Long‐term outcome of giant cell tumors of bone around the knee treated by en bloc resection of tumor and reconstruction with prosthesis

Abstract

Objective:  To study the long‐term outcomes and complications of giant cell tumors around the knee treated with en bloc resection and reconstruction with prosthesis.

Methods:  From January 1991 to March 2005, 19 patients (11 men, 8 women, average age 35.4 years) were treated in our hospital with en bloc resection and reconstruction with domestic prosthesis (15 hinge knee and 4 rotating‐hinge knee). The distal femur was involved in 12 and the proximal tibia in 7 cases. Nine tumors were primary and 10 recurrent. All cases were Campanacci grade III. The affected limb functions were evaluated by the Musculoskeletal Tumor Society scoring system.

Results:  All patients underwent operation successfully with no complications. The mean follow‐up time was 128.9 months (60 to 216 months). Apart from one patient who underwent amputation because of wound infection two years after reoperation, the range of knee motion of 18 patients was 30°–110°. The mean functional score of the affected limb was 22.7 (15 to 27 points). The length of the lower extremities was equal in nine cases; the affected limb was 2–9 cm shorter in the other ten cases. Prosthesis fracture and loosening developed in one, prosthesis aseptic loosening in three, and delayed deep infection and prosthesis loosening in two cases. The prosthesis loosening rate was 31.6%. One patient developed a proximal femur fracture.

Conclusion:  En bloc resection and reconstruction with prosthesis is a feasible method for treating giant cell tumor of bone around the knee. Complications related to the prosthesis, mainly prosthesis loosening and limb shortening, increase gradually with longer survival time.

Introduction

Giant cell tumors of bone (GCT) possess malignant potential and invasive characteristics and mostly occur around the knee joint in young, as well as middle‐aged, patients. There are two surgical treatment methods for the disease, namely intralesional excision and en bloc resection. Patients who receive appropriate treatments are often able to achieve long‐term survival, therefore the long‐term impact of the surgical procedures on limb function must be considered when selecting treatment methods ¹ . En bloc resection and reconstruction by various methods is the widely accepted treatment for GCT with extensive destruction of bone structure, recurrence, pathological fracture, or difficulty in reconstruction after intralesional curettage ² . Among the many reconstruction methods, prosthetic replacement after tumor resection has been widely used because it has the advantages of complete resection, low postoperative recurrence rate, and high rate of excellent short‐term postoperative limb function ³ . However, at present there are very few reports regarding the long‐term outcomes of prosthetic treatment for GCT. To address this issue, in this study 19 patients with GCT and more than 5 years of follow‐up were retrospectively analyzed.

Materials and methods

Data

From January 1991 to March 2005, 19 patients were treated in our hospital by en bloc resection of tumor and reconstruction with a domestic massive prosthesis (Lidakang, Beijing, China). They included 11 men and 8 women whose ages ranged from 23 to 60 years with an average age of 35.4 years. The lesions affected the distal femur in 12 cases and the proximal tibia in 7. Nine cases had primary and 10 had recurrent lesions (curettage and bone grafting of the primary tumor had been done in our hospital in two cases and in other hospitals in eight). Four patients had combined pathological fractures. All cases were classified as Campanacci grade III by X‐ray examination. Rotating hinged knees were used in four cases and hinged knees in the other 15.

Surgical technique

All operations were performed under epidural anesthesia. The patient was placed on the operating table in a supine position and a sandbag placed under the ipsilateral buttock. The conventional anterior medial incision for the knee was adopted. Exposure and removal of the tumor were performed in strict accordance with the principles of tumor‐free surgery. The length of osteotomy was determined based on preoperative X‐ray measurement and the severity of tumor invasion. A length 3 cm longer than the tumor boundary was generally accepted as a standard prosthetic length. Installation of the prosthesis was accompanied by outer cortical bone grafting in which normal autologous cancellous bone was bundled around the post‐excision bone, and a first generation bone cement technique was used for prosthetic fixation. Attention was given to soft tissue reconstruction during surgery. In the case of proximal tibial lesions, a medial gastrocnemius muscle flap was routinely used to cover the anterior side of the prosthesis and provide an attachment for the patellar tendon. The patellar tendon was fixed to this flap to provide maximum retention of flexion function of the knee.

Postoperative treatment

Antibiotics were routinely used for 7 to 10 days postoperatively. For distal lesions of the femur, the drainage tube was removed 2 to 4 days after surgery. The patients were then subjected to continuous passive motion (CPM) exercise; functional exercise with double crutches was started 2 weeks postoperatively. For lesions of the proximal tibia, once the drainage tube had been removed, functional isometric muscle exercise of the lower limbs was started 3 days after surgery; knee and walking exercise with double crutches began 3 weeks postoperatively. The exercise was changed to walking with a single crutch 2 months after surgery, and normal walking was generally achieved by 3 months after surgery.

