- Email: [email protected]
The biomechanical effects of spinal fusion on the sacral loading in adolescent idiopathic scoliosis
Clinical Biomechanics 30 (2015) 981–987
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
The biomechanical effects of spinal fusion on the sacral loading in adolescent idiopathic scoliosis Saba Pasha a,b, Carl-Eric Aubin a,b,c,⁎, Hubert Labelle b,c, Stefan Parent b,c, Jean-Marc Mac-Thiong b,c,d a
Dept. Mechanical Engineering, École Polytechnique Montréal, P.O. Box 6079, Station “Centre-ville”, Montréal, Québec H3C 3A7, Canada Research Center, Sainte-Justine University Hospital Center, 3175, Cote Sainte-Catherine Road, Montréal, Québec H3T 1C5, Canada Department of Surgery, Université de Montréal, C.P. 6128, station “Centre-ville”, Montréal, Québec H3C 3J7, Canada d Division of Orthopedic Surgery, Hôpital du Sacré-Coeur de Montréal, 5400 Gouin Ouest, Montréal, Québec H4J 1C5, Canada b c
a r t i c l e
i n f o
Article history: Received 25 November 2014 Accepted 25 June 2015 Keywords: Scoliosis Biomechanics Finite element model Spinal fusion Sacrum Pelvis Spine
a b s t r a c t Background: Posterior spinal surgical correction is performed to correct spinal deformities in adolescent idiopathic scoliosis. Althoughthe relative spino-pelvic alignment changes after spinal surgery, pelvis remains unfused in idiopathic scoliosis surgery. The impact of the spinal fusion on the transferred load to the pelvis via sacrum is not documented in the scoliotic subgroups. Method: Bi-planar radiographs of 9 scoliotic subjects before and in average 16 months after spinal instrumentation surgery, and 12 controls were selected retrospectively. Patient-specific 3D reconstruction and finite element models of the spine, ribcage, and pelvis were developed. Spinal parameters (Cobb angles, kyphosis, lordosis), sacro-pelvic parameters (pelvic incidence, pelvic tilt, sacral slope), frontal and sagittal balances, the position of the trunk center of mass, and the centroid of the stress distribution on the sacrum superior endplate were measured and computed before operation and in the last follow-up. Findings: The position of the stress distribution centroid on the sacrum superior endplate with respect to the central hip vertical axis was significantly different between pre-operative and post-operative patients p b 0.05. The distance between the anterior–posterior position of the trunk center of mass and thecenter of pressure on the superior sacral endplate significantly decreased after the spinal surgery p b 0.05. Interpretation: The impact of the scoliosis spinal fusion on the transferred load between the spine and pelvis was evaluated. The biomechanical loading of the sacrum endplate was related to the post-operative postural balance and compensatory changes in the spino-pelvic alignment after scoliosis surgery. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Posterior spinal instrumentation and fusion (PSIF) aims to correct and stabilize the spinal deformity in severe cases of scoliosis until fusion occurs (Bridwell et al., 2002; Moen and Nachemson, 1999). Although the clinical impact of the spinal fusion on the geometrical parameters of the spine and pelvis has been studied previously (Masso and Gorton, 2000; Tanguay et al., 2007) the effects of the spinal surgery on the biomechanical loading of the distal un-fused vertebrae are not well documented. More specifically the biomechanical loading of the sacrum, which affects the conducted force between the spine and lower extremities and hence contributes to the standing postural equilibrium (Jiang et al., 2006) is not closely investigated in adolescent idiopathic scoliosis (AIS) subgroups post-operatively. The asymmetrical loading of the spinal vertebrae (Stokes, 2007) is shown to be associated with curve progression, which emphasizes on the importance of the ⁎ Corresponding author at: École Polytechnique Montreal, Department of Mechanical Engineering P.O. Box 6079, Station “Centre-ville”, Montréal, Québec H3C 3A7, Canada. E-mail address: [email protected] (C.-E. Aubin).
http://dx.doi.org/10.1016/j.clinbiomech.2015.06.019 0268-0033/© 2015 Elsevier Ltd. All rights reserved.
considering the vertebral loading of the unfused spine in post-surgical evaluation of the patients. Since the introduction of the “pelvic vertebra” in 1994 (Dubousset, 1994), pre- and post-operative spino-pelvic alignment in scoliosis has become the subject of many studies (Legaye et al., 1998; Pasha et al., 2010; Pasha et al., 2014a; Qiu et al., 2013; Roussouly, et al., 2013). Changes in pelvic alignment and spino-pelvic kinematic interaction were highlighted after spinal surgical correction in scoliosis (Skalli et al., 2006; Tanguay et al., 2007; Yang et al., 2015). The importance of considering the pelvic sagittal alignment with respect to the spine particularly the lumbar lordosis in AIS surgical planning was underlined (Johnson et al., 2012; Roussouly et al., 2013; Tanguay et al., 2007). However the transferred load to the pelvis through sacrum characterized by the compressive stress on the sacrum in post-surgical AIS that may impact the long-term spino-pelvic alignment is still to be investigated. Despite the body of literature examining post-surgical spino-pelvic analysis in AIS, it is not clear to what extent the spinal surgery impacts the biomechanical loading of the pelvis and lower limbs in postoperative AIS subgroups. In order to answer this question the specific objective of the current study was to compute and compare the compressive
982
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
stress on the superior sacral endplate in pre- and post-operative scoliotic subjects with different curve patterns and also compare it to a group of asymptomatic controls using a comprehensive osseo-ligamentous finite element model of the spine and pelvis (Clin et al., 2011). It was hypothesized that even though sacrum remains unfused in AIS spinal surgery the biomechanical loading of the sacrum changes significantly after PSIF in AIS subgroups and becomes more similar to the ones in controls in a balanced spino-pelvic alignment after the surgical correction of the spine. 2. Methods 2.1. Subjects The ethical approval was obtained from the ethical committee of the hospital and the affiliated research institution for this research study. A total number of 9 AIS female subjects (age range [14, 17], average 15 years, SD: 2.4, average weight 50.1 kg, SD 6.2) who had undergone a PSIF between 2006 and 2010 were randomly and retrospectively selected from the database of our institution. This sample size provided 80% statistical power for a paired analysis between the pre- and postoperative patients. All- pedicle- screw construct was used for all the patients except for one who was treated with a hybrid construct (screws and a distal hook). The number of fused vertebrae varied between 6 and 14 vetebrae. No post-operative instrumentation failure or surgical complications during an average follow-up of 16 months [12–18 months, SD: 3] was reported in the studied group. The medical chart and pre- and post-operative bi-planar radiographs of the patients were consulted. A total number of 5 patients had right main thoracic deformity (RT), Cobb angle range [43°, 77°] and 4 patients had right thoracic (RT) [55°, 68°] and left lumbar (LL) [74°, 97°] deformities. The radiographic images of 12 age- and sex-matched asymptomatic female adolescent subjects [age 11–18 years average 14.3, SD 4.0, average weight 54.8, SD 8.3] with no history of spinal disease were examined by an orthopedic surgeon and were added as the control group.
