treatment efficacy instead of focusing on median OS duration, which might represent a more important clinical endpoint.
Patient summary: We reviewed studies on hormonal treatments and chemotherapy used for prostate cancer that has spread outside the prostate gland (metastatic prostate cancer, mPC). We found that the best overall survival was with the hormone agents abiraterone in high-volume mPC and enzalutamide in low-volume mPC. In comparison to the chemotherapy drug docetaxel, median overall survival with abiraterone was 4 months longer among patients with mPC.
© 2021 European Association of Urology. Published by Elsevier B.V. All rights
reserved.

1. Introduction

Several reviews and meta-analyses have focused on thera- pies for metastatic hormone sensitive prostate cancer (mHSPC) [1–13]. According to standard network meta-anal- ysis (NMA) methodology, these reports examined relative measures of efficacy that rely on hazard ratios (HRs), illus- trating a decrease or increase in mortality. In those analyses, androgen receptor axis–targeted therapies (ARAT) invari- ably dominated docetaxel with respect to a relative overall survival (OS) benefit [2,6,7]. However, none of the previous studies and NMAs included median OS to better illustrate the incremental OS differences associated with ARATs versus docetaxel. We addressed this gap by undertaking a two-step systematic review and NMA of currently available reports on first-line ARAT or docetaxel for mHSPC. In the first part of the two-step analyses, we performed a conven- tional systematic review and NMA of all recent controlled phase 3 trials comparing all guideline-recommended sys- tematic management options in mHSPC relative to ADT alone, regardless of study maturity [14–23]. In the second part of the study, we exclusively focused on the controlled phase 3 trials that provided median OS values on the basis of data maturity [19,21,23–26]. Here, instead of relying on conventional HRs, we assessed median OS associated with each combination therapy and provided a rank order of median OS duration instead of HRs.

2. Evidence acquisition

2.1. Search strategy, study selection, and data collection

PubMed, MEDLINE, Embase, Web of Science, Scopus, the Cochrane database, and recent abstracts were searched to identify all available prospective randomized controlled trials (Fig. 1) addressing guideline-recommended combina- tion therapy in mHSPC, published before October 1, 2020 [14,15]. Subsequently, we repeated all of the above steps to focus on studies with median OS available. The following keywords were used in the search strategy to identify potential reports: (prostate cancer OR prostate carcinoma OR prostatic cancer OR prostatic cancer disease) AND (met- astatic OR M1 OR high volume disease OR advanced) AND (hormone sensitive OR hormone naïve OR castration

sensitive) AND (randomized). Two independent reviewers (M.W., C.W.) performed initial screening of the existing literature. Reasons for study exclusions were noted. We then performed a critical evaluation of the studies selected and relevant reports were subjected to a full-text review. The relevance of each study was confirmed on a case-by- case basis. Disagreements were resolved via consensus with all co-authors.

2.2. Outcome measures

In both study parts, the primary endpoint was OS and all results were stratified between CHAARTED-defined low- volume and high-volume mHSPC [24]. This essential dis- tinction was made on the basis of the well-established and generally recognized heterogeneity of these two distinct mHSPC patient populations. In the first part of the study, for which conventional NMA methodology was used, the comparison metrics consisted of HRs associated with each of five treatment alternatives. In the second part of the study, the comparison metrics consisted of median OS values associated with the three assessable treatment alternatives.

2.3. Inclusion and exclusion criteria

In the first part, studies were included if mHSPC patients underwent ADT alone versus ADT plus guideline-recom- mended first-line combination therapy within a prospective randomized controlled trial [14,15,17–23]. In the second part, studies with assessable median OS were included [19,21,23–26]. We excluded case reports, articles in a lan- guage other than English, reviews, editorials, letters, and replies from authors. References in the studies included were used to search for additional studies of potential interest. For part one, the inclusion criteria yielded seven first-line studies that focused on abiraterone, enzalutamide, apalutamide, and docetaxel combination therapies versus ADT alone [16–23]. For part two, five of these seven studies qualified for inclusion on the basis of median OS associated with each first-line combination therapy [19,21,23–26]. Our study search and inclusion criteria were in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [27,28].

E U R O P E A N U R O L O G Y F O C U S X X X ( 2 0 2 1 ) X X X – X X X 3

Fig. 1 – Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart depicting the studies and abstracts included in the two parts of the meta-analysis of overall survival (OS) in metastatic hormone-sensitive prostate cancer (mHSPC).
ADT = androgen deprivation therapy.

The risk of bias was evaluated for each study according to the Cochrane Collaboration tool [29]. This tool assesses selection bias (random sequence generation and allocation concealment), reporting bias, performance bias, attrition bias, detection bias, and other sources of bias (Supplemen- tary Fig. 1).