The patients in this study were regularly followed up in outpatient clinics; outpatient review being performed every 3 months during the first year after operation, every 6 months from the second to the fifth year, and every year after the fifth year. However, patients with changes in their disease condition were scheduled for review promptly. The review checks comprised the condition of the wound, any tumor recurrence, the location of the prosthesis, changes in prosthesis and host bone under X‐ray examination, and changes in limb length. Limb function was evaluated with the Musculoskeletal Tumor Society (MSTS) rating scales, which comprise six items, namely pain, function, emotional acceptance, supports, walking and gait. The highest possible score is 30, 5 points being allocated to each item ⁴ .

Results

The operation was completed successfully in all patients, the incisions healing well without wound infection, skin necrosis, or other complications. The patients were followed up for 60 to 216 months with an average of 128.9 months including three cases for 5 years, two for 5 to 7 years, six for 7–10 years, and eight for more than 10 years.

Evaluation of affected limb function

Apart from one patient who underwent amputation 2 years after revision surgery due to postoperative wound infection, evaluation of limb function was done in 18 patients. The degree of post‐surgery active knee flexion was 30° to 110° including three patients with less than 45°, five with 45°–70°, nine with 70°–90°, and one with more than 90°. MSTS scores were 15–27 points with an average score of 22.7 (75.6%). Eventual limb lengths varied: the length of the affected limb was equal to that of the contralateral normal limb in nine cases, less than 2 cm shorter than the contralateral normal limb in three, 3–5 cm shorter in three, and 5–9 cm shorter in three. Table 1 shows the general patient information, including the follow‐up results.

Table 1.

Patients’ information

No.	Gender	Age (years)	Tumor location	Primary or recurrent tumor	Type of prosthesis	Revision surgery	Limb length, amount of shortening	Degrees of knee flexion	Function	Follow‐up (months)	Prognosis
1	M	37	DF	P	H	No	E	70°	22	68	N
2	F	60	DF	R	H	No	E	30°	15	120	DN
3	F	30	PT	R	H	No	E	90°	24	105	N
4	M	25	DF	R	H	Yes	–	–	–	84	Amputation
5	M	59	DF	P	H	Yes	S, 9 cm	30°	20	216	N
6	F	43	PT	P	H	No	E	70°	21	144	N
7	F	33	DF	P	H	Yes	S, 6 cm	90°	23	168	N
8	M	36	DF	R	H	No	E	90°	24	60	N
9	M	29	PT	R	H	No	E	90°	25	60	N
10	F	42	DF	P	H	No	E	90°	21	168	N
11	F	34	DF	P	H	Yes	S, 6 cm	30°	23	192	N
12	F	27	PT	P	H	No	E	90°	23	96	N
13	M	23	PT	P	H	No	S, 4 cm	90°	21	156	N
14	M	25	PT	P	RH	No	E	90°	24	90	N
15	M	24	PT	P	H	No	S, 2 cm	100°	24	97	DN
16	M	25	DF	R	H	No	S, 2 cm	70°	27	163	N
17	F	40	DF	R	RH	No	S, 3 cm	90°	21	216	N
18	F	52	DF	R	RH	Yes	S, 2 cm	70°	26	84	N
19	M	29	DF	P	H	Yes	S, 9 cm	60°	24	162	N

Notation: DF, distal femur; DN, death from non‐neoplastic causes; E, equal length; H, hinge knee; N, normal; P, primary; PT, proximal tibia; R, relapse; RH, rotating hinge knee; S, limb shortening.

Prosthetic complications

Prosthesis fracture combined with loosening occurred in one case; aseptic prosthesis loosening in three; delayed deep infection combined with prosthesis loosening in two; and prosthesis‐proximal bone fracture in one. The rate of prosthesis loosening was 31.6%. These seven patients all underwent secondary surgical treatment; the six with prosthesis loosening underwent revision surgeries of their tumor prostheses 4–16 years (average 9.2 years) after the initial operation. During revision surgery, the original prostheses (hinge knee in two cases and rotating hinge knee in one) were used in three cases; in the other three cases hinge knees were replaced by rotating hinge knees (Fig. 1). After revision surgeries in two patients with infections, one healed well with no infection nor prosthesis loosening during 6 years of follow‐up, however, there was 6 cm shortening in the affected limb compared to the contralateral healthy limb, with a functional evaluation score of 23. In another case, infection recurred one year after the revision surgery and, although the patient underwent a second debridement, amputation had to be carried out 2 years after the revision surgery due to failure of the incision to heal (Fig. 2a–e). Another four patients were followed up for 3–5 years with no prosthesis‐related complications; the affected limbs shortened by 3–9 cm, and the average limb functional score was 23.8. Prosthesis‐proximal fracture occurred in one patient 13 years after prosthetic replacement surgery. Because there was no loosening of the prosthesis, simple restoration and internal fixation of the bone graft were performed. The bone healed well, with a limb functional score of 27 two years after revision surgery (Fig. 3).

