Rethinking the role of bisphosphonates after denosumab treatment in locally advanced or unresectable aneurysmal bone cysts: A meta-analysis

Abstract

BACKGROUND

Aneurysmal bone cysts (ABCs) are usually treated with curettage or various minimally invasive percutaneous procedures. Patient refractory to these treatments, as well as those with locally advanced or unresectable tumors, present a challenge for orthopedic surgeons and require new treatment approaches. Anti-resorptive drugs inhibit osteoclastic resorption and increase intralesional osteogenesis. Denosumab induces tumor ossification, but this effect may disappear after drug withdrawal due to limited impact on neoplastic cells. Bisphosphonates (BPs) may induce apoptosis of tumor cells and allow for long-term local control. We hypothesized that after denosumab treatment, BPs would better accumulate in the tumor and exert an irreversible antitumor effect.

AIM

To test the hypothesis that the sequential use of BPs after denosumab induction improves treatment outcomes in surgically unsalvageable ABCs.

METHODS

Using data from five electronic databases (Scopus, MEDLINE, EMBASE, PubMed, Web of Science), we aimed to identify all patients who received denosumab therapy (DT) for unresectable ABCs. Among published case reports and case series, we identified patients who discontinued denosumab for various reasons and divided them into two groups: Group 1 included 31 patients without further anti-resorptive therapy and Group 2 included 12 patients who received BPs in the context of rebound hypercalcemia. Local control rates in both groups were analyzed.

RESULTS

As of December 2024, 43 patients have been reported in the literature who received DT for locally advanced/unresectable ABCs. There were 27 males and 16 females with a mean age of 15.8 years. At a median follow-up time of 15.5 months, there were 10 confirmed and two pathologically unconfirmed relapses after denosumab discontinuation. All 10 relapses occurred in patients in Group 1 at a median time of 13.5 months. Among patients in Group 2, with a median follow-up time of 12.5 months after completion of therapy, no local relapses were observed. The difference between local recurrence rates (32% vs 0%) is statistically significant (P value = 0.02). Kaplan-Meier estimates show the same trend with marginal statistical significance (P value = 0.085). Here we put forward a novel treatment algorithm.

CONCLUSION

BPs used in post-denosumab ossifying ABCs appear to improve treatment outcomes, presumably by targeting residual tumor cells. Prospective clinical studies are warranted to validate this promising two-stage conceptual strategy in difficult-to-treat ABC.

Key Words: Aneurysmal bone cyst; Locally advanced/inoperable; Denosumab; Rebound hypercalcemia; Bisphosphonates

Core Tip: Benign bone tumors containing giant cells undergo significant ossification after denosumab treatment. In these conditions, sequentially administered bisphosphonates (BPs) accumulate better in newly formed bone, which may lead to long-term local control, presumably due to a pro-apoptotic effect on residual tumor cells. In this context, we studied a group of patients with inoperable aneurysmal bone cysts (ABCs) who were treated with denosumab and found that among those patients who received BPs for post-denosumab rebound hypercalcemia, there were no local relapses. We assume that BPs could have the same irreversible effect on residual tumor cells in ABCs and propose to continue experimental and prospective clinical studies to confirm this hypothesis.

INTRODUCTION

Aneurysmal bone cyst (ABC) is a rare benign tumor of bone with abundant blood-filled spaces in its classical form, belonging to the group of osteoclastic giant cell-rich tumors[1,2]. ABCs occurs most frequently in patients under 20 years of age, with the metaphysis of long tubular bones and vertebrae being anatomic predilection sites.

Knowledge about ABCs has evolved, especially regarding the molecular biology and neoplastic nature of this lesion[3-5]. A translocation t (16;17) (q22; p13) involving multiple fusion partners[6-11] results in the activation of the USP6 oncogene in fibroblast-like spindle cells of the osteoblastic lineage[4,12-14]. Overexpression of USP6 disrupts the normal process of bone modeling in growing patients. ABCs is characterized by an evolving course consisting of four successive phases: Initiation, active growth, stabilization, and healing[15]. In the early stages, increased cell proliferation[16-18], blockade of neoplastic cell differentiation[13], formation of an aggressive tumor microenvironment (TME)[14,18-22], neo-angiogenesis with hemorrhagic areas[12,22,23], and activation of osteoclastogenesis with the formation of multinucleated giant cells (MNGC)[24-28] are key features. USP6+ cells promote osteoclastogenesis through excessive production of receptor activator of nuclear factor kappa-B ligand (RANKL) at the TME level, particularly its soluble form, M-CSF, and via a deficiency of the natural osteoclastogenesis blocker, osteoprotegerin[24-28]. The RANKL/RANK/OPG cascade plays a key role in MNGC-mediated bone destruction in ABC. At later stages, microcirculation disturbances occur within the bone, leading to increased hydrostatic pressure and the formation of blood-filled cysts[29-31]. Another distinctive feature of ABC tumor biology, as with other USP6 mutated tumors, is its self-limiting growth[32], with a tendency toward spontaneous healing after minimal surgical interventions, such as biopsy[33-37].