A
B
2.2. 3D reconstruction technique and anatomical measurements A self-calibration technique was used to generate the weightbearing 3D reconstructions of the spine, rib cage, and pelvis of the cohort from their bi-planar X-rays before and after surgery using the technique explained by Kadoury et al. (2007) (Fig. 1-A and Fig. 1-B). The reconstruction method consisted of identifying a limited number of points on the radiographs (14 nodes per vertebra, 11 nodes per rib, and 24 nodes on the pelvis) and using a detailed atlas of the spine and pelvis along with a free form morphing algorithm to create the detailed skeletal geometry of the spine, ribcage, and pelvis (Cheriet et al., 2002; Delorme et al., 2003; Kadoury et al., 2007). In the self-calibration process the retro-projection errors of the anatomical landmarks were minimized by changing the radiological setup. The new radiological setup was determined by changing the geometrical parameters of the radiographic system during a non-linear optimization process. The self-calibration and reconstruction methods are described in farther detail in Kadoury et al. (2007). In order to determine the accuracy of the 3D reconstruction model using the self-calibration technique all the 3D measurements were compared to a previously verified calibration and 3D reconstruction technique (Cheriet et al., 2002 and 2007). The 3D reconstructions were generated for 60 patients using both methods and were statistically compared. The average errors in the 3D reconstruction of the vertebral body, originated from the selfcalibration technique, were (1.2 mm, S.D. 0.8 mm) and vertebral pedicles (1.6 mm, S.D. 1.1 mm). The accuracy of the bi-femoral head axis alignment in the frontal plane was 0.44°, S.D. 0.46°. The maximum error in measurement of the pelvic sagittal parameters was 0.99°, S.D. 1.10° (Kadoury et al., 2007). A maximum error of 7° was reported in 2D coronal and sagittal spinal curves measurements using the 3D model when compared to the 2D clinical measurements on the X-ray images (Delorme et al., 2003). An analytical method was used to measure the spinal curvatures in the frontal and sagittal planes (Stokes, 1994) as shown in Fig. 1-B and
C
D
Fig. 1. A) Bi-planar X-rays. An analytical method was used to measure B) thoracic and lumbar Cobb angles, C) kyphosis (T4–T12), and lordosis (L1–S1) from the 3D reconstruction of the spine. The dashed line in Fig. 1-B is the spline connecting the vertebral centroids. D) Sacro-pelvic parameters in the sagittal plane: Sacral slope (SS), pelvic tilt (PT), and pelvic incidence (PI).
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
983
Fig. 1-C respectively. In this method the spinal curvature was defined by a cubic spline generated by interpolating the position of the vertebrae center. The angle between the lines perpendicular to the projection of the spinal curve in the frontal and sagittal planes at its inflection points is used to measure the spinal curvatures in the frontal and sagittal planes respectively (Fig. 1-B). The spinal curvature in the frontal plane was described by the thoracic and lumbar Cobb angles (Fig. 1-B). In the sagittal plane kyphosis included T4 to T12 vertebrae and lordosis was measured between L1 and S1 (Fig. 1-C). Sacro-pelvic parameters were characterized by pelvic incidence (PI), pelvic tilt (PT), and sacral slope (SS) (Fig. 1-D). The measured spinal and pelvic parameters are shown in Fig. 1-B–1-D. The averages of the spinal and pelvic parameters are presented in Table 1 for the patients before and after the surgery and for the asymptomatic controls. Frontal and sagittal balances were calculated as the distance between the C7 plumb line i.e. the vertical line passing through the center of the superior endplate of the C7 and S1 superior endplate center in the frontal and sagittal planes respectively. These measurements are presented in Table 2 for pre- and postoperative subjects and controls.
weight that was applied at each vertebral level in the FEM. A total number of 17 nodes were associated with the weights of each trunk slice from T1 to L5 in the FEM. The 3D position of the trunk's COM (COMtrunk) was determined as the weighted sum of the COM positions at the level of each trunk slice. The gravitational load was applied to the FEM in two steps. First the gravitational force was applied in the opposite direction of the gravity and the spinal loadings calculated at the end of this phase were set to zero. In the second phase, the weight of each trunk slice was applied in the true direction of the gravity and the geometry of the spine after the second phase of the simulation was compared to the geometry of the weight-bearing 3D reconstruction model of the spine. Transverse forces were applied at vertebral levels in the second phase of the simulation. An optimization process minimized the differences between the weight-bearing 3D reconstruction of the spine and the FEM after application of the gravitational force by changing the components of the applied forces (Clin et al., 2011). This method is detailed in Clin et al. (2011). The objective function of the optimization process was defined as:
2.3. Personalized finite element model (FEM)
X17 Fob j ¼ Min i¼1 X f −X i j þ jY f −Y i j þ jZ f −Z i
The weight-bearing 3D geometry of the spinal vertebrae, pelvis, and ribcage from the 3D reconstruction model was used to develop a personalized osseo-ligamentous FEM of the spine, pelvis, and ribcage using ANSYS 11.0 FE package (ANSYS Inc., Canonsburg, PA, USA) for the cohort of subjects (Fig. 2). This model is presented elsewhere in more details (Clin et al., 2011) and is summarized here. The FEM contains a total number of 3163 nodes and 3209 elements. The element types, Young's modulus, and Poisson's ratio of each part were derived from literature and are presented in Table 3 (Aubin et al., 1996; Clin et al., 2010; Clin et al., 2011; Perie et al., 2004). Element sizes were determined from the 3D reconstruction of the bone and were scaled for each patient. Each vertebra was modeled with a 3D beam element and has an octagonal shape cross-section (Fig. 2-C). Zygapophyseal joints were simulated by contact and shell elements. The abdominal wall was approximated by linear interpolation of the nodes of ribcage, pelvis, and vertebral bodies. Two different coordinate systems, global and local, were defined for the FEM. The origin of the global coordinate system was at the central hip vertical axis (CHVA). The Z-axis was ascending vertically, the X-axis was placed posterior-anteriorly, and the Y-axis was medio-laterally (to the left) (Fig. 2-B). The origin of the local coordinate system was fixed at the S1 center (Fig. 2-C). The X-axis is pointing anteriorly and the Y-axis was medio-laterally to the left (Fig. 2-C). The position of the center of mass (COM) at the level of each vertebra in the sagittal plane was determined from literature (Liu et al., 1971; Pearsall et al., 1996). A rigid beam was created in the FEM to connect the COM of each trunk slice to the center of the corresponding vertebra. The COM position in the frontal plane was assumed to be located at the center of the vertebrae in the frontal plane (Clin et al., 2011). The weight and COM of each trunk slice, as a percentage of the whole body weight, were found in the literature (Pearsall et al., 1996). The weight and the COM of the head, neck, and arms were associated with the COM of the T1, and T3-T5 vertebrae levels by the method described by El-Rich and Shirazi-Adl (2005). Table 4 summarizes the percentage of the body
where Xi, Yi, and Zi are the positions of the center of vertebrae from the 3D reconstruction and Xf, Yf, and Zf are the positions of the center of the vertebrae in the FEM after the gravitational forces were included. A difference less than 1% between the two consecutive iterations of the objective function was set as the convergence criterion (Clin et al., 2011). The model was fixed only in the transverse plane at the level of the first thoracic vertebrae (T1) and the 3D rotation and ascending or descending movements of the spine were allowed in the Z direction. The pelvis was fixed, both translationally and rotationally, in its actual position as it was identified from the 3D reconstructions (Clin et al., 2011). The FEM was used to calculate the distribution of the compressive stress on the S1 superior endplate (Fig. 2-C) in the local coordinate system. The compressive stress on the octagonal shape of the simulated sacrum in the FEM was calculated. Given the stress distribution at the 8 corners of the S1 endplate (Fig. 2-C) and the sacrum radius in individuals, the centroid of the stress distribution on the S1 superior endplate (CoPS1) was calculated in the cohort of subjects as it was explained in Pasha et al., 2014b. The position of the CoPS1 is shown in the local coordinate system in Fig. 2-C. The compressive stress was later normalized to the patient's weight and scaled between the minimum and maximum compressive stresses on the sacrum resulting in a normalized stress distribution between [0, 1]. The projections of the positions of the COMTrunk and CoPS1 on the transverse plane were determined in the global coordinate system with respect to the CHVA to make comparison between the two parameters possible (Pasha et al., 2014b). Each patient was described with 9 geometrical parameters (thoracic and lumbar Cobb angles, kyphosis, lordosis, PI, PT, SS, and sagittal and frontal planes balances) and two biomechanical parameters (the position of the COMTrunk in the global coordinate system and the position of the CoPS1 in the global and local coordinate systems).
