Impact of Extraskeletal Metastases on Skeletal-Related Events in Metastatic Castration-Resistant Prostate Cancer with Bone Metastases

The therapeutic landscape of metastatic castration-resistant prostate cancer (mCRPC) has substantially evolved over the last decade. Nonetheless, a better understanding of bone-targeted agents (BTAs) action in mCRPC remains an unmet need. Theuse of BTAs aims to reduce the incidence of skeletal-related events (SREs) in patients with mCRPC. Less frequent BTA schedules are currently being studied to minimize adverse events. In this study, the impact of metastatic compartment (bone and extraskeletal metastases (BESM) vs. bone-only metastases (BOM)) on bone biomarker kinetics, time to first on-study SRE, and symptomatic skeletal events (SSEs) is evaluated. This is a retrospective analysis of the prospective, randomized, multicenter clinical trial of denosumab vs. zoledronic acid in patients with mCRPC and bone metastases. A total of 1901 patients were included, 1559 (82.0%) with BOM and 342 with BESM (18.0%). Bone metastases burden was balanced between groups. Baseline levels and normalization rates of corrected urinary N-terminal telopeptide and bone alkaline phosphatase did not differ between groups. However, BESM patients had a higher risk of SREs (adjusted HR 1.21; 95% CI 1.01–1.46; p = 0.043) and SSEs (adjusted HR 1.30; 95% CI 1.06–1.61; p = 0.014). This difference was more pronounced in the first 12 months of BTA treatment.In mCRPC, strategies of BTA schedule de-escalation may take into account presence of extraskeletal metastases.


Introduction
Prostate carcinoma is one of the most common cancers in men globally and the second cause of cancer death in this gender in Europe [1,2]. In these patients, bone is the most common site of metastatic disease (>90% of patients) [3] and bone metastases are frequently associated with detrimental bone outcomes-collectively referred to as skeletal-related events (SREs; pathological fracture, spinal cord compression, and radiotherapy or surgery to bone), which negatively impact quality of life and

Skeletal-Related Events
After a median follow-up of 20.1 months (IQR 15.9-23.8; balanced between arms), 27.9% (n = 530) of patients developed an on-study SSE and 38.2% (n = 727) an on-study SRE in the overall cohort ( Table 2). The proportion of patients developing SREs was similar between groups (42.1% in BESM vs. 37.4% in BOM group). No substantial differences were found regarding SRE pattern, such as number of SRE and type of SRE. (Table 2).

Discussion
The treatment of mCRPC has substantially evolved over the last decade, namely with the introduction of new therapeutic options. However, bone remains a major metastatic site in mCRPC, with important morbidity, quality of life, and survival implications. Assessment of a high-quality cohort derived from one of the largest clinical trials of mCRPC patients treated with BTAs is a major opportunity to improve knowledge and optimize BTA use in mCRPC.
Visceral metastases have long been considered a negative prognostic factor for survival in mCRPC [7,12,21]. In this large and high-quality cohort of patients with mCRPC and bone metastases receiving BTAs, patients with bone plus extraskeletal metastases had a similar baseline and 3-month normalization patterns of bone remodeling markers as patients with bone-only metastases, but a higher risk of SREs, SSEs, and death irrespective of bone disease volume and disease aggressiveness features. This difference was more evident in the first year after diagnosis of mCRPC with bone metastases.
Bone metastatic niche is a rich source of growth factors and other soluble molecules that are released by bone metabolism activation from mCRPC bone metastases [12,20,22]. Several of these biomarkers, such as uNTX and bALP, are amenable to serum or urine quantification and constitute strong predictors of survival and SRE risk in patients with bone metastases from several tumor types [8,12,20,22]. Indeed, a 2016 study by Lipton et al., using an integrated analysis of three similar phase III trials, showed that patients with bone turnover markers above the median after 3 months of antiresorptive therapy had significantly worse clinical outcomes, including OS and bone disease progression, compared to patients with bone turnover markers below the median [8].
In the present study, bone biomarker assessment, at baseline and 3 months after initiation of BTA therapy, was used to dissect the prognostic implications of different metastatic spread patterns (i.e., metastatic compartments) by analyzing patients with BOM and BESM. While a previous study of patients with breast cancer and bone metastases showed that BESM patients had a numerical trend for persistently higher uNTX levels and an erratic uNTX variation during ZA treatment [20], in the present cohort bone markers did not differ according to metastatic compartment both at baseline and at 3 months. While the nature of the present cohort provides a more definitive assessment on how bone biomarkers differ at baseline and vary over time, the short-term assessment (3 months) restricts the longitudinal understanding of bone marker dynamics at longer time intervals. In addition, despite the fact that all bone metastases activate bone metabolism, bone-forming vs. bone-degrading components are differently impacted in patients with bone metastases from breast and prostate cancer [23,24], a fact thatmay also contribute to explain the observed difference.
Despite similar bone biomarker levels at baseline and comparable short-term variation, patients with BESM had higher SRE and SSE risk compared with patients with BOM, and this seems to be particularly relevant within the first 12 months after diagnosis. The added SRE risk in patients with BESM was also identified by Tanaka and colleagues in a cohort of 534 breast cancer women who developed bone metastases [19]. The present study confirms that result in a large cohort of mCRPC patients, and taken together these findings challengethe current trend towards a reduced BTA scheduling frequency from every 4 weeks to every 12 weeks (especially for ZA) across all mCRPC patients [16,25,26]. Specifically, this study supports the hypothesis that the metastatic compartment has impact in bone outcomes and may be taken into account when considering de-escalation strategies to better tailor such approaches. Therefore, patients with BESM are at higher SRE risk and should receive a more conservative treatment schedule (i.e., every 4 weeks), at least during the first year of treatment. It remains unknown whether humoral factors affecting bone resorption are produced by extraskeletal metastases and/or whether cancer cells can circulate from different metastatic sites (outside of bone to bone and vice versa) to restimulate microenvironment [10,20,21]. Despite similar metastatic bone disease burden and disease biology, this cohort shows an impact of BESM on SREsthat is not captured by corrected uNTX and bALP levels. This may suggest that these biomarkers are not the best surrogates for such humoral factors. While the present study does not directly address these issues, it discloses new research avenues with potential therapeutic impact for optimizing bone outcomes in patients with mCRPC and raises relevant questions to understand cancer dynamics during advanced-stage of disease.
As anticipated, patients with BESM displayed worse survival outcomes. This is consistent with previous studies [21,27,28] and raises awareness of a patient population for whom novel approaches are urgently required.
Despite the high-quality dataand having included a large cohort with long-term follow-up, the present study has intrinsic limitations. It was a retrospective analysis, therefore subject to residual confounding, notwithstanding methodological accuracy. While all models were controlled for disease biology (Gleason score) and burden (bone metastases number), full disease biology is not captured by these features, whichmay partially explain the study findings.

