Giant cell tumour of bone around the knee: a systematic review of the functional and oncological outcomes

Introduction

Giant cell tumour (GCT) of bone is a relatively rare, locally aggressive benign neoplasm associated with a wide pathological spectrum, ranging from latent benign to highly recurrent and, occasionally, malignant metastatic potential. ^1–4 This tumour is commonly observed in the long bone epiphyseal-metaphyseal regions of young adults, with the most common age of onset being between 20 and 40 years old. The clinical symptoms are non-specific and may include local pain, swelling, and a limited range of motion (ROM) in the adjacent joint. Radiographs and magnetic resonance imaging (MRI) are the imaging modalities of choice for the diagnosis of this tumour. ^1–3 One of the most distinguishing clinical features of GCTs is the formation of extensive osteolytic lesions in the long bone epiphyseal-metaphyseal transitional areas. ^3,5 Histologically, mononuclear cells are the primary feature, and they determine the biological behaviour of this tumour. ^2,6,7 GCTs occur mainly around the knee joint, involving the distal femur or the proximal tibia. Nonetheless, this tumour can involve virtually any bone, with the distal radius, proximal humerus, or the proximal femur as other frequent locations. ^5,8

Wide resection and intralesional resection with curettage are two of the most common options for surgical treatment. When curettage is performed, local adjuvant therapies are often used to reduce the probability of a local recurrence, which is usually high. ^5,8 Observational data indicate higher rates of local recurrence with intralesional curettage in comparison to more extensive surgery, even with the addition of an adjuvant. ⁸ As such, and despite the common use of local adjuvants after curettage, the role of these adjuvant therapies remains controversial. It is important to note that GCTs are not sensitive to radiotherapy or chemotherapy, often used against other primary bone tumours. ^1,9

Distant metastases and malignant lesions may also occur in GCTs, but they are rare. ^1,2,4 Additionally, pulmonary metastases do not carry the same connotation as metastases associated with malignant tumours, such as sarcoma. In most cases, the clinical behaviour is benign, and metastatic disease does not lead to patient’s death. ^10,11

Methods

This study is a systematic review of published studies in the English-language literature concerning the outcomes obtained with surgical treatment of GCTs of bone located around the knee, namely those arising in the two most common anatomical sites: distal femur and proximal tibia.

To develop a systematic review, two electronic databases were used: Medline/PubMed and Scopus databases, using a search from 2000 to September 2020. We systematically searched for studies that included the keywords/MeSH words:(((“Osteoclastoma” OR “giant cell tumor” OR “giant cell tumour” OR “GCT”) AND (“proximal humerus” OR “distal radius” OR “proximal femur” OR “distal femur” OR “proximal tibia”)) AND (“outcomes” OR “surgical revision” OR “pain” OR “range of motion” OR “recurrence” OR “metastases” OR “infection”))). Our initial search included GCT of bone from all the most affected locations in the appendicular skeleton, from which we finally retrieved the studies that solely involved the distal femur, the proximal tibia, or both. The last search date was 24 September 2020.

For the inclusion criteria, we applied the Population, Intervention, Comparison, Outcome (PICO) strategy. We defined the following as:

1. Population: adult population (+18 years) with diagnosis of GCT of bone in the distal femur and/or proximal tibia.

2. Intervention: any given surgical technique with or without local adjuvants used to treat GCTs of bone.

3. Comparison: to compare the functional and oncological results obtained after surgical treatment of distal femur and/or proximal tibia GCTs.

4. Outcomes: local recurrence of the disease; ROM; infection; need for surgical revision; metastatic disease; functional status (using the Enneking/Musculo Skeletal Tumor Society (MSTS) score).

We included all prospective and retrospective observational studies, randomized controlled trials, case-controlled studies, and cohort studies. We excluded review articles, case report studies, articles with full text in languages other than English; articles with only abstracts available, and articles where the full text was not accessible. To achieve more consistency in the information retrieved, papers that did not include at least 10 patients and 24 months mean follow-up were also excluded. Whenever needed, the Enneking/MSTS score was converted to its percentage form to provide more homogeneous and comprehensive information for analysis.

Results

Article selection

A total of 851 articles (304 from PubMed/Medline and 547 from Scopus databases) were initially identified using the above-specified criteria. With the software available in rayyan.qcri.org we identified articles with more than 90% duplication, which were eliminated. After excluding the duplicated articles between both databases, 586 articles were pre-selected. The initial selection was followed by a title and abstract analysis for confirmation of the MESH keywords searched, the inclusion and exclusion criteria. The title and abstract analysis allowed us to further exclude 538 articles and, as such, 48 articles remained for full reading and analysis. After further reading we additionally excluded 34 papers due to conflicting information, lack of clarity of the information, extremely heterogeneous series, insufficient follow-up or series numbers, as these would have invalidated a proper analysis. For the final selection, we included 14 articles (Fig.1): two prospective and 12 retrospective studies.

View Full Size

Flowchart illustrating the search strategy and number of records screened and included.