2.4. Statistical analyses

In the first part, we designed an NMA using random models with a Bayesian approach for direct and indirect treatment comparisons with ADT alone or combination therapies in mHSPC [30,31]. In the assessment of OS, contrast-based analyses were applied with estimated differences in the log HR and the standard error calculated from published HRs and confidence intervals (CIs) [32]. The relative treat- ment effects were presented as the HR and 95% credible interval (CrI). In addition, we estimated the relative ranking of different treatments for each outcome using the P score according to previous methodology [33–35]. All analyses were stratified between low- and high-volume mHSPC.

In part two, a weighted median OS (according to sample size) was defined for each available combination therapy and for ADT alone according to recently described method- ology [36]. Again, all analyses were stratified between low- and high-volume mHSPC. In high-volume mHSPC analyses, abiraterone was compared to docetaxel and to ADT alone. In low-volume mHSPC analyses, only docetaxel versus ADT alone could be compared. All analyses were performed with R version 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria).

3. Evidence synthesis

3.1. Descriptive characteristics of the studies included

Seven prospective randomized controlled trials (Table 1) qualified for inclusion in part one (n = 6639) of the study [17–23]. In three studies, docetaxel was compared to ADT alone. In two studies, abiraterone was compared to ADT alone. Finally, one study compared apalutamide or enzalu- tamide to ADT alone. Median follow-up ranged from 23 mo

Table 1 – Summary of prospective randomized controlled studies of combination therapy in patients with mHSPC

Reference Davis et al [17] Kyriakopoulos et al [24]

Gravis et al [25] Clarke et al [19] Fizazi et al [20] James et al [22] Chi et al [18]

Enrollment period 2014–2017 2006–2012 2004–2008 2005–2013 2013–2014 2011–2014 2015–2017

Inclusion criteria mHSPC mHSPC No previous CTx
2 prior docetaxel cycles permitted

mHSPC No previous CTx

mHSPC No previous CTx

High-risk b mHSPC No previous CTx or surgery

mHSPC or node- positive PC or 2 risk factors or high-risk relapse a

mHSPC
6 prior docetaxel cycles permitted

Performance status ECOG 0–2 ECOG 0–2 Karnofsky ≤70 WHO 0–2 ECOG 0–2 WHO 0–2 ECOG 0–1
Primary endpoint OS OS OS OS OS, rPFS OS OS, rPFS Patients 562/563 393/397 193/192 724/362 602/597 502/500 c 527/525
(control/treatment)
Median follow-up (mo) 34 54 84 78 52 73 c 23
Median age, yr (range) 69 (64–75) vs 62 (39–91) vs 64 (58–70) vs 65 (60–71) vs 67 (T9) vs 67 (39–83) vs 68 (43–90) vs
(control vs treatment) 69 (63–75) d 64 (36–88) 63 (57–68) d 65 (62–70) d 67 (T 9) 67 (42–85) a 69 (45–94)
Median PSA, ng/ml (range) – 52.1 (0.1–8056) vs 25.8 (5.0–126) vs 103 (33–338) vs – 56 (0–10 530) vs 4 (0–2229) vs
(control vs treatment) 50.9 (0.2–8540) 26.7(5.0–106) d 97 (38–348) d 51 (0–21 460) a 6 (0–2682)

Gleason score 8–10 (control vs treatment)

57% vs 60% 62% vs 61% 59% vs 55% 68% vs 69% 97% vs 98% 75% vs 74% a 68% vs 67%

RP or RT (control vs treatment) 42% vs 42% 27% vs 27% 24% vs 32% 5% vs 4% 5% vs 3% 3% vs 4% c 15% vs 18%

Bone metastasis (control vs treatment) Visceral metastasis (control vs treatment) High-volume disease (control vs treatment)

82% vs 80% – 81% vs 81% 87% vs 83% 87% vs 98% 47% vs 45% a 100% vs 100%

12% vs 11% 17% vs 14% 12% vs 15% 13% vs 12% 22% vs 22% 6% vs 4% a 15% vs 11%

52% vs 53% 64% vs 66% 47% vs 48% 57% vs 54% 78 vs 82% 48% balanced c 64% vs 62%

Prior CTx (control vs treatment) 52% vs 53% 0% vs 0% 0% vs 0% 0% vs 0% 0% vs 0% 0% vs 0% a 10% vs 11%
OS HR 0.67 HR 0.61 HR 0.88 HR 0.81 HR 0.66 HR 0.60 HR 0.67
(control vs treatment) 95% CI 0.52–0.86 95% CI 0.47–0.80 95% CI 0.68–1.14 95% CI 0.69–0.95 95% CI 0.56–0.78 95% CI 0.50–0.71 c 95% CI 0.51–0.89
OS in low volume mHSPC HR 0.43 HR 1.04 HR 1.02 HR 0.76 HR 0.72 HR 0.55 HR 0.67
(control vs treatment) 95% CI 0.26–0.72 95% CI 0.70–1.55 95% CI 0.67–1.55 95% CI 0.54–1.07 95% CI 0.47–1.10 h 95% CI 0.41–0.76 c 95% CI 0.34–1.32
OS in high volume mHSPC HR 0.80 HR 0.63 HR 0.78 HR 0.81 HR 0.62 HR 0.54 HR 0.68
(control vs treatment) 95% CI 0.59–1.07 95% CI 0.50–0.79 95% CI 0.56–1.09 95% CI 0.64–1.02 95% CI 0.52–0.74 h 95% CI 0.43–0.69 c 95% CI 0.50–0.92

