The treatment of diffuse tenosynovial giant cell tumor (TGCT) requires extensive surgical resection of the hypertrophic synovium and multiple soft tissue masses yet still may result in high rates of local failure. The authors of this report examined their experience in treating patients with advanced/multiply recurrent TGCT with a combination of surgery and external-beam radiotherapy.
Fifty patients who were treated for TGCT with radiotherapy and surgery from 1972 to 2006 were identified. Patient demographics, radiotherapy treatment parameters, surgical treatment, and oncologic and functional outcomes were evaluated. All patients had pathologic review at presentation and required at least 1 year follow-up.
Forty-nine patients had diffuse TGCT with both intra-articular and extra-articular disease (1 had malignant TGCT). Twenty-eight patients (56%) were referred after at least 1 local recurrence. Thirty patients (60%) underwent at least 2 operations before radiotherapy. The mean dose of radiation delivered was 39.8 gray. At a mean follow-up of 94 months (range, 19-330 months), 47 patients (94%) had not developed a recurrence or had stable disease/signal characteristics on serial cross-sectional imaging (for those patients who had gross residual disease at the time of radiotherapy). Two patients required subsequent total hip arthroplasty because of progressive osteoarthritis, and there were 4 cases of avascular necrosis (only 1 post-treatment). Forty-one patients had good/excellent function.
Tenosynovial giant cell tumor (TGCT) (previously named pigmented villonodular synovitis [PVNS]) is a rare monoarticular proliferative process originating from the synovial membranes.1, 2 It is unclear whether it is a reactive2-4 or neoplastic process,5-8 and cases of malignant TGCT/PVNS have been reported.9, 10 It can involve any joint but is located most commonly around the knee. It generally occurs in 2 forms, which probably are different manifestations of the same disease spectrum11: a localized form (LTGCT/PVNS) characterized by a solitary, pedunculated lesion surrounded by normal synovium, and a diffuse form (DTGCT/PVNS), which involves the majority of the synovium of the joint. TGCT/PVNS also can extend to the extra-articular tissues in a destructive manner, often involving local structures, such as muscle, tendon, skin, bone, and neurovascular bundles. We consider multiply recurrent, diffuse TGCT/PVNS as well as TGCT/PVNS with extra-articular extension, particularly when there is additional bone or soft tissue involvement, as forms of the disease at high risk for local tumor relapse.2, 12, 13
Complete or marginal excision of the nodule in LTGCT/PVNS generally is adequate for long-term control.14 However, the infiltrative and more widespread nature of DTGCT/PVNS requires extensive surgical resection of the hypertrophic synovium and multiple soft tissue masses still may result in high rates of local failure.13, 15 Refractory and/or multiply recurrent cases can lead to a serious impairment of functional ability and may necessitate joint replacement and possibly amputation for an essentially benign disease.12 Adjuvant radiotherapy has been used in a few small series in the literature delivered either with external beam16-22 or through the instillation of intra-articular radiocolloids.23-26
We previously reported preliminary results from treatment with external-beam radiotherapy for advanced or multiply recurrent and symptomatic TGCT/PVNS.27 The objective of the current study was to update this experience with extended follow-up and additional patients at high risk for recurrence with functional loss, including instances in which amputation was the sole alternative for symptomatic disease.
