Pre-proof Three-Dimensional Correction of Scoliosis by a Double Spring Reduction system as a Dynamic Internal Brace: A Pre-clinical Study in G¨ottingen Minipigs Three-Dimensional Correction of Scoliosis by a Double Spring Reduction system as a Dynamic Internal Brace: A Pre-clinical Study in G¨ottingen Minipigs, Journal

Background Context: Adolescent idiopathic scoliosis (AIS) is a major skeletal deformity, that is characterized by a combination of apical rotation, lateral bending and apical lordosis. To provide full 3D correction, all these deformations should be addressed. We developed the Double Spring Reduction (DSR) system, a growth-friendly concept that continuously corrects the deformity through two different elements: A posterior convex Torsional Spring Implant (TSI) that provides a derotational torque at the apex, and a concave Spring Distraction System (SDS), which provides posterior, concave distraction to restore thoracic kyphosis. Purpose: To determine whether the DSR-concept is able to correct an induced idiopathic-like scoliosis and to compare correction realized by TSI with correction through DSR. Study Design/Setting: Pre-clinical randomized cohort study. Patient Outcome Measures: Coronal Cobb angle, sagittal lordosis/kyphosis, apical axial rotation, relative anterior lengthening. Methods: All mini-pigs received the TSI with a contralateral tether to induce an idiopathic-like scoliosis with apical rotation (mean Cobb: 20.4°; axial apical rotation: 13.1°, lordosis: 4.9°). After induction, the animals were divided into 2 groups: One group (N=6) was corrected by TSI only (TSI only-group), another group (N=6) was corrected by TSI and SDS (DSR-group). 3D spinal morphology on CT was compared between groups over time. After two months, animals were euthanized. Results: Both intervention groups showed excellent apical derotation (TSI only-group: 15.0° to 5.4°; DSR-group: 11.2° to 3.5°). The TSI only-group showed coronal Cobb improvement from 22.5° to 6.0°, while the DSR-group overcorrected the 18.3° Cobb to -9.2°. Lordosis was converted to kyphosis in both groups (TSI only-group: -4.6° to 4.3°; DSR-group: -5.2° to 25.0°) which was significantly larger in the DSR-group ( p< 0.001). Conclusions: The TSI alone realized strong apical derotation, and moderate correction in the coronal and sagittal plane. The addition of distraction on the posterior concavity resulted in more coronal correction and reversal of induced lordosis into physiological kyphosis. Clinical significance: This study shows that dynamic spring forces could be a viable method to guide the spine towards healthy alignment, without fusing it or inhibiting its growth.


Introduction
Adolescent idiopathic scoliosis (AIS) is a major skeletal deformity with pulmonary and cardiac consequences, [1,2] that is characterized by a combination of apical rotation, lateral bending and apical lordosis. These deformities are largely due to anterior lengthening that is mainly located in the intervertebral disc (IVD). [3][4][5][6] To achieve correction in all planes, all these deformations should be addressed, with specific attention to sufficient posterior concave distraction to accommodate the longer anterior column, so that it may rotate back into the midline.
Current treatments for adolescent idiopathic scoliosis (AIS) either stabilize the curve with bracing or surgically fuse the spine. Bracing can be effective in preventing curve progression, but only with strict patient compliance to wear the brace >16 hours per day. [7,8] Even then, residual curves can be considerable. [7] Spinal fusion is more effective in correcting the 3D morphology of the spine, however at the costs of spinal mobility, which adversely affects long-term quality of life of these patients. [9] Only at very young age, serial (Mehta) casting is known to be able to "cure" the spine, i.e. to resolve the deformity while retaining a flexible spine. [10] However, such permanent rigid casts, which apply large corrective forces, are not tolerated by older children.
To achieve similar results for these patients, an internal brace could possibly overcome many of the disadvantages of serial casts. Most importantly, strategic forces can be exerted on the spine with 100% compliance. By using dynamic and flexible implants, such application of pre-specified forces and torques is possible. For this purpose, we developed the Double Spring Reduction (DSR) implant (Figure 1), which consists of two different spring implants, the torsional spring implant (TSI) and the spring distraction system (SDS). Together, these temporary flexible implants provide continuous apical axial torque (TSI) and posterior distraction forces (SDS) to the spine during the years that it has a chance to mature into a reduced and stable configuration. The implant can accommodate growth, and can therefore be used when the child has not yet reached skeletal maturity. This "growth-friendly" feature 4 not only allows for early correction of AIS curves of older children, it also allows for treatment of "tweeners" aged 9-11, where current "growth-friendly" implant results are often disappointing when compared to results of spinal fusion.
Previous studies have investigated the concept of the TSI, concluding that it has the potential to provide strong apical (de-)rotation with only a very small increase in spinal stiffness. [11,12] A recent pre-clinical study by our group in growing Göttingen minipigs has shown that the TSI, combined with a flexible tether, was able to induce a morphologically idiopathic-like scoliosis whilst retaining mobility and growth. [13] After implant removal, the deformity remained and was shown to reside mainly in the IVD, indicating permanent spinal changes similar to those seen in human AIS.
We performed the current study to determine whether the internal brace concept (DSR) is able to correct the established idiopathic-like scoliosis. In addition, we compared if correction with torsion only is a viable option as compared to combining torsion and posterior distraction.