rPFS HR 0.40 HR 0.61 95% CI 0.50–
0.75 e

HR 0.69 HR 0.66 HR 0.31 HR 0.31 HR 0.48

(control vs treatment) 95% CI 0.33–0.49 95% CI 0.55–0.87 95% CI 0.57–0.76 f 95% CI 0.27–0.36 g 95% CI 0.26–0.37 a,f 95% CI 0.39–0.60

E U R O P E A N U R O L O G Y F O C U S X X X ( 2 0 2 1 ) X X X – X X X 5
Table 2 – Summary of prospective randomized controlled studies in patients with mHSPC receiving docetaxel, abiraterone, or ADT with assessable median overall survival for the control and treatment arms

in the TITAN trial to 84 mo in the GETUG-AFU 15 trial [24]. The number of patients in the ADT and combination therapy arms ranged from 193 to 724 and from 192 to 597, respectively. The proportion of patients with high-volume mHSPC ranged from 47% to 78% in ADT arms and from 48% to 82% in combination therapy arms. It is of note that the study designs differed regarding previous therapy and per- formance status allowed. OS was the primary endpoint in all studies. Unlike in previous studies, the most recent STAM- PEDE trial update was included in the current study [23]. Similarly, unlike in previous studies, the current study also included the most recent LATITUDE trial updates with stratification according to high- versus low-volume mHSPC [20].
Part two (n = 4462) focused on the five prospective randomized controlled trials included in part one that provided assessable median OS (Table 2) [19,21,23–

26]. Two of those studies represent final analyses of CHAARTED and GETUG-AFU 15 [24,25]. The third study focused on the ongoing STAMPEDE trial [19]. The fourth study specifically focused on abiraterone-treated patients according to low- versus high-volume mHSPC definitions [26]. The fifth and final study included consisted of an abstract on the most recent LATITUDE abiraterone update [21]. In three studies, docetaxel was compared to ADT alone in mHSPC patients, while in two studies, abiraterone was compared to ADT alone.

3.2. Part one: NMA of seven studies in high-volume mHSPC regardless of follow-up duration

In the subgroup of high-volume mHSPC and relative to ADT (Fig. 2A), only abiraterone, apalutamide and docetaxel resulted in longer OS. The HR was 0.59 (95% CrI 0.50–

Hazard ratio

Hazard ratio

Fig. 2 – Forest plots of the network meta-analysis of the association between systemic therapy in metastatic hormone-sensitive prostate cancer and overall survival for patients with (A) high-volume and (B) low-volume disease.
ADT = androgen deprivation therapy; HR = hazard ratio; CI = confidence interval.

6 E U R O P E A N U R O L O G Y F O C U S X X X ( 2 0 2 1 ) X X X – X X X

Weighted median of medians Range for observed medians

Weighted median of medians Weighted median of medians
Range for observed medians Range for observed medians

Fig. 3 – Forests plots of weighted medians of overall survival (OS) for (A) ADT versus docetaxel versus abiraterone, (B) all studies including abiraterone,
(C) all studies including docetaxel, and (D) all studies including ADT alone in high-volume hormone-sensitive prostate cancer (mHSPC). ADT = androgen deprivation therapy; NA = not available or not reached.

0.69) for abiraterone, 0.68 (95% CrI 0.50–0.93) for apaluta-
mide, and 0.73 (95% CrI 0.62–0.85) for docetaxel. Enzalu- tamide did not result in longer OS relative to ADT (HR 0.80, 95% CrI 0.59–1.08). In NMA-derived ranking quantifying the highest likelihood of maximal OS benefit in patients with high-volume mHSPC, abiraterone ranked first (P score 0.93), followed by apalutamide (P score 0.65) and docetaxel (P score 0.53).

3.3. Part one: NMA of seven studies in low-volume mHSPC regardless of follow-up duration

In the low-volume mHSPC subgroup, only abiraterone and enzalutamide resulted in longer OS relative to ADT (Fig. 2B). The HR was 0.61 (95% CrI 0.47–0.79) for abiraterone and
0.43 (95% CrI 0.26–0.72) for enzalutamide. Apalutamide (HR 0.67, 95% CrI 0.34–1.32) and docetaxel (HR 0.91, 95% CrI 0.73–1.13) did not result in longer OS relative to ADT. According to NMA-derived ranking of the likelihood of a maximal OS benefit in low-volume mHSPC, enzalutamide ranked first (P score 0.93), followed by abiraterone (P score 0.68).