MATERIALS AND METHODS
After we received Research Ethics Board approval, we retrospectively reviewed our prospectively collected tumor database and the radiotherapy records at our institution. Eligibility included at least 1 year of follow-up after treatment. Between 1972 and 2006, 108 patients were treated for TGCT/PVNS. Fifty-eight patients underwent surgery alone (for either LTGCT/PVNS or DTGCT/PVNS with no extra-articular involvement) and were excluded, and 50 patients received radiotherapy with or without undergoing surgical excision and formed the basis of the study. Forty patients underwent surgery and received radiotherapy at our center, whereas 10 patients received radiotherapy after undergoing surgery at another institution, either because the prereferral surgery was deemed adequate or because further surgery would have been excessively morbid without decreasing the local recurrence risk. All patients had their disease confirmed pathologically at presentation to our center. External-beam radiotherapy was used to deliver the radiotherapy treatment in all patients with the exception of 1, who received 5 millicuries (mCi) of intra-articular 198Au. Radiation treatment used photons consisting of Cobalt-60 irradiation in 17 patients, 4-MV radiation in 4 patients, and 6-MV radiation in 28 patients. The techniques in all patients involved treatment of the entire circumference of the limb with the involved joint for a major component of the treatment course. Smaller volumes were not feasible without omission from the target volume of known disease or tissues at risk. Treatment fields (and, more recently, clinical target volumes) were defined to cover 5 cm superior and inferior to the current or original site(s) of overt disease, as evidenced by clinical, radiographic, and operative descriptions. Select application of a full thickness tissue equivalent bolus of the skin was used in all patients with skin involvement or if it was believed that superficial tissues were at risk when using skin-sparing megavoltage techniques. Radiotherapy treatment courses consisted of a single fraction per treatment session and treated all prescribed fields each day.
Data extracted from the database and radiotherapy records included patient age, sex, whether disease was primary or recurrent, anatomic location, TGCT/PVNS subtype and origin, number and size of lesions, prior treatment, history of trauma, radiation treatment technique, radiation dose and number of fractions, type of surgery, local control, and current disease status. Local control was defined as: 1) absence of clinical or imaging evidence of disease for at least 1 year after treatment in patients without overt disease at the time of radiotherapy or 2) stable disease on serial cross-sectional imaging for at least 1 year after treatment for patients with residual gross disease at the time of radiotherapy (Fig. 1).
Evaluation of function was determined from the clinical descriptions in the medical records. Four functional categories were used: 1) excellent if the joint and effected limb had no objective or subjective alteration related to the disease or treatment; 2) good if minor abnormalities persisted after therapy, including mild intermittent swelling or joint stiffness, or if the function was not described explicitly as excellent; 3) fair if persistent symptoms of discomfort or joint dysfunction existed but did not interfere with recreational or vocational activities; and 4) poor if persistent limb abnormalities interfered with normal life style or required regular analgesia or other medication. A small subgroup of patients also completed the Toronto Extremity Salvage Score (TESS), a patient-completed measure of physical disability that was developed specifically for the sarcoma population, at their latest follow-up.28 They were also evaluated with the clinician-completed Musculoskeletal Tumor Society (MSTS) functional rating systems (MSTS 1987,29 MSTS 199330). The TESS evaluates a patient's perception of difficulties in performing activities of daily living and is scored from 0 to 100, with a higher score denoting less perceived disability. The MSTS 1987 is a 7-item, joint-specific measure of impairment and is scored from 0 to 35. The MSTS 1993 is a 6-item, limb-specific measure and is scored from 0 to 100. For both MSTS scoring systems, a higher score denotes less impairment.
The 30 women and 20 men had a mean age of 37.9 years (range, 18-74 years) (Table 1). The majority of tumors were located around the knee joint (20 of 50 tumors; 40%). Forty-two tumors arose from the joint, and 8 tumors originated from the periarticular tendon sheath. Forty-seven of 50 patients had both intra-articular and extra-articular disease. Forty-nine of 50 tumors also were the diffuse (DTGCT) subtype, the lone exception was the only malignant TGCT/PVNS in the series. Nineteen tumors measured ≥10 cm in greatest dimension (mean, 8.4 cm; range, 1-21 cm). Twenty-eight patients (56%) were referred to our center after at least 1 local recurrence; whereas, for 22 patients (44%), it was their primary presentation. Only 12 patients (24%) reported a prior history of trauma to the involved joint.