Table 1 Spinal and pelvic parameters of the pre- and post-operative patients and asymptomatic controls. Subjects Pre-operative Post-operative Controls
MT(n = 5) RT/LL(n = 4) MT(n = 5) RT/LL(n = 4)
Thoracic Cobb (°)
Lumbar Cobb (°)
Kyphosis (°)
Lordosis (°)
PI (°)
PT (°)
SS (°)
53 SD 33 40 SD 25 26 SD 7 23 SD 18 –
28 SD 22 55 SD 27 15 SD 13 24 SD 20 –
33 SD 25 32 SD 12 29 SD 15 26 SD 13 47 SD 10
44 SD 17 43 SD 13 53 SD 6 50 SD 13 50 SD 7
39 SD 2 42 SD 1 51 SD 6 55 SD 7 44 SD 7
4 SD 2 11 SD 8 12 SD 3 18 SD 9 9 SD 7
34 SD 17 31 SD 5 38 SD 4 36 SD 9 35 SD 4
984
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
Table 2 Frontal and sagittal balances in pre and post-operative subjects and controls (positive measurements are to the left (in frontal plane) and anterior (in sagittal plane). Subjects Pre-operative Post-operative
MT (n = 5) RT/LL (n = 4) MT (n = 5) RT/LL (n = 4)
Controls n = 12
Frontal balance (mm)
Sagittal balance (mm)
−9 SD 17.9 31 SD 19.1 8.8 SD 11.6 15.9 SD 20.3 −1.1 SD 5.3
3 SD 23.7 −5 SD 41.0 −12.0 SD 17.5 −18.7 SD 36.7 −11 SD 6.4
Table 3 The Young modulus of model components in the FEM.
2.4. Statistical analysis All the statistical analyses were performed in PASW statistics 18.0, SPSS Inc., Chicago, IL. A normality test (Shapiro–Wilk test) was performed to determine whether the measured and calculated parameters are normally distributed. The differences between the average of the spinal and pelvic biomechanical and geometrical parameters were determined between the asymptomatic controls and AIS pre- and post-operatively. A Mann Whitney U test was used to compare the clinical parameters before and after the surgery. A two-tailed Pearson's correlation test was applied to determine the relationship between the biomechanical and geometrical parameters of the spine and pelvis in the studied groups before and after operation. Finally a clustering technique (K-mean cluster), with two clusters permitted, was used to divide the distributions of the compressive stress on the S1 in two low and high stress areas in each subject.
Model component
Element
Poisson's ratio
Young's modulus (MPa)
Vertebral body Pedicle Spinal process Sternum Posterior arc Rib Intercostal ligaments Abdominal cavity Intervertebral disks
Linear 3D elastic beam Linear 3D elastic beam Linear 3D elastic beam Linear 3D elastic beam Linear 3D elastic beam Linear 3D elastic beam Non linear spring Linear hexahedral elements Linear 3D elastic beam
0.3 0.3 0.3 0.2 0.3 0.1 – 0.45 0.45
1000 5000 35,000 10,000 1000 5000 35 0.01 3.5–15
3. Results 3.1. Case presentation Fig. 3 presents the bi-planar radiographs and the position of the high and low stress areas on the sacrum endplate in two scoliotic subjects before and after surgery. Spinal and pelvic parameters of these two patients were presented in Table 5. The sacral loading was more contra-laterally asymmetric before operation while it was more equilibrated after operation in both subjects (Fig. 3). 3.2. Comparison between the spinal and pelvic geometrical parameters preand post-operatively
2.5. Sensitivity analysis The normality test suggested that the data is not normally distributed and thus non-parametric statistical analysis was used to compare the mean values. A Mann Whitney U test showed significant decrease in the frontal plane spinal curvature i.e. thoracic and lumbar Cobb angles (p b 0.05), while the spinal parameters in the sagittal plane i.e. kyphosis and lordosis did not change significantly (p N 0.05). Both frontal and sagittal balances were significantly different between pre-operative RT/LL subjects and controls while only frontal balance was significantly different between RT and controls (p b 0.05). No significant difference was observed between post-operative and control subjects when frontal and sagittal balances were compared (p N 0.05). Thoracic Cobb angle was decreased by 51% in RT subjects and 48% in RT/LL group. Lumbar Cobb angles were decreased by 52% in RT subjects
The sensitivity of the FEM was tested for several design parameters. The modulus of elasticity of the intervertebral disk was multiplied by 0.5 and 1.5 to simulate the flexible versus rigid spine respectively (Clin et al., 2010). The effect of the spinal flexibility on the S1 loading and the position of the CoPS1 was studied. Furthermore the sensitivity of the model to the COM position was tested. The COM position in the frontal plane shifted 1 cm to the left and right sides of the vertebral centers at each vertebral level and the sacral loading was calculated for the new position of the COM to evaluate the sensitivity of the FEM to small changes of this parameter. The reaction forces and the bending moment at the T1 vertebra were calculated to determine the required counter balancing forces to stabilize the FEM.
Anterior
COPS1 Y
X ri Si
Posterior
Z X Y
A
B
C
Stressmin
Stressmax
Fig. 2. A) Posterior and lateral view of the FEM of spine, pelvis and ribcage. B) The biomechanical compressive stress of the spine in the FEM after applying the gravitational force. C) Detailed view of the stress distribution on the superior sacral endplate (Si), location of the COPS1, and the sacrum radius ri.
