1 Introduction

Long-term hormone therapy has been the standard of care (SOC) for metastatic hormone-sensitive prostate cancer (mHSPC). However, the program of therapy for mHSPC has changed a lot due to the advent of combination therapy.

Metastatic hormone-sensitive prostate cancer is a heterogeneous disease state, considering burden of disease and site of metastases. [1] High-volume disease had either four or more bone metastases including one or more outside the vertebral body or pelvis, or any visceral metastases, or both. Visceral metastasis, an essential component of HVD, is considered to have the worst prognosis compared to lymph node metastasis and bone metastasis. [2] Thus, we believe patients with mHSPC should be treated personally.

Several randomized controlled trials (RCTs) have demonstrated a significant survival benefit of novel treatment regimens compared with androgen deprivation therapy (ADT) for mHSPC, especially in high-volume disease. These include Enzalutamide (Enza) [3,4,5,6], Apalutamide (Apa) [7, 8], Darolutamide (Daro) [9, 10], Rezvilutamide (Rezvi) [11], Docetaxel (Doc) [12,13,14,15], Abiraterone (Abi) [16,17,18,19], and combinations of them. Thus, we searched all the literatures that provided survival outcomes in mHSPC with HVD, extracted their latest data, and performed a network meta-analysis. In addition, a subset of studies provided separate data on survival outcomes for the visceral metastasis subgroup. Therefore, the relevant meta-analysis is also presented in this article.

2 Methods

This study was registered in the International Prospective Register of Systematic Reviews (PROSPERO; CRD42023416334). We reported this Systematic Review and Meta-Analysis in accordance with the Preferred Reporting Items for Systematic Reviews and Meta analyses reporting guidelines and extension statement for systematic reviews [20, 21].

2.1 Literature search

We searched PubMed, Embase, and Cochrane CENTRAL databases to retrieve published studies of mHSPC. MeSH terms and free words related to prostate cancer, metastasis, Hormone-Sensitive, treatment, and randomized trials were used in this procedure. The detailed strategy is presented in Supplementary Table 1. Both data extraction and quality assessment were achieved by two independent reviewers, and all disagreements were resolved by advice from a senior reviewer.

2.2 Inclusion and exclusion criteria

Inclusion: 1. Patients with metastatic hormone-sensitive prostate cancer; 2. Studies comparing different therapies and reported survival outcomes of HVD or visceral metastasis. 3. Phase III trial.

Exclusion:1. Studies not in English; 2. Review or case report; 3. Patients with metastatic castrate-resistant prostate cancer or local prostate cancer.

The inclusion of literature was done independently by two researchers, and inconsistent results were decided after a consultation process.

2.3 Data extraction

The following information was independently extracted from the included studies by two researchers: trial name, main author’s name, year of publication, number of patients, treatments, control, the primary endpoint, median follow-up time, and outcome definitions. Furthermore, hazard ratios (HR) and 95% credible intervals (CIs) associated with the primary endpoints (OS and PFS) were also collected. Specific results are presented in Table 1. To evaluate the adverse events, the number of positives and the sample size were entered into the Table 3.

Table 1 Study demographics and oncologic outcomes of included trials
Full size table

2.4 Quality assessment

Two reviewers assessed and graded the risk of bias of each included study independently using the Cochrane Collaboration tool by Review Manager. [22,23,24] The result is presented in Fig. 2.

2.5 Data analysis

For meta-analysis, forest plots with HRs and 95% CIs were conducted to analyze the efficacy between systemic therapy and survival outcomes. We also calculated interaction HRs for OS and PFS between mHSPC patients with HVD and those with visceral metastasis and then pooled them. To assess heterogeneity between studies, the I2 was used, with considerable heterogeneity defined as I2 > 50% or p < 0. 1. Both common effect model and random effect model were used.

Bayesian network meta-analysis using fixed-effect model was performed by the Gemtc package in the R program (R 4.2.3). Network graphs were utilized to illustrate the connectivity of the treatment networks. (Supplementary Fig. 1) For studies with multiple publications, only the most up-to-date or complete data for each outcome were utilized. For time-to-event survival outcomes (OS and PFS), the hazard ratio (HR) with its confidence intervals (CI) of each trial was converted to logHR and standard error. Adverse events were analysed based on responders and sample size.

