Introduction: Osteoclast-like multinucleated giant cells (MNGCs) are found in a number of soft tissue sarcomas including malignant fibrous histiocytoma and leiomyosarcoma. The nature of these MNGCs is poorly understood and the cellular mechanisms underlying their accumulation in sarcomas is not known.

Methods: We analysed by immunohistochemistry the expression of osteoclast, macrophage and smooth muscle markers by mononuclear and multinucleated cells in two cases of giant cell-rich leiomyosarcoma. We also characterised the role of mononuclear stromal cells and tumour-associated macrophages in the formation of MNGCs by RT-PCR, cell culture studies and immunohistochemistry/histochemistry for macrophage, osteoclast and smooth muscle markers

Results: MNCGs in giant cell-rich leiomyosarcoma expressed an osteoclast-like phenotype, being negative for smooth muscle actin and CD14 but positive for tartrate-resistant acid phophatase (TRAP), CD45, CD68 and vitronectin receptor (VNR). Scattered mononuclear cells expressing an osteoclast-like antigenic phenotype were also noted. An analysis of 25 conventional (non-giant cell-containing) leiomyosarcomas found isolated CD68+ MNGCs in three cases (approximately 12%); all of these cases were Grade-II/III leiomyosarcomas in which there was a prominent tumour-associated macrophage (TAM) infiltrate. Leiomyosarcoma TAMs isolated from two cases of conventional leiomyosarcoma and cultured in the presence of the osteoclastogenic factors RANKL and M-CSF differentiated into TRAP+/VNR+ MNGCs that were capable of lacunar resorption. RT-PCR studies showed that cultured leiomyosarcoma mononuclear stromal cells expressed RANKL, OPG and TRAIL.

Discussion: These findings show that the MNGCs which are found in leiomyosarcomas are osteoclast-like in nature and that these MNGCs are formed from TAMs by a RANKL dependent mechanism which involves an interaction with RANKL-expressing mononuclear stromal cells. A similar mechanism is likely to account for MNGC accumulation in other soft tissue sarcomas.

We report on a group of 20 metal-on-metal resurfaced hips (17 patients) presenting with a soft tissue mass associated with various symptoms. We describe these masses as pseudotumours.

All patients underwent plain radiography and fuller investigation with CT, MRI and ultrasound. Where samples were available, histology was performed. All patients in this series were female. Presentation was variable; the most common symptom was pain or discomfort in the hip region. Other symptoms included spontaneous dislocation, nerve palsy, an enlarging mass or a rash. The common histological features were extensive necrosis and lymphocytic infiltration. Fourteen of the 20 cases (70%) have so far required revision to a conventional hip replacement and their symptoms have either settled completely or improved substantially since the revision surgery. Two of the three bilateral cases have asymptomatic pseudotumours on the opposite side.

We estimate that about 1% of patients develop a pseudotumour in the first five postoperative years after a hip resurfacing. The cause of these pseudotumours is unknown and is probably multi-factorial, further work is required to define this; they may be manifestations of a metal sensitivity response. We are concerned that with time the incidence of these pseudotumours will increase.

Introduction: We report on a group of 20 metal-on-metal resurfaced hips (17 patients) presenting with a soft tissue mass associated with various symptoms; these masses we termed pseudotumours. All patients underwent plane radiography; CT, MRI and ultrasound investigations were also performed for some patients. Where samples were available histology was performed.

Methods: All patients in this series were female. Presentation was variable; the most common symptom was pain or discomfort in the hip region. Other symptoms included spontaneous dislocation, nerve palsy, a noticeable mass or a rash. The common histological features were extensive necrosis and lymphocytic infiltration. Fourteen of the 20 cases (70%) have so far required revision to a conventional hip replacement and their symptoms have either settled completely or improved substantially since the revision surgery. Two of the three bilateral cases have asymptomatic pseudotumours on the opposite side.

Conclusions: We estimate that about 1% of patients develop a pseudotumour in the first five postoperative years after a hip resurfacing. The cause of these pseudotumours is unknown and is probably multi-factorial, further work is required to define this; they may be manifestations of a metal sensitivity response. We are concerned that with time the incidence of these pseudotumours will increase.

We report 17 patients (20 hips) in whom metal-on-metal resurfacing had been performed and who presented with various symptoms and a soft-tissue mass which we termed a pseudotumour. Each patient underwent plain radiography and in some, CT, MRI and ultrasonography were also performed. In addition, histological examination of available samples was undertaken.