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987 Table 4 Trunk slices' weight as a percentage of total body weight. Vertebra level
Percentage of total body weight
T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 L1 L2 L3 L4 L5 Total
1.1% + 8% (head) 1.1% 1.3% + 4% (arms) 1.3% + 4% (arms) 1.3% + 4% (arms) 1.3% 1.4% 1.5% 1.6% 2.0% 2.1% 2.5% 2.4% 2.4% 2.3% 2.6% 2.6% 50.8%
and 58% in RT/LL group after operation. Kyphosis was decreased by 8% and 14% in RT and RT/LL groups respectively. Lordosis was increased by 20% in RT and 12% in RT/LL group. PI was increased by 15% after operation while SS was decreased by 5%. However, the Mann Whitney U test did not show any significant difference in the magnitude of the PI and the SS before and after operation (p N 0.05). PI was the only pelvic parameter, which was significantly correlated to the lumbar lordosis both before and after spinal surgical correction (r = 0.59, r = 0.64, p b 0.05 pre- and post-operatively respectively). 3.3. Comparison between the spinal and pelvic biomechanical parameters pre- and post-operatively The average positions of the CoPS1 and COMTrunk were presented in Table 6 for the studied groups. The medio-lateral parameters were
985
measured in the frontal plane and the anterior–posterior parameters were measured in the sagittal plane with respect to the CHVA (global coordinate system). The position of the CoPS1 with respect to the CHVA was significantly different pre- and post-operatively and between the pre-operative subjects and controls (p b 0.05) while no such difference was observed between the post-operative subjects and controls. The biomechanical loading of the sacrum varied between the pre- and post-operative subjects; the stress distribution on the sacrum was more symmetric i.e. no significant difference between the compressive stresses was measured on the left and right sides of the sacrum after operation (p b 0.05). The distance between the projection of the COMTrunk and the CoPS1 on the transverse plane was decreased in both medio-lateral (88% RT, 65% RT/LL) and anterior–posterior (55% RT, 57% RT/LL) directions after operation. A Mann Whitney U test showed that the distance between the projection of the COMTrunk and CoPS1 on the transverse plane decreased significantly after operation in the cohort of scoliotic subjects (p b 0.05). The anterior–posterior distance between the CoPS1 and COMTrunk in the transverse plane was significantly correlated to the sagittal balance (r = 0.47, p b 0.05) and its medio-lateral distance was significantly related to the frontal balance (r = 0.74, p b 0.05) in the cohort of the pre- and post-operative scoliotic subjects. A significant relationship was observed between the SS and the biomechanical parameters i.e. the position of the CoPS1 and the COMTrunk; as the SS increased the anterior–posterior distance between the COMTrunk and the CoPS1 decreased significantly in the postoperative subjects (r = 0.6, p b 0.05). 3.4. Sensitivity analysis The result of the sensitivity analysis did not show any significant change in the positions of the CoPS1 when the spinal flexibility varied from stiff to normal and to flexible spine (p N 0.05, Mann Whitney U test). Small changes in the position of the COMTrunk, as was explained in the method section, did not significantly change the positions of the
High stress High stress Low stress Low stress
A
High stress
High stress Low stress
B Fig. 3. Bi-planar radiographs and the location of the high and low stress areas on the superior sacral endplate before and after surgery in a A) MT (Patient1) and B) RT/LL (Patient2) (Table 6). The dash line separates the anterior and posterior parts of the superior sacral endplate.
986
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
Table 5 Pre- and post-operative spinal and pelvic parameters in a patient with a) RT deformity (patient1) and b) RT/LL curve (patient2).
a) Patient1 (RT) b) Patient2 (RT/LL)
Pre-operative Post-operative (16 months) Pre-operative Post-operative (12 months)
Thoracic Cobb (°)
Lumbar Cobb (°)
Kyphosis (°)
Lordosis (°)
PI (°)
PT (°)
SS (°)
43 24
25 8.5
44 32
57 55
50 40
6 13
44 27
68 48
97 54
19 38
61 54
71 67
40 24
31 42
CoPS1. The maximum reaction force at the level of T1 was calculated at 18 N. The maximum bending moment about the Y-axis (My) did not exceed 135 N-mm. 4. Discussion An osseo-ligamentous FEM of the spine and pelvis was used to study the differences between the pre- and post-operative biomechanical loading of the superior endplate of the sacrum in AIS. The effect of the scoliotic surgical correction of the spine on the compressive stress distribution on the S1 endplate was shown in post-operative AIS subgroups. The compressive stress distribution on the S1 superior endplate after spinal operation was more equilibrated contra-laterally. The position of the CoPS1 was related to the sagittal and frontal balance of the pre- and post-operative patients. This result was in line with our hypothesis that the biomechanical loading of the sacral endplate is more similar between controls and post-operative patients despite the fact that the sacrum and pelvis remained unfused. The biomechanical analysis in this study provided a quantitative measure to evaluate the sacral loading and consequently the transferred load between the spine and pelvis in pre- and post-operative AIS patients. Moreover we related the biomechanical loading of the sacrum to the overall postural balance of the AIS. Although the sagittal spino-pelvic alignment in AIS has been studied before, this study is the first to analyze the relationship between the sacral compressive stress and postural balances which can affect the long-term spino-pelvic alignment in AIS. A decrease in the compressive stress asymmetry on the vertebra after surgery can decrease the bone remodeling and subsequent curve progression in scoliosis. The asymmetrical compressive stress of the vertebra has been associated with disk degeneration and curve progression in scoliosis, (Adams et al., 2000) which suggests the importance of considering the vertebral mechanical loading at the unfused spinal sections below the fused spine in AIS particularly at the base of the spine i.e. sacrum to prevent post-surgical curve progression. The correlation between the sagittal and frontal balances and the anterior–posterior distance between the CoPS1 and COMTrunk projections on the transverse plane was shown. An increase in the anterior–posterior distance between the COMTrunk and CoPS1 in the pre-operative patients was accompanied by the pelvic retroversion (decreased SS) and increased distance between C7 and S1 in sagittal plane. In post-operative subjects a posterior shift in the COMTrunk position, in comparison to its pre-operative position, was accompanied by increased SS and an anterior shift in the CoPS1 position. These