Design and Study Population
This was a retrospective analysis of the prospective, randomized, multicenter registration clinical trial of denosumab vs. ZA in patients with mCRPC and bone metastases (NCT00321620) [4]. All participating patients were selected. Figure 4 details patient enrollment in the several preplanned analyses. Bone metastases diagnosis and extraskeletal metastases screening were conducted as prespecified per NCT00321620 protocol.

Study Aim and Outcomes
The main aim of this study was to investigate how metastatic compartment (BOM vs. BESM) impacts clinical outcomes and bone metabolism in patients with prostate cancer receiving BTAs. Specifically, the impact of metastatic compartment (BOM vs. BESM) on uNTX and bALP levels at baseline and normalization rates from baseline to 3 months was studies. We subsequently analyzed its impact on time to first on-study SRE/SSE and on OS. The 3-month-after-antiresorptive-treatment time point was selected to provide adequate therapy response time (no further time points were available). Outcomes were defined as per the study protocol [4].

Bone Marker Determination
In the NCT00321620 trial [4], urine specimens (from secondmorning void) and venous blood samples were collected to measure bone metabolism biochemical markers in all patients. Urinary bone marker to creatinine ratio was assessed for uNTX. Serum bone-specific alkaline phosphatase was also quantified. A central laboratory (Mayo Medical Laboratories, Rochester, MN, USA) performed urine and serum biochemical bone metabolism marker measurements for all patients. Normal uNTX cut-off was <64 nmol BCE/mmol creatinine and bALP cut-off was<22 ng/mL [13].

Statistical Analysis
Descriptive statistics of baseline clinical, pathological, and treatment characteristics were performed for the overall cohort and for BOM and BESM subgroups. Univariate comparisons were performed using a chi-square or t-test, as appropriate.

Study Aim and Outcomes
The main aim of this study was to investigate how metastatic compartment (BOM vs. BESM) impacts clinical outcomes and bone metabolism in patients with prostate cancer receiving BTAs. Specifically, the impact of metastatic compartment (BOM vs. BESM) on uNTX and bALP levels at baseline and normalization rates from baseline to 3 months was studies. We subsequently analyzed its impact on time to first on-study SRE/SSE and on OS. The 3-month-after-antiresorptive-treatment time point was selected to provide adequate therapy response time (no further time points were available). Outcomes were defined as per the study protocol [4].

Bone Marker Determination
In the NCT00321620 trial [4], urine specimens (from secondmorning void) and venous blood samples were collected to measure bone metabolism biochemical markers in all patients. Urinary bone marker to creatinine ratio was assessed for uNTX. Serum bone-specific alkaline phosphatase was also quantified. A central laboratory (Mayo Medical Laboratories, Rochester, MN, USA) performed urine and serum biochemical bone metabolism marker measurements for all patients. Normal uNTX cut-off was <64 nmol BCE/mmol creatinine and bALP cut-off was <22 ng/mL [13].

Statistical Analysis
Descriptive statistics of baseline clinical, pathological, and treatment characteristics were performed for the overall cohort and for BOM and BESM subgroups. Univariate comparisons were performed using a chi-square or t-test, as appropriate.
uNTX and bALP variation from baseline to 3 months according to metastatic compartment was tested using paired t-test. Time-to-event outcomes, survival, and cumulative incidence were plotted using the Kaplan-Meier method. Univariate and multivariate differences between survival rates according to metastatic compartment were tested using standard univariate and multivariate Cox proportional hazards model and the Andersen and Gill model for multiple failure time data. Multivariate analyses were corrected for age, ECOG-PS, Gleason score, total PSA at diagnosis, number of bone lesions, previous SREs, castration type, and previous surgery or radiotherapy. Given the strong association between visceral involvement and chemotherapy, no adjustment was performed for previous or current chemotherapy treatment.

Ethical Statement
Approval from appropriate research Ethics Committees was obtained from each study center. All patients provided written informed consent before any study-specific procedure.

Conclusions
In this large mCRPC registration and high-quality study, despite similar bone marker levels at baseline and at 3 months, patients with bone metastases plus extraskeletal metastases had higher risk of SRE and SSE than patients with BOM. Given the higher risk for adverse outcomes, strategies of BTA schedule de-escalation should consider the impact of metastatic compartment, particularly during the first treatment year.