Ethical approval
This study was performed in the AAALAC certified experimental surgery animal laboratory of BLINDED BLINDED. Ethical approval was granted by the Animal Experiment Committee of BLINDED BLINDED before the start of this study (AVD 115002016804).

Study Design
The current study consisted of 2 phases (Figure 2). Phase 1 was the induction phase, wherein a scoliosis was induced in 12 growing Göttingen minipigs through implantation of a left-sided unilateral, posterior tether combined with a contralateral TSI, tensioned to provide 2 Nm axial torque. [13] After 3 months of induction, the curves were confirmed with CT scans and phase 2, the reduction phase, was initiated. All 12 minipigs were operated again, the induction forces were released and animals were randomized into two groups, each undergoing a different method of scoliosis reduction: (1) Reduction by only de-rotating the curve apex with the torsional TSI (TSI only-group, N=6), or (2) Reduction by combining the TSI with the concave distraction implant SDS (DSR-group, N=6). After the reduction surgery, 2 months of follow-up was allowed for spinal remodeling. Then, the animals were euthanized. Spinal morphology between groups was compared with 3D imaging at several timepoints.

Animal Model
The Ellegaard Göttingen minipig is bred specifically for research purposes, and has a predictable linear growth curve from birth to 2 years of age, which can be translated very well to pediatric spinal growth. [14,15] In addition, spinal anatomy is similar in size and shape to human pediatric anatomy. The animals can be housed in groups, and their small stature makes animal husbandry less cumbersome as compared to larger cattle. Whilst the minipig spine is positioned horizontally and not upright like the human spine, it serves as a representative scoliosis model, since it has been shown that muscle forces in quadrupeds ensure similar axial compressive force vectors comparable to vertical human spinal loading. [16]

Double Spring Reduction implant
The complete internal brace, the DSR implant ( At the apex (T14), the connector between both springs can be pre-tensioned with 2 6 Nm by 45° of rotation and can then be mounted to the apical anchor to deliver the axial torque. The U-shaped loops ( Figure 1D) are designed so that with spinal growth, the combined torsional stiffness of the springs and the loops increase, causing the corrective torque to remain essentially the same in spite of the decreasing pretension angle. This counteracts the decrease in torque that would otherwise take place due to apical derotation over time. The TSI has a growth potential of 100 mm, 50 mm on both the cranial-and caudal side and adds <20% additional spinal stiffness, which is far less than what is found with contemporary correction implants. [11] The SDS ( Figure 1C