Table 1. Patient and Tumor Demographics
No. of Patients (%)
Mean age [range], y
Tumor size, cm
No. of surgeries
Thirty patients (60%) underwent ≥2 surgeries (range, 2-5 surgeries) before receiving radiotherapy. Sites other than the knee generally were managed by open synovectomy. For the knee (n = 20), in the early years, open anterior and posterior synovectomy was undertaken; but, more recently, anterior arthroscopic and posterior open synovectomy have been undertaken because of frequent extracapsular extent of disease. The final type of surgery performed was varied and was determined by the location of disease in relation to the joint, how accessible it was to complete removal, and the anatomic location involved: These included 21 single-incision open synovectomies, 14 double-incision open synovectomies, 6 combined open/arthroscopic synovectomies, 2 two-stage open/arthroscopic synovectomies, 2 arthroscopic synovectomies alone, 1 margin-negative excision, and 4 debulking or stabilization procedures only. Gross, residual, measurable disease remained in 25 patients (50%); whereas 24 patients underwent grossly complete resections. The patient with malignant TGCT/PVNS underwent margin-negative resection of the lesion after open biopsy and preoperative radiotherapy, in keeping with oncologic principles for soft tissue sarcomas. After multidisciplinary discussion, a second patient also received preoperative external-beam radiotherapy, because the lesion was very extensive, and the clinicians believed radiotherapy would help facilitate surgical resection; this patient had a history of multiple surgical corrections for congenital hip dysplasia with resultant limb shortening. The remaining 48 patients received radiotherapy postoperatively (n = 44) or alone in patients who underwent debulking or stabilization only (n = 4).
The mean dose delivered was 39.8 gray (Gy), (range, 24-50 Gy). Most patients (26; 52%) received 35 Gy in 14 fractions (Table 2). The number of fractions to deliver these doses ranged from 13 to 25 and were delivered over a mean of 26.5 days (median, 21 days; range, 18-45 days). The exception was the patient who received 5 mCi of intra-articular 198Au in 1975. One patient was treated before referral with yttrium-90 at an unknown dose for TGCT/PVNS of the knee. She had already developed avascular necrosis of the lateral femoral condyle when she presented at our center. After undergoing a single-incision open synovectomy, she received a subsequent 35 Gy of external-beam radiotherapy in 14 fractions.
The mean follow-up was 94 months (range, 19-330 months). One of 25 patients (4%) who had no evidence of gross disease after resection developed a local recurrence, and 2 of 25 patients (8%) with gross disease after resection demonstrated progression on imaging, which also was defined as a local recurrence. Thus, the overall local control rate was 94% (47 of 50 patients). The 3 local recurrences were all in patients who received 35 Gy (not a statistically significant difference). In 2 patients, both of whom had tendon sheath lesions of the wrist/hand that were radiated with gross residual tumor in situ, small (1 cm), palpable, but asymptomatic lesions were removed by excisional biopsy. Both were confirmed as TGCT/PVNS on pathologic examination. In 1 patient, this occurred 9 years after radiotherapy, and the patient remained without subsequent treatment or progression 17 years later; the second patient underwent a biopsy 16 years after treatment, received no additional therapy, and was well without progression 3 years later. A third patient who had a lesion in the knee underwent 2-stage, open/arthroscopic synovectomy and received postoperative radiation. Four years post-treatment, she developed evidence of a new mass centered in the suprapatellar aspect of the knee joint with diffuse signal changes throughout the joint on magnetic resonance imaging (MRI) (Fig. 2). This mass has remained stable 10 years post-treatment, and the patient is only mildly symptomatic, mainly because of a large, fluctuant joint effusion.
No patients suffered a wound breakdown nor were there any episodes of deep infection. However, 1 patient was hospitalized 2 years after radiation treatment with 30 Gy in 15 fractions because of inflammation and tenderness of the involved knee joint. It is believed that this probably was a case of septic arthritis.