2.6 Sensitivity analysis

Of the 11 studies we included, the control groups in the CHART and ENZAMET studies used first-generation androgen receptor inhibitors (FARI) such as bicalutamide, flutamide throughout the experimental period. In contrast, the control groups in the other studies applied single ADT with luteinizing hormone-releasing hormone (LHRH) agonists or antagonists, and FARI, if used at all, was used only at the beginning to seal the lesion receptors. Although in a portion of the network meta-analysis, FARI + ADT was also attributed to the island of ADT [25]. We believe that this decision may be responsible for differences in the transmission of efficacy in the relevant experimental groups during the calculation process, and may even lead to changes in the order of ranking. We therefore performed another version of the network meta-analysis with the control groups from the ENZAMET and CHART studies as FARI + ADT and placed the final results in the Supplementary Materials.

3 Results

3.1 Study Selection and characteristics

We initially found 805 records through Pubmed, Embase, and Cochrane CENTRAL databases, and 587 publications were retained after removing duplicates. Based on the inclusion criteria, 17 publications from 11 RCTs were identified. The entire process of study selection is shown in the PRISMA flowchart (Fig. 1). Table 1 summarized the demographics of each included study.

Fig. 1
figure 1

PRISMA flowchart summarized the process of study selection

Full size image

We need to note that overall survival (OS) was measured as the interval from randomization to death but progression-free survival (PFS) we used in statistical analyses has been defined differently in studies. ENZAMET provided the outcome of clinical progression-free survival (cPFS), while others were radiographic progression-free survival (rPFS). Notably, a subset of patients was allowed docetaxel in ENZAMET, ARCHES and TITAN, which may have enhanced OS and PFS outcomes. Besides, ARASENS didn't provide any data of PFS, so we used the survival outcomes of “Time to CRPC” instead.

3.2 Quality assessment of the study

Figure 2 showed the risk of bias s assessment results. Although the included studies were meticulously designed randomized controlled trials, a subset of trials were open-label trials such as CHART, CHAARTED, ENZAMET, GETUG-AFU 15, PEACE-1, STAMPEDE arm C, and STAMPEDE arm G, and these are considered to take a potentially high risk of blindness to participants and investigators, which may bias to results.

Fig. 2
figure 2

Assessment of risk of bias of included studies. A Risk of bias summary for each study assessed according to the methods recommended by the Cochrane Collaboration; B Risk of bias graph about each risk of bias item

Full size image

3.3 Efficacy in the high-volume disease

Eleven RCTs included in this review provided the survival outcomes in patients with HVD. The results of meta-analysis showed a longer OS with the addition of NHA (pooled HR: 0.66; 95% CI: 0.60–0.72) and Doc (pooled HR: 0.73; 95% CI: 0.63–0.84) (Fig. 3AB). Furthermore, adding NHA to Doc + ADT significantly improved therapeutic efficacy (pooled HR: 0.70; 95% CI: 0.60–0.81). (Fig. 3C). In terms of PFS, both NHA-based doublet therapy and Doc-based therapy significantly reduced the risk of disease progression. The pooled HRs were 0.47 (95% CI: 0.43–0.52) and 0.60 (95% CI: 0.52–0.70) (Fig. 4AB). Triple therapy also achieved better results (pooled HR: 0.43; 95% CI: 0.37–0.50) compared to doublet therapy (Fig. 4C). Beyond that, the results in this section suggest no significant heterogeneity across studies.

Fig. 3
figure 3

OS for systemic combination therapy for patients with HVD. A OS for the combination therapy of NHA + ADT; B OS for the combination therapy of Doc + ADT; C OS for the triplet therapy

Full size image
Fig. 4
figure 4

PFS for systemic combination therapy for patients with HVD. A PFS for the combination therapy of NHA + ADT; B PFS for the combination therapy of Doc + ADT; C PFS for the triplet therapy