Surgical technique
The surgical induction technique has been described previously, it is summarized here for completeness ( Figure 3). [13] After standard surgical preparation of the minipigs, exposure through a dorsal midline approach was performed spanning levels  For scoliosis induction, an UHMWPE tether (Dyneema, DSM, Geleen, The Netherlands) was looped around the cranial and caudal anchors on the left side and closed tight but without tension. The TSI was mounted on the right side, with bearings that fit on the rails cranially and caudally, leaving the apical connector unlocked. The apical part of the torsional spring ( Figure 1B) was then rotated 45° in the axial plane with a custom wrench (to induce a right-sided scoliosis) and was subsequently locked to the apical rail ( Figure 3A/B). Immediately following surgery, radiographic-and CT imaging was obtained. The animals were returned to their housing units where they were kept in groups and fed ad libitum.
Three months following the induction surgery, the animals were anaesthetized and CT scans were made to visualize scoliosis morphology and signs of implant failure. After exposure, the integrity of the tether was checked before it was released and subsequently removed. On the right side, the apical connector of the TSI was unlocked and returned back to neutral. Mobility of the spine was assessed manually under dynamic fluoroscopic imaging. The TSI was then rotated back 45° to reduce the rotational deformity and was locked again. After randomization, the wound was then either closed immediately (TSI only-group) or an SDS was first implanted on the concave side (DSR-group). The SDS rod was fixated to the apical anchor and could slide through two sliding bearings that were mounted on the cranial and caudal anchor. Then the springs were tensioned to 70±5 N with the buttress.
After two months follow-up, CT and radiography were performed. Following this, the minipigs were euthanized by intracardiac injection of pentobarbital, compliant with the 2020 American Veterinary Medical Association guidelines for the euthanasia of animals. [20] The implants were removed and checked for damage and the spines were manually tested for flexibility.

Radiographic analysis
Coronal Cobb angle and instrumented kyphosis were measured on the coronal and sagittal CT reconstructions, More detailed analysis of axial rotation and relative 8 anterior-posterior and convex-concave lengthening of both the intervertebral discs and vertebral bodies was performed using the ScoliosisAnalysis 4.1 software (Imaging Sciences Institute, Utrecht University). This validated method has been used previously and is detailed in Figure 4. Where 1 and 2 are two similar points on two different endplates (e.g. anterior, posterior, left, right) that are to be compared, is the shortest distance between these points, and 1,2, 1,2 and 1,2 are the respective 3D coordinates. By comparing the upper-and lower endplate of the same vertebra, lengths corresponding to that vertebra can be calculated, while comparing the lower endplate of one vertebra with the upper endplate of the vertebra below yields values corresponding to the IVD space. Axial rotation of an endplate was defined as the angular difference between its AP axis relative to the AP axis of L3. The axial rotation of a vertebral body was determined by averaging the AP axis of the upper-and lower endplate. This yields a difference in rotation between the apical level (T14) and the most cranial (T7) and most caudal (L3) instrumented level. The apical rotation was then obtained by taking the mean of both rotation differences.

Statistical Analysis
Prospectively, a power calculation was performed based on our earlier induction animal study. [13] The study was powered to show a difference in correction capabilities between the DSR-group and the TSI only-group after the introduction of the reduction implants. To detect a difference in coronal Cobb angle between groups of 5° (SD 3.0), with a power of 80% and an alpha of 0.05, six animals per group were needed.
Differences in curve morphology between start of the induction period (immediately following induction surgery) and the end of the induction period (after 3 months, just before reduction surgery) were calculated for all 12 animals and shown as mean ± SD. Following this, paired t-tests comparing both timepoints were performed. If the residuals of differences were non-parametric, the Wilcoxon-signed rank test was performed. For the analysis of the reduction period, 2 way repeated measurement ANOVA was performed comparing both the DSR-and TSI-groups over time. Two-tailed significance for all analyses was set at p=0.05. Statistical analyses and data visualization were performed with GraphPad Prism 9.2.0. (Graphpad Software, San Diego, CA, USA).

