ABSTRACT: The development of effective systemic therapies to reduce the risk of disease recurrence or metastases in early-stage breast cancer remains an important challenge. The use of bone-modifying agents (BMAs), including the bisphosphonates (BPs) and the monoclonal antibody denosumab (Xgeva), is well established for metastatic bone disease. In the adjuvant setting, some studies have shown provocative findings with some of these agents for the prevention of future breast cancer–related events, with improved survival in some subgroups. The most compelling results have been seen with clodronate and zoledronic acid(Drug information on zoledronic acid). In this review we describe the current evidence for use of BPs as part of the adjuvant treatment of patients with early-stage breast cancer.
The bisphosphonate (BP) drugs have a long established role in the management of osteoporosis, hypercalcemia of malignancy, and bone metastases from solid tumors and multiple myeloma. In the breast cancer population, zoledronic acid (ZA) and pamidronate(Drug information on pamidronate) have been widely used for over a decade in patients with bone metastases as an adjunct to radiotherapy, hormonal therapy, and chemotherapy, based on trials that have shown a reduction in skeletal-related events (SREs), including fractures, spinal cord compression, and worsening pain necessitating radiotherapy. In addition, these agents are being suggested for premenopausal patients with bone loss from premature ovarian failure due to chemotherapy and in postmenopausal patients treated with aromatase inhibitors (AIs). There is increasing interest in using these drugs to reduce the risk of breast cancer metastases, both bone and non-bone, independent of the established uses for bone-related indications; this is in light of a plausible mechanism of action based on preclinical evidence, some compelling observational data, and results from several interventional trials suggesting a role in some subpopulations of breast cancer patients. However, trial results with BPs are conflicting, and currently none of these agents have been designated by the US Food and Drug Administration (FDA) as being indicated for reducing breast cancer recurrence or improving disease-free survival (DFS) and overall survival (OS), so any use for this purpose in the US must be considered off-label and experimental.
This manuscript will review the preclinical and clinical data for the use of bone-modifying agents (BMAs) in the adjuvant treatment of early breast cancer. We will examine clinical trial results for these agents and will address issues of toxicity and cost-effectiveness. We will attempt to offer a perspective on their use in the absence of existing guidelines for the primary purpose of reducing breast cancer recurrence.
Bone-Modifying Agents and Early Breast Cancer: Preclinical Observations
In normal bone, particularly in young adults, there is a balance between osteoclastic bone resorption and osteoblastic bone formation. This can be altered by the presence of tumor cells in bone, which accelerate resorption by promoting osteoclast formation and activity through the release of tumor cell–derived factors, such as parathyroid hormone–related peptide (PTHrP), prostaglandin-E, and bone sialoprotein. PTHrP binds to receptors on osteoblasts, enhancing production of receptor activator of nuclear factor κβ ligand (RANKL). This pathway results in the differentiation of osteoclasts and their activation. On osteoclasts, RANKL binds to its receptor, RANK, promoting an increase in bone resorption.
Osteoclast inhibition induced by bisphosphonates (BPs) can alter the bone microenvironment, making it less supportive of tumor cell proliferation, and preclinical data suggest that BPs can reduce the likelihood of developing bone metastases. The mechanisms of antitumor activity of BPs have not been fully elucidated, but it is proposed that these drugs exert both indirect and direct antitumor effects. One potential indirect mechanism of action would be inhibition of bone resorption, in which BPs reduce the release of cytokines and other growth factors. This makes the bone environment less hospitable for tumor cell migration, decreasing the likelihood of cell adhesion and proliferation. Another indirect antitumor effect is the induction of a T cell–mediated immune response, in which BPs may inhibit angiogenesis by activating T-effector cells producing interferon-gamma.