Figure 1.

Radiographs of patient number 19, who underwent en bloc resection of GCT in the distal femur and reconstruction with hinged knee prosthesis 13 years previously. (a) Preoperative X‐ray images show extensive destruction of bone in the distal femur with a fracture which has been fixed with a nail. (b) Postoperative X‐ray images show the hinged knee prosthesis to be in a good position. (c) X‐ray images 13 years postoperatively shows loosening and movement of the prosthesis. (d) X‐ray images 3 years after revision surgery with a rotation hinged knee show good location of the prosthesis, but obvious shortening of the affected limb.

Figure 2.

Radiographs of patient number 4, who underwent en bloc resection of recurrent GCT in the distal femur and reconstruction with rotation hinged knee prosthesis 7 years previously. (a) Preoperative X‐ray image shows extensive destruction of bone in the distal femur with fracture 1 year after intralesional curettage with bone grafting. (b) Postoperative X‐ray image shows the rotation hinged knee prosthesis to be in a good position. (c) X‐ray image 5 years postoperatively shows loosening of the prosthesis. (d) Post‐revision X‐ray image shows good location of the prosthesis. (e) The patient had to undergo amputation because of repeated local infection after stage 1 revision. This picture shows local deep infection with sinus tract.

Figure 3.

Radiographs of patient number 16, who underwent en bloc resection of recurrent GCT in the distal femur and reconstruction with hinged knee prosthesis 14 years previously. (a) Preoperative X‐ray image shows an extensive osteolytic destruction of bone in the distal femur with fracture and the tumor cavity filled with artificial bone. (b) Postoperative X‐ray image shows fine position of the hinged knee prosthesis. (c) X‐ray image of 13 years postoperatively shows proximal femur fracture with no loosening of prosthesis. (d) The radiograph shows bone healing well and no loosening of prosthesis 12 months after open reduction and bone grafting with internal fixation.

Oncological results

No recurrence nor metastasis developed during the follow‐up; two patients died from other diseases at 8 and 10 years after surgery, respectively; while the other 17 patients are still alive.

Discussion

Surgical treatment of giant cell tumor of bone

GCT often occurs around the weight‐bearing joints, almost half of them involving the distal femur and proximal tibia; different treatment procedures directly affect the recovery of the limb function. In surgical treatment, the operative procedures are classified into intralesional curettage and en bloc resection based on the surgical boundary. Curettage of bone lesions of GCT is the basic treatment method, its major drawback being the high recurrence rate; local recurrence rates after simple intralesional curettage combined with bone grafting are 27% to 54% ⁵ . Therefore, many investigators apply chemical (such as zinc chloride, carbolic acid and liquid nitrogen) and physical (such as high‐speed burr) methods for destruction of any bone tumor in the cavity wall in order to reduce the local recurrence rate. However, when there is pathologic fracture, tumor recurrence, extensive destruction of the bone cortex, or insufficient solidity to meet the physical load after tumor excision, en bloc resection of tumor and reconstructive surgery of the bone defect need to be performed ² . In this study, 19 patients were all classified as having Campanacci grade III tumors by X‐ray examination, met the standard selection criteria for en bloc tumor resection and prosthetic replacement, and had an average postoperative follow‐up time of 128.9 months. No recurrence developed during follow‐up, which demonstrates that as long as a reasonable tumor resection procedure is selected, the prognosis of GCT is good with long‐term disease‐free survival.

