Introduction

Pigmented villonodular synovitis (PVNS) (nodular and diffuse lesions) and giant cell tumors of tendon sheaths (GCTTS) are a benign disease of the synovial tissue. Most cases of diffuse PVNS are located in the knee and hip joint, whereas GCTTS are seen mainly in tendon sheaths of the hand [1, 2].

PVNS and GCTTS occur most often in young adults, but the lesion has also been observed in children and adolescents [3]. Symptoms are non-specific. The incidence is estimated to be 9.2 cases per million for GCT, 1.8 [4] per million for PVS. For diagnosis X-rays and MRI are recommended [5, 6, 14].

Complete surgical resection is the treatment of choice [1, 6, 7] and therapy is simple in nodular lesions. Complete resection is somewhat problematic in cases with a diffuse spread, owing to the anatomical conditions of the joints involved. As a result, the rate of recurrence (ROR) is relatively high. However, as the disease is so rare, even large institutions have had very little experience with its treatment. Reports with more than 50 patients are few and, to our knowledge, only one prospective study has compared different therapeutical criteria [8].

Consequently, the Task Force Group for Musculoskeletal Tumors (Arbeitsgemeinschaft Muskuloskeletale Tumoren, DGOOC) of the German Orthopaedic Society initiated a retrospective study to sample a sufficient number of patients suffering from PVNS.

Materials and methods

A standardized questionnaire was sent to all contributing hospitals in Germany and Austria. Questions about location of the disease, duration of symptoms, affected joint or other synovial structure, surgical treatment and adjuvant therapy, outcome and, in particular, number of recurrences and number of patients treated between 1980 and 2002 were asked.

Results

We could sample data from 173 patients treated in 10 orthopedic departments in Germany and Austria. The mean number of patients per institution was 17.3 (range 1–54). In 4 of the 10 institutions, more than 20 patients had been treated. In all, there were 71 (41 %) male patients with a mean age of 35 years and 102 (59 %) female patients with a mean age of 43.1 years. The distribution of age and gender is shown in Fig. 1.

Fig. 1
figure 1

Age and gender distribution of the patients

Full size image

Regarding the anatomical localization, the disease was found mostly in joints (75 %), followed by tendon sheaths (21 %) and bursae (4 %). In joints the disease exhibited more the diffuse (64 %) than the nodular type (36 %). In tendon sheaths the distribution—diffuse versus nodular—was nearly equal (47 vs. 53 %), whereas in bursae the disease was usually diffuse (86 %) rather than nodular (14 %).

Regarding the anatomical regions, the knee joint (56 %) was the most affected, followed by a lesser involvement of the hand (14 %) and the ankle and subtalar joint (9 %) (Table 1). In the hip region, the joint itself was involved in all 11 cases. In the knee and ankle regions, mostly the joint itself was involved (92/97, 15/16). Only in a few patients were tendons and bursae involved in contrast to the hand region and tendon sheaths showed most of the lesions (22/25).

Table 1 Localization of disease
Full size table

Regarding the distribution—diffuse versus nodular—, the ratio was 82 versus 18 % in the hip joint, 62 versus 38 % in the knee joint and 67 versus 33 % in the ankle and subtalar joint. However, in the hand region mostly the nodular form was found (64 vs. 36 %).

Symptoms

The median duration of symptoms was 12.7 years with a wide range (6 days–52 years). Most of the patients suffered from pain (87 %), swelling (66 %) and/or a reduced range of motion (41 %). If an osteolytic lesion was present, the rate of patients suffering from pain increased from 84 to 93 % and the rate of limited range of motion from 58 to 93 %.

Diagnostic procedures

Diagnostically, radiographs were used in 151 cases (87 %) and MRI in 90/161 cases (56 %). Less frequently a CT (10/155) or an arthrography (5/144 = 3 %) was carried out. Owing to the long study period, the use of MRI slowly replaced other diagnostic methods (Fig. 2).

Fig. 2
figure 2

The used diagnostic tools

Full size image

Osseous lesions

Major osseous lesions which were detectable with X-rays were seen in 41/151 patients (27 %); minor lesions, probably detectable only in tomographies (MRI and CT), were not counted.