Two patients with tumors around the hip required total hip arthroplasty because of progressive osteoarthritis post-treatment. The first patient already had osteoarthritic changes and a large amount of bone involvement by tumor at diagnosis at age 33 years. She had a prophylactic pin and plate inserted before radiotherapy (35 Gy delivered in 14 fractions) to reduce the risk of femoral neck fracture. Three years post-therapy, her osteoarthritis progressed substantially and underwent complete synovectomy and total hip arthroplasty. The second patient had extensive osseous and soft tissue involvement at presentation and required excision of approximately 75% of the proximal femur. This was reconstructed with iliac crest autograft, allograft and hip screw, followed by radiotherapy (50 Gy delivered in 25 fractions). Eight years later, this was revised to a total hip arthroplasty, because the patient developed severe and incapacitating osteoarthritis. In neither of these patients was there evidence of radiographic progression of synovial-based disease, although residual TGCT/PVNS was noted in resected specimens from both patients.
One patient presented with very extensive forefoot lesions, which were unresectable without amputation. Two years after radiotherapy, a residual plantar mass was excised, because it was interfering with shoe wear. It had not changed appreciably in size after radiation treatment. Pathologic examination after removal of the mass demonstrated residual TGCT/PVNS. This patient remains only mildly symptomatic 75 months after treatment with stable bulk disease at imaging and clinical follow-up.
Four patients with avascular necrosis (AVN) have been observed. One patient presented to our center with AVN in the femoral condyle (documented on MRI) after receiving treatment elsewhere with yttrium for a local recurrence of TGCT/PVNS, as noted above. She received 35 Gy in 14 fractions and remains disease-free 20 months post-treatment, although her knee joint is painful. A second patient originally underwent a 2-stage arthroscopic/open posterior synovectomy. A substantial local recurrence 4 years later was treated with 1-stage anterior/posterior open synovectomy and postoperative radiotherapy. Radiotherapy was stopped early (26 Gy of a planned 50 Gy) because of MRI evidence of AVN in the distal femur and proximal tibia. Initially asymptomatic, she developed progressive joint collapse and required a total knee arthroplasty 29 months later. A third patient had 3 knee arthroscopies over 10 years before referral without a diagnosis and had moderate arthritic changes in the knee with a very large and diffuse soft tissue mass in both the popliteal fossa and distal quadriceps. The recurrent TGCT/PVNS was resected with an open anterior/posterior synovectomy. Toward the end of radiotherapy (50 Gy delivered in 25 fractions), he developed increasing knee pain, and imaging studies revealed AVN in both femoral condyles. The joint rapidly collapsed, and the patient underwent total knee arthroplasty 5 months after radiotherapy. The fourth patient underwent an anterior/posterior open synovectomy followed by radiotherapy (46 Gy delivered in 23 fractions). Ten months after radiotherapy, MRI indicated early stage AVN in the lateral femoral condyle (Fig. 3). This was managed conservatively. Forty-one months post-treatment, she suffered a complex tear of her lateral meniscus without any evidence of joint collapse; this will be managed arthroscopically.
Functionally, 7 patients were rated as excellent, 34 were rated good, 5 were rated fair, and 4 were rated poor according to the rating system described above. A subgroup of 14 patients completed functional outcome measures, and their mean scores were 90.0 (range, 65.2-99.2) for the TESS, 31.4 (range, 25-35) measured on the MSTS 1987, and 91.7 (range, 63-100) measured on the MSTS 1993. No patients developed a malignancy related to the TGCT/PVNS treatment volume.
