Biomechanical Evaluation of Oblique Lateral Locking Plate System for Oblique Lumbar Interbody Fusion: A Finite Element Analysis

Objective: Oblique lateral locking plate system (OLLPS) with the locking and reverse pedicle track screw conguration is a novel internal xation designed for oblique lumbar interbody fusion(OLIF). It is placed in a single-position through the oblique lateral surgical corridor to reduce operative time and subsequent complications of prolonged anesthesia and prone positioning. The purpose of this study was to verify the biomechanical effect of OLLPS. Methods: The intact nite element model of L1–S1(cid:0)Intact(cid:0) was established based on CT images of a healthy male volunteer. The L4-L5 intervertebral space was selected as the surgical segment. The surgical models were established separately according to the OLIF surgical procedures and the different internal xations: (1) stand-alone OLIF (SA); (2) OLIF with 2-screw lateral plate (LP-2); (3) OLIF with 4-screw lateral plate (LP-4); (4) OLIF with OLLPS (OLLPS); and (5) OLIF with bilateral pedicle screw xation (BPS). After validating the intact model, the physiological loading was applied to the superior surface of L1 to simulate exion, extension, left bending, right bending, left rotation, and right rotation motions. The evaluation indexes included the L4/5 range of motion (ROM), the L4 maximum displacement, and the maximum stress of the superior and inferior endplate, cage, and supplemental xation. Results: In OLIF surgery, OLLPS provided multiplanar stability which was similar to that of BPS. Compared with LP-2 and LP-4, OLLPS had the better biomechanical properties in enhancing the instant stability of the surgical segment, reducing the stress of the superior and inferior endplates of the surgical segment, and reducing the risk of cage subsidence. Conclsions: With the minimally invasive background, OLLPS can be an alternative to BPS in OLIF and has a better prospect of clinical promotion and application.


Introduction
Advances in minimally invasive technology open up a new era of spine surgery. Oblique lumbar interbody fusion (OLIF) is widely used in the treatment of degenerative diseases of the lumbar spine because of its advantages such as minimally invasive indirect decompression, e cient interbody fusion and rapid postoperative rehabilitation. The OLIF procedure uses the natural gap between the anterior border of the psoas major and the abdominal vessel to access the lateral aspect of the vertebral body for interbody fusion, eliminating the need for expensive neurological monitoring and allowing multiple segmental fusions with a single small incision using the "sliding window" technique. 1 The application of large size cage has good results for the correction of spinal force lines, but the high postoperative rate of the cage subsidence is an unavoidable problem. A clinical study by Abe et al. 2 reported a 9.03% rate of cage subsidence. However, Zeng et al. 3 found a subsidence rate up to 19.8% in patients who underwent standalone OLIF surgery during postoperative follow-up. Indirect decompression failure after cage subsidence may cause a range of clinical symptoms, high reoperation rate, huge nancial and health burden. In patients with a high risk of cage subsidence, OLIF with supplementary xation may be a more prudent solution.
Even though OLIF is an effective treatment, there is still no consensus regarding the ideal supplemental xation. Different xation methods have been reported in the literature, for example, posterior xations include pedicle screw xation, modi ed cortical bone trajectory (CBT) screws, transfacet screw xation, and lateral xations contain anterolateral screw xation, lateral plate. [4][5][6][7][8] Bilateral pedicle screw xation has been regarded as the gold standard, but it may bring about adjacent spondylosis or damage to paravertebral structures. 8-10 Furthermore, pedicle screw placement requires turning the patient from a lateral to a prone position during surgery and the additional posterior surgical incision, which increases the perioperative risk. The lateral xation can avoid the induced problems of intraoperative patient ipping and reduce operative time and risk. The shape of lateral plate is more rounded and blunt than the anterolateral screw xation, which is less invasive to the surrounding tissues and more promising for application. However, the biomechanical stability of the 2-screw lateral plate currently used in OLIF has been questioned, and the acceptance of clinical application is low. [11][12][13] We plan to improve the design of the lateral plate to enhance its biomechanical properties. An in vitro biomechanical experiment by DenHaese et al. 12 con rmed the superior biomechanical properties of the 4-screw lateral plate compared to the 2-screw lateral plate. In a clinical study, Sardhara et al. 14 proposed a new theory of reverse pedicle screw xation (RPSF), which relies on the special trajectory of the lateral screws to achieve 3-column xation (Fig. 1). We designed the oblique lateral locking plate system (OLLPS) for the OLIF procedure by using the angular stability of the locking screw structure and the RPSF theory, combining the understanding of the connotations of OLIF (Patent No. ZL 202022949889.1). In this study, the biomechanical properties of OLLPS were evaluated to provide a biomechanical basis for its clinical application by establishing 3-dimensional surgical models with various internal xation and using nite element analysis.