Therapy

The surgical treatment procedures are listed in Table 2. Adjuvant treatment had been used in 26/152 patients (17 %), this usually being radiosynoviorthesis (RSO) (65 %) followed by radiation therapy (RT) (31 %); in 4 % of cases phenol was applied intraoperatively for osteolytic lesions. These procedures were used more often in recurrent (62 %) than in primary forms of the disease (38 %).

Table 2 Type of surgical therapy
Full size table

Recurrence

The most striking result was that the ROR was dependent on the number of cases treated in a single institution. At the end of the study, those institutions which had treated at least 20 patients had a distinctly lower ROR (27.8 %, range 15–37 %) than those with less than 20 patients (56 %) (Table 3).

Table 3 Recurrence depending on the single institution
Full size table

The ROR was also dependent on the type of disease; in nodular lesions the mean ROR was 20 % (in articular lesions it was 24 %, in tendon sheaths 7 % and in bursae 0 %). In diffuse type of disease, the ROR was 38 % in articular lesions, 50 % in tendon sheaths and 0 % in bursae (Table 4). The ROR in cases who had already developed one recurrence was not reduced to zero after a second surgery but, with repeated surgeries, the ROR could be reduced from 38 to 1 %.

Table 4 Recurrence rate depending on the type of disease
Full size table

Regarding the anatomical site, the ROR was found to be highest in the hip joint (50 %) and lowest in the feet (1 %) (Table 5).

Table 5 Recurrence rate depending on the localization
Full size table

Comparison of the two principal surgical procedures in the knee joint (arthrotomy vs. arthroscopy) exhibited that the ROR was distinctly higher after arthroscopy than after arthrotomy (69 vs. 25 %).

The overall rate of application of adjuvant therapeutic tools was relatively small (17 %): those patients (n = 10) who were additionally treated with either RSO or RT after the primary occurrence did not exhibit any recurrence. In contrast, 86 % of those patients with a recurrent disease in whom RSO or RT had been applied, after at least the first recurrence (n = 16), were free from any further recurrence.

Functional results

In nearly 85 % of the patients, an excellent or good functional result could be detected (Fig. 3).

Fig. 3
figure 3

The graph shows the distribution of the functional results

Full size image

Discussion

Owing to the rarity of the disease, the main purpose of the study was to sample a sufficient number of cases to give relevant therapeutical recommendations. The rarity does not allow prospective single- or multi-center studies within an acceptable period of time (Fig. 4).

Fig. 4
figure 4

Clinical case of a circumscript pigmented villonodular synovitis located in the superior recessus of the knee; small inset shows surgical resection specimen

Full size image

Recurrence was the most important concern. Our major result was that the ROR was lower in institutions which had treated more than 20 patients; institutions with less than 20 cases had a distinctly higher ROR. This indicates that this disease should be treated in a center and/or by a surgeon experienced with its treatment.

We found a high ROR in nodular cases as well as in affected tendon sheaths. The ROR should be zero or near to it, at least in joints with a nodular PVNS, and marginal excision is recommended [5, 6, 9–19]. The ROR of GCTTS was 8 % [30]. Sharma and Cheng [20] reported a recurrence-free survival of 91 % for the localized type versus 70 % for the diffuse type after 2 years and a recurrence-free survival of 73 and 48 %, respectively, after 5 years. Kim et al. [6] did not observe any recurrence after a follow-up of 30 months in 11 cases treated arthroscopically. Chiari et al. [10] noted a recurrence-free survival 1 year after surgery in digits rather than in large joints. Recurrence is due to incomplete resection [1, 2, 13, 21].

The diffuse type is more difficult to treat. The necessity of a complete synovectomy is demonstrated by Sharma and Cheng [20] surveilling 49 patients. They compared an arthroscopic partial synovectomy with a complete open synovectomy, using an anterior and posterior approach to the knee, and found a highly significant, better recurrence-free survival for the open procedure in knee joints [20]. Ogilvie-Harris et al. [17] compared partial with complete arthroscopic synovectomies and found a lower ROR after complete as opposed to partial synovectomy.

