Denosumab induction and Zoledronic acid maintenance therapy for recurrent unresectable giant cell tumour of the distal tibia: A case report with sustained tumour control after drug withdrawal

Highlights

• •Denosumab induces considerable but reversible response in GCTB. Bisphosphonates have a limited activity in untreated GCTB.
• •Denosumab-induced morphological changes could improve pharmacodynamics of
• •bisphosphonates.
• •Zoledronic acid use after effective induction may prevents tumour reactivation after Denosumab withdrawal.

Abstract

A 31-year-old woman was diagnosed with a recurrent and rapidly growing giant cell tumour of distal tibia with skin ulceration after intralesional curettage. The patient started on Denosumab 120 mg subcutaneously, once per month with additional loading doses on Days 8 and 15 attempting to avoid below-knee amputation. Twelve doses of Denosumab were administered in 9 months, resulting in resolution of pain, reduction of tumour size and calcification. Hence, the local surgical treatment was delayed and bisphosphonate maintenance therapy was initiated as skin healing was incomplete. The patient was given Zoledronic acid infusions at a dose of 4 mg. After the third infusion, the skin healed. As tumour remained stable, it was decided to continue maintenance. Overall, six doses of Zoledronic acid at 6 months intervals were administrated over 3 years. At the end of the maintenance, the patient experienced no pain and satisfied with her limb function. Since the lesion remained stable over 3 years after Denosumab discontinuation, it was suggested to stop further medical treatment and proceed to active surveillance. The patient’s disease is still stable clinical remission with no serious adverse events 41 months after Denosumab cessation and 10 months after the last bisphosphonate infusion. The present case confirmed the high effectiveness of denosumab as induction therapy in advanced recurrent giant cell tumour. We speculate that following the Denosumab-induced tumour ossification, high Zoledronic acid uptake in lesion may prevent tumour reactivation due to its improved pharmacodynamics with an assumed irreversible anti-tumoral effect on residual mutated cells. This hypothesis requires confirmation in future studies.

1. Introduction

Giant cell tumour of bone (GCTB) is a locally aggressive, benign lesion, which causes extensive bone destruction. Microscopically, GCTB tends to be a densely cellular lesion, with an admixture of three types of cells: mononuclear round to oval polygonal or spindle-shaped stromal cells considered responsible for the neoplastic nature of the tumour, and the reactive component represented by mononuclear monocytes and scattered abundant giant multinucleated cells [1], [2], [3]. Somatic driver mutation of the H3F3A gene underlies the pathogenesis of conventional GCTB [4]. Mutated stromal cells of the osteoblastic lineage acquire a proliferative advantage and stop in differentiation through intrinsic and/or extrinsic mechanisms [5], [6], [7]. They are responsible for unbalanced excessive receptor activator of nuclear factor kappa-Β ligand (RANKL) production that stimulates formation of multinucleated osteoclast-like giant cells (MNGCs) resulting in bone resorption [8]. In turn, MNGCs also suppress osteogenic differentiation of immature stromal cells and maintain them in a RANKL-secreting state through various mediating factors [9], [10], [11]. Thus, a vicious cycle is created at the tumor microenvironment (TME) level, which drives the growth and clinical evolution of GCTB [5], [12].

The current treatment paradigm for GCTB relies on individualized approaches based on the Campanacci stage and tumour location [13]. Stage III and advanced recurrent tumours, located in the foot and ankle bones are difficult to treat due to close proximity to the ankle joint, significant surgical morbidity, and high recurrence rate after curettage, leading to repeated surgical procedures, and even amputations [14], [15]. In this anatomical location, limb-salvage procedures are challenging and associated with a high risk of positive margins, scarce soft tissue coverage, and poor long-term functional results after surgery [16].

Two major classes of drugs are used for treatment of unresectable and inoperable GCTB: bisphosphonates (BPs) and RANKL-inhibitors. Despite the pronounced clinical, radiological and morphological effect of medical treatment with Denosumab (DM) representing the latter class of drugs, tumor control is only temporary and the disease is reactivated after discontinuation of the drug. In the lack of prospective data on unresectable GTCB, current NCCN guidelines suggest that the treatment should be continued until evidence of disease progression in responding disease or toxicity [17].