Materials And Methods
Oblique lateral locking plate system (OLLPS) design The oblique lateral locking plate system (OLLPS) consists of a palm-leaf fan-shaped plate and 4 anglespeci c locking screws (Fig. 2). The OLLPS is easy to assemble and compatible with OLIF cage, mainly for the L2-L5 intervertebral space. 15 The OLLPS design is based on the concept of Biological Osteosynthesis for bone trauma. We may consider the superior and inferior bony endplates of the surgical segment as the two ends of the fracture. Accordingly, intervertebral fusion can be considered as the healing of the fracture. The stability of conventional lateral plates is achieved by the friction between the bone surface and the plate. Most of the lumbar vertebral body surfaces have irregular morphology, and the xable range of the oblique anterior side of the lumbar spine is narrow, resulting in the inability of the conventional lateral plate to closely adhere to the vertebral body. The stress is too concentrated at the contact surface between the screw and the plate, with the risk of the loose screw, broken screw, and internal xation failure. 16 The OLLPS screws are threaded into the lateral plate at a speci c angle, providing angular stability and weight load can be distributed by the screw and lateral plate, which can share the endplate stress. The placement of OLLPS is referenced to a bone trauma locking plate, does not have to closely adhere to the vertebral bone surface, does not require excessive exposure, causes minimal damage to the vertebral periosteum.
The length of the lateral plate is limited by the distance between the superior and inferior segmental artery after the placement of the intervertebral cage. The width of the actual operating area of the surgical corridor is thus a major parameter, determining the width of the plate. The OLIF clinical anatomical studies give us detailed design parameters, and we set the lateral plate width at 22 mm, length at 30 mm-42 mm (in 4 mm increments), and thickness at 5 mm. The overall design of the anatomical lateral plate has a streamlined curved appearance, with a 30° arc on the coronal and axial positions. The screws are designed as 6.0 mm diameter solid or hollow cancellous bone screws, and the length is set at 30 mm-60 mm (in 5 mm increments). The thread of the screw body is far deep and shallow shbone spur type tapered thread, which can effectively increase the contact area of the screw/bone interface and increase the screw holding force.
The special trajectory of the screw is the main point of the OLLPS design. Based on the theory of RPSF, we used a 3-dimensional topology optimization method to set the two ventral screws (screws #1 and #3 in Fig. 2) to reverse the pedicle trajectory, pointing to the cortical bone area at the junction of the contralateral pedicle and vertebral body, but without penetrating the contralateral bone cortex. However, the space in the region is limited and cannot accommodate multiple screws at the same time, While we design OLLPS to jointly use in multiple segments. Therefore, we adjusted the screw trajectory for topological optimization. Speci cally: screws #1 and #4 are at an angle of 25° to the horizontal centerline of the plate (to extend the screw force arm and increase the contact area of the screw/bone interface) and at an angle of 5° to the vertical centerline of the plate (screw #1 points to the ventral cortical bone area of the contralateral pedicle region to prevent overlapping of the trajectories of screws #1 and #3 of the upper and lower xation plates when multiple segments are used together); screws #2 and #3 are at an angle of 0° to the horizontal centerline of the plate (parallel to the endplate, increasing the endplate support) and 15° angle to the vertical centerline of the plate (screw #3 points to the contralateral pedicle area). Any two screw trajectories are not in the same plane and form an angle with each other, forming a multi-dimensional multi-axial locking. The plate and screws of OLLPS are made of titanium alloy (Ti6Al4V). The risk of subsidence is minimal when the cage is placed in zones II and III, and OLLPS is also usually placed laterally anterior to the vertebral body corresponding to this. 17,18 Operators determine the correct orientation of the lateral plate by the non-asymmetric clamping notches on both sides of the plate during the placement operation and use a guide to assist in screw placement based on navigation or uoroscopic assistance. We recommend that the #3 reverse pedicle screw should be rstly placed, followed by the oblique contralateral #2 screw to ensur e that the lateral plate is at against the surface of the vertebral body, and then the remaining screws are placed in sequence.