changes decreased the anterior–posterior CoPS1–COMTrunk distance in the sagittal plane. The results also showed that a posterior shift in the position of the CoPS1 on the sacral superior endplate in the sagittal plane is linked to the decreased SS and increased distance between the C7 and S1 in the sagittal plane. This phenomenon was observed previously by Roussouly et al. (2013), where it was explained that pelvic retroversion should be considered as a compensatory mechanism in subjects with poor sagittal balance which helps with improving the postural equilibrium during quiet stance. The aforementioned pelvic compensatory mechanism is an attempt to bring back the spine to an upright position however it has been associated with pain, fatigue, and disability particularly in adult patients (Ames et al., 2012; Lafage et al., 2009; Roussouly et al., 2013). The results of our study are in line with previous observation in sagittal spino-pelvic alignment and pelvic compensatory mechanism and can quantitatively explain this phenomenon from a biomechanical point of view. The pelvic 3D orientation and sacro-pelvic parameters did not vary in the same way i.e. increase or decrease for all the subjects postoperatively, similar to the results in Skalli et al. (2006). The results of the current study showed that the impact of the spinal operation is more prominent on the biomechanical parameters of the pelvis i.e. CoPS1 than on its geometrical parameters e.g. SS and PI which highlights the importance of this biomechanical notion in spino-pelvic evaluation of the scoliotic subjects after surgery. This study had some limitations that should be undertaken in the future; The study was only conveyed on a limited number of subjects. A multicenter study design that evaluates a higher number of subjects with different curve patterns, different fusion levels, surgical techniques, and instrumentation strategies would permit to better investigate the association between the spino-pelvic parameters and the sacral loading post-operatively. The position of the COMTrunk and the spinal material properties were not personalized in the FEM. However the results of the clustering analysis did not change significantly as a function of spinal stiffness or small changes in the position of the COMTrunk. The changes in the muscles' moment arm with respect to spinal vertebrae before and after the surgery that can impact the in vivo vertebral loading, was not included in this FEM. Despite the limitations of the FEM, the small reaction forces at the proximal boundary level suggest that this model is stable and the reaction forces are physiologically realistic. Including the muscle forces, personalized ground reaction forces, and personalized position of the COM would allow not only studying the normalized-scaled stress distribution on the sacrum but also comparing the magnitude of the compressive stress
Table 6 The average position of the pre- and post-operative biomechanical parameters (COM, CoPS1) in the studied groups. CoPS1 (mm) With respect to the CHVA
Pre-operative Post-operative Controls (n = 12)
MT (n = 5) RT/LL (n = 4) MT (n = 5) RT/LL (n = 4)
COM (mm) With respect to the CHVA
Medio-lateral
Postero-anterior
Medio-lateral
Postero-anterior
−8 SD 2 8 SD 4 5 SD 1 4 SD 1 3 SD 4
−6 SD 1 −13 SD 5 −13 SD 2 −21 SD 5 −18 SD 7
−15 SD 11 25 SD 18 6 SD 9 12 SD 15 8 SD 5
1 SD 5 −5 SD 8 −10 SD 6 −16 SD 10 −23 SD 12
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
before and after operation. Another limitation of the model originates from the simulation of the abdominal components; in the FEM the geometry of the abdominal wall was approximated from the geometry of the ribs and vertebrae (Clin et al., 2011) and it was not compared to the 3D surface scans of the trunk. However it was shown that the vertebral loading in the current model when the pressure from the abdominal cavity was included in the FEM is similar to the ones in the literature (0.1–0.2 MPa) (Driscoll et al., 2009 and Stokes, 2007). Although the different components of the intervertebral disks were not simulated in the current FEM this model has proven to provide realistic information on qualitative vertebral compressive loading when it was compared to other models where the growth modulation (Stokes, 2007) and annulus-nucleus (Driscoll et al., 2009) were simulated (Clin et al., 2010, 2011). The FEM simulation was not different between the patients who were treated by the hybrid or all- screw constructs. The focus of the study was on analyzing the impact of the post-surgical spinal geometrical changes on the sacral endplate loading and other surgical parameters such as the surgical technique, implant type and density, fusion length, and the level of fusion were not considered in the analysis. Such analysis although can highlight the impact of the surgical parameters on the biomechanical loading of the sacrum, will require a larger sample size to include a variety of surgical techniques and should be the subject of another study. The biomechanical parameters of the sacral endplate provided coherent information about the biomechanics of the spino-pelvic alignment after spinal surgery and can be accompanied by the patients' geometrical parameters to better evaluate the long-term results of the spinal surgical treatment in scoliosis. This study focused on characterizing the sacral loading in post PSIF AIS patients and showed the relationship between the compensatory spino-pelvic alignment and the changes in the sacral loading. The relationship between the local sacral biomechanical loading and the overall postural balance in postsurgical AIS was highlighted. 5. Conclusion The impact of the surgical spinal correction on the biomechanical loading of the sacrum was highlighted in AIS subgroups. The notion of the CoPS1 was used to characterize the sacral mechanical loading and relate it to pre- and post-operative postural compensatory mechanisms in AIS subgroups. Acknowledgments The project was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC : PCIPJ346145-11) and Fonds de Recheche du Quebec- Nature et Technologies Recherche sur la Nature et les Technologies (FRQNT). References Adams, M.A., Freeman, B.J.C., Morrison, H.P., Nelson, I.W., Dolan, P., 2000. Mechanical initiation of intervertebral disc degeneration. Spine 25 (13), 1625–1636. Ames, C.P., Smith, J.S., Scheer, J.K., Bess, S., Bederman, S.S., Deviren, V., Lafage, V., Schwab, F., Shaffrey, C.I., 2012. Impact of spinopelvic alignment on decision making in deformity surgery in adults: a review. J. Neurosurg. Spine 16 (6), 547–564. Aubin, C.E., Dansereau, J., de Guise, J.A., Labelle, H., 1996. A study of biomechanical coupling between spine and rib cage in the treatment by orthosis of scoliosis. Ann. Chir. 50 (8), 641–650. Bridwell, K.H., Hanson, D.S., Rhee, J.M., Lenke, L.G., Baldus, C., Blanke, K., 2002. Correction of thoracic adolescent idiopathic scoliosis with segmental hooks, rods, and Wisconsin wires posteriorly: it's bad and obsolete, correct? Spine 27 (18), 2059–2066. Cheriet, F., Remaki, L., Bellefleur, C., Koller, A., Labelle, H., Dansereau, J., 2002. A new X-ray calibration/reconstruction system for 3D clinical assessment of spinal deformities. Stud. Health Technol. Inform. 91, 257–261.