The currently used major reconstruction methods include autogenous graft, allograft and prosthetic replacement of the resected tumor. Prosthetic replacement after en bloc resection of tumor has been widely used in the clinic because it provides immediate restoration of bone stability, allowing the patients to start using the limb earlier. In addition, rejection does not occur and there are fewer early complications. However, the major drawbacks of this technique are the high incidence of prosthesis loosening and the limited joint function resulting from the large bone defects caused by tumor resection, the short length of fixation to the medullary cavity of the host bone, and the lack of attachment points for important muscles. Recently published articles show that the short‐term outcome of prosthetic replacement for the treatment of GCT is satisfactory ³ . However, because patients can now achieve long‐term survival through en bloc resection and prosthetic replacement, the long‐term results of prosthetic treatment have become an issue of concern. Maruthainar et al. treated 25 patients with GCT in the distal femur by prosthetic replacement, and the results of follow‐up with an average time of 12 years (5–18 years) showed no tumor recurrence, revision due to infection, or death ⁶ . However, six cases needed revision surgery due to aseptic loosening, including five cases with hinge knees and one case with a rotating hinge knee. The aseptic loosening rate of prosthesis in patients with hinge knee was 33%, whereas the result with rotating hinge knees was better (10%). Although the patients in their study were young, good long‐term limb function was still achieved. In this study, 19 patients were followed up for an average period of 128.9 months and, except for one patient who underwent amputation due to infection 84 months after surgery, the other 18 patients showed active knee flexion degrees of 30°–110° with an average limb function score of 22.7. These results demonstrate that satisfactory limb function can be achieved through prosthesis treatment for giant cell tumor of bone around the knee.

Complications of prosthetic replacement

Prosthesis loosening, the major long‐term complication after prosthetic replacement, has a reported incidence rate of 7%–30% the incidence gradually increasing along with longer follow‐up ⁷ , ⁸ , ⁹ , ¹⁰ . Unwin et al. reviewed 1001 patients who underwent tumor resection and reconstruction with prosthesis and found aseptic loosening in 71 cases ⁷ . Kawai et al. followed up 40 patients with femoral tumor for an average of 8 years and found that 11 cases needed revision surgery due to aseptic loosening ⁸ . Other results concerning the rates of aseptic loosening and prosthesis survival are similar to the above reports ¹⁰ , ¹¹ . Maruthainar et al. performed long‐term follow‐up on 25 patients with GCT in the distal femur who underwent prosthesis replacement, and found that six cases (24%) needed revision surgery due to aseptic loosening (with hinge knees in five cases and rotating hinge knee in one) ⁶ . Among the 19 patients treated in the present study, revision surgeries due to prosthesis loosening were performed in six cases, two of which were complicated by sinus tracts. The total loosening rate of prosthesis in this study was 31.6%, and the aseptic loosening rate 21.1%; these results are consistent with those of previous reports.

All six patients with prosthesis loosening underwent revision surgeries, at an average of 9.2 years after the primary surgeries; in three cases the original prosthesis was used, while in the other three cases hinge knees were replaced by rotating hinge knees. The two patients with sinus tracts underwent phase‐I revision surgery. In one patient the incision healed well with no infection and loosening in the following 6 years of follow‐up, however, the affected limb was eventually 6 cm shorter than the contra lateral limb, with a limb function score of 23. The other patient again suffered from infection one year after the revision operation, and because the sinus was difficult to cure even after repeated debridement, this patient underwent amputation 2 years after revision surgery. The other four cases were followed up for 3–5 years, during which time no prosthesis‐related complications developed. Other than shortening, limb function was significantly improved compared to that before revision surgery. These results demonstrate that a better outcome can be achieved as long as treatment of loosening of the prosthesis is timely. The advanced technologies now available in the field of artificial joints, such as antibiotic‐containing bone cement, should also be applied in revision surgery to ensure long‐term postoperative efficacy. Moreover, for prostheses which are complicated by sinus tracts, phase‐II revision surgery needs to be performed.

One patient in this study needed a second operation due to prosthesis‐proximal bone fracture that occurred 13 years after the primary surgery. Because there was no loosening between the prosthesis and femur, the patient was treated by simple fixation of the proximal femur and increasing the local bone graft. The fracture had healed well by 3 months after surgery; except for 3 cm of shortening. The limb function was also satisfactory during the subsequent 2 years of follow‐up, with knee flexion of 70°. In this kind of case, it is very important for the outcome to determine whether there is prosthesis loosening before the operation and to select the right surgical technique. If there is prosthesis loosening, treatment of the fracture must be accompanied by revision of the prosthesis.

In this study, limb shortening was a major problem that affected the patients’ limb function after surgery, and the limb shortening gradually worsened over time. In this study, limb shortening occurred in 9 cases, the limb shortening rate being 50% with a maximum shortening of 9 cm (Fig. 1).

The long term results of this study indicate that satisfactory clinical results can be achieved by en bloc resection of tumor and reconstruction with prosthesis for GCT around the knee, prosthesis loosening and affected limb shortening being the main long term problems. In the opinion of the authors, prosthetic replacement for GCT should be used only where there is tumor with extensive destruction of the bone structure, pathological fracture, or difficulty in reconstruction after intralesional curettage. Otherwise, intralesional curettage and reconstruction by various methods is the preferred method.