Giant-cell tumour of the knee

Patients and Methods

Between 1988 and 1998 we treated 14 patients with GCT around the knee by curettage and cementing. The lesions were all located close to the articular cartilage (Fig. 1, Table I). The study was approved by the Lund University Ethics Committee. The patients were asked by letter to participate in the study on a voluntary basis. Four failed to reply and one was excluded because of the presence of established OA before the diagnosis of GCT. The nine remaining patients were included in the study at a mean of 11 years (6 to 16) after their first operation. There were three females and six males with a mean age at diagnosis of 29 years (14 to 44). The tumour was in the proximal tibia in four and the distal femur in five. The mean maximum diameter of the tumour was 7 cm (4 to 11) and the mean distance from the cement to the cartilage was 1.0 mm (0.0 to 3.5) (Fig. 1, Table I). All except one patient had been primarily treated at our hospital and all were free from tumour at the time of follow-up.

After the first operation complications occurred in four patients. The first (case 3) who had a GCT in the distal femur developed pain and a mechanical obstruction to movement in the knee six months later. Small pieces of cement were observed in the joint. No cement had been seen in the joint during operation and he had mobilised normally. Two small pieces of cement (6 mm to 8 mm) were removed arthroscopically and a defect was seen in the cartilage in the lateral tibial condyle. Furthermore, 18 months after the first operation he needed further curettage and cementing for local recurrence.

A second patient (case 5) sustained a minimally-displaced intra-articular fracture of the medial femoral condyle after recurettage and cementing of a local recurrence three months after her primary operation.

The third patient (case 6) had local recurrence 2.5 years after the first operation and was treated by further curettage and cementing. The final patient (case 9) had been treated at another hospital by curettage and bone grafting and was referred 18 months later because of a local recurrence. He, too, was treated by further curettage and cementing.

Each patient was investigated using plain radiography, MRI and measurement of the serum level of COMP. Quantitative measurements were performed in six of the patients using dGEMRIC. In each case the opposite side was used as a reference.

Weight-bearing radiographs, including anteroposterior (AP), lateral and skyline views of the patella, were obtained. Both knees were examined in all nine patients. The minimum distance between the cement margin and the adjacent articular cartilage was measured in the operated knee, as was the minimum joint space in each compartment, in both knees in every patient.

Function was evaluated using the system of Lysholm and Gillquist¹³ (maximum score 100).

MRI of the operated knee was carried out in all the patients using a 3T imager (Philips Intera Release 1.2, Philips Medical Systems, Best, The Netherlands) with a SENSE-flex medium coil (Philips Medical Systems). In six patients dGEMRIC of both knees was undertaken.

The T1 relaxation rate measurements were performed before and after injection of contrast using an inversion recovery pulse sequence with different inversion times (50, 100, 200, 400, 800 milliseconds (ms); and 1600 sagittal slices with thickness 5 mm, TR = 2000 ms, TE = 15 ms, turbo factor 11, field of view (FoV) = 120 mm × 120 mm, matrix size = 256 × 204).

In order to allow the contrast medium to penetrate, the patients were instructed to walk for 15 minutes after a triple dose injection of contrast (0.3 mmol/kg bodyweight) before carrying out the post-contrast measurement 90 minutes after the injection.

The dGEMRIC measurements were made in a region of interest in each knee which was determined by identifying the area with the shortest distance between the cement and the articular cartilage. In the unaffected knee this was sited in the corresponding area of cartilage and used as a reference. The same radiologist (IK) carried out all the studies.

The serum COMP level was measured using a commercially-available sandwich-based ELISA (Anamar Medical Co., Gothenburg, Sweden).

Results

Post-operative function was generally good. The mean Lysholm knee score was 92 (83 to 100) (Table I). All the patients were physically active, had returned to their previous occupations and were able to perform recreational sport or exercise. One patient had a Lysholm score of 83 but no radiological signs of OA, although MRI revealed minor degenerative change in the medial meniscus.