987
Cheriet, F., Laporte, C., Kadoury, S., Labelle, H., Dansereau, J., 2007. A novel system for the 3-D reconstruction of the human spine and rib cage from biplanar X-ray images. IEEE Trans. Biomed. Eng. 54 (7), 1356–1358. Clin, J., Aubin, C.-E., Parent, S., Sangole, A., Labelle, H., 2010. Comparison of the biomechanical 3D efficiency of different brace designs for the treatment of scoliosis using a finite element model. Eur. Spine J. 19 (7), 1169–1178. Clin, J., Aubin, C.E., Lalonde, N., Parent, S., Labelle, H., 2011. A new method to include the gravitational forces in a finite element model of the scoliotic spine. Med. Biol. Eng. Comput. 49, 967–977. Delorme, S., Petit, Y., de Guise, J.A., Labelle, H., Aubin, C.E., Dansereau, J., 2003. Assessment of the 3D reconstruction and high-resolution geometrical modeling of the human skeletal trunk from 2D radiographic images. IEEE Trans. Biomed. Eng. 50 (8), 989–998. Driscoll, M., Aubin, C.-E., Moreau, A., Villemure, I., Parent, S., 2009. The role of spinal concave–convex biases in the progression of idiopathic scoliosis. Eur. Spine J. 18 (2), 180–187. Dubousset, J., 1994. Three-dimensional analysis of the scoliotic deformity. In: Weinstein, S.L. (Ed.), Pediatric Spine: Principles and Practice. Raven Press, New York. El-Rich, M., Shirazi-Adl, A., 2005. Effect of load position on muscle forces, internal loads and stability of the human spine in upright postures. Comput. Methods Biomech. Biomed. Engin. 8, 359–368. Jiang, Y., Nagaski, S., You, M., Zhou, J., 2006. Dynamic studies on human body sway by using a simple model with special concerns on the pelvic and muscle roles. Asian J. Control 8, 297–306. Johnson, R.D., Valore, A., Villaminar, A., Comisso, M., Balsano, M., 2012. Sagittal balance and pelvic parameters—a paradigm shift in spinal surgery. J. Clin. Neurosci. 20 (2), 191–196. Kadoury, S., Cheriet, F., Laporte, C., Labelle, H., 2007. A versatile 3D reconstruction system of the spine and pelvis for clinical assessment of spinal deformities. Med. Biol. Eng. Comput. 45, 591–602. Lafage, V., Schwab, F., Patel, A., Hawkinson, N., Farcy, J.P., 2009. Pelvic tilt and truncal inclination: two key radiographic parameters in the setting of adults with spinal deformity. Spine 34 (17), E599–E606. Legaye, J., Duval-Beaupère, G., Hecquet, J., Marty, C., 1998. Pelvic incidence: a fundamental pelvic parameter for three-dimensional regulation of spinal sagittal curves. Eur. Spine J. 7, 99–103. Liu, YK, Laborde, JM, Van Buskirk, WC, 1971. Inertial properties of a segmented cadaver trunk: their implications in acceleration injuries. Aerosp. Med. 42, 650–657. Masso, P.D., Gorton, G.E., 2000. Quantifying changes in standing body segment alignment following spinal instrumentation and fusion in idiopathic scoliosis using an optoelectronic measurement system. Spine 25, 457–462. Moen, K.Y., Nachemson, A.L., 1999. Treatment of scoliosis: an historical perspective. Spine 24, 2570–2575. Pasha, S., Sangole, A.P., Aubin, C.-E., Parent, S., Mac-Thiong, J.-M., Labelle, H., 2010. Characterizing pelvis dynamics in adolescent with idiopathic scoliosis. Spine 35, E820–E826. Pasha, S., Aubin, C.-E., Sangole, A.P., Labelle, H., Parent, S., Mac-Thiong, J.-M., 2014a. Threedimensional spinopelvic relative alignment in adolescent idiopathic scoliosis. Spine 39 (7), 564–570. Pasha, S., Aubin, C.-E., Labelle, H., Parent, S., Mac-Thiong, J.-M., 2014b. Biomechanical loading of the sacrum in adolescent idiopathic scoliosis. Clin. Biomech. 29 (3), 296–303. Pearsall, DJ, Reid, JG, Livingston, LA, 1996. Segmental inertial parameters of the human trunk as determined from computed tomography. Ann. Biomed. Eng. 24, 198–210. Perie, D., Aubin, C.E., Petit, Y., Labelle, H., Dansereau, J., 2004. Personalized biomechanical simulations of orthotic treatment in idiopathic scoliosis. Clin. Biomech. 19 (2), 190–195. Qiu, X.S., Wang, Z.W., Qiu, Y., Wang, W.J., Mao, S.H., Zhu, Z.Z., Qian, B.P., Wang, S.F., Zhu, F., Qiao, J., 2013. Preoperative pelvic axial rotation: a possible predictor for postoperative coronal decompensation in thoracolumbar/lumbar adolescent idiopathic scoliosis. Eur. Spine J. 22 (6), 1264–1272. Roussouly, P., Labelle, H., Rouissi, J., Bodin, A., 2013. Pre- and post-operative sagittal balance in idiopathic scoliosis: a comparison over the ages of two cohorts of 132 adolescents and 52 adults. Eur. Spine J. Suppl. 2, S203–S215. Skalli, W., Zeller, R.D., Miladi, L., Bourcereau, G., Savidan, M., Lavaste, F., Dubousset, J., 2006. Importance of pelvic compensation in posture and motion after posterior spinal fusion using CD instrumentation for idiopathic scoliosis. Spine 31, E359–E366. Stokes, I.A., 1994. Three-dimensional terminology of spinal deformity: a report presented to the Scoliosis Research Society by The Scoliosis Research Society Working Group on 3-D Terminology of Spinal Deformity. Spine 19 (2), 236–248. Stokes, I.A., 2007. Analysis and simulation of progressive adolescent scoliosis by biomechanical growth modulation. Eur. Spine J. 16 (10), 1621–1628. Tanguay, F., Mac-Thiong, J.-M., de Guise, J.A., Labelle, H., 2007. Relation between the sagittal pelvic and lumbar spine geometries following surgical correction of adolescent idiopathic scoliosis. Eur. Spine J. 16 (4), 531–536. Yang, X., Liu, L., Song, Y., Zhou, C., Zhou, Z., Wang, L., Wang, L., 2015. Pre-and postoperative spino-pelvic sagittal balance in adolescent patients with lenke type 5 idiopathic scoliosis. Spine 40 (2), 102–108.