Standard treatment involves curettage with or without local adjuvants and bone grafting[38]. In rapidly growing, locally advanced, and axial ABC, surgical treatment may be associated with severe intraoperative complications and long-term morbidity[39-41]. Multiple less invasive approaches have been investigated as preoperative or definitive treatment including sclerotherapy (ST)[42], selective arterial embolization (SAE)[43], cryotherapy[44], “curopsy”[37], and bone marrow injection[45]. A small proportion of patients with ABC are refractory to these treatments, necessitating new approaches targeting other pathogenetic mechanisms.

Two major classes of drugs targeting osteoclastogenesis are used for the treatment of locally advanced and inoperable ABC: Bisphosphonates (BPs) and RANKL inhibitors such as denosumab[46,47]. In most cases, such anti-resorptive therapy (ART) results in clinical improvement and tumor ossification. However, several issues are associated with this strategy, i.e. relapses after treatment cessation, acute side effects in children, and long-term toxicity in skeletally mature patients. Additionally, there are no generally accepted standards regarding the optimal management of patients after discontinuation of denosumab therapy (DT).

In a recent paper, we hypothesized that the use of BPs as maintenance following DT induction may have a protective role in giant cell tumor of bone (GCTB) due to the direct antitumor effect on mutated cells[48], and we formulated the concept of a two-stage treatment for locally advanced or inoperable GCTB. In light of the similarities in pathogenesis between GCTB and ABCs, and the frequent use of BPs for rebound hypercalcemia (RH) observed after discontinuation of DT in children, we were intrigued to explore the effect of sequential use of these two drugs to achieve long-term local control of ABC as well. In this paper, we have reviewed the existing literature on DT in ABC, focusing on the potential benefit of the additional use of BPs. The concept is further illustrated by two cases from our own clinical practice.

MATERIALS AND METHODS

Search strategies

A systematic literature search was conducted across five electronic databases: Scopus, MEDLINE, EMBASE, PubMed, and Web of Science, from inception to December 2024. The Medical Subject Headings (MeSH) and keywords used were “Aneurysmal bone cyst” and “Denosumab”.

Data extraction and charting

The systematic searches revealed 29 peer-reviewed articles describing the use of DT in ABCs. Due to the rarity of ABCs and the extreme rarity of their treatment with Denosumab[47], and recognizing the possible impact of this fact on the quality of the analysis, we included all reported cases regardless of patient age, disease stage, and treatment regimen. Among the 68 patients with pathologically confirmed ABC, DT was used in a neoadjuvant setting for 12 patients (17.6%). Nine patients (13%) were reported to have continued DT. In 46 cases (67.6%), DT was discontinued for different reasons, mainly due to stable clinical and imaging findings. Four patients had a progression-free follow-up time of < 6 months and were excluded from the analysis. Hence, 43 patients met the above inclusion criteria and were selected for review (Table 1)[49-64]. Extracted data included age, gender, tumor site, disease status (primary/refractory or recurrent), as well as treatments prior to DT. Upon initiation of denosumab, the following data were extracted: Duration of DT, clinical and radiological response, outcome after DT cessation, relapse-free survival after DT discontinuation, follow-up time, adverse effects in the form of RH, and treatment of this complication.

Table 1 Details in a published series of patients with aneurismal bone cyst treated with denosumab with or without bisphosphonates.

