TESSYS Technique With Small Grade of Facetectomy Has Potential Biomechanical Advantages Compared to the In-Out TED With Intact Articular Process : An In-Silico Study

Jingchi Li West China Hospital/West China School of Medicine for Sichuan University Chen Xu Changzheng Hospital A liated to the Naval Medical University Xiaoyu Zhang A liated Hospital of Integrated Traditional Chinese and Western Medicine for Nanjing University of Chinese Medicine Zhipeng Xi A liated Hospital of Integrated Traditional Chinese and Western Medicine for Nanjing University of Chinese Medicine Mengnan Liu Macau University of Science and Technology Zhongxin Fang Xihua University Nan Wang A liated Hospital of Integrated Traditional Chinese and Western Medicine for Nanjing University of Chinese Medicine Lin Xie (  xielin6508@163.com ) A liated Hospital of Integrated Traditional Chinese and Western Medicine for Nanjing University of Chinese Medicine Yueming Song West China Hospital/West China School of Medicine for Sichuan University

factor of symptoms recurrence and futher disc degeneration [16, [21][22][23]. Besides, ZJ is an important structure during the maintaince of spinal stability [13,14,24]. Instability in the surgical segment caused by facetectomy was also a risk factor for the biomechanical deterioration, resulting degeneration in the surgical FSU and poor long term prognosis [19][20][21][22]25]. These deductions were consist with our published nite element numerical studies [26][27][28]. In which, larger grade facetectomy has been demonstrated to be associated with biomechanical deterioration and lumbar instability, and these changes may shortly related to spinal futher degeneration and symptoms recurrence [27][28][29].
Considering the axial rotation will enhances the vulnerability of posterior annulus and the zygapophyseal joint (ZJ) could restrict the lumbar spinal motion under axial rotation, the iatrogenic injury of SAP in TED may also increase the risk of annulus tear, related symptoms recurernce and resulting acceleration of disc degeneration [17,18,24,30].
To avoid the facetectomy in TED for patients without foramen stenosis, the standard YESS technique could be modi ed. Firslyly, insert the cannula into the disc space trans Kambin triangle. Then, by pressing down the end of cannula and using different sizes of bending forceps, the herniated disc can be removed without any damage of SAP (i.e. without facetectomy), related risk of biomechanical deterioration and resulting postoperative complications can be avoided (Fig. 1). Facetectomy was an important surgical procedure in YESS technique [3,5], Therefore, the modi ed surgical technique was called optimized inout, rather than YESS technique in the following statement. But the optimized in-out surgical strategy still has its original defects and which may also lead to poor clinical outcomes. Speci cally, the risk of the recurrent lumbar disc herniation (RLDH) was reported to increased dramatically with the expansion of annulus tear (which may up to more than one-quarter when the size annulus tear larger than 6mm) [31][32][33][34][35]. More signi cantly, the strength of scar tissue in the annulus outer lamellae is not strong enough to prevent RLDH [36,37]. Considering the diameter of our working cannula was 7.5mm (type WTS127502, Joimax International, Irvine, Calif) and its insertion will lead to iatrogenic annulus injury inevitably, a relatively larger grade of nucleotomy in the modi ed in-out TED seems necessary to prevent RLDH ( Fig. 1) [3], but this surgical strategy also has its limitations.
Postoperative residual nucleus is still important during to maintenance of spinal biomechanical function [19,38,39], larger grades of nucleotomy may lead to the annulus stress pathological distribution and make it vulnerable to fatigue damage under cyclic loading [38][39][40][41], resulting annulus tears will accelerate disc degeneration. Besides, disc collapse could also be accelerated in this pathological process [16,39,40], related risk of lumbar instability will increased for the laxity of soft tissues, and the foramen stenosis incidence will also increased for the decrease of the foramen cross-sectional areas (CSA) during the collapsed of the surgical FSU without facetectomy [24,25,42]. As a result, higher risk of symptom recurrence can be observed in patients with larger grades of nucleotomy [16,23]. Additionally, the collapse of the surgical segment and DD also lead to secondary spinal irregular loading transmission and which has been proved to increase the load of the ZJ cartilage and resulting risk of ZJ degenerative osteoarthritis, hypertrophy of articular process and resulting spinal stenosis [43][44][45]. In consequence, a large grade of nucleotomy, the remedial action to reduce the risk of RLDH caused by iatrogenic annulus tear in the in-out technique may lead to greater potential risk of poor clinical outcome and lower satisfaction of patients after TED [32].
By contrast, as a remedial procedure of iatrogenic annulus injury, a large grade of nucleotomy was not necessary in TESSYS TED commonly. Speci cally, considering the size of original annulus tear was usually smaller than 6 mm and the residual annulus tissue will not lead to serious clinical symptoms generally [46,47]. Nucleotomy could be accomplished along the original annulus tear without any iatrogenic annulus injury. Besides, for patients with contained type of LDH (i.e. LDH with intact annulus), discectomy can be accomplished by using bipolar radiofrequency to make a small slit (less than 6mm) in the annulus, and extensive greades of nucleotomy is also unnecessary. Meanwhile, facetectomy can be limited by endoscopic dynamic drill for which could be accomplished precisely under direct version with its assistance. As a result, in our clinical practice, the grade of facetectomy can be restricted to less than one third generally for patients without foramen stenosis, and the cartilage and capsule of ZJ can also be protected ( Fig. 1). More signi cantly, the controllable risk of postoperative spinal instability and biomechanical deterioration after endoscopic nucleotomy with small grades of facetectomy has been proved in our published studies [26-28].
On the basic of above theoretical and practical foundations, we can make a hypothesis that even if the iatrogenic injury of SAP can be avoided, compared with the in-out TED, TESSYS TED with small grade of facetectomy still has potential biomechanical advantages. To verify this hypothesis, the biomechanical effect of modi ed in-out and TESSYS TED have been computed in a validated three dimensional lumbosacral model. Considering LDH patients were mainly middle-aged and elderly, their original disc degenerative changes may have potentially impacts on the postoperative biomechanical environment [23,48,49], surgical simulations and nite element analyses have been accomplished in models with and without degeneration. To the best of our knowledge, published studies have not adequately clari ed these issues.