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
The biomechanical effects of spinal fusion on the sacral loading in adolescent idiopathic scoliosis Saba Pasha a,b, Carl-Eric Aubin a,b,c,⁎, Hubert Labelle b,c, Stefan Parent b,c, Jean-Marc Mac-Thiong b,c,d a
Dept. Mechanical Engineering, École Polytechnique Montréal, P.O. Box 6079, Station “Centre-ville”, Montréal, Québec H3C 3A7, Canada Research Center, Sainte-Justine University Hospital Center, 3175, Cote Sainte-Catherine Road, Montréal, Québec H3T 1C5, Canada Department of Surgery, Université de Montréal, C.P. 6128, station “Centre-ville”, Montréal, Québec H3C 3J7, Canada d Division of Orthopedic Surgery, Hôpital du Sacré-Coeur de Montréal, 5400 Gouin Ouest, Montréal, Québec H4J 1C5, Canada b c
a r t i c l e
i n f o
Article history: Received 25 November 2014 Accepted 25 June 2015 Keywords: Scoliosis Biomechanics Finite element model Spinal fusion Sacrum Pelvis Spine
a b s t r a c t Background: Posterior spinal surgical correction is performed to correct spinal deformities in adolescent idiopathic scoliosis. Althoughthe relative spino-pelvic alignment changes after spinal surgery, pelvis remains unfused in idiopathic scoliosis surgery. The impact of the spinal fusion on the transferred load to the pelvis via sacrum is not documented in the scoliotic subgroups. Method: Bi-planar radiographs of 9 scoliotic subjects before and in average 16 months after spinal instrumentation surgery, and 12 controls were selected retrospectively. Patient-specific 3D reconstruction and finite element models of the spine, ribcage, and pelvis were developed. Spinal parameters (Cobb angles, kyphosis, lordosis), sacro-pelvic parameters (pelvic incidence, pelvic tilt, sacral slope), frontal and sagittal balances, the position of the trunk center of mass, and the centroid of the stress distribution on the sacrum superior endplate were measured and computed before operation and in the last follow-up. Findings: The position of the stress distribution centroid on the sacrum superior endplate with respect to the central hip vertical axis was significantly different between pre-operative and post-operative patients p b 0.05. The distance between the anterior–posterior position of the trunk center of mass and thecenter of pressure on the superior sacral endplate significantly decreased after the spinal surgery p b 0.05. Interpretation: The impact of the scoliosis spinal fusion on the transferred load between the spine and pelvis was evaluated. The biomechanical loading of the sacrum endplate was related to the post-operative postural balance and compensatory changes in the spino-pelvic alignment after scoliosis surgery. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Posterior spinal instrumentation and fusion (PSIF) aims to correct and stabilize the spinal deformity in severe cases of scoliosis until fusion occurs (Bridwell et al., 2002; Moen and Nachemson, 1999). Although the clinical impact of the spinal fusion on the geometrical parameters of the spine and pelvis has been studied previously (Masso and Gorton, 2000; Tanguay et al., 2007) the effects of the spinal surgery on the biomechanical loading of the distal un-fused vertebrae are not well documented. More specifically the biomechanical loading of the sacrum, which affects the conducted force between the spine and lower extremities and hence contributes to the standing postural equilibrium (Jiang et al., 2006) is not closely investigated in adolescent idiopathic scoliosis (AIS) subgroups post-operatively. The asymmetrical loading of the spinal vertebrae (Stokes, 2007) is shown to be associated with curve progression, which emphasizes on the importance of the ⁎ Corresponding author at: École Polytechnique Montreal, Department of Mechanical Engineering P.O. Box 6079, Station “Centre-ville”, Montréal, Québec H3C 3A7, Canada. E-mail address: [email protected] (C.-E. Aubin).
http://dx.doi.org/10.1016/j.clinbiomech.2015.06.019 0268-0033/© 2015 Elsevier Ltd. All rights reserved.
considering the vertebral loading of the unfused spine in post-surgical evaluation of the patients. Since the introduction of the “pelvic vertebra” in 1994 (Dubousset, 1994), pre- and post-operative spino-pelvic alignment in scoliosis has become the subject of many studies (Legaye et al., 1998; Pasha et al., 2010; Pasha et al., 2014a; Qiu et al., 2013; Roussouly, et al., 2013). Changes in pelvic alignment and spino-pelvic kinematic interaction were highlighted after spinal surgical correction in scoliosis (Skalli et al., 2006; Tanguay et al., 2007; Yang et al., 2015). The importance of considering the pelvic sagittal alignment with respect to the spine particularly the lumbar lordosis in AIS surgical planning was underlined (Johnson et al., 2012; Roussouly et al., 2013; Tanguay et al., 2007). However the transferred load to the pelvis through sacrum characterized by the compressive stress on the sacrum in post-surgical AIS that may impact the long-term spino-pelvic alignment is still to be investigated. Despite the body of literature examining post-surgical spino-pelvic analysis in AIS, it is not clear to what extent the spinal surgery impacts the biomechanical loading of the pelvis and lower limbs in postoperative AIS subgroups. In order to answer this question the specific objective of the current study was to compute and compare the compressive
982
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
stress on the superior sacral endplate in pre- and post-operative scoliotic subjects with different curve patterns and also compare it to a group of asymptomatic controls using a comprehensive osseo-ligamentous finite element model of the spine and pelvis (Clin et al., 2011). It was hypothesized that even though sacrum remains unfused in AIS spinal surgery the biomechanical loading of the sacrum changes significantly after PSIF in AIS subgroups and becomes more similar to the ones in controls in a balanced spino-pelvic alignment after the surgical correction of the spine. 2. Methods 2.1. Subjects The ethical approval was obtained from the ethical committee of the hospital and the affiliated research institution for this research study. A total number of 9 AIS female subjects (age range [14, 17], average 15 years, SD: 2.4, average weight 50.1 kg, SD 6.2) who had undergone a PSIF between 2006 and 2010 were randomly and retrospectively selected from the database of our institution. This sample size provided 80% statistical power for a paired analysis between the pre- and postoperative patients. All- pedicle- screw construct was used for all the patients except for one who was treated with a hybrid construct (screws and a distal hook). The number of fused vertebrae varied between 6 and 14 vetebrae. No post-operative instrumentation failure or surgical complications during an average follow-up of 16 months [12–18 months, SD: 3] was reported in the studied group. The medical chart and pre- and post-operative bi-planar radiographs of the patients were consulted. A total number of 5 patients had right main thoracic deformity (RT), Cobb angle range [43°, 77°] and 4 patients had right thoracic (RT) [55°, 68°] and left lumbar (LL) [74°, 97°] deformities. The radiographic images of 12 age- and sex-matched asymptomatic female adolescent subjects [age 11–18 years average 14.3, SD 4.0, average weight 54.8, SD 8.3] with no history of spinal disease were examined by an orthopedic surgeon and were added as the control group.
A
B
2.2. 3D reconstruction technique and anatomical measurements A self-calibration technique was used to generate the weightbearing 3D reconstructions of the spine, rib cage, and pelvis of the cohort from their bi-planar X-rays before and after surgery using the technique explained by Kadoury et al. (2007) (Fig. 1-A and Fig. 1-B). The reconstruction method consisted of identifying a limited number of points on the radiographs (14 nodes per vertebra, 11 nodes per rib, and 24 nodes on the pelvis) and using a detailed atlas of the spine and pelvis along with a free form morphing algorithm to create the detailed skeletal geometry of the spine, ribcage, and pelvis (Cheriet et al., 2002; Delorme et al., 2003; Kadoury et al., 2007). In the self-calibration process the retro-projection errors of the anatomical landmarks were minimized by changing the radiological setup. The new radiological setup was determined by changing the geometrical parameters of the radiographic system during a non-linear optimization process. The self-calibration and reconstruction methods are described in farther detail in Kadoury et al. (2007). In order to determine the accuracy of the 3D reconstruction model using the self-calibration technique all the 3D measurements were compared to a previously verified calibration and 3D reconstruction technique (Cheriet et al., 2002 and 2007). The 3D reconstructions were generated for 60 patients using both methods and were statistically compared. The average errors in the 3D reconstruction of the vertebral body, originated from the selfcalibration technique, were (1.2 mm, S.D. 0.8 mm) and vertebral pedicles (1.6 mm, S.D. 1.1 mm). The accuracy of the bi-femoral head axis alignment in the frontal plane was 0.44°, S.D. 0.46°. The maximum error in measurement of the pelvic sagittal parameters was 0.99°, S.D. 1.10° (Kadoury et al., 2007). A maximum error of 7° was reported in 2D coronal and sagittal spinal curves measurements using the 3D model when compared to the 2D clinical measurements on the X-ray images (Delorme et al., 2003). An analytical method was used to measure the spinal curvatures in the frontal and sagittal planes (Stokes, 1994) as shown in Fig. 1-B and
C
D
Fig. 1. A) Bi-planar X-rays. An analytical method was used to measure B) thoracic and lumbar Cobb angles, C) kyphosis (T4–T12), and lordosis (L1–S1) from the 3D reconstruction of the spine. The dashed line in Fig. 1-B is the spline connecting the vertebral centroids. D) Sacro-pelvic parameters in the sagittal plane: Sacral slope (SS), pelvic tilt (PT), and pelvic incidence (PI).
