FormalPara Key Summary Points

Why carry out this study?

 Although denosumab treatment reduces skeletal-related event (SRE) risk and delays the time to first SRE in patients with solid metastatic tumors, the benefits of denosumab to bone health have not been established in patients with early-stage breast cancer at high risk of disease recurrence.

 Besides evaluating bone metastasis-free survival (BMFS), we further investigated the potential benefits of denosumab treatment by assessing the prespecified bone endpoints and the associated subgroup analysis in patients with high-risk early breast cancer receiving denosumab versus placebo.

What was learned from the study?

 The hazard ratios of the prespecified exploratory bone endpoints mostly favored denosumab over placebo, suggesting a clinically meaningful effect of denosumab in these patients.

 This study demonstrates that denosumab improves bone-related outcomes for women with high-risk early breast cancer.

Introduction

Female breast cancer is the most common form of cancer and the fourth leading cause of cancer deaths in the USA [1]. Distant relapse occurs most frequently in the bone, with 65–80% of patients with metastatic breast cancer developing bone metastases during the course of their disease [2, 3]. Metastatic bone disease is associated with pathologic fractures, bone pain, spinal cord compression, and hypercalcemia of malignancy [2].

Denosumab is a fully human monoclonal antibody that binds to and inhibits the receptor activator of nuclear factor-kappa B ligand [4]. Treatment with denosumab reduces the incidence of skeletal-related events (SREs) and delays the time to first SRE in patients with solid metastatic tumors, including breast cancer [5,6,7]. Denosumab was superior to zoledronic acid in delaying or reducing SREs in patients with breast cancer that had metastasized to the bone [5].

D-CARE (ClinicalTrials.gov, NCT01077154) is an international, phase 3, randomized, double-blind, placebo-controlled study in women with early-stage breast cancer at high risk of disease recurrence to assess whether denosumab with standard-of-care adjuvant/neoadjuvant chemotherapy would increase bone metastasis-free survival (BMFS) and disease-free survival. The study did not reach its primary endpoint because adjuvant denosumab did not significantly improve BMFS versus placebo [8]. Exploratory bone endpoints were assessed in D-CARE to determine the bone protective effects of denosumab in this setting, specifically pertaining to metastasis to bone, fractures, and on-study SREs occurring after bone recurrence. Here, we summarize the bone health outcomes of adjuvant denosumab in women with early breast cancer in the D-CARE study.

Methods

Details of the D-CARE clinical study have been previously published [8]. Briefly, eligible women (≥ 18 years old) with histologically confirmed stage II/III or locally advanced disease were randomized to receive subcutaneous denosumab (120 mg) or matching placebo every 4 weeks with adjuvant/neoadjuvant systemic therapy for approximately 6 months followed by denosumab or placebo every 3 months for 4.5 years (total treatment duration, 5 years). Patients were not required to undergo bone mineral density assessment and were not excluded on the basis of bone mineral density results. Written informed consent was provided by all patients before participation, and the study was approved by each study site’s institutional review board and independent ethics committee. Participating investigators and institutions and ethics committees were in compliance with all applicable regulatory requirements and International Conference on Harmonisation and Good Clinical Practice (ICH GCP) guidelines.

Menopause status [i.e., postmenopausal (yes/no)] of patients was assessed by investigators at screening. Postmenopausal status was defined in the study as follows: having undergone bilateral oophorectomy, age ≥ 60 years, age 45–59 years with amenorrhea lasting more than 12 months plus an intact uterus and at least one intact ovary or with amenorrhea lasting no more than 12 months and follicle-stimulating hormone and estradiol in the postmenopausal range, and patients who had received adjuvant or neoadjuvant chemotherapy must have met at least one of the aforementioned criteria for postmenopausal status prior to that chemotherapy. Patients underwent mandatory imaging (mammography, full-body radioisotope bone scans, and computed tomography/magnetic resonance imaging of the chest, abdomen, and all other known or suspected sites of disease) at screening, at the end of year 1, and annually thereafter. Images and scans were assessed by blinded independent central review (BICR) and local investigators. An SRE was defined as any fracture (vertebral or nonvertebral), radiation therapy to bone (including the use of radioisotopes), surgery to bone, or spinal cord compression following the development of bone metastasis. Prespecified exploratory bone endpoints (defined in Table 1) included time to first bone metastasis, time to bone metastasis as site of first recurrence, time to first symptomatic bone metastasis, time to first on-study fracture before bone metastasis, and time to first on-study SRE following bone metastasis.