Citation: EFORT Open Reviews 6, 8; 10.1302/2058-5241.6.200154

• Download Figure
• Download figure as PowerPoint slide

Discussion

Since GCTs arise most often in the long bone epiphyseal-metaphyseal regions, articulations are often involved, and even though GCTs of bone can arise in virtually any anatomical location, the distal femur and proximal tibia are by far the most common regions affected. ^5,8,29,30 Our choice for the subject of this systematic review was based on the above-mentioned facts, fed by the need to summarize the functional and oncological outcomes provided by the current practice.

Depending on the involvement of the articular surfaces, the tumour can be removed either by wide resection (Fig. 2) or curettage with or without local adjuvants (Fig. 3). ^1,9,31,32 As expected, and based on historic series, optimal outcomes arise when a tumour is removed with free margins, minimal morbidity and an acceptable functional outcome. ^13,33 In this setting, several studies report higher rates of local recurrence with intralesional curettage, even with the introduction of extensive curettage and local adjuvants. ^{8,30,32,34–39} In a large retrospective study of 384 patients with a decade of follow-up, recurrence rates were higher in patients treated with intralesional curettage versus those treated with wide excision (33% vs 2%). Nonetheless, among those patients treated with curettage, the addition of bone cement as an adjuvant decreased the local recurrence rate to 22%. ⁸ The use of a wide range of adjuvants after curettage, such as the local application of phenol, PMMA, liquid nitrogen or a combination of several options, is the consequence of a systematic attempt to minimize recurrence rates. The use of phenol and PMMA is probably the most established treatment to use following curettage, but without clear superiority compared to other options. ^30,34–37

View Full Size

Proximal tibia giant cell tumour (GCT) treated with wide excision and an allograft-prosthesis composite. (A) Pre-operative anteroposterior (AP) view; (B) pre-operative lateral view; (C) post-operative AP view; (D) post-operative lateral view.

Citation: EFORT Open Reviews 6, 8; 10.1302/2058-5241.6.200154

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Proximal tibia giant cell tumour (GCT) treated with intralesional curettage plus subchondral bone graft plus polymethyl methacrylate (PMMA) and internal fixation. (A) pre-operative anteroposterior (AP) view; (B) pre-operative lateral view; (C) post-operative AP view; (D) post-operative lateral view.

Citation: EFORT Open Reviews 6, 8; 10.1302/2058-5241.6.200154

• Download Figure
• Download figure as PowerPoint slide

Local recurrence is a devastating event, and, as such, all oncological series take this into account. Some papers in this systematic review documented low recurrence rates with curettage-based techniques. Abdelrahman et al are an example of this finding, reporting a 3.6% local recurrence rate in a prospective study including 28 patients. ²⁵ Another example can be found in the paper by Yu et al, who operated on 16 patients performing curettage and tumoral cavity filling with PMMA, complemented with internal fixation and oral bisphosphonates, without local recurrences detected. ²⁰ The scientific basis behind using oral bisphosphonates in this scenario would be the ability of this drug to stabilize local and even metastatic GCTs. However, the level of evidence for the efficacy of this practice is admittedly low and cannot be recommended as current practice. ^40,41 Saibaba et al also reported a very low recurrence rate despite the 16 Campanacci grade III GCTs in their series. ²⁸ Wu et al also found a similar result with a 3.7% local recurrence rate. ²¹ Taking into account all these findings, and excluding the report by Suzuki et al, the overall recurrence rate associated with intralesional curettage supplemented by adjuvants was below 15%, which contradicted higher recurrence rates often reported in literature. ^32,39,42 These results can be explained by the heterogeneity of patients included in the published series, since we often find different tumour grades, primary, and recurrent GCTs among them. The effort towards building a more homogeneous series in recent years can explain the progressively lower recurrence rates documented.

The results herein summarized for wide resections of GCTs regarding local recurrence support an extremely low local recurrence rate with this technique. The series reported by Natarajan et al included 143 patients with 129 tumours classified as Enneking grade III. Nonetheless, and after wide resection and reconstruction using a custom-made knee endoprosthesis, the recurrence rate was 0.7%. ¹⁷ Similar findings were reported by Li et al, with no recurrences after performing wide resections. ¹⁸

To manage local recurrences, an equal distribution between curettage-based techniques and wide resection could be observed. Regarding this topic, Balke et al reported acceptable re-recurrence rates after new curettage, burring and additional PMMA, and also Raskin et al, in their revision, considered additional intralesional curettage after recurrence, an appropriate procedure. ^32,43 Van der Heijden et al also support the possibility for curettage after recurrence at the expenses of a reasonable re-recurrence rate. ³¹

This review confirms GCT metastatic disease as a rare phenomenon. Only two studies reported on lung metastasis associated with GCTs, and in both cases, the metastatic rate was below 3%, with an overall metastatic rate estimation of 0.8%. In none of the cases, deaths were reported as being related to the metastatic disease.