A potential direct mechanism of BPs is the inhibition of angiogenesis, as shown in particular with ZA.[5,6] The BPs also decrease the adhesion, invasion, and proliferation of tumor cells. Moreover, they may have synergistic antitumor activity with cytotoxic agents, mainly with taxanes and anthracyclines. For instance, the combination of ZA and doxorubicin(Drug information on doxorubicin) produced synergistic antitumor activity in a primary breast cancer murine model without extraskeletal disease. Additionally, antitumor effects appear to be dose-dependent, and in some in vitro studies BPs may directly and dose-dependently induce apoptosis in breast cancer cell lines.[8-11]
The hypothesis of the “seed and soil” described by Paget may further elucidate the antitumor properties of BPs. The “seed” represents the tumor cell and the “soil” represents the bone marrow, which would have the ideal microenvironment to receive tumor cells. Supportive stromal cells, growth factors, and other components of the extracellular matrix, including the vasculature, make the bone marrow microenvironment a favorable milieu for cancer cell proliferation and propagation. Some preclinical evidence also suggests that the bone marrow is a “premetastatic niche,” and the engraftment of tumor cells onto a favorable bone marrow microenvironment would begin the evolution of micrometastasis, to macrometastasis, to distant, non-bone metastasis. Thus, early intervention such as the use of BPs in the adjuvant setting may interfere with the formation of this premetastatic niche.
Bone-Modifying Agents and Early Breast Cancer: Clinical Observations
Studies of clodronate as adjuvant therapy have reported mixed results, with variable effects on DFS and OS. In a prospective, non–placebo-controlled study by Diel et al, patients with primary breast cancer stage T1-T4,N0-N2, who had immunocytochemical evidence of one or more tumor cells in a bone marrow aspirate, were randomly assigned either to treatment with clodronate at 1600 mg orally per day for 2 years or to standard follow-up care. A total of 302 patients were accrued between February 1990 and April 1997. The majority were stage T2, node-positive, estrogen receptor (ER)–positive, and postmenopausal. A total of 246 (81%) received adjuvant systemic treatment. During the median observation period of 36 months, the Kaplan–Meier curves showed significant differences between the two groups in metastasis-free survival (P < .001) and OS (P = .001), in favor of the clodronate group. The differences were also significant considering the proportion of patients developing bony metastases (P = .003) and visceral metastases (P = .003). After 103 months of follow-up, the significant improvement in OS was maintained in the clodronate group (P = .049); however, a significant difference in the incidence of bony(P = .770) and visceral metastasis (P = .222) between the clodronate and control groups was no longer seen.
Powles et al conducted a double-blind, multicenter trial that accrued 1069 patients with operable breast cancer given oral clodronate, 1600 mg daily, or placebo for 2 years. Mortality was significantly reduced for patients randomized to clodronate (hazard ratio [HR] = 0.77; P = .047). Five-year OS was 82.9% for clodronate and 79.3% for placebo (P = .047). This trial showed a decrease in the subsequent occurrence of bone metastases, but this reduction was only significant during the medication period (P = .016). In contrast to the Diel trial, there was no reduction in the incidence of visceral metastasis in the clodronate-treated group.
However, a study by Saarto et al did not find a similar benefit for clodronate. In this trial, 299 pre- and postmenopausal women with T1-T3,N1-N2 breast cancer were randomized to clodronate for 3 years or control. Unlike in the other studies, the development of visceral recurrence was significantly higher in the clodronate group compared with controls: 43% vs 25% (P = .0007). The OS and DFS rates were also significantly lower in the clodronate group compared with controls (OS, 70% vs 83%, P = .009; DFS, 56% vs 71%, P = .007, respectively). The negative impact of clodronate was more prominent in patients with more advanced disease (large tumors, higher number of positive nodes, ER-negative disease). Especially in ER-negative patients, the DFS difference was large (29% in favor of the control group). In a multivariate analysis, after adjusting for progesterone(Drug information on progesterone) receptor status, the adverse effect of clodronate on OS lost its significance, but there was still a significant adverse effect on DFS. These negative results may be explained in part by an imbalance between the treatment arms, with more hormone receptor–negative patients in the clodronate arm. After 10 years of follow-up, however, there was no significant compromise in OS (54% for the clodronate group vs 62% for controls, P = .13), but DFS remained significantly lower in the clodronate group (45% with clodronate vs 58% for controls, P = .01). This difference in DFS was especially observed in ER-negative patients (25% vs 58%, P = .004).