The patients in this series were a highly selected group, either having failed prior treatment (56% presented to our center with at least 1 local recurrence) or presenting with an extensive and infiltrating form of TGCT/PVNS with bone and extra-articular soft tissue involvement and, hence, with a need for adjuvant radiotherapy. All patients were considered to be at high risk for local recurrence and were limited in future surgical treatment options that would preserve a highly functional limb. For those patients who present primarily with diffuse disease that is limited to the joint space or with the localized form of TGCT/PVNS, our policy is to withhold radiotherapy and manage the patients with synovectomy alone. It is important to note that subtotal macroscopic removal of DTGCT/PVNS through partial synovectomy (in the absence of adjuvant therapy) reportedly produced a higher risk of recurrence than total synovectomy.14, 15 Early in this series, open synovectomy usually was used for lesions at the knee (the most common location for TGCT/PVNS) in patients with intra-articular disease, but this has been supplemented by arthroscopic resection, using both anterior and posterior portals when necessary, for complete synovectomy and disease removal. MRI studies are critical for identifying patients with extra-articular disease extension, because failure to identify and remove these sites of soft tissue extension is likely to be a contributing factor to local recurrence for TGCT/PVNS.31
The local control rate of 94% described in this series compares favorably with other recent reports in the literature. Berger et al20 described a small series of 7 patients who achieved a 100% local control rate at mean follow-up of 29 months after open or arthroscopic synovectomy and postoperative radiation. Blanco et al21 used partial arthroscopic synovectomy and low-dose postoperative radiation therapy (26 Gy) in a series of 22 patients with DTGCT/PVNS of the knee and achieved a local control rate of 86% (19 of 22 patients) at mean follow-up of 33 months. Horoschak et al18 treated 18 sites in 17 patients and reported an overall local control rate of 67% (12 of 18 patients) at a mean follow-up of 46 months, which rose to 75% for patients who underwent surgical excision or synovectomy before radiation (12 of 16 patients). Lee et al19 described 7 patients who underwent open synovectomy and received postoperative radiation for DTGCT/PVNS of the foot and ankle and reported a 100% local control rate at mean follow-up of 24 months. Wu et al22 treated 9 patients with DTGCT/PVNS around the knee with a 2-incision open synovectomy technique and postoperative radiation and reported 1 asymptomatic local recurrence at mean follow-up of 67 months. Although the mean follow-up times are relatively short, these outcomes lend support to the use of adjuvant radiotherapy for difficult cases of TGCT/PVNS. However, it is unclear in the studies from Blanco et al21 and Horoschak et al18 whether patients with only intrarticular involvement were included; whereas, in the study by Berger et al,20 only 1 patient had extra-articular disease. We reserve radiotherapy for patients who have multiply recurrent intra-articular disease or extra-articular spread and patients in whom a local recurrence would lead to major functional morbidity when re-excised: all scenarios we consider to be high risk.
In 2001, concern about a local recurrence in a patient who had received 35 Gy in 2.5 Gy fractions resulted in a reconsideration of the dose for treating TGCT/PVNS. It was believed that an augmented dose using a shrinking field technique was not realistic using 2.5 Gy per fraction because of concern about poor functional outcome if we proceeded to a higher dose range. Therefore, we commenced a new treatment protocol with the dual goals of enhanced protection of normal tissues from injury through smaller fraction size use and the potential to enhance local control. The result was a 2-phase approach: a wide initial phase of 40 Gy in 20 fractions conforming to the previously described volume used for the entire course and a subsequent, coned down boost of 8 Gy to the regions at highest risk (where overt disease had resided before excision or where gross disease still remained). The phase 2 boost volume was intended to avoid full circumferential irradiation of the limb and reduce the risk of limb edema and periarticular fibrosis. The total dose of 48 Gy in 24 fractions was chosen to be within the established tolerances of most late responding tissues, especially bone, which may manifest a small risk of radiation-induced fracture after doses of 50 Gy in 2-Gy fraction treatments.32
Radiation toxicities in this study were relatively mild. No episodes of grade 2 or greater fibrosis or edema were noted, which was not unexpected given the moderate doses used, nor were there any episodes of wound breakdown or documented deep infection. Many of the patients did have radiographic evidence on follow-up of at least mild arthritic change in the affected joint, although this is more likely because of the combined effect of the disease process and multiple surgical interventions rather than the radiotherapy itself, an impression echoed by other authors.18 Both patients with TGCT/PVNS affecting the hip that required joint replacement presented to our center with pre-existent bone involvement and degenerative arthritis. Although adjuvant radiotherapy likely delayed further disease-associated joint destruction, the already extensive bone and cartilage involvement was largely irreparable. Early intervention by surgery with or without radiation before the TGCT/PVNS burden leading to established arthritis is important in preventing symptomatic joint degeneration, especially in the hip, where the normal physiologic loads and rotational forces are more severe than at other anatomic locations.