Construction of an intact lumbar nite element model A healthy male volunteer (34 years old, weight 70 kg, height 175 cm, no previous lumbar spine disease) was recruited for this study, and 439 images were obtained using GE 64-slice spiral CT for continuous thin-section scanning of the L1-S1 vertebrae (slice thickness 0.625 mm) after signing an informed consent. The CT images in Dicom format were sequentially imported into Mimics 23.0 (Materialise Inc., Leuven, Belgium), 3-Matic (Materialise Inc., Leuven, Belgium) software to establish models and perform the smooth restoration. The smoothed model was processed using SolidWorks 2017 CAD (SolidWorks Corporation, Concord, MA, USA) to construct the endplates, annulus brosus, nucleus pulposus, and facet joints. The solid model was meshed using HyperMesh (Altair Technologies Inc., Fremont, CA). Finally, ANSYS (Ansys Inc., Canonsburg, PA, USA) was used for material property de nition, model assembly, and nite element analysis.
The nite element model ( Fig. 3) includes the L1-S1 vertebral body, intervertebral disc and ligament  20 and the material properties of the components are shown in Table 1. Table 1 Material properties used in the nite element models from the literature. 11 Construction of the surgical nite element models Cage and internal xation were performed with a tetrahedral element (Solid187). The L4-L5 intervertebral space was used as the surgical segment, and the annulus brosus, nucleus pulposus, and cartilage endplates were removed. On this basis, a stand-alone OLIF (SA) model was constructed based on the CLYDESDALE (Medtronic Sofamor Danek USA, Inc.), with 6° of anterior convexity, 50 mm in length, 18 mm in width, and 12 mm in anterior height, and made of Polyetheretherketone (PEEK). The internal xation models were all constructed based on the SA model (Fig. 4). The OLIF with 2-screw lateral plate (LP-2) model was constructed based on the Pivox Oblique Lateral Spinal System (Medtronic Sofamor Danek USA, Inc.), in which the plate is 34.6 mm length, 12 mm width, 5.4 mm thickness, and the screws are 45 mm length and 5.5 mm outer diameter (15° angle between the screw and the horizontal centerline of the plate, 0° angle between the screw and the vertical centerline of the plate). The OLIF with 4-screw lateral plate (LP-4) model was constructed on the basis of the LITe plate system (Stryker USA, Inc.) with the plate length of 28 mm, the width of 21 mm, the thickness of 4.5 mm and the length of the screws of 45 mm, 5.5 mm outer diameter (the two screws on the ventral side were at an angle of 20° to the horizontal centerline of the plate, with a vertical centerline angle of 0° and the two screws on the backside were parallel to the endplate with a vertical centerline angle of 0° ). The OLIF with oblique lateral locking plate system (OLLPS) model was constructed based on the oblique lateral locking plate system with the plate length of 32 mm, the width of 22 mm, the thickness of 5 mm, and the length of the screws of 45 mm, 6.0 mm external diameter (the screws tilt angle as described previously). The OLIF with bilateral pedicle screw xation(BPS) model was constructed based on the CDH SEXTANT II (Medtronic Sofamor Danek USA, Inc.) with the pedicle screws diameter of 6.5 mm, length of 45 mm and the diameter of the rods of 5.5 mm, length over the upper and lower pedicle screw spacing. All materials of internal xation were titanium alloy (Ti6Al4V), and the material properties of the implants are shown in Table 1.

Contact, boundary and loading conditions
The connection between the constructed model disc/cage and the superior and inferior endplates was made by means of the no separation contact. The contact surface of the cage and the endplate had a toothed anti-dislocation structure with a friction coe cient of 0.8. 24 Frictional contact existed between the facet joints, the screw/bone interface, and the interspace between the screws and threaded holes (no sliding between the screws and the threaded holes in the OLLPS model), and the friction coe cient was set at 0.2 to simulate the immediate postoperative state. The inferior surface of the S1 vertebra was xed, which means that all nodes of the inferior endplate of the S1 vertebra were constrained from moving in any direction. A 150 N axial compressive preload was set on the upper surface of the S1 vertebra to simulate physiological load in upright state, and a pure moment of 10 N·m was applied to simulate the model in six directions: (1)  Therefore, the results of the lumbar spine in exion, extension, left bending, right bending, left rotation, and right rotation were recorded separately in this study.

