The FEA method can simulate the human body and analyze the biomechanical changes associated with different surgical methods on various parts of the spine and is widely used in the field of spinal surgery [20–23]. The L4-5 segment is the segment with the highest incidence of LDH, so in the present study, the L4-5 segment was used as the target segment, and the FEA method was used to establish a normal three-dimensional finite element model of the L1-S1 segment. Based on this model, an intraoperative arthroplasty model was established to investigate the effects of intraoperative arthroplasty on lumbar spine biomechanics. An arthroplasty model was established on the basis of this model to investigate the effect of intervertebral foramenoscopy on the biomechanics of the lumbar spine in patients after arthroplasty. The analysis of the effect of surgery on the biomechanics of the lumbar spine and the acquisition of the mechanical information of the corresponding motion segments are conducive to tracing the causes of postoperative lumbar pain and can provide a theoretical basis for guiding the design of intraoperative surgical protocols and the formulation of individualized treatment plans.
The first step after establishing a 3D finite element model is to verify the validity of the model, which is usually performed by comparing the experimental results with similar 3D finite element models from previous literature [24–26]. A comparison of our established finite element model with the results of Shim et al. [19] shows that the ROM of our model is within the error range reported in the literature and is comparable, indicating the validity of the finite element modeling method and material assignment in this paper, which can be used for biomechanical analysis. The modeling methods and loading constraints used in this study are basically the same as those used in previous studies, with only the individual samples differing, and the results are plausible from the point of view of a qualitative comparative study. The diversity of the individual morphology and material properties of human spinal segments means that finite element models do not exactly match the results of computer simulations with those of in vivo experiments. The reason for this difference is the inconsistency in the sources used to construct the finite element models, and although the ideal approach would be to use the same in vitro experimental subject and finite element model object, this approach is essentially impossible for any experiment due to ethical issues. The advantage of using cadaveric samples for modeling is that the samples can be dissected and directly validated for each tissue, and the disadvantage is that the metrics in the physiological state are not available. Although the reconstruction of the human spine model can be achieved using cadaveric samples, it cannot be validated against homologous cadaveric samples, so the method of validating the finite element model is mainly to compare it with the test data of previous cadaveric specimens.
During TESSYS, depending on the size of the intervertebral foramen and the location of the herniated disc, arthroplasty on one side is usually required to enlarge the foramen, a procedure that can injure the FJ to varying degrees. Previous studies have reported on the effect of conventional open surgery on FJs. Shah et al. [27] reported that 33–35% of FJs are injured during lumbar nailing via the transosseous interspace approach, whereas Monshirfar et al. [28] analyzed the probability of injury to the FJ during pedicle screw placement via the posterior median approach, and the presence of an injury to the FJ was detected on postoperative radiographs in approximately 15% of the patients. Because the TESSYS technique has not matured for a long time in China, no study has systematically analyzed its effect on lumbar spine biomechanics after performing arthroplasty during TESSYS. After we established a finite element model of TESSYS intraoperative arthroplasty, we analyzed the degree of displacement of the L4 vertebral body under different working conditions and found that there was no significant difference in L4 vertebral body displacement among the three groups of models, which suggests that TESSYS intraoperative arthroplasty is relatively safe and generally does not cause lumbar spine slippage in the postoperative period.
Stress analyses of the L4-5 FJ in the three models revealed that after arthroplasty, the stresses in the L4-5 bilateral FJ in the left lateral flexion condition increased compared with those in the preoperative condition in all patients and that the 8.5 mm arthroplasty had a greater effect on the right synchondrosis than did the 5 mm arthroplasty. In both arthroplasty models, the stress in the bilateral FJ of L4-5 was significantly greater in left lateral flexion or left rotation than in right lateral flexion or right rotation, and the magnitude of the increase was greater than that in the unoperated model, which suggests that arthroplasty of one side of arthroplasty with an ipsilateral disc injury will lead to increased stress in the contralateral FJ and that the larger the amplitude of arthroplasty is, the greater the increase in the stress in the contralateral FJ, which suggests that arthroplasty is best performed by tightly applying media. This finding suggests that it is better to keep the synchondrosis as close to the medial side as possible when performing synchondroplasty and to minimize the removal of the synchondrosis under the premise of adequate decompression. Stress analysis of the L4-5 intervertebral discs of the three models showed that L4-5 intervertebral disc stress was greater in the M3 model than in the M2 model under conditions of posterior extension and rightward rotation, which indicated that a large range of arthroplasty might aggravate the degeneration of the intervertebral discs of the segments compared with a small range of arthroplasty. In contrast, there was no significant difference in the distribution of L3-4 or L5-S1 interbody disc stresses among the three models under various working conditions, suggesting that arthroplasty had no significant effect on disc degeneration in adjacent segments. This is an advantage of nonfusion surgery over fusion surgery, as lumbar fusion surgery mostly accelerates the degeneration of neighboring segments [29–32].
This study also has the following shortcomings. First, the finite element models established in this study are similar to those that have been well validated in previous studies, but they all simplify the physiological contraction force of the lumbar spine to varying degrees, especially simplifying the muscles connected to the lumbar spine and the weight of the upper body, which is still somewhat different from that of a real human body. Second, this FEA is a one-time loading study, which cannot systematically analyze the effect of fatigue loading on the biomechanics of the lumbar spine. The actual clinical situation is complex and variable, and the repeated accumulation of stress in the lumbar spine after disc removal and arthroplasty may accelerate the degeneration of the injured area.
Overall, in this study, a normal L1-S1 segmental finite element model was developed, on the basis of which an L4-5 disc herniation model, a 5 mm arthroplasty model, and an 8.5 mm arthroplasty model were developed. The model was successfully validated, and the predicted results are credible and can be used for biomechanical analyses and simulation of surgical situations. Compared with 5 mm arthroplasty, 8.5 mm arthroplasty on the left side increases the stress on the FJ and segmental discs under certain working conditions, but it does not cause degeneration of the discs of the adjacent segments.