Ref.	Age	Gender	Localization	Status	Pre-denosumab treatment	DT duration	Response to DT	Group	Outcome after DTS	RFS	REC	Follow-up
Ghermandi et al[49], 2016	42	Male	Spine	PD	SAE	9	CR IR	1	SD	20	0	20
Ghermandi et al[49], 2016	16	Male	Spine	PD	SAE	7	CR IR	1	SD	16	0	16
Ntalos et al[50], 2017	35	Female	Pelvis-sacrum	R	SAE SURG DT 60 mg	17	CR IR	1	SD	15	0	15
Kurucu et al[51], 2017	16	Male	mandible	R	ST IT	11	CR IR	1	SD	20	0	20
Kurucu et al[51], 2017	17	Female	Pelvis	P		14	CR IR	1	REC CLIN ST	3	1	3
Kurucu et al[51], 2017	5	Female	Spine	P		12	CR IR	1	SD	12	0	12
Kurucu et al[51], 2017	6	Male	Pelvis	P		9	CR IR	1	REC CLIN HYST negative SD	6	0	20
Kurucu et al[51], 2017	8	Male	Humerus	R	SURG	12	CR	1	SD	12	0	12
Kurucu et al[51], 2017	16	Female	Spine	P		12	CR IR	1	REC DRCH	17	1	21
Kurucu et al[51], 2017	10	Female	Spine	P		6	CR PIR	1	REC SURGERY	6	1	24
Patel et al[52], 2018	16	Male	Spine	R	ST	12	CR IR	1	SD	12	0	12
Palmerini et al[53], 2018, Evangelisti et al[54], 2024	16	Male	L5-S1	P	SAE 5	41	CR IR	1	SD	24	0	24
Palmerini et al[53], 2018, Evangelisti et al[54], 2024	42	Male	Spine C7	P	SURG SAE 2	26	CR IR	1	REC DRCH ongoing SD 24 mos.	15	1	103
Palmerini et al[53], 2018, Evangelisti et al[54], 2024	25	Male	Spine Th10	P	ND	36	CR IR	1	REC MSC SD 26 mos.	20	1	56
Palmerini et al[53], 2018, Evangelisti et al[54], 2024	19	Male	Spine L3-L4	P	ND	10	CR IR	1	SD	33	0	33
Palmerini et al[53], 2018	12	Male	Proximal ulna	P		8	CR PIR	1	SD	12	0	12
Raux et al[55], 2019	8	Male	Spine	P		12	CR IR	1	REC DRCH ongoing SD 12 mos.	10	1	10
Raux et al[55], 2019	7	Female	Femur	R	SURG	4	CR IR	1	SD	32	0	32
Dürr et al[56], 2019	6	Male	Sacrum	P		12	CR IR	1	SD	6	0	6
Dürr et al[56], 2019	15	Female	Distal radius	PD	SURG ST	6	CR IR	1	SD	36	0	36
Dürr et al[56], 2019	16	Female	Distal femur	P	SURG SAE	24	CR IR	1	SD	24	0	24
Dürr et al[56], 2019	30	Female	Pelvis	R	SURG DT SURG	8	CR IR	1	REC DRCH ongoing SD	36	1	18
Dürr et al[56], 2019	18	Male	Sacrum	R	ST SAE	24	ND	1	REC DRCH ongoing SD 18 mos.	12	1	12
Dürr et al[56], 2019	16	Female	Talus	R	SURG	12	SD	1	REC as Ganglioma	24	0	24
Sydlik et al[57], 2019	6	Male	Femur	P		24	CR IR	1	SD	12	0	12
Kotaka et al[58], 2023	38	Male	Spine L3	P		3	CR IR	1	REC DRCH ongoing SD 37	17	1	37
Vanderniet et al[59], 2023	13	Male	Spine	P		12	IR	1	SD	36	0	36
Vanderniet et al[59], 2023	12	Female	Spine	P	SURG SAE	8	IR	1	SD	6	0	6
Evangelisti et al[54], 2024	20	Male	Spine C6-C7	P	SURG SAE 1	12	CR IR	1	SD	43	0	43
Kulkarni et al[65], 2019	14	Female	Spine	R	SURG	6	CR IR	1	SD	24	0	24
Evangelisti et al[54], 2024	26	Female	Spine C4	P	MSC 1	17	CR IR	1	REC DRCH ongoing 20 mos.	10	1	43
Kurucu et al[51], 2017	12	Male	Pelvis	P		14	CR IR	2	SD	10	0	10
Kurucu et al[51], 2017	16	Male	Sacrum	R	SURG	14	CR PIR	2	SD	12	0	12
Upfill-Brown et al[60], 2019	10	Female	Pelvis	R	SURG	11	CR IR	2	SD	13	0	13
Raux et al[55], 2019	8	Male	Spine	R	SURG SAE	17	CR IR	2	SD	6	0	6
Sydlik et al[57], 2019	11	Male	Sacrum	P		16	CR IR	2	SD	12	0	12
Harcus 2020[61],	13	Male	Proximal tibia	R	SURG n 3	27	CR	2	SD	16	0	16
Del Sindaco et al[62], 2021	8	Male	Spine	P		12	CR IR	2	REC 9 mos. DRCH DTS SD	15	0	15
Deodati et al[63], 2022	10	Male	Pelvis	P		10	ND	2	SD	19	0	19
Vanderniet et al[59], 2023	12	Male	Spine	P	SURG SAE ST	18	IR	2	SD	42	0	42
Vanderniet et al[59], 2023	10	Female	Spine	P		18	IR	2	SD	24	0	24
Vanderniet et al[59], 2023	13	Male	Spine	P	SURG SAE ST	12	IR	2	SD	12	0	12
Gandolfi et al[64], 2023	10	Female	Sacrum	R	SURG SAE	25	CR	2	SD	6	0	6

DT: Denosumab treatment; DTS: Denosumab treatment stopping; RFS: Recurrence-free survival time; REC: Recurrence; 0: No, 1: Yes; PD: Progressive disease; P: Primary tumor; R: Recurrent tumors; SAE: Selective arterial embolization; SURG: Surgery; IT: Immunotherapy; ST: Sclerotherapy; MSC: Mesenchymal cell therapy; CR: Clinical response; IR: Imaging response; PIR: Partial imaging response; ND: No data; SD: Stable disease; CLIN: Clinical recurrence; HYST: Histology; DRCH: Denosumab re-challenge.