Finite Element Model validation
The L1-S1 segment range of motion (ROM) for different motions of the intact model under a 150 N axial compression preload and a 10N·m moment load was measured and compared with the outcomes of in the vitro experiment conducted by Yamamoto et al. 26 (Table 2). The total L1-L5 ROM of the intact model in exion-extension, lateral bending, and rotational mobility was measured by applying a 7.5 N·m moment load and compared with the nite element model investigated in Dreischarf et al. 27 (Fig. 5A). The L4/5 intervertebral disc pressure (IDP) was tested with pure compressive forces of 300 and 1000 N, which was compared with in vitro experimental data of Brinckmann & Grootenboer 28 and the calculated results of the nite element model of Zhang et al. 29 , respectively (Fig. 5B). More segments and approaches were used in our validation than in other similar studies, and the results obtained were in good agreement with those reported in the literature, ensuring the validity of the intact model. Considering that the most frequent lumbar motions were exion and extension, BPS had the best overall performance in restricting lumbar motion. Despite this, the restrictive effect of OLLPS on the operated segment was signi cantly improved relative to LP-2 and LP-4, and the ability to restrict lumbar motion in all directions was higher than 60%, which was comparable to BPS in restricting lateral bending and rotational activities.

L4 Maximum Displacement
The maximum displacement nephogram of L4 with various xation options in six motion modes are shown in Fig. 7. The maximum spatial displacement of the superior vertebral body of the surgical segment relative to the inferior vertebral body indirectly re ected the stability of the surgical segment. It may be simply interpreted as the smaller the displacement degree is, the more stable it is. However, the results of relative displacement contained not only the displacement change of the surgical gap but also the displacement increment caused by the vertebral deformation, which was less reliable than ROM, but can be used as an auxiliary criterion for ROM evaluation of lumbar spine stability. 30 Compared with the intact model, the L4 vertebral displacement in the surgical models was decreased in six motion modes. The data of the three lateral xation modalities showed a stepwise performance, in short, OLPPS was better than LP-4 and LP-4 was better than LP-2. OLPPS was slightly better than BPS in limiting lumbar motion during lateral bending and rotational movements of the lumbar spine, whereas BPS was better in exion and extension (Fig. 6B).
Endplate Stress Fig. 8 describes the maximum stress in the superior and inferior endplates of the surgical segment. Fig. 9 visualizes the stress distribution of the L5 superior endplate.  Fig. 8 shows that the stress of the L5 superior endplate was generally higher than that of the L4 inferior endplate under the same internal xation and motion conditions. This implied that the risk of cage subsidence is higher in the L5 superior endplate than in the L4 inferior endplate. This result was consistent with the clinical study by Hu et al. 31 , who found that the subsidence probability of the superior endplate was signi cantly higher than that of the inferior endplate during the clinical radiographic follow-up of the surgical segment. Fig. 9 clearly shows that the stress of the L5 superior endplate was concentrated in the epiphyseal ring and cortical compact of the vertebral endplate in contact with the cage.
Cage Stress Fig. 10A shows the cage stress with various xation options in six motion modes. The cage stress of SA was maximum in all models, especially in left and right rotation, which was reduced after the implantation of any internal xation. The cage stress of LP-2 was slightly lower than that of SA, and the cage stress of LP-4 was lower than that of LP-2. Among all surgical models, BPS had the lowest cage stress in the exion-extension motion state, and OLLPS had the lowest cage stress in the left-right bending and left-right rotation states. In the comparison of BPS, OLLPS had 8.86%, 12.56%, 3.76%, and 13.85% lower cage stress in left bending, right bending, left rotation, and right rotation, respectively. Fig.  11 shows the nephogram of the cage stress with various xation options. The stress distribution was concentrated at the periphery of the cage in all motion states (especially in the dorsal region), which corresponded to the high-stress region of the endplate.

