The Efficacy of Bisphosphonate in the Treatment of Giant Cell Tumour of the Bone: A Systematic Review and Meta-Analysis

Introduction Anti-osteoclastic mechanism of Bisphosphonate (BP) is crucial to treat Giant Cell Tumour of the Bone (GCTB), however no established guidelines of its use have been published. This systematic review and meta-analysis is the first to summarise recent clinical studies on the subject. Materials and methods A systematic search was performed based on PRISMA guidelines for clinical trials of BP administration in GCTB. Baseline data including BP regimen, dose and timing was summarised. The primary outcomes assessed were recurrence rate, metastases, survival rate, functional outcome, clinical outcome, radiological outcome, and adverse effect. Results We identified 8 articles from 2008-2020. Most studies administer 4mg of Zoledronic acid post-operatively, with five studies mentioning pre-operative administration and six studies describing post-operative administration. There was a total of 181 GCTB cases analysed in this study. The BP group presented lower recurrence rate than control group (three studies; Odds Ratio [OR] 0.15; 95% Confidence Interval [CI], 0.05 – 0.43; p<0.05; heterogeneity, I2=0%). As for survival rate, BP group is comparable to control group (two studies; OR 1.67; 95% CI, 0.06 – 48.46; p=0.77; heterogeneity, I2=65%). Conclusion Bisphosphonate therapy offers satisfactory recurrence rate, functional outcome, clinical outcome, radiological outcome, survival rate and metastases rate in patients with GCTB, with minimal adverse effects. Pre- and post-operative administration of bisphosphonates in combination might be the most beneficial in minimalising the recurrence rate.


INTRODUCTION
Bisphosphonate (BP) application in Giant Cell Tumour of the Bone (GCTB) as systemic adjuvant treatment is still controversial. In vitro and animal studies showed its effect in inducing stromal cell inhibition, apoptosis and osteogenic differentiation. However, studies with high level of evidence are still needed to strengthen the recommendation of routine BP application in GCTB treatment. This systematic review and meta-analysis is the first to summarise recent clinical studies on the subject.
The mainstay of treatment for Giant Cell Tumour of the Bone (GCTB) is surgical resection, either en bloc resection or curettage, with or without local adjuvants. However, high risk of recurrences after this primary treatment creates the need for systemic adjuvant treatment such as BP. As part of a reliable treatment regimen for osteoporosis, metastatic bone disease, and Paget's disease, the use of BP for GCTB and its efficacy has not been discussed a lot in literature. As an analogue of pyrophosphate, BP with its anti-osteoclastic actions seems to be promising in improving the outcomes of patients with resectable as well as unresectable GCTB 1 .
There are still no established guidelines on the indication, dose, regimen, and efficacy of BP for patients with GCTB. Most trials administer BP as a part of post-operative treatment, yet pre-operative application is an interesting subject to explore, where it has the capability of reducing tumour size before surgery. This provides the basis for the study question: what is the efficacy of bisphosphonate therapy for GCTB, and what is the proper dosing and time of administration? Through this systematic review, we aim to discuss the details of BP treatment in previous studies, while the meta-analysis part will summarise quantitative evidence of BP effect in terms of survival and recurrence rate.

MATERIALS AND METHODS
This systematic review was performed in accordance to Preferred Reporting Items for Systematic Reviews and Metaanalysis (PRISMA) guidelines (Fig. 1). We performed systematic search through MEDLINE, Embase, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Web of Science, Cochrane Central Register of Controlled Trials (CENTRAL), and ClinicalTrials.gov. Language was limited to English. Our search strategy was as follows: keywords such as "Giant Cell Tumour of Bone" and "Bisphosphonate" and "Outcome" were used. Those data were then manually scanned and reviewed by all authors based on inclusion and exclusion criteria according to PICO (Population, Intervention, Comparison, Outcome) as depicted in Table I.
The data extraction was collected under basic characteristics and outcomes. In each study, mean difference (MD) for continuous outcome and odds ratio (OR) for dichotomous outcome with a 95% confidence interval (CI) was calculated using Review Manager (RevMan) [Computer program, Version 5.3. Copenhagen: The Nordic Cochrane Centre, the