General outcomes and complications
Mean age of the animals during the induction surgery was 7.4 months and mean weight was 20.3 kg, with no significant difference between groups. Weight increased in 5 months to 34.3 kg according to their normal growth charts. [14] During both the induction-and reduction phase, there were no major complications or malpositions of pedicle screws. All tethers functioned as expected and were intact and removed during the reduction surgery. The spines remained flexible in the instrumented segment and the axial torque was successfully inverted in all minipigs. One of the minipigs suffered a deep surgical site infection following reduction surgery. Subfascial 10 pus collections were seen on the CT scan obtained pre-euthanasia, although no clinical symptoms of infection were observed in the months before. Tissue and pus cultures obtained post-euthanasia showed infection with Trueperella pyogenes. All TSI implants were intact upon removal with no signs of substantial wear of the bearings. In one animal in the DSR-group, the SDS PEEK sliding rod buckled out which negatively influenced the distraction force. Curve morphology results of this minipig were included in all analyses.

Radiographic outcomes
Significant changes were induced in all evaluated radiological parameters during the induction period (Table 1). Cobb angle increased from 6.2° immediately postoperatively, to 20.4° after 3 months. Instrumented kyphosis changed from 6.2° to -4.9°. Axial rotation of the apical level increased from 6.5° to 13.1° at the end of induction. The anterior spine lengthened during induction, with modest but significant changes in the vertebral bodies (A-P ratio from 0.98 to 0.99), and larger changes in the IVD (A-P ratio from 1.12 to 1.19).
Changes after the 2 month reduction period for both groups are shown in Table 2. For the TSI only-group, Cobb angle reduced from 22.5° to 6.0°. For the DSRgroup, the curve was overcorrected from 18.3° to -9.2°. The change in coronal curve was significantly larger in the DSR-group. For the sagittal plane, the instrumented kyphosis in the TSI only-group changed from -4.6° (i.e. lordosis) at the end of induction to 4.3° (i.e. kyphosis) at the end of reduction. In the DSR-group, a change from lordosis to kyphosis was seen as well, from -5.2° to +25.0°. The induced kyphosis was significantly greater in the DSR-group. In the axial plane, the mean apical axial rotation (i.e. the relative rotation of level T14 compared to the mean rotation of the most cranial and caudal instrumented level) for the TSI only-group decreased from 15.0° to 5.4°. For the DSR-group, this rotation similarly decreased from 11.2° to 3.5°. Figure 5 shows the mean rotation per level for each of the groups.
In both groups, axial rotation can be observed immediately following induction surgery. In terms of distribution of rotation at the end of induction, a gradual increase is observed from the non-apical areas towards the apical area. In both groups, the rotation appears to be symmetrically distributed between the cranial and caudal part of the spine.
Both the TSI only-group and DSR-group were able to (partly) reduce the anterior lengthening of the IVD during the reduction period ( Figure 5). However, only in the DSR treatment, did we also find a significant posterior lengthening of the vertebral body, indicating asymmetrical growth (A-P ratio 0.99 to 0.97).