Model construction
The intact nite element (FE) model from L3 to S1 has been constructed in our published studies [26,28,50]. Bone structures include cortical, cancellous, and posterior structures, nonbony components include the intervertebral disc and ZJ cartilages. IVD consist of the nucleus core, the surrounding annulus, and cartilage endplates [51,52]. The thickness of the cortical and the endplates was set as 0.8 mm [48, 49,53]. Ligaments and ZJ capsules were constructed by cable elements [50,54]. Facet cartilages were de ned by surface-surface contact elements and the gap between cartilages was set as 0.5 mm [48,55]. In the construction of models with disc degeneration in segments adjacent to the surgical segment, the disc height was reduced to 67 %, the cross-sectional area of annulus was expanded by 40%, and material properties of annulus and nucleus were modi ed based on anterior published studies [23,48,49].

Boundary and Loading Conditions
Current models were set to be symmetric in the sagittal plane to increase the computational e ciency by computing the bending and axial rotation loading conditions of intact models unilaterally [27]. Different sizes of hybrid elements were set in FE models and the mesh was re ned in thin structures and the structures with large deformation [27, 50,56]. To ensure the computational credibility, a mesh convergency test was performed on the intact model by evaluating the change of maximum annulus shear stress, and the model was considered to be converged if the change of computational value less than 3%. All of freedom degrees were xed under the inferior of S1, stress and moments were applied on the superior of L3 [50,57], and the contact between facet cartilages was de ned as frictionless [54,57].