For the majority of orthopaedic oncologists, the decision between curettage and wide resection in GCTs is based on the Campanacci classification, since there is a historical tendency to perform more aggressive surgery for grade III tumours. ^1,9,13,31,32 However, a grade III tumour does not preclude the curettage technique, and even though Campanacci grade III tumours seem to recur more often after curettage, the recurrence rates herein presented are often low. In this setting, it is important to highlight a trend toward pathologic fractures at presentation with Campanacci grade III tumours, as published by Saibaba et al, who reported five pathologic fractures in 36 patients. ²⁸ Chen et al also found five pathologic fractures at presentation, all associated with Campanacci grade III tumours. ¹⁵

One of the most difficult balances when operating on patients with GCTs is the surgical morbidity and future function of the joint. Herein, the best functional results were obtained when curettage-based techniques were used. ^22,23 Also, the limited analysis on the ROM achieved within distal femur and proximal tibia GCTs showed a mean ROM of 123° (range, 98–130°) for the distal femur and 124° (range, 96–136°) in the proximal tibia, which meant a similar outcome.

The MSTS scores published in the different series very often reflected good or excellent functional outcomes. ^24,25 Nonetheless, the worst MSTS scores were seen after wide resections. ¹⁹ In the end, the overall MSTS score found for the curettage group was 90.1%, while the wide resection group had a mean MSTS score of 85.6%. Additionally, a slightly better MSTS score was found for the proximal tibia when compared with the distal femur (90.6% vs. 87.2%).

This systematic review highlights a low overall infection rate with a 4.1% infection rate recorded for the curettage group and 5.3% in the wide resection group. These data were extracted based in only 10 papers, since four did not present information regarding infection. Among them, Zheng et al published the largest series on intralesional curettage procedures included in this systematic review, reporting a 5.8% infection rate. ¹⁴ Similarly, Natarajan et al reported the largest series of wide resections included in this review and observed a 6.3% infection rate. ¹⁷ Overall, and based on the findings gathered in this review, it is not possible to attribute a higher infection rate proportion towards wide resections, when compared with intralesional curettage. Additionally, a correlation between infection and anatomical location was also impossible to establish due to the low number of infections and the lack of information on this particular topic.

The need for surgical revision was common (Table 2and 3). Most often, the underlying causes were local recurrence, infection, or implant-related complications. In the study published by Yu et al, almost 50% of patients presented aseptic loosening and infections related to the endoprosthesis. ¹⁹ Natarajan et al also highlighted an important number of patients with similar complications (Table 3). ¹⁷ For Suzuki et al, the need for surgical re-interventions reached more than 40% of the series (13 patients), particularly due to local recurrence and fractures. ²⁷ In their large series, Teng et al found a considerable number of mechanical failures associated with reconstructions after curettage and local adjuvants. ¹⁶ These authors also found that reconstructions in the distal femur are three times more at risk for mechanical failure, when compared with the proximal tibia. The explanation could lie in the rotational force that is applied to the femoral condyles during the last 20° of knee flexion and extension. ¹⁶

Based on the available data, we estimated a minimum of 18% for new surgical procedures, which is an underestimation. This revision rate was higher among patients treated with wide resections (22.7%).

This systematic review represents a sum of the functional and oncological outcomes published regarding GCTs in the distal femur and proximal tibia. Overall, among 14 papers included, only two were prospective case series, with the remaining 12 being retrospective studies. Despite the limitations, this review allows us to gain important insights. First, there is a trend towards fewer local recurrences when wide resections are performed to treat GCTs. Nonetheless, the recurrence rates after curettage-based techniques seem to be lower than previously reported in the literature, and, as such, could be an acceptable approach even for Campanacci or Enneking grade III tumours and in the presence of pathologic fractures. Second, there is also a trend towards better function in the knee when curettage and local adjuvants are used. However, the results obtained with wide resection are also good, reporting acceptable ROMs and MSTS scores. In this particular setting, slightly better MSTS scores were achieved for GCTs that arose in the proximal tibia. Third, this systematic review did not find any evidence of higher infection rates associated with aggressive resections and reconstructions. Additionally, intralesional curettage and local adjuvants presented significant failures due to subchondral fractures and articular collapse, while reconstructions with tumoral endoprosthesis were prone to aseptic loosening or periprosthetic fractures. In the end, each surgical option has its pros and cons, but based on this review, we should promote, whenever possible, techniques that allow maximal preservation of the knee articulation.

Conclusions

The optimal treatment for GCTs arising around the knee remains controversial. Despite the trend towards better oncological control of this tumour after wide resections, the recurrence rates achieved through a more conservative procedure such as the curettage-based interventions seem to be acceptable. Additionally, curettage and local adjuvants allow for overall better function of the knee, in particular with proximal tibia GCTs.

There is a clear lack of high-quality studies regarding GCTs of bone to support a gold-standard option. As such, prospective studies comparing groups of more homogeneous patients, tumours, and treatment options, should be designed, to achieve more conclusive results regarding the optimal strategy to follow.