Clinical response was defined as pain relief with or without reduction in tumor volume. Imaging response was defined as an increase in peripheral or internal bone mineralization in a radiograph or computed tomography (CT) compared with baseline values. Recurrence of the disease was usually defined as an increase in the size of osteolytic areas on radiography, CT, or magnetic resonance imaging (MRI) with or without the reappearance of clinical symptoms.

In the context of our meta-analysis, after discontinuation of DT, patients were divided into two groups. The first cohort consisted of patients who had no indications for BP use (Group 1). The second cohort included cases with RH who received BPs as a part of therapy (Group 2).

The significantly lower age in Group 2 (Table 2) can be explained by the occurrence of RH mainly in children.

Table 2 Demographics, tumor characteristics, treatment details and outcomes in two groups, n (%)/median (range).

Feature	Group 1 (n = 31)	Group 2 (n = 12)	P value
Gende			0.3
Male	18 (58)	9 (75)
Female	13 (42)	3 (25)
Age (years)	16 (5-42)	10.5 (8-16)	0.03
Site			0.2
Axial	23 (74)	11 (92)
Extremities	8 (26)	1 (8)
Primary/failed initial treatment/poor response	22 (71)	7 (58)	0.4
Recurrent	9 (29)	5 (42)
DT duration (months)	12 (4-41)	15 (10-27)	0.05
Clinical and/or imaging response to DT			0.9
Yes	27 (90)	10 (91)
No	3 (10)	1 (9)
Post-DT withdrawal follow-up (months)	16 (3-43)	12 (6-42)	0.4
Clinical and/or imaging and/or histological relapse			0.02
Yes	10 (32)	0 (0)
No	21 (68)	12 (100)

DT: Denosumab treatment.

Local control was assessed using the reported clinical, imaging, and in some cases pathological examinations. In cases of tumor recurrences following DT, additional data were extracted regarding the clinical management approaches used. To illustrate the role of ART in difficult-to-treat patients, we present two cases from our own clinical experience regarding the treatment of locally advanced ABCs, that failed to show clinical or radiological response after multiple cyst decompression procedures.

Statistical analysis

The medians and range of demographic information and outcomes were calculated and reported. The χ² test was used to compare recurrence rates between the two patient cohorts, with significance set at P < 0.05. The Kaplan-Meier method was used to estimate the survival function from lifetime data in both groups.

RESULTS

Study characteristics and demographics

There were 27 males and 16 females with a mean age of 15.8 years. (range 5-42), Tables 1 and 2. In 35 cases (81%), the tumor involved the axial skeleton, and in eight cases (19%), the tumor localization was in the extremities. Among axial lesions, the most common was the spine (n = 27), followed by the pelvis (n = 7) and mandible (n = 1). Twenty-nine patients had untreated tumors (17 cases) or were refractory to ongoing therapy (debulking surgery, SAE, and ST) (12 cases). The remaining 14 patients had recurrent lesions after initial intralesional surgery alone (7) or combined with SAE, ST, or DT (7). Thirty-one patients were included in Group 1, while Group 2 included 12 patients (Figure 1). Details of patient demographics, clinical features, treatment, and outcomes of both groups are presented in Table 2.

Figure 1 Cumulative survival estimates of local-progression free survival in ABC after denosumab withdrawal. Blue line: Group 1; Broun line: Group 2, (P value = 0085). The tables provide details of time, events, censored cases, and survival rates in both groups. A: Local-progression free survival; B: Details time, events, censored cases, and survival rate.

DT

DT was administered according to the GCTB protocol (denosumab 120 mg SC every 4 weeks with loading doses on study days 8 and 15) (57%) or monthly without loading doses (43%)[47]. The duration of the DT varied from 3 to 41 months (median 14). A good clinical and/or imaging response was documented in 37 of 41 evaluated patients (90%), while four patients (10%) had a partial response or stabilization of the disease (Table 1). The difference between the response rates in both groups was statistically insignificant (P = 0.9). In one case from our clinical experience, a 19-year-old girl with locally advanced pelvic ABC failed to respond to multiple decompressions and was treated with ART. Monthly DT for one year resulted in pain relief, pronounced tumor ossification, no pelvic dysfunction, and a good quality of life, Figure 2.

Figure 2 Aneurysmal bone cysts of the left pelvic bones. A: Frontal computed tomography (CT) scans before denosumab therapy (DT); B: Frontal CT scans 12 months after initiation of DT demonstrate more pronounced ossification and demarcation of the tumor margin; C: Planar bone scintigraphy after DT showing intensive tracer uptake in the ossified tumor (arrow). Bladder accumulating excreted ^99mTc-BP is dislocated to the right.