On the other hand, the absence of a substrate for the accumulation of BPs in an untreated tumor reduces their clinical effectiveness as first-line option. The purpose of this case report is to present a case of locally advanced, ulcerating and unresectable recurrent GCTB of the distal tibia in which long-term disease control with skin healing was achieved using a two-stage approach consisting of sequential administration of both groups of drugs. Such induction therapy with DM and maintenance therapy with zoledronic acid (ZA) have seemingly resulted in a long-term disease control maintained also after withdrawal of medical treatment.

2. Case report

2.1. History and physical examination

A 31-year-old woman was diagnosed with aneurysmal bone cyst (ABC) of the left distal tibia in April 2019 and the tumour was removed via intralesional curettage. According to medical records, the initial X-Ray examination showed an expansile osteolytic lesion extending to the subchondral bone. The histopathological features were found to be consistent with primary ABC. Three months later, a palpable mass was detected at the surgical site with further rapid growth leading to extensive skin ulceration. At the first presentation in our department in November 2019, a painful recurrent tumour was revealed (Fig. 1A).

Fig. 1.

Recurrent GCTB of distal tibia before treatment. A – Locally advanced tumor with significant skin ulceration and inflammation; B – Coronal view on CT scans showing an expansile osteolytic lesion with minimal internal septation, cortical breakthrough and soft tissue extension; C – T2-weighted axial MRI demonstrates a bulky extraosseous component containing large secondary ABCs with fluid–fluid levels; D – In hematoxylin and eosin staining, multinucleate giant cells were visible, surrounded by neoplastic stromal cells, confirming the diagnosis of a giant cell tumor of bone (20X). E – Strong immunoreactivity of neoplastic mononuclear stromal cells for the H3F3A G34W mutation-specific antibody (20X).

2.2. Imaging and classification

CT and MRI showed an osteolytic lesion with cortical bone destruction and soft tissue expansion with multiple secondary ABCs (Fig. 1B and 1C). Chest CT showed no evidence of lung metastases. After revision of histological material involving immunohistochemistry, the tumour was re-classified as GCTB due to positive staining of tumour cells for H3F3A G34W mutation (Fig. 1D and 1E).

2.3. Antiresorptive induction treatment with Denosumab

Given the diagnosis, disease extent and its rapid local progression, the only surgical option at that time was below-knee amputation. Based on previously published data on the effectiveness of DM in GCTB and after obtaining the patient‘s informed consent, we initiated anti-resorptive therapy to downstage the lesion. The patient started with DM in a dose 120 mg subcutaneously, once per month with additional loading doses on Days 8 and 15, as well as oral vitamin D, and calcium supplementation. Twelve doses of DM were administered from December 2019 to September 2020.

2.4. Response to Denosumab induction therapy

After DM induction a considerable tumour response consisting of pain relief, partial skin epithelialization, tumour size reduction was observed with calcification of the lesion, reduction of cystic component, and disappearance of fluid–fluid levels (Fig. 2A-C).

Fig. 2.

Tumor response after induction treatment with Denosumab. A – Reduction of tumor size and partial skin epithelization with diminution of perifocal inflammation; B – Coronal view on CT scans showing a significant mineralization of internal septa and incomplete peripheral bony rim; C – T2-weighted axial MRI demonstrates partial tumor regression, bone formation within the lesion, decreasing of cystic component and disappearance of fluid–fluid levels.

2.5. Maintenance therapy with Zoledronic acid

Because of the expected risk of complications following limb salvage surgery at that time and to avoid long-term DM toxicity, we decided to consolidate the achieved response and delay local treatment by switching to ZA maintenance until skin healing. In January 2021, four months after DM discontinuation, the patient received the first ZA infusion at a dose of 4 mg with the same oral vitamin D and calcium supplementation. After the third infusion of ZA (October 2021) the skin had healed. Taking into consideration that the tumour remained stable, quality of life was good and having the distal tibia resection with endo-prosthetic replacement as salvage option, we recommended the patient to continue maintenance treatment. Until April 2023, three additional doses of ZA were administrated every 6 months. At the end of the third year of maintenance, without serious adverse events or complications, the patient was pain-free and satisfied with good limb function and cosmetic appearance, despite minimal leg deformity (Fig. 3A). The MRI showed further decreasing of cystic component (Fig. 3B). At the same time, the normalization of serum C-telopeptide level was documented (0.154 ng/mL). 18FDG- PET/CT in June 2023 revealed continued bone formation inside the tumour and at its peripheral rim with a complete metabolic response (Fig. 3C). Since the lesion continued to remain stable over 2.5 years after DM withdrawal and showed no evidence of increased metabolic activity, the decision was made to recommend discontinuation of ZA maintenance and proceed to active surveillance. The patient agreed to undergo close monitoring every 3 months with possible limb-sparing salvage surgery in the case of tumor reactivation.

