Robot-assisted core decompression combined with bone grafting can improve the early clinical results in femoral head necrosis: a prospective cohort study

Background: To explore the clinical effect of robot-assisted core decompression combined with bone grafting in the treatment of femoral head necrosis involved in early stage. Methods: This study is a prospective cohort study. The study included 49 patients (78 hips) who attended the Department of Orthopedics and Joint Surgery of the Second Aliated Hospital of Xi'an Jiaotong University from August 2019 to February 2021. All the patients suffering Association Research Circulation Osseous (ARCO) II stage of femoral head necrosis underwent core decompression and bone grafting. Among the patients undergoing surgery, 30 patients (54 hips) were treated with traditional surgical methods, and 19 patients (24 hips) were assisted by the made-in-China orthopedic robot system. All operations were performed by the same operator. The baseline data of the two groups of patients, the time of unilateral operation, the number of unilateral X-ray uoroscopy, the Harris hip score (HHS) at the last follow-up after surgery, the visual analog score (VAS), and the collapse rate at the last follow-up were collected and compared. Results: A total of 41 patients (70 hips) were followed up, including 24 cases (42 hips) in the traditional surgery group and 17 cases in the robot-assisted group (28 hips). The average follow-up time of all cases was (13.9±3.4) months (range: 8-18 months). At the last follow-up, a total of 13 patients (13 hips) suffered femoral head surface collapse, including 11 patients in the traditional surgery group (11 hips) and 2 patients in the robot-assisted group (2 hips). The rate of femoral head collapse between the two groups had statistical difference. The average operation time of unilateral hip in the traditional operation group was (22.5±5.5) min, and (18.2±4.0) min in the robot-assisted group, with signicant difference. The number of X-ray uoroscopy of unilateral hip in the traditional operation group was (14.7±3.1) times, and (10.1±3.1) times in the robot-assisted group,

At the last follow-up of the two groups, there were signi cant differences in VAS, and signi cant difference between the two groups was also detected.
Conclusion: Core decompression combined with bone grafting have a de nite effect in the treatment of femoral head necrosis at early stage. Compared with traditional surgery, robot-assisted surgery can achieve better short-term results and head preservation rate. Background Osteonecrosis of femoral head (ONFH) remains a major challenge to orthopedic surgeons due to its unclear pathogenesis (1). Clinically, most patients with ONFH have progressed to the stage of collapse at the time of treatment, and the normal structure of the hip joint is destroyed, resulting in secondary hip arthritis. The vast majority of patients need to receive joint replacement therapy, which brings serious burden to the family and society (2). Therefore, how to intervene ONFH before the collapse of the femoral head, delay or even reverse the process of femoral head necrosis, and delay the time of joint replacement are urgent clinical problems to be solved.
Core decompression is an important means for early intervention of femoral head necrosis (3). In order to enhance the decompression effect, a large number of studies have focused on the implant, but the improvement of the surgical technique itself has been neglected (4,5). With the gradual application of intelligent robots and navigation technologies in clinical practice, some achievements have been made in the eld of orthopedic surgery assisted by robots(6). Due to the congenital advantages of orthopedic robots in the operation of bone tissues, they have been widely used in the elds of trauma, spine suegery and joint replacement, and have obtained good results(6, 7). The Second A liated Hospital of Xi 'an Jiaotong University is one of the few hospitals owning made-in-China orthopedic robots. The robots have helped complete a large number of minimally invasive fracture surgery, pedicle screw implantation and other operations. Since August 2019, we have completed a number of core decompression and bone grafting of femoral head using the orthopedic robots, and conducted a retrospective study on the patients. The results are reported as follows.

Research object
This study was a retrospective analysis. A total of 49 patients (78 hips) who underwent core decompression and bone grafting for femoral bone necrosis at the International Society of Bone Circulation (ARCO) II stage in the Department of Bone and Joint Surgery, the Second A liated Hospital of Xi 'an Jiaotong University from August 2019 to February 2021 were included. All hip joints were examined by X ray, computed tomography(CT) and MRI before operation, which indicated different scope and degree of femoral head necrosis. According to the treatment, the hip joints were divided into two groups.
Inclusion criteria: patients with femoral head necrosis diagnosed by history, symptoms, signs and MRI, all patients were fully informed of the risks of surgery and signed surgical informed consent. Exclusion criteria: a. patients diagnosed with ARCO stage III or above; b. patients with serious cardiovascular and cerebrovascular diseases or tumors; c. people with mental illness; d. refusal to operate. Ultimately, 41 patients (70 hips) were followed up, including 24 patients (42 hips) in the conventional surgery group and 17 patients (28 hips) in the robot-assisted group. There was no signi cant difference in preoperative data between the two groups (p<0.05).