Model calibration and validation
During the model calibration process, the stiffness of ligamentum structures were slightly modi ed within the physiological range to reduce the difference between the computed biomechanical indicators with which from widely cited in-vitro studies [51,[57][58][59][60][61][62]. Then, to ensure the reliability of the calibrated model, multi-indicators model validation has been accomplished by comparing the computed range of motion (ROM), the intradiscal pressure (IDP) and the value of disc compression (DC) with which from in-vitro studies under different loading conditions [63-66].

Simulation of the modi ed in-out and TESSYS TED
The right side of L4-L5 segment was selected for the simulation of TED. TESSYS TED with a small grade of facetectomy was simulated according to the reported surgical technique and our clinical experience [4,50]. In which, a 3 mm incision was made on the annulus to simulate the annulus tear and its width was set as 1 mm. One-sixth of the nucleus around the incision was removed to simulate a small grade of nucleotomy. The vertex of facetectomy was located on the cranial tip of SAP, one-third of SAP and the ligamentum avum (LF) were excised during this procedure (Fig. 2) [6, 7,26,50]. During the simulation of the in-out TED with intact SAP, the original annulus tear was set as the center of working cannula insertion, annulus in the surrounding 7.5 mm area was completely deleted to simulate the iatrogenic injury. One-third (twice the range of nucleotomy in the model with TESSYS technique) of the nucleus around the annulus tear was removed to simulate a larger range of nucleotomy (Fig. 3). Pathological changes caused by DD were simulated in the L3-L4 and L5-S1 segments, and the surgical simulation keep consistent in models with and without DD (Fig. 2). To simplify the following statement, FEA models in this study were named from model 1 to model 6 (Table. 1). Table 1 Summary

Model validation
The concept of computational accuracy (ACC) was presented in the gure. 4. In this study, ACC was greater than 90% except for the disc compression in L3-L4 segment. In which, the value of ACC was 85.2% and the difference between our computational result and the average value from in-vitro study was obviously less than one standard deviation [59]. Additionally, DD in segments adjacent to the surgical segment lead to the increase of IDP and the decrease of facet contact force (FCF) slightly in the surgical segment, this tendency was consistent with published studies [23,49]. So we believe our models make good representation of real biomechanical environment.

Biomechanical change in different models
Special emphasis should be placed on FCF. FCF was not been recorded in the exion condition. Besides, cartilages in the bending side were contact, and the opposite side of cartilages were contact in the axial rotation condition. In other words, FCF under the left lateral bending, was the force on left facet cartilages, and when it comes to the FCF under left axial rotation, the force was recorded on the right side, and vice versa.
Biomechanical indicators in different models were presented in gure.5 to gure. 8. Generally, DD in adjacent segments will lead to the deterioration of biomechanical indicators in the surgical segment, however, the decrease of FCF and ROM can also be observed and the original biomechanical change tendency in different surgical models was not varied or ampli ed obviously in degenerated models. Compared with intact models, change in models after TESSYS TED was slightly under most of loading conditions. Most importantly, biomechanical deterioration can be observed in models after in-out TED compared with intact models and TESSYS TED models especially in the extension, bending and rotation conditions to the surgical side (i.e. right lateral bending and axial rotation).