Discussion

The possible adverse effect of acrylic cement on articular cartilage after curettage and cementing of a GCT has not been fully assessed. Earlier published series of GCTs treated by curettage and cementing record only a few cases of degenerative change which were not necessarily related either to the surgery or to the cement.^2,8 In a recent report Lackman et al⁴ described 63 patients treated for GCT by curettage and cementing, of whom 46 had lesions either in the distal femur or in the proximal tibia. After a mean follow-up of nine years one patient had developed OA. Moreover, patients with radiological signs of OA pre-operatively did not deteriorate substantially after their surgery. However, no weight-bearing radiographs had routinely been obtained and evaluation of OA was therefore somewhat basic.

In our study, OA was observed on weight-bearing radiographs in one patient, in whom an intra-articular fracture had occurred after surgery for local recurrence.

The use of dGEMRIC has been suggested as a possible method of detecting early OA.⁹ However, it is not routinely employed in clinical practice and more studies are needed to establish its efficacy. Our results were not easy to interpret and the data obtained varied to a large extent without any obvious relation to other factors. In a study of dGEM-RIC in early OA of the knee Tiderius et al,⁹ chose a region of interest in the midline of the condyles of the knee where the cartilage is best outlined and supposedly thickest. Giant-cell tumours tend to be located eccentrically in the distal femur or proximal tibia. Consequently, in our study we measured the T1 relaxation rate at the point where the distance between the bone cement and the articular cartilage was least. In most cases this was at the edge of the condyles where the cartilage was thin and thus hampered the selection of a region of interest which included only cartilage; this may explain the variation in the values. Apart from the results in one patient, in whom dGEMRIC may have indicated degeneration of cartilage, this method did not confirm or exclude the possible presence of early OA.

The level of the serum COMP has been used as a marker for evaluating degeneration of cartilage and OA.^10,11 Petersson et al¹⁰ noted that the baseline levels between groups of individuals without radiological signs of OA did not differ. However, after three years, patients who had developed signs of OA also had a raised serum COMP level. Diurnal variation in the COMP level occurs.^14,15 However, when sampling was performed at four-hour intervals throughout a day and night, it was noted that COMP levels were constant during the day.¹⁵ Therefore, samples can be reproducibly analysed during the daytime when the patients are mobile, as in the present study.

In our study we had only baseline values of the serum COMP at the outset. There was no longitudinal follow-up and the level of activity before sampling was not standardised. We do not, however, believe that these factors significantly affected our results. A level of serum COMP of around 15 U/L or more is above the 95% percentile in a population of blood donors and in this relatively young group, the levels in four patients are towards the high end.¹⁶ One patient apparently had an ongoing process in the cartilage which led to the morphological changes observed. It is possible that another who showed cement protruding into the joint cartilage, had altered cartilage turnover. Further follow-up will be of interest in two patients as to whether an altered cartilage metabolism or elevated levels indicate the early phase of degenerative change. The long interval between surgery and follow-up without clinical signs of OA suggests that if there is a pathological process leading to OA it is well compensated for by continuous remodelling. Apart from the results in one patient we were unable to distinguish a pattern in which the levels of the serum COMP could be correlated with any other measurement such as age, time from diagnosis or surgery, morphological findings and function.

The role of bone in the development of OA is not clear. It has always been assumed that the subchondral sclerosis seen in OA is the result of increased bone density and stiffness. This must now be questioned given that there appears to be a reduction in bone density and in other aspects of mineralisation of tissues, e.g. cartilage and trabecular bone.¹⁷

A finite-element study of the femoral head and neck,¹⁸ suggested that stiffening of the subchondral plate plays an important mechanical role in the degeneration of cartilage. Frassica et al¹⁹ have shown that in dogs, replacing subchondral bone with cement does not increase its stiffness. The cement, however, does serve as a stable subchondral platform which may be an important factor for avoiding cartilage degeneration. In our series even the patients with little or no space between the cement and cartilage did not appear prone to develop OA, despite the altered mechanical and physiological environment. Another factor is the effect of altering the blood supply but this is likely to be of less significance since the cartilage is principally nourished by the synovium and, as long as this remains intact, cartilage nutrition will be maintained.

This is a small series with a long follow-up. We have found no evidence to support the view that the treatment of GCT by curettage and cementing induces degeneration of cartilage leading to OA. In our opinion it remains a safe method of treatment even if the tumour involves the subchondral bone of the knee.