Surgical Procedure
All surgical procedures were performed by the same team of doctors. Surgical operations in the robot-assisted group: Preoperative CT scan of the hip joint (thickness: 0.625mm) was completed, and DICOM data were obtained and imported into the robot, and surgical channel planning was carried out in advance. The patient was supine on a traction bed with both lower limbs abducted and the affected side was elevated. The required equipment was placed in place and connected. The surgical area was routinely sterilized and covered with sterile surgical towels to fully expose the affected surgical area. After install sterile of C arm, the hip X-ray was taken under the two different positions, and the Xray data was imported to the host to t with the preoperative CT reconstruction data. When achieving the condition of the optimal tting, according to the preoperative planning, the surgical path and the nail entry point were ne-tuned, then the data were transmitted back to the manipulator controller. On the operating table, computer was used to simulate the movement process of the manipulator arm. After con rming that the movement process and direction were correct, the manipulator arm were operated. The robot arm automatically run the guide needle and cannula to the skin surface according to the planned direction and the needle feeding point, and the screw could be inserted into the patient's body through this point and direction. The surgeon made an incision at the skin of the needle insertion point, bluntly separated the soft tissue, inserted the kirschner wire along the guide needle trocar, and then inserted the hollow drill bit with a diameter of 10 mm along the direction of the kirschner wire to below the bone cortex. After the length and position were determined by perspective, the kirschner wire and the hollow drill bit were pulled out. After fully scraping the surrounding dead bones with trephine and scraping spoon, allogeneic bone grains were taken for pressure bone grafting. After con rming the bone grafting amount under uoroscopy, 10mm nano-bone rods were screwed in. After con rming the position through uoroscopy again, the wound was rinsed, hemostatic thoroughly, and the incision was sutured. The operation time, the number of uoroscopy and the amount of blood loss were counted. The surgical procedure is shown in Figure 1.
In the conventional group, the patients were supine in traction bed, and the lower limbs were abducted in traction bed. Locating the greater trochanter on the surface through C-arm uoroscopy. Longitudinally cut the skin in length outside the hip joint along the lower edge of the greater trochanter of femur. Blunt dissected the subcutaneous fascia, muscle and periosteum. Con rming the direction of kirschner wire drilling by C-arm uoroscopy. The kirschner wire was inserted into the bone about 4cm, and then the hollow brick with a diameter of 10mm was drilled into the femoral head to the position below the femoral head cortex. After the uoroscopic positioning of the length and position was satisfactory, kirschner wire and the hollow bit were pulled out. After fully scraping the surrounding dead bones with trephine and scraping spoon, allogeneic bone grains were taken for pressure bone grafting. After con rming the bone grafting amount under uoroscopy, 10mm nano-bone rods were screwed in. After con rming the position through uoroscopy again, the wound was rinsed, hemostatic thoroughly, and the incision was sutured. The operation time, the number of uoroscopy and the amount of blood loss were counted.

Postoperative Management
All patients were treated with prevention of infection and analgesia after operation. Postoperative X-ray and CT were reviewed 3 months after surgery. All patients in the two groups exercised half weight bearing activities with crutches until 3 months after surgery.

Statistical analysis
All data were processed by statistical software SPSS 22.0. The measurement data were expressed as Mean±SD and t test was used. Counting data were compared by R×C contingency table chi-square test, p<0.05 was considered signi cant difference.

Basic Information
Forty-one patients (70 hips) were nally included in this study. Baseline data of the robot-assisted and convention groups were collected and compared. No signi cant difference was detected in age