Side effects after denosumab discontinuation

In the context of our analysis, we focused on the incidence of RH during or after DT. RH alone or in combination with other side effects was observed exclusively in 15 (35.7%) skeletally immature children (80% male) with a median age of 10 years (range 6-16), Table 3. All episodes of RH occurred after discontinuation of denosumab or during DT over the extended treatment time.

Table 3 Side effects during and after denosumab therapy, n (%).

Side effects	Frequency
Hypercalcemia	12 (29)
Hypercalcemia + Growth plate ossification	1 (2)
Hypocalcemia + Hypercalcemia	1 (2)
Hypocalcemia + Hypophosphatemia + Hypercalcemia	1 (2)
Hypocalcemia + Hypophosphatemia	1 (2)
Vomiting	1 (2)
Not reported	26 (60)
Total	43

BPs therapy for RH

BPs therapy was a part of intensive care for RH, which also included diuretics, corticosteroids, calcitonin, intravenous infusions of 0.9% saline, 5% dextrose, and other solutions. As shown in Table 4, pamidronate (PAMI) was used alone in one patient and combination with zoledronic acid (ZA) in two others. ZA alone was administered in one patient and combined with risedronate in three patients. One patient was treated with neridronate and another four with unspecified BP regimens. The duration of BP exposure ranged from one to six months (median 2.5). RH was always resolved and the duration of BP therapy was relatively short (Table 4) with no side effects described as seen with long-term BP use in cancer or osteoporosis.

Table 4 Bisphosphonate therapy regimens of rebound hypercalcemia during and after denosumab treatment.

Ref.	Age/gender	Site	BPs therapy	BPs therapy in months
Kurucu et al[51], 2017	12/Male	Pelvis	BPs 1 dose	1
Kuruku et al[51], 2017	16/Male	Sacrum	BPs 1 dose	1
Raux et al[55], 2019	8/Male	Spine	BPs 3 doses	6
Sydlik et al[57], 2019	11/Male	Sacrum	Neridronate 2 doses (2 mg/kg)	1
Upfill-Brown et al[60], 2019	10/Female	Pelvis	BPs 1 dose	1
Harcus et al[61], 2020	13/Male	PT	PAMI 2 IV doses (0.25 mg/kg and then 0.5 mg/kg, 24 hours apart). PAMI one IV dose (0.5 mg/kg). ZA one IV dose (0.05 mg/kg)	2
Del Sindaco et al[62], 2021	8/Male	Spine	ZA six IV doses (0.05 mg/kg)	6
Deodati et al[63], 2022			PAMI one IV dose (1 mg/kg) in 1 week. PAMI five IV doses every 15 days (0.5 mg/kg)	3
Vanderniet et al[59], 2023	12/Male	Spine	ZA two IV doses (0.025 mg/kg), oral risedronate up to 6 months	Up to 6
Vanderniet et al[59], 2023	10/Female	Spine	ZA two IV doses (0.025 mg/kg), oral risedronate up to 6 months	Up to 6
Vanderniet et al[59], 2023	13/Male	Spine	ZA one IV dose (0.025 mg/kg), oral risedronate up to 6 months	Up to 6
Gandolfi et al[64], 2023	10/Female	Sacrum	PAMI two IV doses (1 mg/kg). ZA one IV dose (0.03 mg/kg)	1

PAMI: Pamidronate IV infusion; ZA: Zoledronic acid IV infusion; PT: Proximal tibia; BP: Bisphosphonates.

Another of our patients, a 16-year-old girl with recurrent and unresectable ABC of the metacarpal bones, also failed to respond to decompressions, Figure 3A. In order to convert the tumor into a resectable form, neoadjuvant therapy with BP was initiated. After six monthly PAMI infusions, peripheral ossification of the tumor allowed us to perform a function-preserving surgery without intraoperative complications and to achieve a long-term relapse-free interval, Figure 3B and C.

Figure 3 Rapidly growing recurrent aneurysmal bone cysts of metacarpal bones II-III after multiple decompressions and intralesional surgery with cement spacer placement. A: Radiographs before bisphosphonates (BPs) therapy with pamidronate; B: Radiographs after six BP infusions showing treatment-induced ossification; C: Radiographs after the removal of an ossified tumor with the reconstruction of metacarpal bones with free fibula autografts and microvascular anastomoses.