Supplemental Fixation Stress
The maximum stress of supplemental xation is shown in Fig. 10B MPa, respectively, which were higher than that of LP-2, LP-4 and BPS. The maximum stress of LP-2 appeared in LR and RR, and the maximum values of LP-4, OLLPS and BPS appeared in FL and RR. Further analysis reveals that the maximum stress of LP-2 and BPS occurred at the interface between the superior screw and the bone, and the maximum stress of OLLPS occurred in the plate (Fig. 12). OLLPS was endured higher stress than other internal xations, but was far below than the fatigue strength 310-610 MPa and yield strength 789-1013 MPa reported in the literature for titanium internal xation. 16 The maximum stress of BPS was smooth in all motions, and there was no sudden increase in stress in one motion mode, suggesting a balanced biomechanical performance in all directions of motion of the lumbar spine.

Discussion
Biomechanical properties are the important element of the novel internal xation research. There are two mainstream biomechanical assessment methods: one is in vitro cadaveric biomechanical experiments; the other is nite element analysis. In vitro cadaver biomechanical experiments are di cult to implement because of the strict requirements for cadaver conditions and laboratory equipment. With the advancement of nite element analysis technology, the reconstruction of nite element models for the lumbar spine based on normal human CT images has become the accepted method for biomechanical analysis. Through the simulation of different internal xation models, the comparative analysis of the activity of surgical segments and the stress characteristics of each structure after applying physiological load can help us understand the biomechanical characteristics of the current internal xation methods. In the study, the biomechanical properties of the combined application of OLLPS and OLIF were discussed for the rst time. We innovatively introduced the concept of Biological Osteosynthesis for bone trauma into vertebral fusion and introduced the angle stabilization structure and reverse pedicle screw trajectory design on the basis of conventional lateral plate xation to improve the stability of OLLPS.
The retrospective study by Silvestre et al. 15 rstly named and discussed the OLIF procedure. 179 patients who underwent OLIF were included in his study, the results indicate that a mean intraoperative bleeding of 56.8 ± 131.3 mL, a mean operative time of 32.5 ± 13.2 min, and postoperative complications mainly consisting of incisional pain (4 cases, 2.2%) and lower extremity symptoms due to sympathetic nerve chain injury (3 cases, 1.7%). Compared with PLIF in traditional posterior open surgery, OLIF had signi cant advantages such as a smaller surgical incision, shorter anesthesia time, less intraoperative bleeding, less postoperative pain, and faster postoperative recovery. 32 Compared with MIS-TLIF, a posterior minimally invasive procedure, OLIF can provide better correction of sagittal parameters and clinical outcomes, more satisfactory restoration of vertebral space height, and earlier intervertebral fusion. [33][34][35][36] During the clinical applications of lumbar interbody fusion, the importance of the posterior lumbar ligamentous complex for maintaining spinal stability has been increasingly recognized. 37 The reports of postoperative intractable low back pain caused by the injuries of bony structures, muscles, and ligaments in the posterior lumbar region were also common. Stand-alone OLIF can minimize structural damage to the posterior lumbar spine. 8,38,39 However, cage subsidence has been a problem in the stand-alone technique. The clinical application effects and complications of stand-alone OLIF can be explained in biomechanical studies. According to the mechanism of fracture healing, the relative stability of the fracture end is a prerequisite for healing. Our study suggested that SA had the least decrease of ROM in all motion modes and internal xation modalities, which resulting in the least ability to maintain lumbar spine stability. Our results were consistent with the in vitro biomechanical ndings by Laws et al. 40 The reduction of ROM in SA exion-extension and lateral bending mode is better in rotation mode, which is associated with the design of the fusion cage convex angle and the transverse placement of the lumbar coronal line.
The purpose of supplemental xation for lumbar fusion is to stabilize the lumbar spine, reduce the mobility of the surgical segment, create a stable external environment for fusion, share the stress of the endplate, and reduce the incidence of cage subsidence. 11 The results of this study suggested that all xation modalities enhanced the stability of the lumbar spine structure compared to the SA model, but the degree of stability varied considerably in the different supplemental xation modalities. BPS had the greatest ability to maintain the stability of the lumbar spine in the immediate postoperative period.