All the patients were women and their presentation was variable. The most common symptom was discomfort in the region of the hip. Other symptoms included spontaneous dislocation, nerve palsy, a noticeable mass or a rash. The common histological features were extensive necrosis and lymphocytic infiltration. To date, 13 of the 20 hips have required revision to a conventional hip replacement. Two are awaiting revision.

We estimate that approximately 1% of patients who have a metal-on-metal resurfacing develop a pseudotumour within five years. The cause is unknown and is probably multifactorial. There may be a toxic reaction to an excess of particulate metal wear debris or a hypersensitivity reaction to a normal amount of metal debris. We are concerned that with time the incidence of these pseudotumours may increase. Further investigation is required to define their cause.

The aim of this study was to observe the macroscopic and microscopic appearance of the Coracoacromial ligament and Subacromial bursa during Subacromial decompression and correlate it with the outcome at 3 months. Twenty patients with Subacromial Impingement without Rotator Cuff tear and five patients with large/massive irreparable Rotator Cuff tears who underwent a Subacromial Decompression. Patients with other shoulder pathology were excluded. Patients completed an Oxford Shoulder Score pre-operatively and their injection history was noted. At operation the shape of the acromion was noted. The macroscopic appearance of the CA ligament and the Subacromial bursa was classified as normal, mild/moderate and severe. Biopsies of the Subacromial bursa and CA ligament were taken and were analysed using histological and contempory immunocytochemical techniques. A histological analysis was performed using Mayer’s Haemotoxylin and Eosin, Toluidine Blue and Congo Red. Sections were stained with primary antibodies against PCNA (Proliferating cell nuclear antigen), Mast Cell Tryptase, CD3 (T-cell), CD20 (B cell), CD 34 (QBEnd 10), CD45 (Leucocyte Common Antigen), CD68 and D2–40 (Lymphatic Endothelial Marker). Post operatively the patients completed an Oxford Shoulder Score at 3 months. All the patients demonstrated an improvement in their Oxford Shoulder Score. The histological analysis demonstrated thickening of the synovial membrane and increased vascularity within the bursa and ligament. Increased numbers of inflammatory cells were present within the ligament and bursa of patients with impingement compared with massive rotator cuff tears. There was a relationship between outcome and the appearance of the bursa and ligament.

Purpose of study: We report the results of a prospective case series of 10 patients who developed tumour-like masses following resurfacing arthroplasty

Method: Ten subjects were referred to the tumour service at the Nuffield Orthopaedic Centre with symptomatic masses around the hip, all had previously received a resurfacing arthroplasty.

We report the clinical, radiographic and histologic features of these cases.

Results: MRI and ultrasound scanning was preformed, which demonstrated masses with solid and cystic components.

Biopsy was performed and subsequent histological examination revealed a profound plasma-cell lymphocytic response associated with metal wear debris.

There were no infections in this series.

Three subjects required revision surgery.

Conclusion: Over 50,000 resurfacing arthroplasties have been implanted worldwide over the past ten years. Although the early clinical results are encouraging little is known about the long term consequences of large head metal on metal bearing surfaces. Despite this, these devices are being widely marketed and are often implanted in younger patients. Resurfacing arthroplasties are associated with high serum and urine metal ion concentrations, metal particles have also been shown to migrate along the lymphatic system. In addition, there is now evidence that high local metal ion concentrations can induce haempoietic cancers.

This study suggests that resurfacing arthoplasty can also induce a local hypersensitivity reaction in response to metal wear debris. It therefore raises new concerns regarding the long-term safety of this procedure.

Antero-medial osteoarthritis of the knee displays a well recognised pattern of cartilage damage on the medial tibial plateau. Anteriorly there is a full thickness cartilage defect, with transition to a partial thickness defect, becoming full thickness in the posterior third of the plateau. The retained posterior cartilage is macroscopically normal, but no previous study has assessed its histo-logical features. This study characterises the histological changes, to examine if antero-medial OA of the knee represents a model of progressive osteoarthritic cartilage damage.