3.4. Part two: meta-analyses of studies with assessable median OS in high-volume mHSPC

Two studies reported on abiraterone-exposed high-volume mHSPC (Fig. 3), for which median OS ranged from 49.7 to
51.0 mo, resulting in a combined weighted median OS of
50.1 mo. Three studies reported on docetaxel-exposed high- volume mHSPC, for which median OS ranged from 39.8 to
51.2 mo, resulting in a combined weighted median OS of
45.9 mo. Five studies reported on ADT-exposed high-vol- ume mHSPC, for which median OS ranged from 33.3 to

35.1 mo, resulting in a combined weighted median OS of
34.0 mo.

3.5. Part two: meta-analyses of studies with assessable median OS in low-volume mHSPC

Two studies reported on docetaxel-exposed low-volume mHSPC (Fig. 4), for which median OS ranged from 63.5 to
76.0 mo, resulting in a combined weighted median OS of
69.5 mo. Two studies reported on ADT-exposed low-volume mHSPC, for which median OS ranged from 61.0 to 83.4 mo, resulting in a combined weighted median OS of 67.7 mo. Owing to immature data for abiraterone in low-volume mHSPC, no weighted OS median could be computed.

3.6. Discussion

The topic of all guideline-recommended combination ther- apies in mHSPC was addressed in three previous NMAs [4,6,7]. All three relied on HRs according to the standard methodology. Within the three NMAs, the rank order attrib- uted to various first-line systemic treatment alternatives changed according to study maturity. For example, in the first NMA report, enzalutamide ranked first in high-volume mHSPC [6]. Conversely, abiraterone ranked first in the second NMA for high-volume mHSPC [7]. The third NMA did not stratify according to low- versus high-volume mHSPC [4]. It is of note that several studies that have not reached final follow-up maturity were included in those analyses. For example, the first reported NMA assessed abiraterone studies (LATITUDE and STAMPEDE arm G) with less mature follow-up [22,37]. As a consequence, previous NMAs that relied on data with suboptimal maturity may have under- or overestimated the effect of the agent of

E U R O P E A N U R O L O G Y F O C U S X X X ( 2 0 2 1 ) X X X – X X X 7

Weighted median of medians Range for observed medians

Weighted median of medians Range for observed medians

Fig. 4 – Forests plots of weighted medians of overall survival for (A) all studies including docetaxel and (B) all studies including ADT alone in low- volume hormone-sensitive prostate cancer (mHSPC).
ADT = androgen deprivation therapy; NA = not available or not reached.

interest on OS according to final analyses that relied on median OS based on preplanned study designs. This obser- vation is emphasized by the fact that HRs for the same trial may change according to the follow-up duration (eg, the LATITUDE trial). To address this limitation, we conducted a systematic review and an NMA that relied on the most mature studies that examined mHSPC (part one). Moreover, since final reports for some of the first-line combination therapies examined were not available at the time of the current study (eg, ENZAMET and TITAN), we also conducted a separate meta-analysis (part two) that exclusively included mature studies with assessable median OS in order to eliminate the potential bias related to suboptimal data maturity. This approach was not used in previous meta- analyses. However, use of studies with assessable median OS may provide the most bias-free efficacy comparisons.
Unlike previous studies, we used stratification between low- and high-volume mHSPC. This distinction was initially defined in the CHAARTED trial and illustrated very impor- tant OS differences between these two patient groups [38]. These differences were subsequently validated in other prospective and retrospective analyses that evidenced equally important OS differences. Taken together, those findings imply that low- versus high-volume mHSPC strat- ification results in two distinct mHSPC patient populations that ideally should not be grouped or interpreted together [26,39]. This criterion (low- vs high-volume mHSPC) has been endorsed and integrated in treatment guideline recommendations [14,15]. In light of these considerations, all analyses performed in the current study were stratified according to high- versus low-volume mHSPC. Accordingly, no interpretations or comparisons of OS were made for the combined group of patients with low- and high-volume mHSPC, as previously done [40]. Our analyses revealed several important observations.
In part one, we analyzed seven randomized controlled trials that compared ARATs, docetaxel, and ADT alone in the setting of high- versus low-volume mHSPC. Our methodol- ogy used standard NMA considerations, with efficacy and rank order defined according to OS HRs for each combina- tion therapy relative to ADT alone. In part one, the NMA methodology applied to high-volume mHSPC ranked abir- aterone first as the most effective for OS, followed by