Table 1 Summary of prespecified exploratory bone endpoints
Full size table

Statistical Analyses

The full analysis set, which included all patients who were randomized to the study, was used for all efficacy endpoints, including exploratory bone endpoints. For time to bone metastasis as site of first recurrence and time to first symptomatic bone metastasis, a cumulative incidence function (CIF) [9] was estimated with non-bone site(s) of first recurrence and first asymptomatic bone metastasis, respectively, as competing events. Comparison of the CIF between denosumab and placebo was based on Gray [10] and Zhou et al. [11] and stratified by the randomization stratification factors. For other time to event endpoints, the Kaplan-Meier method was used to estimate the survival function for each treatment group. The hazard ratio (HR) of denosumab versus placebo and the corresponding two-sided 95% confidence interval (CI) were estimated using a Cox proportional hazards model stratified by the randomization stratification factors. A log-rank test, stratified by the randomization stratification factors, was performed to compare the treatment groups. Post hoc subgroup analyses for the exploratory bone endpoints were conducted by age group (< 50 years vs ≥ 50 years) and menopause status (premenopausal vs postmenopausal). Post hoc sensitivity analyses were conducted for the exploratory bone endpoints. In the sensitivity analysis for each exploratory endpoint, patients who received on-study bisphosphonate therapy prior to their event/censoring time of the original analysis were censored at the start date of the bisphosphonate use.

Results

Detailed study outcomes have been previously reported [8]. Overall, 4509 patients were randomized to receive either denosumab or placebo and were included in the primary analysis (Table 2). The median age was 51 years [569 (13%) aged 18–39, 1491 (33%) aged 40–49, and 2449 (54%) aged ≥ 50], 2360 (52%) patients were premenopausal, and 3492 (77%) patients had estrogen receptor-positive and/or progesterone receptor-positive status. Neoadjuvant and adjuvant therapies were received by a total of 1091 (24%) and 3418 (76%) patients, respectively. Of the 3424 patients with estrogen receptor-positive breast cancer, 1961 (57%) received tamoxifen, 1898 (55%) received aromatase inhibitors, and 392 (11%) underwent ovarian ablation as the prior and/or on-study hormonal therapy (excluding the usage after disease recurrence). Overall, 27/2256 (1%) patients in the denosumab arm and 70/2253 (3%) in the placebo arm used on-study bisphosphonate before bone metastasis or death. For patients with confirmed bone metastasis, 2/155 (1%) patients in the denosumab arm and 6/189 (3%) in the placebo arm used on-study bisphosphonate prior to the event. The median (range) time on study before the primary analysis cutoff was 67 (0–87) months in the denosumab arm and 67 (0–86) months in the placebo arm.

Table 2 Baseline demographics and clinical characteristics
Full size table

For the exploratory bone endpoints (Table 3), the time to first bone metastasis was similar for denosumab and placebo (HR 0.82; 95% CI 0.66–1.02; p = 0.068; Fig. 1a). Improvements in bone metastasis-related exploratory endpoints were seen with denosumab versus placebo for time to bone metastasis as site of first recurrence (non-bone site of first recurrence as a competing risk [HR 0.76; 95% CI 0.59–0.97; p = 0.031]; Fig. 1b) and time to first symptomatic bone metastasis (asymptomatic bone metastasis as a competing risk [HR 0.60; 95% CI 0.39–0.93; p = 0.024]; Fig. 1c). Subgroup analysis showed improvements in the bone metastasis-related exploratory endpoints for denosumab versus placebo in women < 50 years old and in premenopausal women, but not in women ≥ 50 years old or in postmenopausal women (Table 3).