Local control

As is shown in Table 2, the groups were unbalanced in terms of age and duration of DT, with a predominance of children and longer DT in the BPs group. At a median follow-up time of 15.5 months (range of 3-43), there were 10 confirmed and two pathologically unconfirmed clinical and/or imaging relapses among the 43 patients who discontinued DT, Tables 1 and 2. All local relapses occurred in patients who did not receive BPs after DT withdrawal. Thus, the rate of local relapses in Group 1 and Group 2 was 33% (10/31) and 0% (0/12), respectively (P = 0.02). The median time to relapse was 13.5 months (range 3-36). Seven relapsed patients (70%) received denosumab re-challenge with stable tumors during ongoing DT. One patient was treated with mesenchymal stem cells and was progression-free during 56 months of follow-up. Two additional patients underwent surgery and ST for local recurrence. Among the 31 patients in Group 1, 21 patients were relapse-free at a median follow-up time of 20 months (range 6-43). Thus, all patients in Group 1 achieved disease control, seven of whom continued long-term DT at the time of publication of the articles. Interestingly, no local recurrences were observed among children treated with BPs for post-denosumab RH. Disease control in this group was maintained at a median follow-up time of 12.5 months (range 6-42). The difference in local control rates between the two groups was statistically significant when the χ² test was used. Kaplan-Meier estimates show the same trend, but due to the small sample size and short follow-up in Group 2, this difference did not reach statistical significance (Figure 1).

DISCUSSION

Locally advanced and rapidly growing ABCs are difficult to treat surgically and require new approaches to facilitate surgery or achieve definitive control in inoperable cases. Percutaneous procedures can fail in non-stabilized ABCs that are highly cellular, contain numerous MNGCs, and have minimal blood-filled cystic components. Pathogenetic treatment targeting the RANKL-RANK axis in ABC has recently been reviewed[47,65-67]. DT is used in difficult situations such as axial tumors often associated with high surgical morbidity and in cases refractory to standard treatments. However, several issues are associated with DT, including limited, if any, direct antitumor effect on mutated cells, as documented in GCTB[68,69], and frequent disease reactivation after treatment discontinuation. Moreover, there are no established guidelines regarding the optimal duration of DT and potential toxicity with long time use is a concern. In addition, RANKL inhibition in children is associated with a high risk of RH and metaphyseal sclerosis, making prolonged DT highly problematic in skeletally immature patients. This may imply shorter treatment duration and careful considerations of the risk-benefit ratio as well as modalities to prevent RH[47,59].

Vanderniet et al[59] reported that the majority of ABC patients (65%) discontinued DT for various reasons, mainly due to complete clinical and imaging response. In the current meta-analysis, we observed that among patients receiving DT alone, the relapse rate after drug withdrawal was 32%. In cases of ABCs relapse, re-challenge with denosumab is reported to provide additional tumor control[54] and disease stabilization rates reached 90%; however, this comes at the cost of long-term DT and the risk of late side effects[70,71]. To avoid long-term toxicity, various strategies have been proposed for DT discontinuation, such as drug holidays until relapse, extended dosing intervals, second-look surgery or mesenchymal stem cells therapy[54].

Thus, in a significant proportion (68%), of ABC patients (Table 2), long-term local control was achieved exclusively after RANKL inhibition, but the exact mechanism of this effect remains unclear. These positive results should be interpreted with caution since the post-denosumab follow-up period was relatively short. We cannot exclude that DT may delay relapse rather than prevent it, and with longer follow-up, relapse rates may increase. Residual USP6+ cells[72] are likely responsible for tumor reactivation after DT discontinuation in ABC. This scenario was also described in GCTB, which demonstrates significant but reversible responses to denosumab[70,71,73], and a longer follow-up is needed to draw a definitive conclusion in ABC.

BPs have been used as first-line ART in patients with locally advanced/inoperable ABCs of the axial skeleton. To date, several case reports and case series have described this approach as definitive treatment with positive results[46,74-76]. Simm et al[75] treated an 8-year-old boy with locally advanced lumbosacral ABC with ZA infusions at 0.04 mg/kg. The patient received seven doses at 4-month intervals over 24 months. Dramatic improvement in symptoms was noted after the first infusion. Although residual cysts were visible on MRI after two years of treatment, the patient remained asymptomatic at 12-month follow-up after stopping the infusions. Seven patients with axial ABC were reported in two overlapping papers[46,74]. After BP therapy, which lasted from 3 to 16 months (median 8 months), significant symptomatic and radiological improvement was achieved. Notably, at a median follow-up of 14 months (range 0-42), no local relapses were reported. Kumar et al[76] reported a 16-year-old male with recurrent ABC of the Th9 vertebra. He underwent SAE alone, following which his radicular pain markedly improved. Subsequently, he was also treated with BPs therapy, intravenous ZA at a dose of 0.04 mg/ kg every 4-month interval, for 1 year (3 doses). At 2-year follow-up with MRI and CT studies, there was complete bone formation within the lytic areas and pain relief.

