Pedicle screw placement has been widely used in spine surgery. It has gradually developed from free hand method in open surgery to percutaneous insertion in minimally invasive methods with different auxiliary equipment applied during the procedures of screw placement. However, in any case, the safety and accuracy of pedicle screw placement are essential. Thus, different instrumental techniques have been applied to increase intrapedicular accuracy and reduce cranial facet violation.[6] Recently, various robots including SpineAssist/Renaissance/Mazor X robots (Mazor Robotics Ltd., Caesarea, Israel), ROSA (Zimmer Biomet Robotics, Montpellier, France), and TiRobot system (TINAVI Medical Technologies Co. Ltd., Beijing, China) were designed to improve surgical accuracy and reduce radiation exposure.[2, 12] However, the outcomes concerning the superiority in intrapedicular accuracy and cranial facet joint protection of RA over FP and FH are conflicting. [2, 6, 11–17] The studies regarding TiRobot RA versus FP and FH are rare in previously published literature. Meanwhile, the thoracolumbar segments are the most common areas for spinal fractures and the main levels, where different pedicle screw placement techniques can be applied. Therefore, the study was designed to evaluate the rate and risk factors of the pedicle screw placement accuracy and cranial FJV between RA, FP, and FH techniques in thoracolumbar pedicle screw implantation.
For intrapedicular accuracy, Han et al.[12] and Feng et al.[9] conducted RCTs with TiRobot system. They concluded that the RA techniques were associated with significantly higher accuracy rates of 95.3% and 98.2% than the FH methods with the rates of 86.1% and 93.1%, respectively. Meanwhile, Yang et al.[11] revealed that Renaissance RA method showed remarkably increased inserting accuracy rate of 93.8% compared with the FP technique with the rate of 73.8%. However, some studies indicated that RA techniques performed no advantage over the conventional pedicle screw placement. Kim et al.[13] reported that no significant difference existed between Renaissance RA and FH groups regarding intrapedicular accuracy in an RCT, the outcome of which was similar to that in another prospective RCT conducted by Hyun et al.[14] Furthermore, Ringel et al.[15] found that the intrapedicular accuracy in FH was superior to that in RA technique. We have found in the present study that no significant differences existed between RA and FH techniques in terms of optimal accuracy, but significantly lower rates of optimal accuracy of FP than RA and FH were discernible. In addition, no remarkable differences were found among the three techniques in terms of the clinically acceptable accuracy. The results might be explained as follows. On the one hand, with the 3D images that provide more intuitive anatomical landmarks, the RA system can automatically formulate the optimal screw entry point and trajectory in accordance with the specific pedicle shape in different segments to reduce the manual errors and improve the insertion accuracy. Comparatively, although no auxiliary imaging was found in the FH method, the anatomical landmarks can be clearly revealed during the operation. Thus, the surgeons can identify the ideal entry point and insert screws through optimal trajectory in the surgical field of open version. Therefore, the RA and FH methods showed high intrapedicular accuracy. However, when inserting the screws under fluoroscopy guidance in the FP method, the surgeon only relied on 2D images and limited tactile feedback to determine the entry point and trajectory. Meanwhile, the pedicle shadow was difficult to identify on the anteroposterior and lateral images, thereby easily causing the entry point of screws to the inside and reducing the abduction angle of the screw. As a result, the screws were more likely to penetrate the cortex of the pedicles in the FP technique compared with the RA and FH techniques.
Cranial FJV has been regarded as a crucial risk factor of ASD.[3, 6, 8] Moreover, the facet violation contributes to the relative displacement and angular deformity of the vertebrae, thereby resulting in the postoperative back pain and the instability of spine.[4, 6] Many studies have demonstrated that the FP technique showed higher rate of FJV than FH techniques.[8, 18, 19, 28] Babu et al.[19] compared the effects of FP versus FH on facet violation and found that 40.2% of the screws in FP group caused FJV. This finding was significantly higher than the 34.1% in FH group. Meanwhile, Teles et al.[8] performed a multivariate regression analysis showing that the risk of FJV caused by FP method was 3.31 times that of FH method. Moreover, the majority of literature reported that RA methods caused less cranial FJV than FP technique although both methods were percutaneous minimally invasive techniques.[10, 13, 16, 18, 19] Han et al.[12] performed an RCT with 1116 pedicle screws, revealing that none of the screws in the RA group violated the cranial facet joint. This finding was remarkably lower than that of 12 screws (2.1%) in the FH group. In addition, Yang et al.[11] and Archavlis et al.[7] found that RA showed evident advantages in facet joint protection with the rates of 94.9–95.0% over FP with the rates of 78.0–84.4%. Our comparative analysis revealed a significantly lower rate of cranial FJV in the RA group than in the FP and FH groups. However, no significant difference was found between the FP and FH methods.
