Retrospective analysis of 51 intralesionally treated cases with progressed giant cell tumor of the bone: local adjuvant use of hydrogen peroxide reduces the risk for tumor recurrence

Giant cell tumor of the bone (GCT) is a challenging orthopedic disease representing 10–15% of all benign and 4–5% of all primary bone tumors [1]. GCT has unpredictable biological behavior with locally destructive growth, frequent local recurrences, multifaceted histological appearance, and potential of pulmonary metastases [2‐4]. Sixty to eighty percent of all GCT are found in the third and fourth decade of life. Women have a higher risk than men [5]. At the time of initial diagnosis, pathological fracture with clinical symptoms of instability is often already evident [2, 6]. Systemic therapy with denosumab has been introduced as a new encouraging tool to limit activity and growth of GCT but has also been discussed controversially because of unclear benefits and risks [4, 7‐9]. Surgical treatment remains the gold standard but new surgical treatment strategies have not been established for a long time, and the literature reports highly heterogeneous study results with inconsistent data [5, 7‐13]. Avoiding local recurrence and achieving the best functional outcome remains a hard challenge. Despite recurrence rates up to 65%, intralesional resection with joint preservation and vigorous curettage is most often preferred due to less morbidity compared to extralesional resection [3, 5, 14‐16]. Most authors prefer polymethylmethacrylate (PMMA) bone cement filling after curettage instead of cancellous bone, but evidence is inconsistent [13‐21]. Cell-toxic local adjuncts such as phenol can reduce the risk for recurrence [22]. Multiple other local adjuvants have been described such as liquid nitrogen, alcohol, iodine, cyclophosphamide, and especially hydrogen peroxide, but due to their heterogeneous use, the effectiveness on recurrence rates and recurrence-free survival has never been proven for one of these individual substances, so far [5, 18, 22, 23].

The aim of the study was to retrospectively analyze recurrence rates, cumulative recurrence-free survival rates, metastases, complications, and functionality and satisfaction after surgical treatment of progressed but intralesionally treated GCT cases. As a novelty of this study, we focused on the additional effect of adjuvant local hydrogen peroxide treatment. For the first time, the present study presents proof of principle in a clinically relevant in vitro analysis of hydrogen peroxide effects on GCT cells together with retrospective clinical data of progressed GCT cases either treated with hydrogen peroxide or not.

The present series analyzed 51 advanced GCT cases of the upper and lower extremities with a majority of Campanacci grade III tumors. Contrary to other studies, which could not prove the additional effect of hydrogen peroxide due to its heterogeneous use, we for the first time could demonstrate that additional adjuvant local hydrogen peroxide application after curettage significantly reduced recurrence rate, increased cumulative recurrence-free survival, and was a significant risk factor in multivariate analysis. Further in vitro analysis confirmed hydrogen peroxide-induced apoptotic cell death of GCT cells as the underlying mechanism.

The literature describes heterogeneous use of local adjuvants to further clean the GCT cavity and kill potentially remaining GCT cells before filling the defect with bone cement or cancellous bone [5, 18, 22, 23]. Phenol was mainly used but is abandoned due to serious systemic toxicity and carcinogenic potential also for the surgeon by inhalation. Multiple other adjuvants have been described with, however, limited evidence due to rare and heterogeneous usage. Besides the filling with bone cement itself, local adjuvants can be divided into thermal adjuvants (cryotherapy with liquid nitrogen; cauterization) and toxic chemical substances such as alcohol, iodine, cyclophosphamide, phenol, and hydrogen peroxide [5, 18, 22, 23]. Hydrogen peroxide is a cheap and easily available adjunct which has been reported as an alternative for phenol with proven in vitro effects against GCT cells [26, 27], but its clinical value has not been proven so far [5, 18, 22, 23]. The present study documents statistically significant clinical effects of hydrogen peroxide.

In the present study, we for the first time tried to transfer the clinical intraoperative scenario of high-dose short-time hydrogen peroxide application also to in vitro analyses. Hydrogen peroxide is expected to kill potentially remaining tumor cells in the tumor cavity after aggressive curettage. These tumor cells are not macroscopically visible but potentially scattered and attached at the surface of the cavity. Hence, in vitro-isolated GCT cells attached at the well plate might be an appropriate target to test whether short-time application of highly concentrated hydrogen peroxide solution can diminish viable tumor cells in order to prevent later recurrence. After initial short-time treatment of the attached cells with hydrogen peroxide, cells were further incubated in DMEM for 24 h without hydrogen peroxide to allow recovery. This may resemble potential later tumor cell recovery after hydrogen peroxide is diluted and removed from the tumor cavity. Apart from general limitations based on the artificial setup of in vitro studies, our data support detrimental hydrogen peroxide effects on GCT cells. We found reduced cell viability and activated apoptotic pathways both after high-dose short-time application and after longer incubation times with lower hydrogen peroxide concentrations.