Discussion
The objective of this study To evaluate the risk biomechanical deterioration and resulting postoperative complications in the surgical segment caused by the optimized in-out technique with intact SAP and TESSYS TED with a small grade of facetectomy, intact lumbo-sacral models with and without disc degeneration and corresponding models after these operations have been constructed and biomechanical indicators shortly related to lumbar degenerative diseases (LDD) have been computed and evaluated. Considering the importance of biomechanical environment for postoperative clinical outcomes has been repeatedly demonstrated [23,50,52], investigations for the biomechanical change caused by these two different surgical techniques in TED was of great signi cance for the reference of surgical strategy selection.
Notable points in the model construction process Adjacent segments, rather than the surgical segment were selected for the construction of DD models. This model construction strategy was based on our clinical practice. As is mentioned above, DD was very common in TED patients, This nature degenerative change may not lead to serious clinical symptoms, but will adversely affect the biomechanical environment in adjacent segments [23,48,49]. Hence, the simulation of disc degeneration was meaningful for the evaluation of real postoperative biomechanical environment. Disc collapse during the process of DD lead to a reduction in the cross-sectional area in Kambin triangle, and the risk of exiting nerve root injury in a degenerated disc will increase during the insertion of working cannula without facetectomy [67, 68]. As a result, LDH with narrow disc space can be seen as a contraindication for the application of modi ed in-out TED and the surgical segment has been excluded from the construction of degenerative change models.
Besides, although ZJ degeneration was shortly related to DD [43,45,58,69], and some FE studies construct ZJ degenerative models by reducing the facet gap [48, 57], we still give up the construction of ZJ degeneration. The gap thickness of ZJ should be seen as a re ection of the cartilage wear, sclerosis and hyperplasia of subchondral bone [60, 70,71], and this pathological process was di cult to simulate in the model construction. Speci cally, the decrease of the facet gap by thickening the facet cartilage was completely contrary to the pathological change of ZJ degeneration. Besides, if the gap was reduced by thickening the bone tissue of articular process, the de nition of material properties for sclerosis subchondral bone structures was also inaccurate for which was obviously differ from normal bone tissues [15,72,73], and the casual de nition of material properties without reliable data will decrease the credibility of this study. Hence, we chose to construct DD models without the change of facet gap [49,57].
Additionally, the grade of facetectomy in TESSYS TED models and discectomy in in-out TED models were set as one-third, this grade was consistent with the maximum one in our clinical practice. This modeling strategy was selected for facetectomy and nucleotomy were assumed to be main reasons for poor clinical outcomes after these operations respectively. Therefore, larger grades of these two procedures should lead to more pronounced biomechanical deterioration and which could provide us a clear reference for the evaluation of these two techniques.

Clinical signi cance of biomechanical indicators
Disc collapse and DD acceleration in the surgical segment and resulting several kind of secondary pathological changes were most signi cant causes for the poor clinical outcome for patients after nonfusion lumbar surgery [33, 74,75]. As is reported by Adam et. al, the injury of endplates and annulus can be seen as two different separately pathways in the DD process [21]. The maximum von-Mises stress and the strain energy of endplates were recorded to evaluate the risk of DD caused by the endplate lesion and ossi cation. Speci cally, endplates play a key role in the pressure distribution, post operative abnormal stress concentration on endplates increases the risk of endplate lesions [21,76,77], resulting in ammatory response, autoimmune reaction and disc innervation can be seen as signi cant triggers for DD acceleration and increase risk of low back pain (LBP) [51,[78][79][80].
Besides, IVD was an avascular structure, and the most important pathway for its metabolism was the trans-endplate diffusion [81,82]. According to the Wolff's law, the concentration of strain energy, a kind of compensatory reaction for the endplate stress concentration can be seen as a predicted factor for its ossi cation [83,84], and resulting occlusion of trans-endplate diffusion pathway will lead to DD acceleration [43,80,85,86]. Moreover, endplate injury caused by the abnormal stress concentration was closely associated with the disruption of the annulus and can also be re ected by the deterioration of biomechanical indicators on the annulus, especially in the post and post lateral part of annulus [21,29,82]. The concentration of shear stress and compressive stress have been proved to be related to different kinds of annulus tear and resulting DD, more signi cantly, to resulting discogenic LBP and RLDH [22,23,85]. Hence, we can speculate that above biomechanical indicators can be seen as credible predictors for the assessment of postoperative prognosis.
Meanwhile, the foramen stenosis was another vital reason for the clinical outcome deterioration, and special attention should be taken for models after in-out TED with intact SAP for the risk of foramen stenosis will get worse with disc collapse caused by a larger grade of discectomy in the surgical segment without foraminoplasty [74,75,87]. Noteworthy, the increase of FCF can not only be seen as a risk factor for the cartilage wear and resulting degenerative osteoarthritis of ZJ [15,44,73], a trigger of LBP [22,25,45], but also for the foramen stenosis considering a larger load will promote the osteogenic activity [43,83,88]. More importantly, disc collapse and degeneration in the surgical segment will lead to pathological stress concentration on ZJ cartilages and resulting further degenerative osteoarthritis and osteophytes formation in which [15,45]. Besides, lumbar instability was also an important cause of prognosis deterioration after non-fusion surgery, and which have been proven to be related to LBP and further DD [86,89,90]. Therefore, ROM can be used not only as an indicator for model validation, but also for the assessment of postoperative complications. In a word, there was close interactions between different biomechanical indicators, and biomechanical deterioration will lead to a series of clinical symptoms and deterioration of prognosis.

