Knee joint-preserving tumor resection for malignant bone tumors around the knee allows better function and helps reduce leg length discrepancy in minor patients in the distant future by preserving the growth capacity of the epiphysis[16, 17]. However, no standard reconstruction modality exists because of the rarity of cases that qualify for joint preservation. To provide a reference for the future selection of treatment modalities for knee joint preservation surgery, we retrospectively analyzed reconstruction using a customized prosthesis and reconstruction using a liquid nitrogen-inactivated autologous bone graft in terms of surgical difficulty, complications, and MSTS score.
Safety is the most important indicator of limb salvage surgery for bone malignancies[18]. The resection area that reaches a safe surgical margin is of utmost importance in reducing local tumor recurrence. According to Kawaguchi, the surgical margin of high-grade sarcomas must be greater than 3 cm in the absence of neoadjuvant therapy or when neoadjuvant therapy is ineffective, whereas it must be 2 cm when neoadjuvant therapy is effective[19]. All patients in this study, except for three patients with chondrosarcoma, underwent standardized neoadjuvant chemotherapy. Therefore, the resection site was designed to be > 2 cm from the tumor at the end and distal to the articular side. However, on the articular side, the minimum distance between the resection site and tumor was designed to be only 1 cm to preserve the thicker residual bone. Deng reported that in knee preservation limb salvage surgery for osteosarcoma, the difference between the resection position and the preoperatively planned position was 8.3 ± 6.0 mm when the resection was performed with bare hands, whereas the error was only 2.0 ± 1.6 mm when the resection was performed with positioning using a computerized navigation system[20]. Studies on pelvic resection and special resection surfaces have also confirmed that the accuracy of resection with computerized navigation systems is much higher than that of bare-handed resection[21, 22]. Wong also reported the use of computer-assisted positioning in joint preservation surgery for resection 1 cm from the tumor to obtain safe margins [6]. To ensure the safety of the resected bone on the articular side, all patients in this study underwent preoperative radiography, enhanced CT, enhanced MRI, and electroconvulsive therapy. The tumor margins were determined by combining the results of all examinations. During the surgery, a computer-assisted navigation positioning system was used to determine the resection position. After the intraoperative resection, the resected ends were observed to be safe with the naked eye. Tissues of the resected ends were obtained during surgery for pathological examination, and the results were all negative. There were no cases of recurrence at the end of resection during the follow-up period. There was only one case (4.3%) of soft tissue recurrence, which is comparable to the recurrence rates reported for other tumor resections around the knee (2.4–10.5%)[23, 24, 25, 26]. Based on our experience, within a median follow-up period > 60 months, joint preservation tumor resection is a safe and acceptable approach from the perspective of oncology.
In addition to tumor safety, the precise resection position is associated with reconstruction. The two reconstruction modalities do not have the same requirements for resection position accuracy. For reconstruction with a custom-made endoprosthesis, the resection surface close to the joint affects the position and orientation of the endoprosthesis, position of the lateral steel plate, and orientation and length of the fixation screws. Therefore, the resection position must be sufficiently accurate to perfectly match the endoprosthesis. In contrast, the only requirement for the inactivation technique is that the resection position exceeds the designed safety range, making surgical resection much easier.
For joint preservation surgery, it is challenging to stabilize and fix the residual bone close to the joint to an endoprosthesis or inactivated bone and to achieve long-term stability. Reconstruction with a short-stem interpositional endoprosthesis can be used in cases of long residual bone on the articular side. Tsuda et al. reported that a custom-made endoprosthesis with a short intramedullary stem and lateral steel plate can be used for reconstruction when the resected bone is 5 cm away from the joint, and a hydroxyapatite (HA) coating is used on the surface of the endoprosthetic stem and steel plate to minimize the risk of aseptic loosening[27]. Kong reported the use of a customized prosthesis and lateral plate to fix the residual bone, and HA coating was used to enhance osseointegration at the bone-prosthesis interface[17]. Zhao et al. treated five patients with an average residual bone of 2.65 cm at the proximal or distal tibia after tumor resection with a 3D-printed porous endoprosthesis. All patients achieved early biological fixation, with an average time to clinical osseointegration of three months at the bone-prosthesis interface of 3.2 months[11]. In this study, HA coating and 3D printing technology were not used on the customized prostheses, residual bone contact surfaces, or extracortical steel plates, but there was no displacement of the residual bone from the prosthesis and no fixation failure during the follow-up period, and the five-year survival rate of the prostheses was 100%. This may be attributed to the precise design and manufacturing of the prosthesis after the surgeon has determined the resection position based on preoperative imaging, and the position and length of the internal fixation screws were determined before surgery. Moreover, a computerized navigation system was used to determine the resection position during surgery so that the endoprosthesis perfectly matched the residual bone. Six screws on both sides also provided immediate postoperative stabilization. If endoprosthesis manufacturing companies offer HA coatings for resection surfaces or 3D-printed porous structures for contact surfaces in the future, better early- and long-term stability may be achieved.