A further clinical advantage of hydrogen peroxide use was improved cleaning of the cancellous bone margins with better visibility compared to only cleaning with salt solution. Hence, the surgeon may detect and resect residual tumor tissue more easily. As a result of the present study, we generally recommend adjuvant local use of easily available hydrogen peroxide to finally clean the tumor cavity before filling the defect.

In agreement with most other studies [5, 16‐19, 21], our data also support that bone cement filling reduces recurrence rate compared to cancellous bone filling, which can be explained by heat destruction of potentially remaining tumor cells during the exothermic polymerization of polymethylmethacrylate. Another advantage is better radiological identification of potential recurrences after bone cement filling [5]. The beneficial effect of bone cement filling, however, was not confirmed by some other studies [15, 22, 23]. The majority of our cases (45) had bone cement fillings, whereas only six cases had cancellous bone fillings, which limits statistical analysis. Statistical significance was only achieved for the recurrence rate analyzed by chi-square tests. Differences in cumulative recurrence-free survival did not reach statistical significance.

Age, gender distribution, localizations, and recurrence rates of our series are compatible with the literature. Our overall high recurrence rate can be explained by the majority of initially complicated and advanced cases in our university institution, specialized in orthopedic oncology, as well as by the cases initially treated in other surrounding hospitals and referred to us for further follow-up due to their complexity. Another reason will be our philosophy of avoiding extralesional wide resection with endoprosthetic reconstruction in the young GCT population as much as possible and even in cases with partial joint-line infiltration, due to long-time hazards related to a tumor mega prosthesis. Some authors, however, prefer wide resection in advanced Campanacci grade III tumors [4] to minimize recurrences and avoid multiple surgeries [5, 16]. Improvements of survival and function of today’s tumor endoprostheses need further long-time analysis.

Histologically, osteoclast-like multinuclear giant cells and mononuclear neoplastic stroma cells influence tumor activity [1, 28], but histological classification of GCT was not reliable in predicting prognosis [5, 15, 29]. We also evaluated mitosis rate of the mononuclear cell population and found that higher mitosis rate was associated with significantly lower recurrence-free survival and higher hazard ratio in multivariate analysis, which is compatible with another report [21].

Molecular pathogenesis of GCT [30, 31] will influence future therapeutic strategies. Adjuvant or neo-adjuvant systemic treatment with denosumab, a RANKL inhibitor, is increasingly used for cases where surgical resection would cause inappropriate morbidity [32]. The present study did not analyze denosumab treatment which might improve outcome. From the authors’ experience, however, denosumab will be more appropriate for advanced tumors in the pelvis and spine where complete surgical tumor removal is not possible without mutilation. Even higher postoperative recurrence rates in cases with neo-adjuvant denosumab treatment have been reported [33], which might be related to tissue remodeling with less obvious tumor margins after denosumab treatment making complete tumor removal more difficult for the surgeon. These effects, however, are controversial as beneficial reduction of tumor size was reported after denosumab treatment [34] and randomized trials have been initiated for further evaluation [8]. Optimal duration of denosumab treatment also remains unclear as well as the risk for the induction of secondary osteosarcomas [7, 8, 33, 34].

We observed higher recurrence rates and lower cumulative recurrence-free survival in cases, where osteosynthesis was avoided. A possible explanation might be more vigorous curettage and improved visibility in cases with additional stabilization via osteosynthesis, as the surgeon might have been less concerned about intra- or postoperative fracture.

Important limitations of the present study have to be acknowledged. Compared to some other studies, absolute numbers of cases were low and depending on the different parameter, data was not available for all cases. Analysis was retrospectively performed and hydrogen peroxide was heterogeneously used with potential bias. As an additional cell-toxic adjunct will be more likely used in complicated cases, the beneficial effect on recurrence rate will be, however, rather underestimated than overestimated. For final evaluation, further studies are needed with higher patient numbers, potentially prospective study design and specific selection criteria for hydrogen peroxide use. Nevertheless, the strength of the present study is proof of principle for hydrogen peroxide effectiveness by both in vitro cell culture studies and clinical analyses with sufficient numbers of equally treated patients with and without hydrogen peroxide use.

For the first time, this study proves the clinical effectiveness of adjuvant local hydrogen peroxide to clean the cavity before defect filling. Hydrogen peroxide treatment significantly reduced recurrence rate and increased cumulative recurrence-free survival. Reduction of viable GCT cells by induction of apoptosis analyzed in vitro supports this.