The cases of AVN are concerning, but an etiology as a direct consequence of radiotherapy is unlikely. All cases developed very early after surgery (2 were confirmed on MRI early during radiotherapy, and the third was confirmed on MRI 10 months postradiotherapy) and correspondingly manifested within the established safe dose ranges for radiation effects on bone, excepting growth suppression in childhood bone or late malignant induction at any age. Like the development of osteoarthritis noted in many patients, it is more likely a combination of the multiple surgical procedures and the local aggressiveness and advanced state of the disease process. However, it is interesting that 3 of these patients underwent open anterior/posterior synovectomy. It is possible that, for some patients with TGCT/PVNS who have very extensive extra-articular soft tissue masses, surgical disruption of the extraosseous periarticular vasculature, which occurs as part of gross tumor removal, may put some patients at risk of developing AVN.
Intra-articular radiation with radioactive colloids like yttrium-90 has been used as an alternative to external-beam radiotherapy.23-26, 33, 34 Used as an alternative treatment for rheumatoid arthritis, yttrium-90 emits beta radiation with good penetration of the synovium and has a short half-life of 64 hours with low systemic uptake, thereby reducing total body irradiation.23, 33, 35, 36 Although they are effective in some patients with arthritis, we remain uncertain whether the radiocolloids can effectively treat all tissues involved, especially in those patients who have diffuse extra-articular and/or bony involvement. Adjuvant radiocolloids may be suited best for patients who have disease confined to the intra-articular compartment. However, in these situations, we tend to use total synovectomy alone and reserve the use of radiation therapy for patients who have intra-articular disease with multiple recurrences as well as those who have extensive extra-articular disease. It has been reported that certain anatomic sites, such as the ankle, are susceptible to unacceptably high complication rates after yttrium-90 instillation, with the development of severe pain and full-thickness skin ulcers that eventually required skin grafts and muscle flaps for wound management.34
Recently, investigations into the molecular mechanisms of TGCT/PVNS have outlined the potential utility of targeted molecular therapy as an emerging treatment option. Geldyyev et al postulated that high levels of expression of bone degrading proteins RANKL, osteopontin, and bone sialoprotein in TGCT/PVNS play a critical role in the osteolysis observed in patients with severe disease.37 West et al identified a key molecular pathway in the development of TGCT/PVNS involving colony-stimulating factor 1 (CSF1) and its receptor, CSF1R.38 Those authors reported that only a small number of cells in the tumor mass actually contain a translocation and are neoplastic, and those cells recruit reactive, non-neoplastic cells that express CSF1R, thereby promoting their proliferation and differentiation.38 Inhibitors of the various components of this pathway may provide therapeutic options,38, 39 as indicated in the report by Blay et al, who treated 1 patient with imatinib for recurrent TGCT/PVNS and achieved excellent results.40 However, given the excellent control rates observed with surgery and radiotherapy, these targeted approaches generally are applicable to situations in which surgery and radiotherapy are not feasible and permanent control of the disease is considered an implausible outcome.
Like in other benign yet locally aggressive tumors, such as fibromatosis, in which there is no risk of systemic spread yet local growth can cause significant destruction and functional impairment, the addition of moderate-dose adjuvant radiotherapy to surgical resection is quite effective in providing excellent local control while maintaining good function with low treatment-related morbidity.41-43 For those patients with extensive or multiple local relapses or in those for whom surgery alone would result in a large burden of residual disease or would not be possible without major functional loss, combined-modality therapy remains our treatment protocol of choice.