Table 3 Exploratory bone endpoints
Full size table
Fig. 1
figure 1

Time to a first bone metastasis, b bone metastasis as site of first recurrence with non-bone site of first recurrence as a competing risk, c first symptomatic bone metastasis with asymptomatic bone metastasis as a competing risk, d first on-study fracture (before bone metastasis), and e first on-study SRE (following bone metastasis). CI confidence interval, HR hazard ratio, SRE skeletal-related event. *Based on subdistribution hazards (Gray [10] and Zhou et al. [11]) stratified by the randomization stratification factors

Full size image

Skeletal morbidity was reduced for time to first on-study fracture (before bone metastasis [HR 0.76; 95% CI 0.63–0.92; p = 0.004]; Fig. 1d) and time to first on-study SRE (after disease recurrence in bone [HR 0.52; 95% CI 0.35–0.78; p = 0.001]; Fig. 1e). Subgroup analysis (Table 3) revealed improvements in time to first on-study fracture for denosumab versus placebo in women ≥ 50 years old and in both premenopausal and postmenopausal women, with similar relative risk reduction. Improvements in time to first on-study SRE after disease recurrence in bone for denosumab versus placebo were seen in women < 50 years old and in premenopausal women, but not in women ≥ 50 years old or in postmenopausal women.

The sensitivity analysis in which patients who received on-study bisphosphonate therapy prior to their event/censoring time of the original analysis were censored at the start date of the bisphosphonate use and showed similar results to the original analysis, further confirming the observed results (Table 4).

Table 4 Exploratory bone endpoints: sensitivity analysis—censoring at on-study bisphosphonate use
Full size table

Discussion

No statistically significant improvements in the primary endpoint of BMFS were observed with denosumab treatment versus placebo overall or in any subgroup in the D-CARE study. However, BMFS is a composite endpoint that includes relapse in bone with and without recurrence at other anatomic sites and death from any cause. Approximately 40% of the patients died before experiencing bone metastasis. Thus, the composite BMFS endpoint may have limited the ability to identify a clinically meaningful effect of denosumab. However, for the exploratory bone endpoints, denosumab treatment was associated with a delay to bone metastasis as the site of first recurrence and first symptomatic bone metastasis, suggesting beneficial biological effects on the bone microenvironment. Additionally, bone events reflecting the expected structural benefits of denosumab on bone, such as time to first on-study SRE and time to first on-study fracture (before bone metastasis), were prolonged by denosumab when compared with placebo, and confirmed the expected bone health benefits of denosumab. As reported previously, our results indicate that denosumab, which functions within the bone microenvironment, was unable to modify the metastatic process if recurrences first occurred outside bone or in extraosseous sites, subsequently leading to bone involvement.

Subgroup analyses suggested improvements in time to first on-study fracture (before bone metastasis) in women who were premenopausal at study entry that were of similar magnitude to the expected reduction in fractures affecting postmenopausal women. As the risk of fragility fractures in premenopausal women is low, this appears somewhat counterintuitive. However, it may reflect the very rapid bone loss known to be associated with chemotherapy-induced menopause that would have occurred in many of the premenopausal women aged 40–49 years [12]. Unfortunately, information on chemotherapy-induced menopause was not collected in this trial; thus, this potential explanation cannot be explored further.