Five unicompartmental resection specimens of patients with idiopathic single compartment antero-medial osteoarthritis were assessed. The samples were stained with H& E and Saffinin-O stains and reviewed using the Mankin system, an established method for scoring osteoarthritic changes in cartilage (range 0 [normal] to 14 [grossly osteoarthritic]) Digital images of the histology were reviewed by two observers to exclude inter and intra observer error. Each specimen was assessed at 4 interval points (A,B,C,D) along the A-P axis starting from the most posterior aspect of the exposed bone to the area of macroscopically normal cartilage. Three repeat measurements were taken from the macroscopically normal region (D1,D2,D3). The scores were compared to historical age matched controls of non-osteoarthritic cartilage, where a Mankin grade of < 3 suggests normal cartilage.

From anterior to posterior the H& E staining showed a consistent decrease in structural integrity and cellularity of the cartilage, matched by a qualitative decrease in GAG content (Saffinin-O staining). Mean Mankin scores showed a progressive decrease in score; A = 14.0 (95% CI 0), B = 5.8 (95%CI 2.4), C = 4.4 (95%CI 2.5), D = 1.0 (95%CI 0.9) {p=0.04 ANOVA}. Repeated measurements at the macroscopically normal area showed the Mankin grade was maintained; D1= 1.0 (95%CI 0.9), D2 = 0.6 (95%CI 0.5), D3 = 0.6 (95%CI 0.6).

The results show that the retained posterior cartilage in antero-medial arthritis has a consistently normal Mankin grade. We suggest the defect represents a model of progressive cartilage damage from near normal (posterior) to the grossly osteoarthritic state (anterior).

The aim of this study was to observe cellular and vascular changes in different stages of full thickness rotator cuff tear.

Biopsies of the Supraspinatus tendon in 40 patients with chronic rotator cuff tears undergoing surgery were analysed using histological and contempary immunocytochemical techniques. Sections were stained with primary antibodies against PCNA (Proliferating cell nuclear antigen), CD34 (QBEnd 10), CD45 (Leucocyte Common Antigen), CD68, D2-40 (Lymphatic Endothelial Marker) and Mast Cell Tryptase. A histological analysis was performed with Mayer’s Haemotoxylin and Eosin, Congo Red and Toluidine Blue.

The reparative response and inflammatory component (figure 1) of the tissue was seen to diminish as the rotator cuff tear size increased. This was evidenced by increasing degeneration and oedema, reducing fibroblast proliferation, reduced thickening of the synovial membrane and reducing vascularity. Macrophage, other leucocyte and mast cell numbers also reduced as tear size increased. Large and massive tears revealed a higher degree of chondroid metaplasia and amyloid deposition when compared to smaller sized tears. There was no association with the patient’s age or duration of symptoms.

Small sized rotator cuff tears retain the greatest potential to heal and have a significant inflammatory component. Tissue from large and massive tears is of such a degenerate nature that it may never heal and this is probably a significant cause of re-rupture after surgical repair in this group. Selection of patients for reconstructive surgery should take into account the composition and healing potential of tendon tissue and its relationship to tear size in chronic tears of the rotator cuff.

A prospective study was carried out to determine if recognised histological features seen at surgery could help predict those rotator cuff tendon repairs which re-ruptured. 40 rotator cuff tendon edge specimens from 40 patients’ shoulders were analysed histologically following routine mini-open rotator cuff repair. 32/40 underwent Ultrasonography, at a mean time of 35 months post-operatively, to determine repair integrity. The histological features seen at surgery were then compared to the repair integrity of the tendon from which it had been taken. Rotator cuff repairs that remained intact demonstrated a greater reparative response, in terms of increased fibrobast cellularity, cell proliferation and a thickened synovial membrane, than those repairs which reruptured. Larger tears which remained intact showed a higher degree of vasacularity and a significant inflammatory component than those that re-ruptured. Good tissue quality at the time of surgery allows the repair the best chance of remaining intact despite the size of the lesion. Routine histological analysis of the tissue biopsy, preformed in the post-operatively, can now aid the clinician in terms of early management and repair prognosis.

Aim of Study: Assess clinical outcome and function of planned marginal excision of low grade liposarcoma of the forearm.

Material and Methods: Between 1997 and 2005 15 of 27 soft tissue sarcomas of the forearm were liposarcoma.

13 presented in the extensor compartment and 2 flexor compartment at the level of the distal radius. All presented with a painless mass. 5 patients with neurological symptoms. 4 involving the post interosseus nerve and 1 radial nerve. MRI was the diagnostic imaging technique of choice, 2 had biopsies where there was atypical imaging features.