Limitations
Firstly, ligaments were constructed by cable elements, and the simulation for the LF excision was accomplished by reducing its CSA. Cable elements can not stimulate the fold, the hypertrophy and the calci cation of ligaments, and there pathological changes were reported to be vital risk factors for the spinal stenosis and nerve compression.
Meanwhile, as a common issue of FE studies, the proliferation of scar on the annulus and its biomechanical effects can not be evaluated. Considering that the size of annulus breakage was an important variable in this study, biomechanical changes caused by the formation of annulus scar tissue (although its strength proved not strong enough to stop RLDH) may also have its potential biomechanical impact on patients prognosis. Hence, current computational results should be recognized and discussed on the basic of the awareness of this defect and follow-up clinical study is still necessary for more de nitive conclusions.

Conclusions
Based on above theoretical foundations and our computational results, it was not hard to make following conclusions. Biomechanical deterioration can be observed in models with optimized in-out TED without facetectomy (i.e. TED with intact SAP), especially in the extension, right bending and rotation conditions. Although the facetectomy was avoided, iatrogenic annulus breakage caused by the insertion of working cannula and a larger grade of nucleotomy for the prevention of RLDH provide this surgical method a potentially poor clinical outcome, and which was consistent with the report that larger grades of nucleotomy will lead to lower patients' satisfaction [32]. By contrast, the promotion of endoscopic dynamic drill in modi ed TESSYS TED (TED with a small grade of facetectomy), a kind of surgical design without extra injury of annulus and a large grade of discectomy, was proved to be advantageous biomechanically (Fig. 9).
Moreover, biomechanical deterioration can be observed in models with DD generally, and which was consistent with published reports [48,49]. Although DD in adjacent segments did not obviously exacerbate the biomechanical deterioration in the surgical segment, the reported vicious cycle of DD can also be observed and which provide us the signi cance of this FE study from another perspective [21,91].
Speci cally, in this vicious circle, the deterioration of biomechanical environment caused by the unsuitable surgery may be continuously ampli ed and lead to disastrous prognosis. Therefore, the selection and optimization of surgical techniques based on the biomechanical FE study was of great signi cance. Approval for the current study protocol (including the lumbar CT scan) was obtained from the ethics committees of A liated Hospital of Integrated Traditional Chinese and Western Medicine for Nanjing University of Chinese Medicine (2019LWKY015).We con rm that the subject signed the informed consent and submitted it to the ethics committee for review before the examination, andall methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication
Not Applicable

Availability of data and materials
All the data of the manuscript are presented in the paper.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was funded by Foundation for leading talent in traditional Chinese medicine of Jiangsu province (2018SLJ0210) Key project of jiangsu province social development (BE2019765), and the innovative education project of Jiangsu Province Postgraduated student (SJCX20_0616). The funds provided by the above projects are used for the CT scan, the compensation of volunteer, the polishing of English manuscripts in the mother tongue, the conference affairs and travel expenses of studying courses related to three-dimensional spinal modeling and nite element analysis and the publication charges of this manuscript.   The schematic for the optimization of TED by two different strategies (ligamentum structures have been hided for the sake of brevity in schematic diagrams)

Figure 2
Intact 3D models in the current study  Model validation