Full size image

In network meta-analysis, all tested treatment combinations resulted in longer OS and PFS relative to ADT alone. Among the OS results, two trails involving triplet therapy were pooled, including ADT plus docetaxel plus darolutamide and abiraterone (HR: 0.50, 95%CI: 0.39–0.63; HR: 0.52, 95%CI: 0.38–0.72). Also, six trials were pooled in OS and rPFS analyses of doublet therapy with ADT plus NHA, including abiraterone, enzalutamide, apalutamide, and rezvilutamide (LATITUDE, STAMPEDE arm G, TITAN, CHART, ARCHES and ENZAMET trials). There were three trials involving ADT plus docetaxel (GETUG-AFU 15, STAMPEDE arm C and CHAARTED trials).The HRs for OS were 0.58 (95% CI: 0.44–0.77), 0.61 (95% CI: 0.53–0.71), 0.70 (95% CI: 0.56–0.88), 0.72 (95% CI: 0.62–0.84) and 0.72 (95% CI: 0.62–0.85) for Rezvi + ADT; Abi + ADT, Apa + ADT; Doc + ADT and Enza + ADT, respectively (Fig. 5A). The HRs for PFS were 0.25 (95% CI: 0.19–0.31), 0.28 (95% CI: 0.21–0.38), 0.44 (95% CI: 0.33–0.58), 0.46 (95% CI: 0.40–0.52), 0.48 (95% CI: 0.41–0.56), 0.52 (95% CI: 0.41–0.67) and 0.60 (95% CI: 0.52–0.70) for Daro + Doc + ADT; Abi + Doc + ADT; Rezvi + ADT; Abi + ADT, Enza + ADT; Apa + ADT and Doc + ADT, respectively (Figs. 5B). The ranking of all treatment programs is shown in Table 2.

Fig. 5
figure 5

Network meta-analysis for patients with HVD. A Impact of combination therapy on OS outcomes in patients with HVD. B Impact of combination therapy on PFS outcomes in patients with HVD

Full size image
Table 2 Treatment rankings based on surface under the cumulative ranking analysis
Full size table

3.4 Efficacy in the visceral metastasis

In the subgroup of visceral metastasis patients, 7 studies published the results of OS and 4 studies published the outcome of PFS. The NHA-based therapy reduced the risk of overall mortality (pooled HR: 0.80; 95% CI: 0.66–0.99) and tumor progression (pooled HR: 0.58; 95% CI: 0.46–0.72) (Figs. 6, 7). As for Doc plus ADT, Only the CHAARTED trial provided the data on OS in mHSPC patients by presence of visceral metastasis. The HR was 0.53 (95% CI: 0.30–0.92). From ARASENS trial, there were no significant differences between patients with triplet therapy and those with Doc + ADT for OS (HR: 0.79; 95% CI: 0.55–1.14).

Fig. 6
figure 6

OS for NHA-based doublet therapy for patients with visceral metastasis

Full size image
Fig. 7
figure 7

PFS for NHA-based doublet therapy for patients with visceral metastasis

Full size image

Triple therapy (Daro + Doc + ADT) still ranked highest for improvement in OS outcomes (HR: 0.41; 95% CI: 0.21–0.81), according to network meta-analysis. In contrast to the previous results in HVD, Docetaxel plus ADT ranked highest among all the doublet therapies in OS (HR: 0.52; 95% CI: 0.30–0.92). From the data available, Abi + ADT is the only treatment regimen among all doublet therapies that showed improvement in both OS (HR: 0.59; 95% CI: 0.41–0.83) and PFS (HR: 0.53; 95% CI: 0.37–0.76) (Fig. 8). The HRs calculated from the results of the current phase of the RCTs, the efficacy of the other doublet therapy was not statistically significant. Also, it is important to note that ADT alone ranked even higher than the addition of Enza and Rezvi in terms of prolonging overall survival time in prostate cancer patients.

Fig. 8
figure 8

Network meta-analysis for patients with visceral metastasis. A Impact of combination therapy on OS outcomes in patients with visceral metastasis. B Impact of combination therapy on PFS outcomes in patients with visceral metastasis

Full size image

3.5 Adverse events

We also recorded data related to adverse events in each study and summarised them in Table 3. Unfortunately, with the exception of ARASENS and CHART, none of the trials provided data on adverse reactions in the HVD group alone. The CHART study included patients with high-volume disease, so the results are all about that population. In the ARASENS trial, no significant difference was seen in the rate of adverse events in the HVD group compared to the overall patient population.

Table 3 Results of adverse events of included trials
Full size table

We then performed a Bayesian network meta-analysis of grade ≥ 3 adverse events. (Fig. 9) With the forest plot we can see that all combination therapies, except Apa, lead to more adverse events. Triplet therapy has the highest risk ranking, which is also in line with our conventional knowledge. Docetaxel, although ranked highly, was not statistically significant compared to drugs acting on the androgen pathway outside of Apa.