Treatment and Results: All treated by planned marginal excision in view of proximity of neurovascular structures. The majority of tumours of the extensor compartment of the forearm were either involving or abutting the post interosseus nerve or neurovascular conduit.

All underwent planned marginal excision preserving juxtaposed peripheral nerve. There were no radial, spiral or PIN nerve palsies. One patient presented with PIN palsy had partial resolution of symptoms and function. I wound infection

Conclusion: Low grade lipoma-like liposarcomas have low metastatic potential. In the forearm a wide margin would mean ablation of critical neurological structures and planned marginal excision results in good function and to date no evidence of local recurrence at 2–9 year follow up.

Osteomyelitis commonly causes bone destruction and is most frequently due to infection by Staphylococcus aureus. S. aureus is known to secrete a number of surface-associated proteins which are extremely potent stimulators of bone resorption in the mouse calvarial assay system. The precise cellular and humoral mechanisms whereby this stimulatory effect is mediated, in particular whether osteoclast formation or activity is directly promoted by these factors, have not been determined by this study. Surface-associated material (SAM)(0.001ug/ml)obtained from 24 hour cultures of S. aureus was added to cultures of mouse and human osteoclast precursors (RAW 264.7 cells and human peripheral blood mononuclear cells respectively). These cultures were incubated in the presence and absence of receptor activator of nuclear factor kappa B ligand (RANKL) and macrophage colony stimulating factor (M-CSF). It was found that independent of RANKL, SAM was capable of inducing osteoclast formation in cultures of RAW cells and human monocytes. This was evidenced by the generation of tartrate-resistant acid phosphatase-positive multinucleated cells, which formed lacunar resorption pits when these cells were cultured on dentine slices. In cultures where M-CSF, RANKL and SAM were added, osteoclast formation was increased, but did not exceed the osteoclast formation in cultures with M-CSF and RANKL. These findings indicate that S. aureus produces a soluble factor which can promote osteoclast formation. Identification of this factor may help to develop therapeutic strategies for treating bone destruction due to Staphylococcal osteomyelitis.

Introduction: Osseous metastases from melanoma are relatively common (7% of cases), and occur most often in the axial skeleton. Bone destruction in skeletal metastases of solid tumours is due to stimulation of osteoclast formation and bone resorption. Osteoclasts are formed by the fusion of marrow-derived mononuclear phagocyte precursors which express RANK (receptor activator of nuclear factor κB) which interacts with RANKL expressed by osteoblasts/bone stromal cells in the presence of macrophage colony-stimulating factor (M-CSF). Osteoclast formation by a RANKL-independent, tumour necrosis factor α (TNFα)-induced mechanism has also been reported. Tumour-associated macrophages (TAMs) are present in both primary and secondary tumours and TAMs are known to be capable of osteoclast differentiation. Our aim in this study was to determine the role of TAMs and the humoral mechanisms of osteolysis associated with melanoma metastases.

Materials and Method: In this study we isolated TAMs from extraskeletal primary melanomas and lymph node metastases. TAMs were cultured for up to 21 days in the presence and absence of M-CSF and RANKL or TNF. In a separate experiment, conditioned medium was extracted from the melanoma cell line, SK-Mel-29, and cultured with human peripheral blood mononuclear cells in the presence of M-CSF.

Results: TAM-osteoclast differentiation, as evidenced by the expression of tartrate-resistant acid phosphatase, vitronectin receptor and lacunar resorption pit formation, occurred by both RANKL-dependent and RANKL-independent mechanisms. Osteoclast formation induced by RANKL-independent mechanism was not abolished by the addition of osteoprotegerin or RANK:Fc, decoy receptors for RANK. Conditioned medium from SK-Mel-29 could support osteoclast differentiation in the absence of RANKL. This effect was not abolished by antibodies to RANKL, TNFα, TGFβ, IL-8 or gp130.

Discussion: These results indicate that melanoma TAMs are capable of differentiation into osteoclasts and that both RANKL-dependent and RANKL-independent (TNFα) mechanisms are involved. Melanoma tumour cells also secrete a soluble factor that supports osteoclastogenesis.

Conclusion: Inhibitors of osteoclast formation targeting TAM-osteoclast differentiation and osteoclast activity and identification of the osteoclastogenic factor produced by melanoma cells may have a therapeutic potential in controlling tumour osteolysis.