The outcomes might be attributed to the following factors. The first factor is the guidance equipment, the selection of which can directly affect the clarity of the anatomical landmarks during the operation. This factor influences the rates of insertion accuracy and violation of facet joint. Different from the FP method mainly relying on the 2D images, the RA system can automatically identify the entry point and preplan insertion trajectory in accordance with the 3D images, which show anatomical structures clearly.[21] Thus, the screws easily pass through or rub the facet joints. The second factor is the resistance from soft tissues around the spine. RA and FP can protect the screws from strength of muscles with sleeves, and the pressure from soft tissue had a greater impact on FH, which affected the selection of entry point and direction of the trajectory.The entry point in the FH group tends to be inward, and the abduction angle is reduced. As a result, the FH technique performed no advantage over the RA technique; but it showed similar outcomes to the FP technique. The third factor is the intrapedicular accuracy, where the RA methods performed well in the optimal implantation of pedicle screws according to the preplanned trajectory. As a result, the cranial FJV might be avoided as the preoperative planning. The fourth factor is the distance of pedicle screws from the facets. Kim et al.[13] reported that the mean distance of pedicle screws from facets in the FH group (2.7 ± 1.6 mm) was remarkably closer than in the RA group (5.2 ± 2.1 mm), indicating a lower possibility of the RA system in violating the cranial facet joints. The fifth factor is the learning curve effect. Kam et al.[29] found that RA pedicle screw placement had a very short (almost no) learning curve, showing that the RA technique was less demanding in surgical experience and skills of surgeon due to the automated and precise procedures. However, a higher level of skills is required for surgeons to increase the rate of intrapedicular accuracy and decrease the rate of FJV in the FH and FP methods.
The characteristics of the adverse events among the three techniques were reported. In the FH and FP groups, most deviations of screws were medial, thereby resulting in severe complications, such as spinal cord and nerve root injuries. However, in the RA group, most deviations were lateral, thereby reducing the possibility of severe consequences and revealing a safer choice for screw implantation. In addition, more than 90% cranial FJVs were caused by the screw head in three groups. This finding reminded the surgeons to insert screws in less depth on the premise to ensure spinal stability for facet joint protection. Moreover, for the complications, no patients underwent postoperative revision caused by screw malposition, and 3.3% of patients received intraoperative revision caused by screw malposition in the FH and FP groups. These results were higher than those of the RA group. Furthermore, no patients suffered from wound infections after surgery in three groups.
The factors that potentially affected the intrapedicular accuracy and cranial FJV need further investigation. Kim et al.[5] retrospectively evaluated 488 percutaneous pedicle screws in 110 consecutive patients and determined the obesity, measured by BMI, as the risk factor of screw malposition, because the obese or older patients would have hard muscles and higher resistance of soft tissues. Thus, the difficulty of selecting the best entry point and trajectory is increased.[19, 30] The statistical analysis also demonstrated that the FP pedicle screw placement technique was the risk factor for accurate insertion. The surgeons failed to accurately identify the landmarks with limited visual and tactile feedback, depending on the 2D images. Moreover, the L3 segment instrument was a protective factor for intrapedicular accuracy. We assumed that the diameters of pedicles were larger in the lower segments than in the T10-L2 levels, thereby reducing the possibility of pedicle cortex breach in L3 segment. Meanwhile, the pedicle angle and the distances from skin to insertion point were larger in lower lumbar segments. This finding explains the protective factors of larger pedicle angle and distances from skin to insertion point for intrapedicular accuracy and cranial facet joint protection. Furthermore, we found that the facet osteoarthritis was the risk factor for cranial FJV due to the fact that the osteoarthritis of facet joints distorted the contour of the anatomical landmark in pedicle screw placement, thereby ultimately violating the cranial facet joint.[31]
The following limitations should be interpreted in this retrospective study. First, the clinical outcomes among the three techniques were not reported in the current study, and the relationship between the different grades of malposition and clinical outcomes need further investigation for the future work. Second, the clinical results were not discussed in this study, and the radiographic outcomes were obtained in short term after surgery without long-term follow-up. However, we assumed that the short-term results concerning the safety and accuracy of pedicle screw placement were crucial issues, thereby providing reference and reflection for surgeons. Third, the number of patients and screws included in the study is limited and the prospective trials with large sample size and high quality are needed in the future.