Fig. 3.

Evolution of tumor response during ZA maintenance. A – Reduction of tumor size and skin healing; B –– Axial view on MRI T2-weighted scans demonstrating further increasing of mineralization of solid components of tumor and significant reduction of cysts; C – Coronal view on 18FDG PET-CT fused scans showing an increased degree of tumor ossification and a complete metabolic response (SUVmax = 1.0) 3 years after DM discontinuation and 2 months after last ZA dose; D – Radiograph at last follow-up (February 2024) demonstrates ongoing bone remodeling with cancellous bone formation within the lesion (ten months without antiresorptive treatment). E – Whole body 99mTc-pirfotex scintigraphy demonstrates slightly increased and diffuse radioactivity in the affected area, suggesting ongoing bone remodeling at minimal levels 39 months after DM discontinuation.

2.6. Outcomes

As of the time of writing, 41 months after DM discontinuation and 10 months after the last ZA infusion, the tumour remains stable on MRI with continued bone healing on X-ray (Fig. 3D) and minimal bone remodeling on 99mTc-pirfotex bone scintigraphy (Fig. 3E).

3. Discussion

In difficult-to-treat GCTBs conservative approaches have been conceptualized on the basis of breaking the vicious cycle established at the TME level in GCTB [5], [12]. BPs like ZA are, besides their osteoclast inhibitory effects, believed to impair tumour growth by targeting both cellular components, MNGCs, and mutated H3F3A G34W + cells [18], [19], [20], [21], [22], [23], while RANKL inhibitors suppress only the osteoclastogenesis with limited direct effects on tumour cells [3], [22], [24].

To date, BPs are the only class of drugs directly affecting the GCTB neoplastic stromal cell population. In vitro studies have shown that single treatment with pamidronate (PAM), farnesyltransferase inhibitor-277, and GGTase I inhibitor-298 all reduced cell viability and inhibited the proliferation of GCTB stromal cells in a dose-dependent manner [25]. A microscopic view has demonstrated a reduced cell density in GCTB stromal cell culture after ZA treatment; but no obvious changes were seen after DM treatment as compared with the control [26]. By MTT assay, ZA was found to exhibit dose-dependent inhibition in cell viability in all three GCTB stromal cell lines tested [26]. In an in vivo mouse model, ZA reduced tumour cell viability in GCTB [27]. Another study with ZA demonstrated higher cytotoxic effect on GCTB stromal cells and renal cell carcinoma than on multiple myeloma, and decrease in number of viable cells was seen in a dose-dependent manner [28].

Shibuya et al. reported that ZA demonstrated an inhibitory effect on proliferation of neoplastic cells at 2 μg/mL (a concentration at which an inhibitory effect on osteoclast differentiation was seen) [22]. Moreover, as concentration increased further, ZA showed a dose-dependent inhibitory effect on neoplastic cells.

In different in-vitro and animal studies, it was demonstrated that nitrogen-containing BPs did cause substantial apoptotic effect in GCTB stromal cells [25], [26], [29]. This effect can be enhanced by melatonin, which acts synergistically with ZA [23]. Using flow cytometry following fluorescent Annexin-V labeling, it was observed that statistically significant increase of tumour cell apoptosis in GCTB primary cultures started at 30 µM of ZA or PAM with maximal effects at 150 µM [29]. In another study, the same PAM dosage significantly induced cell apoptosis 3.7-fold over the vehicle control [25].

In addition to its apoptotic effect, it was speculated that BPs might have the ability to induce directly or indirectly the differentiation of GCTB stromal cells. In different in-vitro and animal studies, it was shown that ZA induced neoplastic stromal cell osteogenic differentiation, which was associated with transcriptional activation of RUNX2, osterix (Sp7 transcription factor) and osteocalcin genes [30]. Wang et al explored the effects of different melatonin and ZA concentrations on the mRNA and protein expression of osteogenesis markers (OPN, OCN, RUNX2 and ALP) in GCTB cells and found that different concentrations of either melatonin or ZA promoted the mRNA expression of osteogenesis markers, and the effect of a higher concentration was more obvious [23].