The benefit of exploratory bone endpoints is similar to that observed in other denosumab studies in breast cancer. In the phase 3 ABCSG-18 trial (NCT00556374) [7], 3425 postmenopausal women with early hormone receptor-positive breast cancer (breast cancer with lower relapse risk than in D-CARE) received twice-yearly denosumab (60 mg) or placebo with aromatase inhibitors. Denosumab significantly delayed the time to first clinical fracture (HR 0.50; 95% CI 0.39–0.65; p < 0.001) and reduced the incidence of new vertebral fractures (odds ratio, 0.53; 95% CI 0.33–0.85; p = 0.009) versus placebo [7]. In the phase 3 AZURE trial (NCT00072020) [13] in 3360 women with stage II or III breast cancer (45% premenopausal), treatment with zoledronic acid (4 mg every 3–4 weeks for 6 doses, then every 3 months for 8 doses, followed by every 6 months for 5 doses) reduced the risk of fracture at 5 years (3.8% for zoledronic acid vs 5.9% for standard adjuvant systemic treatment; HR 0.69; 95% CI 0.53–0.90; p = 0.005), although most of the benefit was associated with fewer fractures after disease recurrence in bone rather than effects on the normal skeleton [13]. In AZURE, treatment with zoledronic acid also significantly reduced the incidence of bone metastases as the first recurrence (adjusted HR 0.78; 95% CI 0.63–0.96; p = 0.020) or at any time (adjusted HR 0.81; 95% CI 0.68–0.97; p = 0.022).

The benefit of denosumab for exploratory bone endpoints observed in breast, prostate, and other cancer studies may be explained by modification of the disease process in bone and suppression of bone resorption and osteoclastic activity. In a phase 2 trial [14], 111 patients with solid tumors (e.g., prostate, breast) or multiple myeloma were treated with a bisphosphonate every 4 weeks or denosumab (180 mg) every 4 or 12 weeks. The percentage of patients experiencing a first on-study SRE by week 25 was 8% (6/75) in the denosumab arm and 17% (6/35) in the zoledronic acid arm (odds ratio, 0.31; 95% CI 0.08–1.18). Denosumab treatment resulted in greater suppression of bone turnover markers and an osteoclast biomarker (TRAP-5b) than bisphosphonate treatment.

In a phase 3 trial [5], 2049 patients with breast adenocarcinoma that had metastasized to the bone were randomized to receive either denosumab 120 mg or zoledronic acid 4 mg every 4 weeks. Denosumab treatment was significantly more effective at delaying or reducing SREs and resulted in greater suppression of bone turnover markers compared with zoledronic acid. Specifically, denosumab versus zoledronic acid reduced the risk of developing multiple SREs (time to first and subsequent SREs) by 23% (rate ratio, 0.77; 95% CI 0.66–0.89; p = 0.001) and reduced the mean skeletal morbidity rate (i.e., the ratio of SREs per patient divided by the patient’s time at risk) by 22%. By week 14, denosumab treatment resulted in an 80% median decrease in urinary N-telopeptide for creatinine levels versus 68% with zoledronic acid (p < 0.001). Denosumab treatment also resulted in a 44% median decrease in bone-specific alkaline phosphatase levels versus 37% with zoledronic acid (p < 0.001). The authors hypothesized that the greater inhibition of osteoclast-induced bone resorption by denosumab, as suggested by the increased suppression of bone turnover markers, results in improved treatment outcomes for denosumab relative to zoledronic acid.

Limitations

Several limitations are present in the study. Our study results may have been limited by the high proportion of patients who withdrew consent (more than 10% of randomized patients), which possibly introduced bias in the clinical outcomes. Moreover, the favorable bone metastasis as site of first recurrence exploratory endpoint may have been diluted by the deaths that occurred before bone metastasis and/or preferential spread of tumor cells to other metastatic sites. Inclusion of patients with variable bone health and treatments may also have influenced our findings.

Conclusions

Overall, the benefits demonstrated by these exploratory D-CARE bone endpoints confirmed the expected bone health aspects of denosumab and probable effects on bone metastasis as a first site of metastasis. The lower rates of fractures and SREs with denosumab treatment likely reflect the structural benefits of suppression of osteoclast activity and change in the balance of bone remodeling in favor of bone formation.