In our study, a combined anterior and posterior open synovectomy was found to be followed by a distinctly lower ROR than after arthroscopic synovectomy. In addition, Blanco et al. [8] observed after 1 year a clinically, ultrasound-detected ROR of 14 % following an anterior arthroscopic partial synovectomy combined with postoperative radiation therapy. Recurrence was most frequent in the knee and in patients with recurrences [19]. The cumulative ROR was significantly higher after incomplete as opposed to complete synovectomies [19].

The ROR also seems to depend on the anatomical region. We found the highest ROR in the hip joint and lowest one in the foot. To achieve a complete synovectomy in the hip joint, a surgical dislocation or even the implantation of a total hip arthroplasty is necessary [21–23]. Thus, implantation of a total arthroplasty is recommended and results in a lower ROR than after an incomplete synovectomy without disarticulation [7, 19, 21, 24].

After arthroscopic resection in the knee joint, the ROR was distinctly higher than after open synovectomy. Similar statements are given by Tyler et al. [25] and De Ponti et al. [26]. They noted a lower ROR and had better clinical results after an extended arthroscopic synovectomy using 4–5 portals. Schwartz et al. [19] reported a cumulative ROR of 7.1 % after 1 year, 15 % after 5 years, 27 % after 10 years and 31 % after 15 years. Recently, Liu et al. [27], after an average follow-up time of 22 months, found a ROR in PVNS treated with arthroscopic resection of 13.6 %. Sharma and Cheng [20] reported a recurrence-free survival of 95 % following open synovectomies, as compared to a recurrence-free survival of 62 % after 2 years and 71 versus 41 % after 5 years following arthroscopy.

In our study, only 17 % of all cases underwent adjuvant therapy. Radiotherapy (RT) and RSO [28–30] are recommended, the overall number of patients treated with either RSO or RT remains small.

For RSO mostly 90-Yttrium [31–33] has been used. Chin et al. [34] applied 165-dysprosium-ferric-hydroxia macro-aggregates. Shabat et al. [33] and Chin et al. [34] reported the largest number of 10/40 patients who received RSO. Nine of these patients had suffered no recurrence of the tumor and, in one patient, the disease had remained stable. Chin et al. [34] treated 30 patients who received postoperatively 165-dysprosium, 40 % suffered from recurrence. Similar results have been reported by Ward et al. [35] on RSO using 32-P. Zook et al. [36] found three recurrences of nine patients also treated with 32-P. Complications were reported by Bickels et al. [37]: From seven patients, two suffered from skin necroses.

RT has the advantage of being also applicable in patients with extra-articular spread. Chin et al. [34] reported on three groups of patients. The ROR was zero in patients treated with surgery alone, 17 % in those who received RSO and 40 % in those treated with RT.

RT, like RSO, seems to have a similar influence on the ROR. There are a few papers with a small number of patients, the groups being heterogenous regarding history, localization, primary or recurrent disease and treatment [29, 34, 38, 39]. The ROR in all groups ranged from 0 % [38] to 40 % [34]. Only two papers reported on a larger group: Blanco et al. [8] had treated patients with only a partial anterior arthroscopic synovectomy with additional small radiation dose and found a ROR of 13.7 %. O’Sullivan et al. [29], reporting on 14 patients suffering from an intra- and extra-articular disease observed quite good results after a mean follow-up of 69 months. Similar results were found in a multi-center study [40]. This study analyzed 41 patients on whom either a primarily incomplete resection was performed or who suffered from recurrence. Local control was achieved in 95.1 %.

Functional results have been reported by several authors. Comparison is difficult owing to different scores and heterogenous groups. Arthroscopic treatment has been reported by Liu et al. [27]. The Lysholm score and the IKDC score increased significantly after treatment. Dines et al. [12] reported on a long-term follow-up (65.8 months) of 10/26 patients who had been treated either arthroscopically or with an arthrotomy for nodular PVNS in the knee with six excellent and one good result (Lysholm-Score). After open synovectomy and adjuvant RT, Wu et al. [39] reported a distinct postoperative improvement. Similar results using different scores were reported by Horoschak et al. [41], Heydt et al. [40], and de Visser et al. [11].