apalutamide and docetaxel, in that order. Because its CrI intersects with the null value (HR 1.0), the effect of enza- lutamide on OS could not be considered as significantly better than ADT alone.
In part one of the current study, the NMA methodology applied to low-volume mHSPC ranked enzalutamide first and abiraterone second. Apalutamide and docetaxel did not exhibit a significant OS advantage relative to ADT alone, as the CrI intersects with HR of 1.0. These results rely on the most mature data and agree with the most recently reported NMA by Chen et al [7]. Conversely, our findings do not agree with the second most recent and second most mature NMA by Sathianathen et al [6], in which enzaluta- mide was ranked first for both high- and low-volume mHSPC. Differences in outcome rankings from previous NMAs may be directly linked to data maturity and fol- low-up duration.
Despite reporting the most mature data, only five of seven studies included in the first part of the analyses provided median OS. Therefore, the ranks obtained in the current NMA may result in a measure of efficacy that may exaggerate or underestimate OS outcomes according to trial results with assessable median OS. Thus, in the second part of the current study we exclusively focused on trials with assessable median OS and also stratified according to high- versus low-volume mHSPC.
Data addressing high-volume mHSPC with assessable median OS only allowed comparisons for abiraterone versus docetaxel versus ADT alone. The combined median OS was
50.1 mo for abiraterone versus 45.9 mo for docetaxel versus
34.0 mo for ADT alone. Two studies reported median OS with abiraterone use, which ranged from 49.7 to 51.0 mo. This narrow median OS range for the two trials indicates that they can be considered as very homogeneous and highly comparable. Therefore, it can be postulated that abiraterone efficacy is highly reproducible, at least in Euro- pean cohorts, according to the study locations. Docetaxel data came from three studies. Two of these (Clarke et al [19], n = 148; Gravis et al [25], n = 92) involved European patients and demonstrated equally homogeneous findings, as evi- denced by a very narrow median OS range of 39.8–40.1 mo. Conversely, the third study involved North American patients (Kyriakopoulos et al [24], n = 263) and had

8 E U R O P E A N U R O L O G Y F O C U S X X X ( 2 0 2 1 ) X X X – X X X

substantially longer median OS of 51.2 mo. Its effect con- tributed to a weighted median OS that shifted to 45.9 mo for all three trials combined. This shift in median OS away from median OS values from the two European trials (39.8 and
40.1 mo) suggests heterogeneity of docetaxel median OS findings between Europe and North America and may indi- cate that European and North American data should not be combined. This assumption is supported by the AFFIRM trial (enzalutamide), in which stratified outcomes differed with respect to median OS between European and North Ameri- can patients with metastatic castration-resistant prostate cancer [41]. The differences may be explained by differences between European and North American patients regarding disease characteristics, such as stage at diagnosis [42]. Nonetheless, interpretation is at the discretion of clinicians treating mHSPC patients. It is also of interest that median OS values from five different ADT treatment arms, of which four were European and one was North American, also demonstrated impressive homogeneity (range 33.3–
35.1 mo). On the basis of these observations, it can be postulated that European median OS data for abiraterone, docetaxel, and ADT are in strong agreement. It is also of note that North American observations for median OS after ADT alone are in virtually perfect agreement with European data. The only exception to this remarkable homogeneity is the North American docetaxel arm [24], which reported a 10- mo median OS advantage relative to European data. This discrepancy suggests differences in docetaxel response rates between North American and European patients. However, it is also possible that population differences contributed to the strong observed median OS discrepancy. However, these hypotheses remain speculative, since they represent inter-trial comparisons. Nonetheless, it should be noted that no existing statistical methodology, including NMA methodology, is capable of adjusting for population or treatment efficacy differences that underlie heterogeneity. For low-volume mHSPC, there was no numerical differ- ence in weighted median OS between docetaxel and ADT alone after combining two docetaxel arms and two ADTarms. It is noteworthy that the individual treatment arms for both docetaxel and ADTalone in this setting exhibited very impor- tant heterogeneity. For example, Kyriakopoulos et al [24] reported median OS of 63.5 mo with docetaxel for North AmericanpatientswithlowvolumemHSPC, while Clarkeetal
[19] reported median OS of 76 mo for their European counter- parts. This difference in median OS of more than 12 mo favors the European setting. The discrepancy might be attributable to a difference in the low-volumemHSPCphenotypebetween Europe and North America. Alternatively, the reason could be a better response to abiraterone among European patients relative to their North American counterparts. The same may apply to the difference in median OS with ADT alone in low- volume mHSPC between European and North American cohorts. For both docetaxel and ADT, this heterogeneity in median OS may mean that it is not methodologically appro- priate to consider weighted OS medians. Finally, it is also of interest that North American docetaxel data [24] revealed markedly shorter median OS than for European patients [19]. In summary, very important heterogeneity in median