RESULTS
We identified 8 articles from 2008-2020, with a total of 181 cases. Two studies were Level I evidence, two others were Level III, and four others were Level IV (Table II). Critical appraisal of all studies included were conducted based on Joanna Briggs Institute Scoring System, showing no study had more than three invalid parameters (Fig. 2).
From the table of study characteristics, it appeared that GCTB equally affected male and female in control group (30 males vs 27 females) and slightly more female than male in bisphosphonate group (56 males vs 68 females), with the sample age ranging from 19 to 75 years old. The most common locations were distal femur (42 cases) and proximal tibia (33 cases), as well as sacrum (48 cases). Other locations found were proximal femur, proximal tibia, proximal humerus, and distal radius. Appendicular bones were more commonly affected (129 cases) than axial bones (55 cases) (Table III).
Most studies administered 4mg of intravenous Zoledronic acid, but other bisphosphonates had also been used, such as Pamidronate, alendronate, clodronate, and sodium ibandronate. Timing of administration differed, with five studies mentioning pre-operative administration and six studies describing post-operative administration. In terms of surgical treatment choice, intralesional curettage and cementation was the most performed procedure (36 cases).
Other commonly performed procedures were wide resection and bone grafting (35 cases), intralesional curettage and bone grafting (23 cases), wide resection and cementation (22 cases), and nerve sparing surgery for sacrum GCTB (10 cases). A study by Singaravadivelu et al (2020) described 10 cases of GCTB around knee managed by extended curettage and structural support by Fibula Cortical Struts 2 . Phenol and Polymethyl Methacrylate (PMMA) were also used as adjuvant therapies in more than 20 cases, on their own as well as in combination. The follow up period ranged from 3 to 192 months (Table IV).
We considered two parameters for the forest plot. The BP group presented lower recurrence rate than control group (three studies; OR 0.15; 95% CI, 0.05 -0.43; p<0.05; heterogeneity, I2=0%) (Fig. 3). A study by Tse et al (2008), using pre-and post-operative bisphosphonate treatment, contributed the most to the overall final Odds Ratio for recurrence rate (55.8%) 3 . As for survival rate, BP group was comparable to control group (two studies; OR 1.67; 95% CI, 0.06 -48.46; p=0.77; heterogeneity, I2=65%) (Fig. 4). The metastases rate was low in patients treated with BP, with lung as the most common location. In terms of functional outcome, BP seemed to offer better urinary and bowel function, and most patients were able to return to their presurgery functionality. Pain also improved in most patients, with one literature stated that the mean Musculoskeletal Tumour Society (MSTS) Score was 92%. Radiologically, BP was also proven to increase mineralisation and calcification, resulting in stable or decreased size, with better delineated border. Though some side effects have been reported, mostly they were minor and transient, such as fever, headache, or flu-like symptoms (Table V).