Long-term survival of liquid nitrogen-inactivated autologous bone grafts requires healing of the host bone. Commonly used inactivation methods include freezing, pasteurization, and irradiation[28, 29, 14, 30]. The liquid nitrogen inactivation method used in this study is simple, inexpensive, and does not rely on special equipment, such as radiotherapy or strict thermal control. Inactivated bone retains good biomechanical strength[31], preserves bone morphogenetic protein (BMP)[32], and retains osteoinductive capacity; however, the healing time of inactivated bone remains much longer than that of normal fractures. Wu reported that 84 patients who underwent surgery using the liquid nitrogen inactivation method achieved 79.8% healing at 18 months after surgery[14]. Araki reported that the healing time of 37 patients after surgery using the liquid nitrogen inactivation method was 3–24 months[30]. In the inactivated autograft group, the longest healing time was 16 months on the articular side and 26 months on the diaphyseal side. Therefore, it is necessary to provide an internal fixation with sufficient strength before healing. Frisoni‘s study[33] of reconstruction of backbone defect with bone allograft suggested that steel plate fixation was superior to intramedullary fixation for increasing the strength of reconstruction, which is supported by the findings of Chen et al. using inactivated bone grafts[34]. In order to provide stronger internal fixation, for all the patients in the inactivated autograft group in this study, the inactivated bone marrow cavity was filled with bone cement and then fixed with steel plates on both sides. No type 2 or 3 complications due to insufficient bone strength occurred during the follow-up.
In this study, there was no significant difference in the MSTS scores at the final follow-up visit between the two groups, which were 91% and 94% higher than the MSTS scores for conventional prostheses (71–88.6%)[35, 36, 24, 9]. Previous reports on joint preservation surgery have shown that higher MSTS scores (90–96.7%) are associated with maximum intraoperative preservation of the cruciate ligaments, medial and lateral collateral ligaments, and contralateral articular surfaces[37, 4, 6].
Based on our clinical experience, we have become familiar with the advantages and disadvantages of using custom-made endoprostheses and liquid nitrogen-inactivated autologous bone grafts for the reconstruction of preserved joints. The endoprosthesis group was not limited by the degree of osteolysis and had a shorter operation time, faster postoperative recovery, early weight bearing, and good postoperative function. However, the design and manufacturing of prostheses requires professional skills, specialized equipment, and good medical–industrial integration. Bone cement is currently used to fix prosthetic stems, which have a high long-term risk of loosening. Inactivated bone grafts require simple equipment, have a high healing rate, function well, have a low complication rate, and can be restored using an endoprosthesis after failure. However, they are not suitable for patients with severe osteolysis or pathological fractures. Moreover, there are many steps for inactivation, rewarming, and fixation of the graft during surgery, resulting in a long duration of surgery, which might increase surgical complications. Inactivated bone grafts also have slow postoperative recovery and can only bear weight after complete healing. Therefore, liquid nitrogen-inactivated autologous bone grafts are preferred for patients suitable for joint preservation surgery. Reconstruction with a custom-made endoprosthesis was performed if 50% of the cortical bone was lysed or a pathological fracture was present.
This study had some limitations. First, this was a retrospective study with a small sample size, and the two patient groups were not randomized. Second, as the liquid nitrogen inactivation method started in 2015 at our institution, all patients underwent reconstruction with a custom-made endoprosthesis before that. Third, reconstruction with a bone allograft is another possible treatment modality for joint preservation in addition to endoprosthesis and inactivated autologous bone grafts for joint preservation surgery. However, as there is no bone bank in our medical institution, matching bone allografts were difficult to obtain; therefore, reconstruction using bone allografts was not included in this comparative study. Fourth, aseptic loosening of the endoprosthesis and fracture of the inactivated bone graft increase with time, which may result in reconstruction failure[38]. The current follow-up period is still relatively short, but we will continue to follow-up and obtain 10-year survival data for patients and grafts.