However, in our opinion, the effect of the combination of OLIF with BPS is 1+1 <2. As mentioned, the damage to the posterior lumbar structures associated with posterior internal xation plays a subtractive role. Re-positioning to the prone position is traditionally required for placing BPS, which extends operative time by an average of 45 to 60 minutes. 41 Prolonged anesthesia in the prone position leaves patients at risk for complications, such as increased blood loss, peripheral nerve injuries, di culties in airway access and postoperative infection. 42,43 The design of OLLPS is based on the idea that 1+1>2. OLLPS is placed in a single-position through the oblique lateral surgical corridor to give full advantage to the OLIF procedure for non-injury to the posterior lumbar structures and reduce operative time and subsequent complications of prolonged anesthesia and prone positioning. The lateral plate has an inherent de ciency in achieving bilateral equalization, but the big size cage crossing the epiphysis ring can diminish this de ciency. Previous studies showed that lateral plate increased the stiffness of the lumbar spine in bending and rotation but had little effect in exion and extension. 11,40,44 In our study, lateral xations of LP-2, LP-4, and OLLPS were more restrictive for lumbar lateral bending and axial rotation. In the treatment of long bong fracture, locking plate systems have been proved to have signi cant advantages in stability and bone regeneration. 45 Our ndings suggest that the idea of angular stability is also bene cial in lateral plate. OLLPS demonstrated superior stability in lateral bending and axial rotation following implantation. The junctions between the screws and the plate of LP-2 and LP-4 are equivalent to a rotatable portal axis, and as the lumbar spine move, the screws and the plate can rotate relative to each other, especially in exion and extension. The locking structure of OLLPS combines the screws with the plate as a single unit, with all the screws and plate acting synergistically to counteract lumbar motion, and the plate does not need to be pressed against the vertebral bone surface, thus protecting the periosteal blood supply. The multidimensional angular stabilization layout of the OLLPS screws and the plate/screws locking structure an improved its ability to limit lumbar exion and extension compared with the existing lateral internal xation LP-2 and LP-4. The screws of OLLPS have a thicker diameter than that of LP-2 and LP-4, and the special inclined trajectory and surface threads of the screws increase the contact area of the screw-bone to provide a tighter bond with cancellous bone, which improves the holding power and xation of the OLLPS on the vertebral body. Screw-locked structural stability was not signi cantly decreased in the osteoporotic patients. 46 The different biomechanical behavior depends on the direction of loads acting on the screws. 46 For motion in exion/extension, the longitudinal axis of the OLLPS screws intersects the lumbar rotation axis at a small angle, while the BPS is orthogonal to the rotation axis. The torque of OLLPS screws against lumbar exion and extension is less than the BPS. Therefore, the OLLPS is less capable of limiting lumbar exion/extension than the BPS. For lateral bending and axial rotational motion, the longitudinal axis of the OLLPS screws is orthogonal to the rotation axis, and the reverse pedicle screw trajectory gives it a longer lever arm for resisting lumbar motion. Therefore, the OLLPS is comparable to the BPS in limiting lumbar lateral bending and rotational activity. In our study, the results of the maximum displacement of the superior vertebral body corroborated with the ROM results of the surgical segment.