DISCUSSION
The locally aggressive nature of GCTB and its ability to metastasize presents a challenge in terms of recurrence and mortality. Surgical resection alone is often insufficient for disease control, and the addition of anti-osteoclastic agent may serve as an alternative to enhance outcomes4. Some adjuvant therapies have been on trial in previous studies, including BP which was already well-known for its efficacy. This systematic review and meta-analysis is the first to summarise recent clinical studies on the subject, focusing on the effectiveness of its regimen and the related outcome.
Bisphosphonates, one of a reliable treatment regimen for osteolytic cancers and metastases, has also been proven to be beneficial in treating osteoclast-mediated bone loss. Bisphosphonates work by binding to hydroxyapatite on the bone surface and inhibit adhesion of osteoclasts to the mineralised bone surface. Bisphosphonates also have a direct effect on stromal cells of GCTs through mevalonate pathway, blocking protein prenylation and promotes the activation. Furthermore, Bisphosphonates inhibit osteoclast-like giant cell formation from immature precursors as well as induces apoptosis in mature osteoclasts. Some literatures have reported the use of bisphosphonates for GCTB with the result of increased mineralisation of lesion as well as replacement of pathological bone lesion into normal bone structure; however, those studies were mostly noncomparative one-arm studies with relatively small amount of samples 5 . Chang et al (2004) in their study proved the inhibitory effect of Bisphosphonates on proliferation and apoptosis of osteoclasts by affecting the osteoprotegerin (OPG)/ Receptor Activator of Nuclear Factor Kβ-Ligand (RANKL) mRNA expression of stromal cells. Furthermore, Zoledronic acid displayed higher efficacy (10-20 times) in apoptosis and decrease in the live-cell rate compared to pamidronate 6 . The ability of BP to lower the amount of osteoclasts and inhibit osteoclastic resorption, especially amino BP, enables it as a potential treatment for GCTB. When applied postoperatively, BP could also eradicate the remaining cells 7 .
Another adjuvant therapy recommended for GCT is Denosumab. As a RANKL inhibitor, Denosumab was proven to be beneficial in tumour growth inhibition and reduced morbidity. A case series study by Goldschlager et al (2015) proved that Denosumab demonstrated beneficial radiological and histological response in most patients with spinal GCT 8 . However, in a study by Lau et al (2013) comparing Denosumab and Zoledronic Acid, Zoledronic Acid was proven to reduce cell growth, causing apoptosis in most cell lines, and significantly inhibiting mRNA expression of RANKL and Osteoprotegerin. These features were not found in Denosumab, raising a concern that tumour recurrence might happen after drug withdrawal9. Therefore, there were still some controversies in the treatment of choice between the two, as Denosumab failed to prove a permanent apoptotic effect on the neoplastic stromal cell population 6,8 . A study by Gouin et al (2014) stated that the local recurrence rate of GCTB in appendicular bones treated by extensive curettage with or without local adjuvant treatment was lower than axial GCTB, where it was around 15% for appendicular bones, 19% for axial bones, and 53% for sacral bones specifically 7 . Regarding the surgical treatment performed, a study by Zhang et al (2011) reported cases where interventional blood vessel embolism of the tumour reduced the blood supply, further disturbed the delivery of bisphosphonate to the lesion, and reducing the effectivity of bisphosphonate therapy. Some other adjuvant therapies might also alter the blood supply to GCTB, therefore limiting bisphosphonates' efficacy, such as radiation and thermotherapy 5 .
In terms of location, sacral GCTB warrants special attention. Despite being one of the most commonly affected bones, the treatment for sacral GCT remains challenging, as sacrificing sacral nerve roots is associated with severe morbidity, such as the disturbance of gait and foot plantar flexion (S1 nerve roots) as well as bowel and bladder dysfunction (S2-3 nerve roots). Even after a successful nerve-sparing surgery, the high recurrence rate (25-35% in most cases and up to 50% in some studies) demands an additional therapy to minimise it, potentially by the use of bisphosphonate therapy 9,10 .    Though some literature supports the efficacy of Bisphosphonates, its use is not without consequences. Bisphosphonate therapy has been reported to relate to some adverse effects in approximately 15% -30% of cases, though mostly mild such as fever, headache, chest pain, arthralgia, nausea, bone pain, altered taste, urticaria, gastritis, fatigue, flu-like symptoms, and periodontal disease. In more severe cases, osteonecrosis of the jaw could be found in <1% case per year of treatment, however regular dental assessment and avoiding invasive dental procedures are beneficial in preventing this adverse effect 7 .
Another topic still in debate is the timing of bisphosphonate administration.
Pre-operative administration of bisphosphonates has been proven to increase peripheral mineralisation, therefore better containing the lesion, more clearly delineating the borders, and making complete removal easier to achieve. However, post-operative administration has also been described in literature, where it was closely related to the recurrence rate. Due to its ability in inducing apoptosis, bisphosphonates can clear the residual microscopic tumour tissue after surgical procedures 2,11 . In our analysis, a study by Tse et al (2008) 3 was proven to have the highest contribution to overall odds ratio in the recurrence rate, leading to a conclusion that pre-and post-operative administration of bisphosphonates in combination might be the most beneficial in minimising the recurrence rate. In terms of survival and metastases rate, bisphosphonates were also comparable to control, making it a considerable choice in the treatment of GCTB, with appropriate dosing and time of administration. This study has several limitations: (1) Different generations of BPs were used, different surgical techniques (curettage and resection), different protocols (pre-or post-operatively, dose, period of treatment), different adjuvant therapies were applied, and different stages of the disease were treated (primary, recurrent, metastatic). This might contribute to a possible bias of analysis. (2) Due to the scarcity of qualified studies in this field, studies included are mostly of Level IV evidence. However, we have ensured the quality of included studies by using quality and bias assessment. (3) Some studies have short follow up time (3 months), which may contribute to the low rate of adverse effects shown by these studies. Despite these limitations, this study still serves as an important update in the treatment of GCTB, as this is the first meta-analysis study to objectively describe the efficacy of Bisphosphonate therapy. It is hoped that this study will be beneficial in considering adjuvant therapy for GCTB, as well as influential in conducting further well-designed studies with bigger amount of samples.

CONCLUSION
Current systematic review and meta-analysis suggest that Bisphosphonate therapy offers satisfactory recurrence rate, functional outcome, clinical outcome, and radiological outcome, as well as comparable survival rate and metastases rate compared to control in patients with GCTB, with minimal adverse effects. The combination of pre-and postoperative administration of bisphosphonates in combination might be the most beneficial in minimising the recurrence rate.

CONFLICT OF INTEREST
Authors declared no conflict of interest.