We are presenting the outcome of a young adult with extensive epithelioid hemangioendothelioma of the femur treated with wide excision and vascularised fibular graft.

An 18-year-old builder was referred with an aggressive primary bone tumor of the right femur. Initial staging showed no evidence of distant disease but tumor confined to a 26.5cm diaphyseal segment of the femoral shaft. The patient’s pre-operative Oxford knee score was 28 and the AKSS scores were 74 (observational) and 65 (functional). True cut open biopsy confirmed low grade angiosarcoma. The patient underwent a wide excision of the lesion through a lateral approach leaving a generous cuff of bone and muscle tissue around the tumor. Clear resection margins were assessed intraoperatively. Histologically, the tumor was found to be epithelioid hemangioendothelioma. The 29.5cm defect was filled with a vascularised bone graft of the ipsilateral fibula. The graft was secured with a 22-hole DCS bridging plate and screws at both ends. Intraoperative knee range of motion was from 0 to 125 degrees without recurvatum and graft movement.

The patient had an unremarkable recovery. At the latest follow-up, one year after his operation, the patient had made an excellent functional recovery with non-symptomatic full weight bearing and had also returned to his work as a builder. He demonstrated a knee range of motion of 0 to 115 with a slight genu varum. The patient’s post-operative Oxford knee score was 40 and the AKSS scores were 70 (observational) and 90 (functional). Radiographs showed excellent union at the distal aspect of the graft and a healing stress fracture of the fibula graft at the proximal aspect.

Vascularized fibular graft with plating is a safe reconstruction limb salvage option for defects of long bones after tumor resection.

Purpose: To quantify the functional outcome of patients who were known to have sciatic nerve involvement pre-operatively and went on to have nerve preserving surgery utilising a planned marginal excision with epineurectomy.

Materials and Methods: Ten patients with large volume posterior thigh soft tissue sarcoma with known sciatic nerve involvement were reviewed between 1997 and 2004. Nine underwent surgery with extended epineurectomy of the sciatic nerve and planned marginal excision.

All patients underwent staging and follow up at Sarcoma Clinic with functional assessment and TESS evaluation.

Results: There were seven low and two high grade posterior thigh tumours of which nine were liposarcoma and 1 haemangiopericytoma. Two were recurrent and eight primary. There were five men and five women with a mean age of 77.

Nine patients underwent planned marginal excision. Sciatic nerve involvement was 13–30cm in eight cases and in one case the sciatic nerve was abutting the tumour throughout its length. There was soft tissue reconstruction in three cases using fascial adductor or gracilis graft for sciatic nerve cover and one with superficial femoral nerve and vein resection requiring ipsilateral saphenous reconstruction. The remainder underwent direct primary reconstruction.

Four patients underwent radiotherapy 46–60 Gy.

There was no local recurrence of disease within 14 – 96m follow-up. There was one patient with post radiation wound breakdown that resolved.

Three patients have died of unrelated causes. To date there has been no evidence of local recurrence of disease at FU.

Conclusion: Planned marginal excision of low grade large volume posterior thigh sarcomas with extensive sciatic nerve involvement can be successfully treated with preservation of the sciatic nerve without significant morbidity and resultant good limb function.

Aim: To review our series of mid foot sarcomas with regard to excision of tumour, tolerance of radiotherapy and preservation of function.

Methods and results: We identified 6 patients with mid foot sarcomas treated in our unit. Synovial sarcoma was the commonest diagnosis. All the patients had stage 1 disease with no evidence of pulmonary metastases at presentation. Patients judged to have resectable tumour but preserving sufficient foot to be functional were spared amputation. They had excision of the sarcoma and immediate reconstruction using fascio-cutaneous free flaps. Complete excision was achieved in all cases. One flap was lost and repeated. In all patients, subsequent radiotherapy was well tolerated without significant complications. All patients remain disease free. All patients have returned to pre-operative functioning including walking and jogging. All except one have returned to work.

Conclusion: Patients and feet treated by wide local excision of mid foot sarcomas and reconstructed by free fascio-cutaneous flaps tolerate post-operative radio-therapy well, and return to near normal function.

Between 1972 and 2002 74 patients were treated under the combined care of the orthopaedic oncology service and lymphoma clinic with primary bone lymphoma. We reviewed the seventeen cases affecting the upper limb (23%). Of the seventeen patients nine remain alive. Assessment of the patient’s clinical presentation, histopathological definition, treatment and function outcome was made. The nine survivors were assessed clinically and with the Oxford shoulder score and the Toronto extremity salvage score.