It had been reported that cage subsidence was closely related to the endplate stress and cage stress of the fused segment. 47,48 According to the stress-growth curve of the vertebral body cells, the higher the compressive stress is, the more inhibited the growth of vertebral body cells is. 49 The bone density of the epiphyseal ring at the periphery of the vertebral endplate was much higher than that of the central region of the endplate, which providing the greatest resistance to cage subsidence. 50 During the fusion process of the OLIF, the epiphyseal ring and cortical compact have the effect of supporting the cage, and the cancellous bone contributes to the fusion with the intervertebral cage. A longer and wider cage has larger contact area, more dispersed stress and lower subsidence risk. [51][52][53] Although compared with traditional banana-shaped or bullet-shaped cages, the cage of stand-alone OLIF across the epiphyseal ring of endplate reduces the risk subsidence. However, without supplementary xations, the weight load may be distributed directly on the cage and endplate surfaces, increasing the possibility of endplate collapse and cage subsidence, and the protentional complications associated with endplate collapse and cage subsidence remain a concern, especially in osteoporotic patients. 13,54 The results of Zhang et al. 11 showed that the lateral plate can reduce the stress of the fusion cage and the endplate when the lumbar lateral bending, but it has little effect in other motion modes. A recent FEA article reported that LP-2 signi cantly reduced cage and endplate stress in lateral bending and axial rotation. 13 Our results showed that when LP-2 was implanted, the endplate and cage stress decreased in a stepwise manner and the risk of cage subsidence was reduced compared with the SA model. The L5 endplate stress of OLLPS decreased by 43.82%-91.42% of SA, which was better than that of LP-2 and LP-4. The decrease of OLLPS was similar to that of BPS in lateral bending and rotational modes, and the risk of cage subsidence was low, but the stress sharing in exion and extension was less than that of BPS. According to the stress transmission mechanism of angle-stabilized structures, the locking screws form a xed angle with the plate, which facilitates the transfer of stress from the screws to the plate. 55 Compared with traditional screws, all locking screws acted synergistically to resist stresses like a single-beam structure, which could distribute the load uniformly on the plate. 56 Our results show that the OLLPS internal xational stress is larger than other internal xations, mainly concentrated in the titanium plate, rather than the screw-bone contact, and is far lower than the titanium fatigue strength and yield strength of titanium alloy. Our model simulates the immediate postoperative state, and the internal xation stress would be further reduced with the completion of fusion.
There is no lateral locking plate designed for the OLIF procedure on the market, and the related research is still in a blank state. The OLLPS we studied is minimally invasive and inserted through the OLIF approach to avoid soft tissue injuries such as dorsal muscles, ligaments, and articular capsules. Sardhara et al. 14 has demonstrated the feasibility of RPSF using the 2-screw lateral plate in clinical cases. Gragnaniello et al. 57 also reported good clinical results by applying OLIF combined with the 4-screw lateral plate in patients with good bone density. Compared with LP-2 and LP-4, OLLPS could better maintain the immediate stability of the surgical segment, effectively reduce the stress of the superior and inferior endplates of the surgical segment, and reduce the risk of cage subsidence. In lateral bending and axial rotation, OLLPS is slightly better than BPS in reducing the stress of the surgical segment endplate and the fusion cage. The ability of OLLPS to limit lumbar exion and extension is not good as that of BPS. In clinical practice, we routinely recommend that patients wear a lumbar brace after lumbar interbody fusion surgery. Some studies have reported that lumbar braces can restrict lumbar sagittal motion well, which could compensate for the OLLPS limitation of lumbar mobilities and expand the clinical applicability of OLLPS. 58,59 The OLLPS angular stabilization structure transfers body loads directly to the lateral plate via screws, allowing early weight-bearing of the surgical segment with no risk of xation failure at the plate/screw junction and facilitating early patient recovery. 60 Certainly, the superiority of OLLPS is balanced with its relatively di cult implantation technology. The use of multi-axis screws can be considered to facilitate the implantation operation through their adaptability and multi-directional screw trajectory.
Our study had some limitations. Firstly, our model did not reconstruct the paravertebral soft tissues to assess the effect of muscle on spinal biomechanical function, which was a common problem faced by all nite element analysis. Secondly, we did not assess the effects of osteoporosis and bone loss, because different degrees of BMD may have different results. Moreover, the model was limited to detect the instant features of static biomechanics after surgery. Despite the limitations existed, in this study we applied more comprehensive methods to validate the validity of the nite element model than in previous studies, and simulations of different internal xation modalities were performed under the same experimental conditions. Therefore, our model was valid to evaluate the biomechanical properties of OLLPS.

Conclusions
Our study con rms that the locking structure and the reverse pedicle track screw can enhance the biomechanical properties of the lateral plate. OLLPS was superior to LP-2 and LP-4 in maintaining postoperative lumbar stability and reducing endplate and cage stress. OLLPS provided multiplanar stability similar to that of BPS and had a slightly better ability to reduce endplate and fusion stress during lumbar lateral bending and rotational activities than BPS. OLLPS is designed to save more patients from additional injury, which has good prospects for promotion and research, but extensive in vitro and clinical trials will need to demonstrate the application effect. In addition, more innovative techniques and instruments will need to facilitate the OLLPS placement operation, and its mechanical properties, such as long-term stability and fatigue resistance, need to be further investigated.