The sustained tumor control during and after BPs treatment suggests that these drugs have a broader spectrum of action than just osteoclastic inhibition, potentially through direct antitumor action on neoplastic cells. In vitro and in vivo studies have shown that BPs, and ZA in particular, are the only class of drugs directly affecting the biology of the neoplastic stromal cell population in GCTB[77-81]. In addition to its apoptotic effect, it is speculated that BPs may also induce, directly or indirectly, differentiation of GCTB stromal cells[82,84] and stimulate osteogenesis through the formation of apoptotic bodies, which have anabolic effects on bone in vivo[84,85], possibly via reverse signaling through the vesicular RANK-mRANKL axis and/or MNGC-derived apoptotic bodies[86]. Supporting these experimental data, a recent meta-analysis showed that ZA reduced the recurrence rate after surgical treatment of GCTB and was therefore recommended after aggressive extended curettage[87].

The currently available experimental and still limited clinical experience with the use of BPs for the treatment of ABC has demonstrated the ability of these drugs to improve the clinical condition and induce significant and sustained tumor ossification, likely, by promoting the apoptosis of USP6+ cells. In light of the substrate-dependent effectiveness of BPs, the significant imaging responses even in these purely osteolytic lesions warrant further explanation. While this aspect plays a role in GCTB, it is unclear whether the same applies to ABC.

Although DT is well tolerated in most cases, lifelong treatment is not considered an optimal and safe option in benign bone tumors. Some concerns are the possible side effects associated with the prolonged treatment such as arthralgia, chronic muscle pain, peripheral neuropathy, fatigue, skin rash, electrolyte disturbances, osteonecrosis of the jaw, atypical bone fractures, and malignant transformation[70,71,88]. Various approaches have been used to reduce the cumulative toxicity of long-term DT in ABC and GCTB, including treatment discontinuation[71], “drug holidays”[89], and increased dose intervals[90-94]. Another approach in this context consists of using BPs alongside denosumab. This was described in our previous case report of a patient with GCTB, who achieved long-term disease control after denosumab induction and ZA maintenance[48]. In that study, we hypothesized that after denosumab-induced tumor ossification, higher local concentrations of an amino-BP may play a protective role by accumulating substrate to target its biological effect on mutated residual cells.

Given the limited data regarding the use of RANKL inhibitors in combination with BPs in ABC, we sought additional evidence in the current literature to support the viability of this concept. We identified a specific subgroup of ABC patients who discontinued DT and subsequently received BPs for treatment or prevention of RH. Interestingly, we found no reported relapses among patients who received BPs after DT discontinuation. Notably, the majority of patients in this group (75%) were followed for at least one year. Del Sindaco et al[62] reported a case of an 8-year-old male who had a relapsed, non-resectable ABC of C4-C7. Denosumab 70 mg/m² was administered for one year. After complete ossification of the tumor, DT was discontinued. Six months later, the patient developed severe RH that required the administration of one dose of ZA. Tumor evaluation performed 9 months after the end of denosumab showed an asymptomatic recurrence and DT was resumed for 2.5 years. Two episodes of hypercalcemia that occurred during the re-challenge were again treated with ZA. In order to prevent RH, six-monthly ZA were administered. At the last follow-up, 15 months after the last denosumab injection, the patient remained relapse-free. Vanderniet et al[59] reported another three post-DT cases with BP therapy aimed to prevent RH. The authors proposed to cease denosumab after 6 months if there was normalization of tumor metabolism on positron emission tomography/CT and spinal stability. They recommended that after DT discontinuation, BPs should be administered for at least 6 months. Progression-free intervals ranging from 12 to 42 months were achieved. A recent review published by Maximen et al[67] noted a statistically significant difference in relapse rates between adults and children. We speculate that the use of BPs in the pediatric population with RH may also have influenced these results. In the context of our analysis, this strategy can be rethought as a strategy to achieve the goal of definitive non-surgical control of inoperable ABC.

In support of this, we present a clinical observation from our own practice, which demonstrates that ossified post-DT ABCs accumulate considerable amounts of ^99mTc-BP as seen in diagnostic bone scintigraphy, Figure 2C. Such a case could conceptually be regarded as a good candidate for BPs maintenance. It is conceivable that high concentrations of an amino-BP in the ossified lesion may provide an anti-tumor effect on residual USP6+ cells, leading to sustained or definitive local control.

We view these results with caution and acknowledge that additional cases with longer follow-ups are needed to strengthen this promising trend and draw definitive conclusions.

Regarding the duration of BP therapy or its discontinuation, we believe that at least two scenarios can be considered today:

Long-term therapy with BPs at subsequently extended intervals (every 6 or 12 months with daily calcium/vit. D3 supplementation).

Discontinuation of BP treatment in case of complete clinical and imaging response with mandatory follow-up examinations every 3 months during the first two years of observation.