Average time from first presentation to diagnosis was 7 months. All seventeen were diagnosed as a B –cell non-Hodgkin’s lymphoma, fifteen cases were high grade and two cases were low grade. The scapula was involved in six, humerus eight and clavicle three cases. Seven patients sustained pathological fractures three of which were at presentation; of these two were treated surgically. Eight patients have subsequently died of their disease. Functional outcome in surviving patients after medical treatment was very good with average TESS score of 79% (52%–99%) and OSS of 27 (12–52).

The presentation of lymphoma of the shoulder girdle may mimic benign shoulder conditions and lead to a delay in radiological and histopathological diagnosis. Pathological fracture is a common presentation and complication of treatment, however these fractures have a high chance of healing with medical treatment alone. Although shoulder stiffness remains a problem following medical treatment, overall upper limb function is good. There is little evidence that these patients require surgery in the short to medium term.

Purpose: Bone destruction occurs due to the growth of primary malignant bone tumours (sarcomas) that are often not amendable to surgery. Bone resorption is carried out by osteoclasts which are formed from cells of the mononuclear phagocyte system. Primary malignant bone tumours contain tumour-associated macrophages (TAMs) in addition to neoplastic cells. The aim of the study was to determine the cellular and humoral conditions required for TAM-osteoclast differentiation and to assess the affect of an anti-osteolytic agent on osteoclastic bone resorption.

Methods: TAMs were isolated form bone and soft tissue sarcoma by collagenase digestion and cultured in the presence of RANKL and M-CSF on coverslips and dentine slices for up to 21 days. The extent of osteoclast formation and resorption was determined by expression of osteoclast markers (TRAP, VNR, cathepsin K) in cell cultures on coverslips and the extent of lacunar resorption in cell cultures on dentine slices.

Results: Osteoclast formation occurred only when RANKL and M-CSF were added to the TAM cultures. This resulted in the formation of numerous mononuclear multinucleated cells which were strongly TRAP, VNR and cathepsin K positive. In cell cultures on dentine slices, it was noted that these cells were capable of extensive lacunar resorption with formation of multiple large lacunar resorption pits. The addition of the bisphosphonate zoledronate to the cell cultures resulted in inhibition of osteoclast formation and complete absence of lacunar resorption.

Conclusion: These findings indicate that sarcoma-associated macrophages are capable of differentiating into osteoclasts and that both RANKL and M-CSF are required for this to occur. This process is likely to contribute to tumour osteolysis associated with the growth of sarcomas in bone. Further assessment of the use of inhibitors of osteoclast formation/resorption, is also indicated by our results.

Aims: To determine the pathological changes in the femur following resurfacing hip arthroplasty and identify possible causes of early failure. Methods: Bone samples from 8 femoral heads at several levels were examined histologically following removal of cemented femoral head surface replacement following aseptic early failure: 4 neck fractures (no history of fall), 3 persistent severe pain and 1 cup loosening. Intra-operatively no obvious macroscopic causes of failure (including notching the neck) were noted. In all patients, the initial diagnosis had been osteoarthritis. None had known risk factors for osteonecrosis. Results: In the patients who had recent fracture, the bony changes were suggestive of relatively longstanding osteonecrosis with degenerative, necrotic and þbrotic changes in the bone marrow and loss of osteocyte nuclei in the trabeculae. There was appositional new bone formation at the surface of the necrotic bone trabeculae. The changes were consistent with osteonecrosis of more than 2 weeks duration and probably preceded the fracture in all cases. In the patients who underwent revision for non-fracture, some osteonecrosis was seen, but this was a lot less than when a fracture had occurred. Conclusion: Osteonecrosis of the femoral head is seen following resurfacing hip arthroplasty and may be a predisposing factor in patients who subsequently fracture.

Giant Cell Tumour of the Tendon Sheath is a benign tumour of synovial origin most frequently affecting the upper limb. Up to 11% exhibit radiographic evidence of cortical erosion and intra-osseous expansion. In the upper limb recurrence rates of between 10–50% following excision have been reported. However, GCT-TS is rarely described in the foot and ankle and its behaviour is ill understood.

17 cases of this rarely described tumour in the foot and ankle are presented, describing their clinical presentation, histopathology, treatment and outcome.