Additional mechanisms may be involved in the activation of osteogenesis after BPs therapy, including the deprivation of TME from MNGCs, which leads to impairment of their ability to block osteogenic differentiation of GCTB stromal cells [9], [10], [11]. BP-induced osteoclast apoptosis results in formation of apoptotic bodies, which have anabolic effects on bone in vivo [31], [32]. Reverse signaling from MNGCs can contribute to osteogenesis and induce mineralization via vesicular RANK–mRANKL signaling and/or MNGC-derived apoptotic bodies [33].

However, tumour cells are unable to acidify the bone surface to release sufficient amounts of BP and/or because they have a far lower endocytic capacity than osteoclasts [34]. To the best of our knowledge, there is still no evidence that neoplastic cells can internalize sufficient amounts of BPs in vivo to have the same inhibitory effects on protein prenylation as in vitro. Therefore, despite strong in vitro activity, and effectiveness in adjuvant setting [35], the clinical and radiological treatment effects of BPs are not impressive when used as preoperative treatment in such purely osteolytic tumors, neither for localized nor in the metastatic setting of GCTBs [18], [21], [36], [37], [38]. Histologically, Tse et al were unable to document any objective generalized increase in mineral content or osteoid deposition within the lesion following 2 months of bisphosphonate therapy [38]. Focal increase in osteoid and mineralized bone was observed in 12 (50 %) patients, typically in sections taken at the margins of the lesion [38].

It is believed that low in vivo activity of BPs is caused by the lack of a significant amount of osteoid and amorphous calcium phosphate containing ossifying collagen-rich stroma in untreated tumour, necessary for drug absorption and further internalization by cells of the TME.

Yang et al. suggested that BPs might have a dose-dependent adjuvant effect on GCTB [30]. High concentrations of BPs found at the bone surface especially at the resorptive excavation induces apoptosis of the tumour cells. Osteoblastic lineage cells attaching to the bone surface could internalize concentrated ZA, resulting in direct cytotoxicity. On the other hand, long-term low to moderate concentrations of unbound form of ZA released from the bone surface into the bone marrow space decrease the proliferation and promote osteoblast differentiation of stem/osteoprogenitor cells.

We hypothesize that during and early after DM treatment, when tumour mineralization is very intense and phosphate uptake is significantly increased (Fig. 4), tumor cell apoptosis is the main direct anti-tumour mechanism of BPs. As mature lamellar bone forms in the lesion, when the intensity of bone remodeling and the concentration of BP in the TME decreases (Fig. 3E), the mechanisms of differentiation of the remaining bone-forming tumour cells come to the fore.

Fig. 4.

Recurrent GCTB in a 17-year-old girl. A-C – Baseline planar bone scintigraphy, CT and SPECT images show a pure osteolytic lesion with central photopenia and peripheral tracer uptake (“donut sign”) due to absence of mineralized bone structures. D-F - planar bone scintigraphy, CT and SPECT images after four months of DM therapy (7 doses), demonstrating new bone formation with a dramatic increase in tracer accumulation in the central part of the lesion due to abundant deposition of amorphous calcium phosphate in the tumour stroma.

In contrast to BPs, numerous data indicate that RANKL inhibitors, such as DM, do not require a substrate for accumulation to promote a significant reduction of MNGCs and their precursors [1], [39]. This represent almost half of all cell populations at the TME level in GCTB [40]. The elimination of proliferative and differentiation blocking signals from MNGCs to tumor cells leads to the shift of stromal cells with pre-osteoblastic phenotype toward the more mature phenotype able to produce and mineralize large amounts of bone matrix [41].

The clinical impact of DM therapy consists of tumour down-staging facilitating function- or limb-sparring surgery [42]. The faster and more significant deposition of new bone induced by DM compared with ZA [43] supports its widespread use as first-line therapy for stage III GCTB [17]. Its main disadvantage as a monotherapy is the lack of direct cytotoxic effect on mutated cells, rendering tumour reactivation the rule after drug withdrawal [44], [45]. The reported rates of local failures after DM discontinuation ranged between 40 % and 67 % [46], [47], [48]. Thus, RANKL inhibition is considered disease-combating, but not curative [49]. Although the occasional use of ZA as a salvage treatment in DM non-responders has been reported [50], the position of ZA in the treatment of GCTB next to DM remains unexplored. The reverse sequence has been studied, but ZA was most often ineffective as a first-line option, while DM allowed for long-term disease control [51], [52], [53].