OS values was identified in low-volume mHSPC. This hetero- geneity may undermine the validity of attempts to combine data from different studies. Moreover, the presence of such important heterogeneity emphasizes the need for more mature data from low-volume mHSPC cohorts. Unfortu- nately, the more favorable phenotype of low-volume mHSPC will require substantially longer follow-up to reach maturity than the duration required for high-volume mHSPC.
Taken together, prospective controlled randomized stud- ies and NMAs using HRs for relative OS benefits need to be interpreted with caution. Particular caution is required when suboptimal maturity may apply to study endpoints and median OS values may not have been reached because of insufficient event numbers or short follow-up. To obviate such limitations, comparative assessments, as well as direct comparisons, should ideally be based on studies with optimal maturity. Optimal maturity may consist of preplanned num- bers and/or proportions of events reached according to final analysis definitions. In the current study, we carried out the first analysis for seven studies with variable maturity, for which HRs were used as the endpoint metric. Our alternative approach used a more limited number of studies, for which the endpoint metric was median OS. The two approaches resulted in different observations. More robust observations may be generated from studies with the most mature follow- up, for which median OS is available. Conversely, use of less maturestudiesforwhichmedian OS isnotavailablemayyield observations that disagree with final analyses. This scenario applies to several NMAs that examined the effect of first-line combination therapies in mHSPC. However, even when the most mature studies with assessable median OS data are used, important heterogeneity may be observed for each combination therapy. This scenario applied to low-volume mHSPC findings, for which the North American docetaxel median OS differed from the European data. In addition, even European median OS values for ADT arms differed from each other, depending on the study country (UK [19] vs France [25]). Conversely, the more robust data available for high- volume mHSPC showed a substantially higher degree of agreement. Here, abiraterone use resulted in a 4-mo OS benefit relative to docetaxel and a 16-mo benefit relative to ADT. The weighted median OS benefit with abiraterone increased to 10 mo relative to docetaxel when only homoge- neous European data were used.
Our systematic review, NMA, and meta-analysis have limitations that have to be considered in the interpretation of our findings. First, differences in patient characteristics and study designs among the randomized controlled trials included are important to consider in the interpretation of the current and all previous NMAs. For example, two of the studies we included (LATITUTE and TITAN) not only consid- ered OS as a primary study endpoint but also focused on radiographic progression-free survival (PFS). Second, study maturity differed extensively between the trials, with median follow-up ranging from 23 to 84 mo. Third, the ARCHES trial (enzalutamide) could not be included in this study since it does not provide OS data that were stratified according to high- versus low-volume mHSPC. Moreover, studies that did not provide median OS data (ENZAMET or

E U R O P E A N U R O L O G Y F O C U S X X X ( 2 0 2 1 ) X X X – X X X 9

TITAN) could not be included in analyses of weighted OS medians [16–18]. Finally, we could not compute weighted median PFS for the combination therapies examined, since most studies did not provide median PFS duration. However, we believe that clinicians invariably choose the combination therapy withlonger OS, regardless of PFS [43,44]. Fourth, only two studies could be included for computation of the weighted median of medians for abiraterone, while three studies could be included for docetaxel. However, the total number of patients included was higher for abiraterone than for docetaxel. Finally, the methodology for weighted medians fromseveral trialsdoes notallow formaltests of the statistical significance of differences. However, this methodology is highly valuable for interpretation of median OS for the pur- pose of clinical decision-making.

4. Conclusions

Conventional NMA using HRs revealed differences in the relative efficacy of the combination therapies examined; abiraterone dominated the alternatives in high-volume mHSPC. In NMA of trials for which median OS is available, comparison of abiraterone versus docetaxel revealed a 4- mo difference in OS in high-volume mHSPC. Conventional NMA may have overestimated the importance of treatment efficacy instead of focusing on median OS duration, which might represent a more important clinical endpoint.

Author contributions: Mike Wenzel 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: Wenzel, Würnschimmel, Karakiewicz. Acquisition of data: Wenzel, Würnschimmel, Nocera, Collà Ruvolo. Analysis and interpretation of data: Wenzel, Würnschimmel, Nocera, Collà Ruvolo, Karakiewicz.
Drafting of the manuscript: Wenzel, Würnschimmel, Karakiewicz. Critical revision of the manuscript for important intellectual content: Shar- iat, Saad, Briganti, Tilki, Graefen, Kluth, Roos, Mandel, Chun.
Statistical analysis: Wenzel, Würnschimmel, Tian.
Obtaining funding: None.
Administrative, technical, or material support: None.
Supervision: Shariat, Saad, Briganti, Tilki, Graefen, Kluth, Roos, Mandel, Chun.
Other: None.

Financial disclosures:Mike Wenzel 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: None.

Funding/Support and role of the sponsor: None.

Appendix A. Supplementary data

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10. 1016/j.euf.2021.04.003.