Discussion
The current study investigated the potential of reducing scoliosis with instrumented apical derotation alone (TSI only-group) or in combination with posterior distraction (DSR-group). We used the same TSI implant to first induce scoliosis, a method that we previously showed to generate a very predictable idiopathic like spinal deformity that remained also months after removal of the implants. [13] By applying torque only, with the TSI, almost complete reduction of axial rotation could be achieved within 2 months. This resulted in correction of the coronal and sagittal plane as well, likely as a consequence of coupled motion in the spine, [21,22] but complete reduction could not be achieved. By adding a distraction force (SDS) to the TSI (thus utilizing the DSR concept), considerably more correction and even over-correction was obtained. This correction could be related directly to reduction in the typical relative anterior lengthening that is mainly present in the IVD, a phenomenon that we described extensively for human scoliosis. [3][4][5] This finding shows that by combining axial torque and concave distraction (DSR) we have a powerful tool to correct scoliosis simultaneously in all planes. However, it also indicates that the forces should be tailored to the specific condition as overcorrection is a risk. One way to mitigate this is to selectively release one of the forces when full correction in that plane has been achieved. Due to the position of the implant and the familiarity with this approach, this would be a minimally invasive procedure.
Besides full correction, we observed that normal growth and mobility of the instrumented segment was maintained.
In our opinion, this brings us one step closer to our goal of curing scoliotic spines in adolescents. This would require a paradigm shift in scoliosis care, namely that patients be treated surgically at an earlier age, perhaps already in the range of curves which are currently braced (30-50°). We propose that DSR may replace those brace treatments that are likely to fail or those that will likely end with considerable remaining curves. Compared to bracing, DSR has the obvious drawback of requiring a surgical intervention. However, the implant can be inserted less invasively and after insertion, the burden for both patients and caregivers will be much lower as there are no mobilization restrictions and no compliance issues. Furthermore, the transmission of forces through DSR's internal brace concept is superior to that of an external brace, especially for correction of the axial rotation.
DSR is not the first surgical technique to attempt gradual correction of AIS curves in the growing spine. In recent years, an increasing body of evidence has been generated wherein growth modulation has been achieved through anterior vertebral body tethering (AVBT). However, we believe AVBT has several disadvantages when compared to DSR. It halts growth on the convex side, and is limited in the amount of correction that can be achieved, in particular in the axial plane. [23] In contrast, DSR is able to continuously correct all planes simultaneously, whilst stimulating (not halting) the shorter concavity of the curve.
In addition to its use in AIS patients, DSR could also be used in growing EOS patients, as DSR has considerable advantages over current "growth-friendly" implants.
Especially in older EOS patients (i.e. "tweeners"), curve correction is often poor and complication rates are high, which has led to several studies concluding that spinal fusion in these patients may be more effective than "growth-friendly" treatment. [24,25] DSR allows for increased apical derotation, while its flexibility decreases stress-shielding of the spine, which may ultimately lead to reduced implant stresses and implant complications. [11,26] However, since the time interval 13 until skeletal maturation is longer than for AIS, DSR treatment of EOS patients may be somewhat unpredictable, an issue that is currently also observed in younger patients (Sanders 1-2) treated with AVBT. [23,27] Limitations of the current study are mainly related to the use of an animal model. Although the scoliosis morphologically resembles human scoliosis more than any other animal model, we do not know exactly how a human scoliotic spine will react to these dynamic forces. Based on our clinical experience using only spring distraction forces, especially idiopathic curves can be very difficult to correct or even control, suggesting that the etiological mechanism remains to be overcome. [18,19] Future fundamental and clinical studies will teach us more on this important aspect of the technology.

Conclusion
In our representative idiopathic-like, scoliotic animal model, correction with only axial torque was able to correct rotation of the apex, in addition to partially correcting the coronal curve, apical lordosis and anterior lengthening of the IVD. However, adding posterior distraction to the axial torque (DSR), resulted in stronger correction in the coronal and sagittal planes, in addition to posterior vertebral growth modulation.

Declarations of Competing Interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Funding
The current study was funded through a K2M (now Stryker Spine) Research Grant (R4198). The funder had no influence on the design of the study, collection, analysis and interpretation of the data, writing of the manuscript, nor on the decision to publish.   with which the tethers are fixated can be seen on the left, the tether itself is radiolucent. D: After 3 months, the scoliosis can be seen, including axial rotation. Also note the spinal growth seen as translation of the U-loops relative to the cranial and caudal anchors. E: During the reduction surgery, the torsional spring force is reversed and two distraction springs on a (radiolucent) flexible PEEK are fixated to the left-side.
F. After 2 months, the scoliosis was reduced in the axial and sagittal plane, and was even overcorrected in the coronal plane. The relative length (anterior length/posterior length) for both groups and for the total spine (i.e. vertebral bodies + intervertebral discs, left), only the vertebral bodies (middle) and only the intervertebral discs (right). A value > 1 indicates a segment that is longer anteriorly than posteriorly, as is common in human scoliosis.