In the present case, highly effective DM induction allowed us to downstage the tumour and avoid amputation. The rationale for ZA administration was to consolidate the response to DM until complete skin healing. As no disease progression was observed during this phase, it was recommended to continue with ZA and close surveillance until an eventual local failure. The rationales to continue conservative treatment included the challenging location of tumor in distal tibia and the expected high morbidity of surgical resection with endo-prosthetic replacement in a young lady (infection, instability etc). Accordingly, the decision to adopt the “wait and see” strategy and delay an eventual surgery seemed to be reasonable. In this case, surgery (wide resection/amputation) is always a salvage procedure; an ultimate “refugium”.

Fortunately, throughout the period of maintenance treatment, which lasted 31 months after DM withdrawal and over 10 additional months after ZA discontinuation, no progression and increasingly cancellous bone formation was observed on radiographic images. In addition, bone scintigraphy showed minimal remodeling in the lesion area at last follow-up. It can be assumed that during the ongoing bone remodeling by osteoclasts, BPs are continuously released from calcified stroma [54], thereby creating sufficiently high local concentrations necessary to have an impact on residual cells embedded in new bone. Studying the pharmacodynamics and osteogenic mechanism of BPs in newly formed bone and in vivo visualization of mineralization in the post-induction period needs further investigation and the clinical value is worthy of further observation.

Our interesting and somewhat unexpected finding may be considered a promising strategy of DM discontinuation in GCTB patients, who are facing potentially mutilating surgical and potential toxic, costly therapy. Although DM is well tolerated in most cases, lifelong treatment is not considered an optimal and safe option for unsalvageable GCTB. Some concerns are the possible side effects associated with the prolonged treatment such as arthralgia (grade 1–2: 50 %), chronic muscle pain (33 %), peripheral neuropathy (11 %), fatigue (17 %), skin rash (9 %), electrolyte disturbances (4 %), osteonecrosis of the jaw (ONJ), (5–9 %), atypical bone fractures (4 %) and malignant transformation (1 %) [39], [48], [47]. Various approaches have been proposed to reduce the cumulative dose-related toxicity and high treatment costs [43] of long-term DM use, including treatment discontinuation [48], “drug holidays” [55] and increased dose intervals [50], [56], [57], [58], [59].

Comparisons of side effect profiles during long-term use of DM and BPs are difficult to make due to very limited clinical experience with the latter class of drugs [18], [60], [61]. In these studies, none of the patients experienced complications related to daily oral or monthly intravenous BPs treatment. The risk of ONJ is relatively high (1–15 %) in patients with cancer metastatic to bone treated with high doses of ZA [62], [63], [64]. In contrast, the incidence of ONJ in patients with osteoporosis is approximately 0.01 % to 0.001 % [62]. The risk of atypical femoral fractures (AFF) increases with duration of BPs use and rises after 3 to 4 years [65]. In our patient, over a time limited to 27 months, we used a BPs regimen that was closer to the treatment of osteoporosis than metastatic cancer and did not observe ONJ or AFF. Optimal BPs treatment duration remains to be defined. In our opinion, consideration may be given to stopping ZA maintenance and switching to active surveillance once sustained disease control is achieved.

To the best of our knowledge, the sequential treatment with DM as an up-front induction therapy followed by ZA as maintenance in GCTB here presented, has not previously been published. This approach allowed us to achieve local control in a difficult-to-treat tumour location, making it possible to at least delay limb-sparing surgery, and at best, avoid it. Our hypothesis-generating case report provides a rationale for initiating cooperative studies in GCTB patients, who are currently treated with DM alone.

4. Conclusions

The present case confirmed the high effectiveness of DM as an induction therapy in a locally advanced relapse of GCTB of the distal tibia, which avoided the patient a leg amputation. As is well known, GCTBs often tends to reactivate after cessation of RANKL inhibition. We speculate that after a good response to DM, ZA maintenance might play a protective role in local control, delaying or even avoiding tumor excision. This is probably due to a change in the pharmacodynamics of BPs, which consists in their better accumulation in a well-mineralized newly formed bone matrix and higher internalization by cells of TME, sufficient to provide a direct and irreversible effect on residual tumour cells. Our hypothesis requires confirmation in future experimental and clinical studies.

CRediT authorship contribution statement

Gennady N. Machak: Writing – original draft, Visualization, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Øyvind S. Bruland: Writing – review & editing, Methodology, Data curation, Conceptualization. Tamara N. Romanova: Visualization, Formal analysis, Data curation. Alexey V. Kovalev: Visualization, Investigation, Formal analysis, Data curation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.