References

[1] Loblaw DA, Virgo KS, Nam R, et al. Initial hormonal management of androgen-sensitive metastatic, recurrent, or progressive prostate cancer: 2006 update of an American Society of Clinical Oncology practice guideline. J Clin Oncol 2007;25:1596–605.
[2] Wallis CJD, Klaassen Z, Bhindi B, et al. Comparison of abiraterone acetate and docetaxel with androgen deprivation therapy in high- risk and metastatic hormone-naïve prostate cancer: a systematic review and network meta-analysis. Eur Urol 2018;73:834–44.
[3] Tan PS, Aguiar P, Haaland B, Lopes G. Addition of abiraterone, docetaxel, bisphosphonate, celecoxib or combinations to andro- gen-deprivation therapy (ADT) for metastatic hormone-sensitive prostate cancer (mHSPC): a network meta-analysis. Prostate Cancer Prostat Dis 2018;21:516–23.
[4] Marchioni M, Di Nicola M, Primiceri G, et al. New antiandrogen compounds compared to docetaxel for metastatic hormone sensi- tive prostate cancer: results from a network meta-analysis. J Urol 2020;203:751–9.
[5] Kinsey EN, Zhang T, Armstrong AJ. Metastatic hormone-sensitive prostate cancer: a review of the current treatment landscape. Cancer J 2020;26:64–75.
[6] Sathianathen NJ, Koschel S, Thangasamy IA, et al. Indirect compar- isons of efficacy between combination approaches in metastatic hormone-sensitive prostate cancer: a systematic review and net- work meta-analysis. Eur Urol 2020;77:365–72.
[7] Chen J, Ni Y, Sun G, et al. Comparison of current systemic combina- tion therapies for metastatic hormone-sensitive prostate cancer and selection of candidates for optimal treatment: a systematic review and Bayesian network meta-analysis. Front Oncol 2020;10:519388.
[8] Barata PC, Sartor AO. Metastatic castration-sensitive prostate can- cer: abiraterone, docetaxel, or Cancer 2019;125:1777–88.
[9] McNamara M, Sweeney C, Antonarakis ES, Armstrong AJ. The evolving landscape of metastatic hormone-sensitive prostate cancer: a critical review of the evidence for adding docetaxel or abiraterone to andro- gen deprivation. Prostate Cancer Prostat Dis 2018;21:306–18.
[10] Kassem L, Shohdy KS, Abdel-Rahman O. Abiraterone acetate/andro- gen deprivation therapy combination versus docetaxel/androgen deprivation therapy combination in advanced hormone-sensitive prostate cancer: a network meta-analysis on safety and efficacy. Curr Med Res Opin 2018;34:903–10.
[11] Aoun F, El Rassy E, Sleilaty G, Assi T, Bakouny Z, Kattan J. The optimal treatment of metastatic hormone-naive prostate cancer: abirater- one acetate or docetaxel? Future Oncol 2017;13:2785–90.
[12] Rydzewska LHM, Burdett S, Vale CL, et al. Adding abiraterone to androgen deprivation therapy in men with metastatic hormone- sensitive prostate cancer: a systematic review and meta-analysis. Eur J Cancer 2017;84:88–101.
[13] Preisser F, Chun FKH, Banek S, et al. Management and treatment options for patients with de novo and recurrent hormone-sensitive oligometastatic prostate cancer. Prostate Int. In press. https://doi. org/10.1016/j.prnil.2020.12.003.
[14] Mottet N, Cornford P, van den Bergh RCN, et al. EAU-EANM-ESTRO- ESUR-SIOG guidelines on prostate cancer. Arnhem, The Netherlands: European Association of Urology; 2020. In: https:// uroweb.org/guideline/prostate-cancer/
[15] Mohler JL, Antonarakis ES, Armstrong AJ, et al. Prostate cancer, version 1.2020, NCCN clinical practice guidelines in oncology. National Comprehensive Cancer Network; 2020.
[16] Armstrong AJ, Szmulewitz RZ, Petrylak DP, et al. ARCHES: a ran- domized, phase III study of androgen deprivation therapy with enzalutamide or placebo in men with metastatic hormone-sensi- tive prostate cancer. J Clin Oncol 2019;37:2974–86.