Comparison Of Postoperative Follow-up Data
The Harris hip score, VAS score and the number of collapsed hip joints of all enrolled patients at the last follow-up were further analyzed. The results showed that at the last follow-up, a total of 13 patients suffered hip collapse, including 11 patients in the traditional surgery group and 2 patients in the robotassisted group. There was a statistical difference in the collapse rate between the two groups (χ ²=4.031). At the last follow-up, there was no signi cant difference in Harris score between the two groups, while there was a statistical difference in VAS score (t=3.920). (Table 3)  in robot-assisted group. The collapse rate of femoral head in traditional surgery group was in agreement with other previous reports(8), and the collapse rate of femoral head in robot-assisted group was signi cantly lower than that of other previous studies. In addition, robotic surgery was superior to traditional surgery in terms of operation time, number of intraoperative X-ray uoroscopy and postoperative pain improvement. At the last follow-up, Harris hip score was better in both groups than before, but there was no signi cant difference between the traditional surgery group and the robotassisted group.
There are many classi cation criteria for femoral head necrosis. In this study, the 2019 revised ARCO staging system was used to regularly classify the included patients. In the 2019 revision of the ARCO staging system, stage 0 was removed, and the location and size of necrotic lesions were not used as staging criteria (9). In this study, a total of 13 hips progressed to ARCO III with collapse at the last followup. After case summary of these 13 hips, it was found that the area of femoral head necrosis of 9 hips exceeded 30%, which was classi ed as ARCO IIc of 1994 edition. Eight patients received hormone therapy again during the follow-up period due to primary disease progression. For the hip joint which necrotic area of the femoral head exceeds 30%, the necrotic bone cannot be completely removed, which leaded to the lack of bone grafting through the core decompression canal. The size of necrotic area in the femoral head before operation and the amount of necrotic bone remaining during operation will affect the success rate of surgery. Landgraeber et al. suggested that a better surgical e ciency could be achieved when the postoperative residual dead bone was less than 1000mm 3 (10). Compared with traditional surgery, robot-assisted surgery can reach the necrotic area more accurately and remove the necrotic bone to the maximum extent. However, at present, the related instruments of robot-assisted surgery are still scarce, which cannot give full play to the advantages of precision of robots (11). In addition, for the hip joint with large necrosis area of the femoral head, a simple working path for decompression cannot completely remove the dead bone. In this case, robot-assisted decompression combined with arthroscopic removal of dead bone and bone grafting of the hip joint or open surgery are required to ensure better surgical results (12,13).
In this study, the postoperative Harris score and VAS score of patients in the two groups were signi cantly improved compared with the preoperative score, while the VAS score was signi cantly different at the last follow-up between the two groups, and the Harris score was not signi cantly different. Patients with early necrosis of the femoral head usually present with pain in the hip joint without signi cant range of motion limitation. Therefore, for the evaluation of the treatment effect of early femoral head necrosis, the improvement of VAS is more signi cant. Among the 41 patients included in this study, a total of 29 patients underwent simultaneous surgical treatment for bilateral femoral head necrosis. For this part of patients, Harris score could not evaluate the function of a single hip joint very well, which affected the nal results. The VAS score can be used to evaluate the unilateral hip joint without being affected by the contralateral hip joint. At present, there is a lack of independent comprehensive evaluation criteria for bilateral hip joint.
The robot used in this study is an orthopedic surgery robot made entirely in China, which is mainly used in trauma, spine and other elds. Till now, there is a lack of relevant research on decompression treatment of femoral head necrosis. As such robots has natural advantages for bone tissues, through a combination of preoperative CT data and intraoperative real-time X-ray image, the robot can help complete core decompression precisely and convenient with a more minimally invasive surgical incision.
According to the follow-up data, robot-assisted operation can achieve superior outcome, and effectively reduce the number of intraoperative radiation exposure, intraoperative bleeding, reduce the operation time, and complications related to drilling. By tting the preoperative CT-based reconstruction image with the intraoperative X-ray image, the robot system can realize the drilling point of femoral under the intraoperative three-dimensional eld of vision. At the same time, for some necrotic areas that are di cult to reach, traditional surgery requires repeated puncture to determine the drilling path, which increases the risk of vessel and nerve damage and steel needle fracture. Traditional surgery usually requires a 5cm surgical incision below the greater trochanter and partial dissection of the gluteus medius, which may result in decreased hip abductor strength. The robot-assisted system has been proved to be able to perfectly avoid these complications in other clinical operations (14,15), and no related complications occurred in this study.
Traditional core decompression combined with bone grafting is usually performed under direct vision or uoroscopy with low grafting strength. The virtual wall technology of the robot can effectively and accurately implant the graft material into the necrotic decompression area, which will ensure the support strength and avoid iatrogenic collapse. In terms of implant material, we chose the commonly used nanobone rod which had been introduced in our former research. This bone rod can provide excellent biocompatibility, mechanical compatibility and bioactivity to avoid the possible rejection of other endoplants in our previous study(16), while good ability of bone ingrowth is one of the important factors for successful hip preservation (17,18). At the last follow-up, the collapse rate of the robot group was signi cantly lower than that of the traditional surgery group, suggesting that accurate removal of dead bones and high strength mechanical support can bring better clinical results. In this study, 2 patients with post-operative collapse chose hip replacement, during which a layer of brous tissue was found between the implanted rod and the bone, and no bone ingrowth occurred. How to induce better osteogenesis and bone ingrowsion is an important research direction for achieving better results of core decompression surgery (19,20).
There are still some problems in this study. First of all, since the robot needs to use intraoperative X-ray and preoperative CT reconstruction data for tting, even slight changes in posture will lead to deviation of the results after the X-ray image of hip joint is obtained. Secondly, there is still a lack of robotic core decompression equipment, which can not give full play to the characteristics of robotic surgery. Thirdly, the number of cases included in this study is still small, the follow-up time is not long, and the medium and long term results are still to be observed.

Conclusions
The results of this study suggest that robot-assisted core decompression and bone grafting can reduce the operative time, reduce the number of intraoperative X-ray uoroscopy, relieve pain, and possibly improve the surgical e ciency. Surgical demonstration of robot-assisted core decompression and bone grafting. a: After obtaining the preoperative CT data of the patient's hip joint, the decompression path was planned in the coronal position; b: Verify decompression path after 3D reconstruction of CT data; c: Intraoperative hip X-ray; d: After tting intraoperative X-ray and preoperative CT data, the decompression path was ne-tuned, and the decompression was assisted by a mechanical arm.

Figure 2
Imaging data of a patient with femoral head necrosis undergoing robot-assisted core decompression and bone grafting. a,b,c: X-ray, CT scan and MRI of the lesion side; d: X-ray of the lesion side 8 months after operation; e: CT-scan of the hip 8 months after operation.