Analysis of all cases of histopathologically proven GCT-TS of the foot and ankle from the Oxford Tumour Registry, was conducted between the periods of January 1984 to December 1999.

22 cases were identified of which 17 cases had adequate records to allow analysis of patient demographics, duration of symptoms, preoperative investigations, presumed diagnosis, precise site of origin, post operative complications and recurrence rates

The mean age of presentation was 28 (8–53). 10 cases were female and 7 male. 76% cases occurred in the foot, all of which arose adjacent to the phalanges or heads of the metatarsals. 14% occurred in relation to the ankle or sub-talar joint.

82% presented with a painless swelling. The duration of symptoms ranged from 6 months to 8 years. Only one patient complained of sensory symptoms.

Pre-operative investigations included radiographs in 64% with 3 cases having an additional MRI scan. The MRI scans of GCT-TS have characteristic changes on T1 and T2 images. The presumed preoperative diagnosis was incorrect in 82%.

36% of radiographs taken showed changes including cortical erosion and speckled calcification.

A local excision was performed in 15 cases, an amputation in one and a wide local excision in one case only. There have been no recurrences during the follow up period of between 1–12 years.

GCT-TS of foot and ankle is rare and is commonly misdiagnosed. Despite only a local excision being performed in more than 80% of this series there were no recurrences.

Plain radiographs may show cortical erosion or speckled calcification in up to 36% and MRI is helpful in further defining the anatomy of the lesion, allowing planned excision and reducing the risk of recurrence.

Aseptic loosening is generally associated with the presence of wear particle-associated macrophages in the pseudomembrane commonly formed around failed prosthetic implants. The extent of the macrophage response evoked by the wear particles has been shown to correlate with the amount of periprosthetic osteolysis. Numerous studies have shown that wear particle-associated macrophages contribute to osteolysis by (i) releasing inflammatory cytokines and/or (ii) differentiating into bone resorbing osteoclasts. Although macrophages and macrophage polykaryons are the main inflammatory cells found in periprosthetic tissues, numerous fibroblasts are also present in the connective tissue pseudomembrane. The recently identified molecule, RANKL has been shown to play a central role in the osteoclast formation and bone resorption observed in aseptic loosening. We have shown that arthroplasty macrophages, which express RANK, the receptor for RANKL, are capable of osteoclast differentiation; this process is inhibited by osteoprotegerin (OPG), the soluble decoy receptor for RANKL. As fibroblasts are known to express RANKL, the aim of the present study was to determine whether fibroblasts isolated from periprosthetic tissues could induce the generation of bone resorbing osteoclasts that would contribute to the osteolysis commonly seen in the periprosthetic loosening.

Fibroblast-like cells were isolated from pseudomembrane from patients (n=5) undergoing hip revision due to aseptic loosening, by routine collagenase enzyme digestion. The isolated cells were seeded in flasks for 2–4 weeks before being passaged for a further 3–4 times. Generated fibroblast-like cells (10⁴) were then co-cultured with 5x10⁵ normal human peripheral blood monocytes (n=5) on glass coverslips and dentine slices in the presence of (i) no added factors, (ii) macrophage colony stimulating factor (M-CSF) and (iii) M-CSF plus OPG. All cultures were maintained for 1,17 and 21 days. The extent of osteoclast differentiation was then determined by the expression of specific osteoclast markers including tartrate-resistant acid phosphatase (TRAP) and vitronectin receptor (VNR) and evidence of lacunar resorption.

In the absence M-CSF, no osteoclast formation was noted in 24 hours, 17 or 21 days in fibroblast/monocyte cultures. However, in the presence of M-CSF alone, large numbers of TRAP⁺ and VNR⁺ multinucleated cells capable of lacunar resorption were noted in these co-cultures. The addition of OPG, which is known to inhibit RANKL-mediated osteoclast formation, significantly reduced the extent of osteoclast formation and lacunar resorption in these co-cultures.

These results indicate that one means whereby peri-prosthetic osteolysis may occur is by fibroblasts in the arthroplasty pseudomembrane inducing macrophage-osteoclast differentiation. Fibroblasts express RANKL and interact with arthroplasty macrophages, which express RANK and function as osteoclast precursors. These findings indicate that suppression of osteoclast formation by OPG may be a possible form of therapy for reducing prosthetic loosening.