10 E U R O P E A N U R O L O G Y F O C U S X X X ( 2 0 2 1 ) X X X – X X X

[17] Davis ID, Martin AJ, Stockler MR, et al. Enzalutamide with standard first-line therapy in metastatic prostate cancer. N Engl J Med 2019;381:121–31.
[18] Chi KN, Agarwal N, Bjartell A, et al. Apalutamide for metastatic, castration-sensitive prostate cancer. N Engl J Med 2019;381:13–24.
[19] Clarke NW, Ali A, Ingleby FC, et al. Addition of docetaxel to hor- monal therapy in low- and high-burden metastatic hormone sen- sitive prostate cancer: long-term survival results from the STAM- PEDE trial. Ann Oncol 2019;30:1992–2003.
[20] Fizazi K, Tran N, Fein L, et al. Abiraterone acetate plus prednisone in patients with newly diagnosed high-risk metastatic castration- sensitive prostate cancer (LATITUDE): final overall survival analysis of a randomised, double-blind, phase 3 trial. Lancet Oncol 2019;20:686–700.
[21] Fizazi K, Tran N, Fein LA, et al. Final analysis of phase III LATITUDE study in patients (pts) with newly diagnosed high-risk metastatic castration-naïve prostate cancer (NDx-HR mCNPC) treated with abiraterone acetate + prednisone (AA+P) added to androgen depri- vation therapy (ADT). J Clin Oncol 2019;37(7 Suppl):141.
[22] James ND, de Bono JS, Spears MR, et al. Abiraterone for prostate cancer not previously treated with hormone therapy. N Engl J Med 2017;377:338–51.
[23] James ND, Rush H, Clarke N, et al. 611O Abiraterone acetate plus prednisolone for hormone-naïve prostate cancer (PCa): long-term results from metastatic (M1) patients in the STAMPEDE randomised trial (NCT00268476). Ann Oncol 2020;31(Suppl 4):S509.
[24] Kyriakopoulos CE, Chen Y-H, Carducci MA, et al. Chemohormonal therapy in metastatic hormone-sensitive prostate cancer: long- term survival analysis of the randomized phase III E3805 CHAARTED trial. J Clin Oncol 2018;36:1080–7.
[25] Gravis G, Boher J-M, Joly F, et al. Androgen deprivation therapy (ADT) plus docetaxel versus ADT alone in metastatic non castrate prostate cancer: impact of metastatic burden and long-term sur- vival analysis of the randomized phase 3 GETUG-AFU15 trial. Eur Urol 2016;70:256–62.
[26] Hoyle AP, Ali A, James ND, et al. Abiraterone in ‘high-’ and ‘low-risk’ metastatic hormone-sensitive prostate cancer. Eur Urol 2019;76:719–28.
[27] Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group. Preferred reporting Items for Systematic Reviews and Meta-Analyses: the PRISMA statement. J Clin Epidemiol 2009;62:1006–12.
[28] Hutton B, Salanti G, Caldwell DM, et al. The PRISMA extension statement for reporting of systematic reviews incorporating net- work meta-analyses of health care interventions: checklist and explanations. Ann Intern Med 2015;162:777–84.
[29] Higgins JPT, Altman DG, Gøtzsche PC, et al. The Cochrane Collabora- tion’s tool for assessing risk of bias in randomised trials. BMJ 2011;343:d5928.
[30] van Valkenhoef G, Lu G, de Brock B, Hillege H, Ades AE, Welton NJ. Automating network meta-analysis. Res Synth Methods 2012;3:285–99.
[31]
Dias S, Sutton AJ, Ades AE, Welton NJ. Evidence synthesis for decision making 2: a generalized linear modeling framework for pairwise and network meta-analysis of randomized controlled trials. Med Decis Making 2013;33:607–17.
[32] Woods BS, Hawkins N, Scott DA. Network meta-analysis on the log- hazard scale, combining count and hazard ratio statistics account- ing for multi-arm trials: a tutorial. BMC Med Res Methodol 2010;10:54.
[33] Mori K, Mostafaei H, Pradere B, et al. Apalutamide, enzalutamide, and darolutamide for non-metastatic castration-resistant prostate cancer: a systematic review and network meta-analysis. Int J Clin Oncol 2020;25:1892–900.
[34] Rücker G, Schwarzer G. Ranking treatments in frequentist network meta-analysis works without resampling methods. BMC Med Res Methodol 2015;15:58.
[35] Salanti G, Ades AE, Ioannidis JPA. Graphical methods and numerical summaries for presenting results from multiple-treatment meta- analysis: an overview and tutorial. J Clin Epidemiol 2011;64:163–71.
[36] McGrath S, Zhao X, Qin ZZ, Steele R, Benedetti A. One-sample aggregate data meta-analysis of medians. Stat Med 2019;38:969– 84.
[37] Fizazi K, Tran N, Fein L, et al. Abiraterone plus prednisone in metastatic, castration-sensitive prostate cancer. N Engl J Med 2017;377:352–60.
[38] Sweeney CJ, Chen YH, Carducci M, et al. Chemohormonal therapy in metastatic hormone-sensitive prostate cancer. N Engl J Med 2015;373:737–46.
[39] Francini E, Gray KP, Xie W, et al. Time of metastatic disease presen- tation and volume of disease are prognostic for metastatic hormone sensitive prostate cancer (mHSPC). Prostate 2018;78:889–95.
[40] Sydes MR, Spears MR, Mason MD, et al. Adding abiraterone or docetaxel to long-term hormone therapy for prostate cancer: directly randomised data from the STAMPEDE multi-arm, multi- stage platform protocol. Ann Oncol 2018;29:1235–48.
[41] Merseburger AS, Scher HI, Bellmunt J, et al. Enzalutamide in Euro- pean and North American men participating in the AFFIRM trial. BJU Int 2015;115:41–9.
[42] Gallina A, Chun FKH, Suardi N, et al. Comparison of stage migration patterns between Europe and the USA: an analysis of 11 350 men treated with radical prostatectomy for prostate cancer. BJU Int 2008;101:1513–8.
[43] Karakiewicz PI, Zaffuto E, Kapoor A, et al. Kidney Cancer Research Network of Canada consensus statement on the role of adjuvant therapy after nephrectomy for high-risk, non-metastatic renal cell carcinoma: a comprehensive analysis of the literature and meta- analysis of randomized controlled trials. Can Urol Assoc J 2018;12:173–80.
[44] Bex A, Albiges L, Ljungberg B, et al. Updated European Association of Urology guidelines regarding adjuvant therapy for renal cell carci- noma. Eur Urol 2017;71:719–22.RP56976