In 2007 a meta-analysis of the three studies described above demonstrated no statistically significant difference in OS between patients treated with adjuvant clodronate and those receiving no treatment (HR = 0.75; 95% confidence interval [CI], 0.31–1.82). Similarly, there was no difference in time to appearance of bone metastases (HR = 0.68; 95% CI, 0.38–1.23) or delay in the occurrence of visceral metastases (HR = 0.89; 95% CI, 0.40–1.98).
The final analysis of National Surgical Adjuvant Breast and Bowel Project (NSABP) B-34, a prospective, randomized trial assessing 3323 patients with stage I to III breast cancer who received oral clodronate at 1600 mg daily for 3 years compared with placebo, given alone or in addition to adjuvant chemotherapy or hormone therapy, was presented at the 2011 San Antonio Breast Cancer Symposium. More than two-thirds of the patients were older than 50 years of age and had ER-positive tumors. The median follow-up was 8.41 years, and the long-term results showed no difference in DFS at any point in time (HR = 0.913; P = .266). Treatment compliance was poor, with only 42% of patients completing assigned study therapy. The greatest drop in compliance occurred during the first 6 months of the study when patients received concurrent chemotherapy, which may partially explain the negative result. Examination of the secondary endpoints did show a significant difference in favor of clodronate with respect to the non–bone- metastasis-free interval (NBMFI) (HR = 0.743; P = .046). In addition, prespecified subgroup analyses that included stratification by patient age at study entry showed that women older than 50 consistently derived more benefit from clodronate than did younger women. DFS remained similar, but analysis of secondary endpoints showed clear separation between treatment groups for recurrence-free interval (HR = 0.76; P = .05), bone metastasis–free interval (HR = 0.61; P = .024), and NBMFI (HR = 0.63; P = .015), all favoring clodronate for older patients. A post-hoc analysis with further age stratification showed that patients older than 60 derived the most benefit from clodronate, including an almost 60% reduction in skeletal metastases and a 40% to 50% reduction in nonskeletal metastases. Therefore, the findings of the B-34 trial were consistent with those observed in older postmenopausal women in other BP studies (Tables 1–3).
In the German Adjuvant Intergroup Node-Positive (GAIN) study, investigators found a similar lack of effect of ibandronate (Boniva) on DFS in patients with node-positive, early breast cancer treated with dose-dense chemotherapy. In this trial, 3023 patients were randomized to a standard regimen of epirubicin(Drug information on epirubicin), paclitaxel(Drug information on paclitaxel), and cyclophosphamide(Drug information on cyclophosphamide) or to lower doses of the same three drugs plus capecitabine(Drug information on capecitabine) (Xeloda), and were further randomized 2:1 to 2 years of treatment with ibandronate at 50 mg orally daily or to observation. The primary endpoint was DFS, with separate analyses for the chemotherapy regimens and ibandronate vs observation. Although the protocol specified 5 years of follow-up, an interim analysis after 50% of the required events had occurred showed that the futility boundary had been crossed with respect to the comparison of ibandronate vs observation. Follow-up continues for the two chemotherapy regimens. At 39 months, patients in the ibandronate arm had a 3-year DFS of 87.6% vs 87.2% for the observation arm (HR = 1.059; 95% CI, 0.861–1.301). Analysis of 3-year OS (secondary endpoint) showed a similar lack of difference (HR = 0.961; 95% CI, 0.705–1.31): 94.7% for ibandronate and 94.1% for observation.
Results from two nonrandomized trials, both presented as abstracts, suggest that adjuvant pamidronate may lead to a significant reduction in bone metastases. Kokufu et al developed a trial of 90 breast cancer patients with four or more positive nodes who were assigned by patient preference to receive IV pamidronate (at 45 mg for 4 cycles, q 2 wks) or to serve as controls. After a median follow-up of 5.4 years, the incidence of bone metastases was significantly lower in the pamidronate arm (12.1% vs 38.6%, P = .008), with no difference in OS.
Jung et al analyzed data on 429 stage I to III patients treated over a 4-year period with standard local and systemic therapy, and about 60% received adjuvant pamidronate (15 mg IV q 4 wks and 100 mg orally daily during adjuvant chemotherapy). The patients were perimenopausal, and the median follow-up was 3.5 years. The incidence of bone metastases was significantly lower in the pamidronate-treated patients (2.3% vs 8.7%), but there was no difference in DFS or OS.