Given the therapeutic benefit of both groups of ART drugs, it is necessary to explore individual treatment strategies for inoperable ABC, taking the patient’s age, tumor location, skeletal stability, developmental phase, and clinical and imaging response into consideration. Based on the currently available clinical experience, we propose a potential treatment algorithm for locally advanced, progressive, and inoperable ABCs tailored to the age of the patient and clinical aggressiveness (Figure 4). Given their favorable safety profile when used in children, BPs can be regarded as a definitive treatment, except in cases when emergency surgery is required due to instability. In actively growing tumors with a minimal cystic component on MRI and when a rapid clinical response is required, treatment can be initiated with denosumab for a limited time, followed by BP maintenance. Continuing DT with prolonged intervals or stopping treatment with re-challenge in case of relapse seems less preferable due to the risk of complications. In skeletally mature patients with aggressive ABC, induction DT can initiate rapid tumor ossification. After approximately one year of treatment and/or when complete response is achieved based on clinical, neurological, imaging, metabolic, and biochemical data, maintenance therapy with BPs can be initiated.

Figure 4 Tentative algorithm of anti-resorptive therapy for locally advanced or inoperable aneurysmal bone cysts. ¹Immediate surgery is required in spinal tumors with instability or progressive neurological deficit; ²bisphosphonates (BPs) as definitive treatment; ³Denosumab for ≤ 6 months in an actively growing tumor; ⁴BPs to prevent rebound hypercalcemia and achieve long-term local control; ⁵Denosumab for ~ one year in an actively growing tumor; ⁶BPs to achieve long-term local control; ⁷Second-look surgery if deemed operable without high morbidity.

The optimal duration of BP therapy remains to be determined by further studies. In our opinion, maintenance therapy could be administered for approximately 6 months, which would avoid the side effects of long-term antiresorptive treatment.

During continuous ART, periodic reassessment of tumor resectability and surgical risks is recommended. To further stimulate ossification of residual ABC elements and increase the hardness of newly formed bone, ART can be used after minimally invasive procedures such as cyst decompression, “curopsy”, ST, SAE, etc., Figure 3.

At the end of our analysis, we formulate and put forward for discussion some unanswered questions: How does the self-limiting nature of ABC interact with the pathogenetic effects of ART and how can these two elements be optimally combined to achieve the best effect? What are the optimal time intervals for RANKL inhibition and BPs administration? What is the best post-DT treatment option? BPs maintenance, continuation of DT at extended intervals, or DT drug holidays with re-challenge in case of relapse.

Our review and meta-analysis have several limitations. First, the eligible studies comprising our literature review mainly consisted of case reports and small case series with considerable heterogeneity in terms of patient population and treatment regimens. The majority of patients in the BPs group were children who received longer-term DT. At the same time, it should be emphasized that both of these factors did not have a significant effect on the clinical and imaging response to DT, which makes it possible to exclude their influence on the results of BP maintenance. Secondly, there are currently no standardized protocols for the use of BPs in RH. In addition, there are differences in the affinity of BPs for bone tissue or in the dosing regimens of BPs, which may affect the outcome. Third, most local recurrences occur during the first two years of follow-up after DT withdrawal. Given the median follow-up time in the BPs group of 12 months, the difference in local control rates requires cautious interpretation.

The strengths of our review include the identification of all published cases of ABC treated with denosumab to date, with careful analysis of complications, in particular RH, and their treatment with BPs. In addition, all patients with a short follow-up period were excluded from the analysis. To the best of our knowledge, this is the first meta-analysis to examine relapse rates after discontinuation of DT and the role of BPs for the treatment of RH in the context of their impact on local control in ABC.

The novelty of our study lies in having a “critical look” into previously published papers/clinical series to scrutinize the support for our combination strategy previously published for aggressive GCBT[48] as a novel option for severe ABC cases. The trend toward achieving local control in denosumab-treated ABC using BPs suggests that these drugs might play a protective role, probably due to better accumulation of last-generation (nitrogenous) BPs in the mineralizing newly formed bone matrix. The internalization by cells of TME may be sufficiently high to provide a direct and irreversible effect on residual USP6+ cells.

CONCLUSION

This review and meta-analysis highlight the outcomes after DT cessation in all patients with ABCs not eligible for surgery reported in the scientific literature to date. A significant proportion of patients benefit from DT alone when other treatments have failed. The mechanisms of disease control by RANKL inhibitors in ABC remain unclear. However, approximately one-third of patients experience tumor reactivation and require either denosumab re-treatment or second-attempt surgery. BPs used in post-denosumab ossifying lesions appear to play a protective role in ABC, presumably by acting directly on tumor elements containing residual USP6+ cells. ART should be adapted to the patient’s age, clinical aggressiveness, and the degree of stability of the involved skeletal compartment. The optimal approach for withdrawal from standard doses of denosumab remains to be determined. Our hypothesis-generating analysis provides a rationale for further experimental and clinical studies aimed at incorporating this promising two-stage combined ART strategy into the clinical management of difficult-to-treat ABCs.