Fig. 9
figure 9

Network meta-analysis for patients with grade ≥ 3 adverse events

Full size image

3.6 Sensitivity analysis

In addition to the above results, we extracted FARI + ADT as a separate group and reported it in the Supplementary Material as a sensitivity analysis. This transform had a significant impact on rezvilutamide's ranking since there was only one study using Rezvi, the CHART study, and this study's control group was Bica + ADT. In this analysis, Rezvi + ADT ranked first among all doublet regimens just behind triplet therapy. (HR for OS: 0.49; 95% CI: 0.32–0.75; HR for PFS: 0.38 95% CI: 0.24–0.58) (Supplementary Fig. 2). Since androgen receptor antagonists were also used in the control group in the ENZAMET trial, Enza + ADT’s ranking was also improved. (Supplementary Table 2).

4 Discussion

We conducted a systemic review and network meta-analysis and compared the clinical efficacy of different combination therapies in patients with high-volume disease and visceral metastasis based on several well-designed RCTs. In patients with HVD, all combination regimens can reduce the risk of death and tumor progression. According to the network meta-analyses, all triplet therapies ranked higher than other doublet therapies and Rezvilutamide, a new androgen receptor inhibitor, plus ADT ranked highest among all doublet therapy. The combination of novel hormonal agents and chemotherapy inhibits prostate cancer progression through two different mechanisms. Although the addition of docetaxel can benefit patients, it still ranked relatively low compared to other therapies. However, in the network meta-analysis, docetaxel's importance as a pivot for comparing doublet therapy to triplet therapy cannot be overstated, which prevented Daro + Doc + ADT from becoming an island. In addition, we can conclude from the sensitivity analysis that settings different groups can lead to differences in the NMS results and even changes the final rank.

It's worth noting that the ranking changed in patients with visceral metastases compared to HVD. Docetaxel and abiraterone most improved Clinical Efficacy in OS and PFS, respectively. Several researches have indicated that the treatment effect of Docetaxel for patients with mHSPC may vary by volume of disease. [26] Visceral metastases considered a risk factor for poor prognosis in prostate cancer patients. [27] This fact may be caused by more androgen receptor–independent cells [28], which leads to decline in the effectiveness of hormone therapy, especially in patients with visceral metastasis. Thus, even though many network meta-analyses have ranked combination regimens for patients with high-volume disease, patients with visceral metastases may not fit into this ranking. Classification of mHSPC into high and low risk is another important categorization method (According to LATTITUDE standard), which used Gleason score as an important criterion. In practice, the Gleason score is also an important reference index to guide drug therapy, especially for prostate cancer with high Gleason score, chemotherapy may be more effective than endocrine therapy. Although both HVD and high Gleason score can respond to tumors with high aggressiveness, this is not taken into account by differentiating drug therapy according to tumor burden.

From the OS and PFS results, there is no statistical evidence that the addition of Apalutamide and Rezvilutamide have an improved prognosis for patients with visceral metastases. And triplet therapy still ranked first for OS improvement, with no clear benefit from Enzalutamide plus ADT. Therefore, we recommend the preferential application of triplet therapy for patients with HVD, as well as the addition of chemotherapy and novel hormone therapy to ADT, provided that adverse effects can be tolerated. In patients who have developed visceral metastases, we suggested prioritizing the application of docetaxel as the first choice of a doublet regimen. If the patient does not have visceral metastases, even if he is already at high tumor burden, we still believe that the concomitant use of NHAs should be considerded before chemotherapy.

In addition to the above results, there are some limitations in this study. Firstly, the data we extracted on adverse events were largely unrestricted to high-volume disease, and data from this subgroup have rarely been reported separately. This can lead to bias. However, we conducted this study primarily to assess differences in efficacy between treatment regimens, and adverse events were only an ancillary part of this. Secondly, most trials do not report detailed information on visceral metastases, and there may also be differences in efficacy for metastases in different organs. More detailed pooled analyses of lesion location may be of further value in guidance. Thirdly, the use of nonsteroidal antiandrogen such as bicalutamide, although minimal in terms of prognostic change compared to ADT treatment alone, is still a factor that can potentially lead to erroneous conclusions in the NMA. In particular, in ENZAMET, the control group involved the use of multiple related medications, and it was difficult to eliminate the effect of this factor in NMA even with the control group in CHART(Bica + ADT). Finally, although NMA can indirectly compare the therapeutic efficacy of two treatment regimens, large-sample, multicenter, multisubgroup randomized controlled trials are still irreplaceable.

5 Conclusion

We performed a meta-analysis by extracting the most recent data from each RCT and gave a ranking of each treatment regimen among HVD and VM patients. For HVD and VM, a population with high tumor burden, triple therapy is more advantageous. In patients with visceral metastasis, chemotherapy needs to be prioritized compared to endocrine therapy.