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
983
Fig. 1-C respectively. In this method the spinal curvature was defined by a cubic spline generated by interpolating the position of the vertebrae center. The angle between the lines perpendicular to the projection of the spinal curve in the frontal and sagittal planes at its inflection points is used to measure the spinal curvatures in the frontal and sagittal planes respectively (Fig. 1-B). The spinal curvature in the frontal plane was described by the thoracic and lumbar Cobb angles (Fig. 1-B). In the sagittal plane kyphosis included T4 to T12 vertebrae and lordosis was measured between L1 and S1 (Fig. 1-C). Sacro-pelvic parameters were characterized by pelvic incidence (PI), pelvic tilt (PT), and sacral slope (SS) (Fig. 1-D). The measured spinal and pelvic parameters are shown in Fig. 1-B–1-D. The averages of the spinal and pelvic parameters are presented in Table 1 for the patients before and after the surgery and for the asymptomatic controls. Frontal and sagittal balances were calculated as the distance between the C7 plumb line i.e. the vertical line passing through the center of the superior endplate of the C7 and S1 superior endplate center in the frontal and sagittal planes respectively. These measurements are presented in Table 2 for pre- and postoperative subjects and controls.
weight that was applied at each vertebral level in the FEM. A total number of 17 nodes were associated with the weights of each trunk slice from T1 to L5 in the FEM. The 3D position of the trunk's COM (COMtrunk) was determined as the weighted sum of the COM positions at the level of each trunk slice. The gravitational load was applied to the FEM in two steps. First the gravitational force was applied in the opposite direction of the gravity and the spinal loadings calculated at the end of this phase were set to zero. In the second phase, the weight of each trunk slice was applied in the true direction of the gravity and the geometry of the spine after the second phase of the simulation was compared to the geometry of the weight-bearing 3D reconstruction model of the spine. Transverse forces were applied at vertebral levels in the second phase of the simulation. An optimization process minimized the differences between the weight-bearing 3D reconstruction of the spine and the FEM after application of the gravitational force by changing the components of the applied forces (Clin et al., 2011). This method is detailed in Clin et al. (2011). The objective function of the optimization process was defined as:
2.3. Personalized finite element model (FEM)
X17 Fob j ¼ Min i¼1 X f −X i j þ jY f −Y i j þ jZ f −Z i
The weight-bearing 3D geometry of the spinal vertebrae, pelvis, and ribcage from the 3D reconstruction model was used to develop a personalized osseo-ligamentous FEM of the spine, pelvis, and ribcage using ANSYS 11.0 FE package (ANSYS Inc., Canonsburg, PA, USA) for the cohort of subjects (Fig. 2). This model is presented elsewhere in more details (Clin et al., 2011) and is summarized here. The FEM contains a total number of 3163 nodes and 3209 elements. The element types, Young's modulus, and Poisson's ratio of each part were derived from literature and are presented in Table 3 (Aubin et al., 1996; Clin et al., 2010; Clin et al., 2011; Perie et al., 2004). Element sizes were determined from the 3D reconstruction of the bone and were scaled for each patient. Each vertebra was modeled with a 3D beam element and has an octagonal shape cross-section (Fig. 2-C). Zygapophyseal joints were simulated by contact and shell elements. The abdominal wall was approximated by linear interpolation of the nodes of ribcage, pelvis, and vertebral bodies. Two different coordinate systems, global and local, were defined for the FEM. The origin of the global coordinate system was at the central hip vertical axis (CHVA). The Z-axis was ascending vertically, the X-axis was placed posterior-anteriorly, and the Y-axis was medio-laterally (to the left) (Fig. 2-B). The origin of the local coordinate system was fixed at the S1 center (Fig. 2-C). The X-axis is pointing anteriorly and the Y-axis was medio-laterally to the left (Fig. 2-C). The position of the center of mass (COM) at the level of each vertebra in the sagittal plane was determined from literature (Liu et al., 1971; Pearsall et al., 1996). A rigid beam was created in the FEM to connect the COM of each trunk slice to the center of the corresponding vertebra. The COM position in the frontal plane was assumed to be located at the center of the vertebrae in the frontal plane (Clin et al., 2011). The weight and COM of each trunk slice, as a percentage of the whole body weight, were found in the literature (Pearsall et al., 1996). The weight and the COM of the head, neck, and arms were associated with the COM of the T1, and T3-T5 vertebrae levels by the method described by El-Rich and Shirazi-Adl (2005). Table 4 summarizes the percentage of the body
where Xi, Yi, and Zi are the positions of the center of vertebrae from the 3D reconstruction and Xf, Yf, and Zf are the positions of the center of the vertebrae in the FEM after the gravitational forces were included. A difference less than 1% between the two consecutive iterations of the objective function was set as the convergence criterion (Clin et al., 2011). The model was fixed only in the transverse plane at the level of the first thoracic vertebrae (T1) and the 3D rotation and ascending or descending movements of the spine were allowed in the Z direction. The pelvis was fixed, both translationally and rotationally, in its actual position as it was identified from the 3D reconstructions (Clin et al., 2011). The FEM was used to calculate the distribution of the compressive stress on the S1 superior endplate (Fig. 2-C) in the local coordinate system. The compressive stress on the octagonal shape of the simulated sacrum in the FEM was calculated. Given the stress distribution at the 8 corners of the S1 endplate (Fig. 2-C) and the sacrum radius in individuals, the centroid of the stress distribution on the S1 superior endplate (CoPS1) was calculated in the cohort of subjects as it was explained in Pasha et al., 2014b. The position of the CoPS1 is shown in the local coordinate system in Fig. 2-C. The compressive stress was later normalized to the patient's weight and scaled between the minimum and maximum compressive stresses on the sacrum resulting in a normalized stress distribution between [0, 1]. The projections of the positions of the COMTrunk and CoPS1 on the transverse plane were determined in the global coordinate system with respect to the CHVA to make comparison between the two parameters possible (Pasha et al., 2014b). Each patient was described with 9 geometrical parameters (thoracic and lumbar Cobb angles, kyphosis, lordosis, PI, PT, SS, and sagittal and frontal planes balances) and two biomechanical parameters (the position of the COMTrunk in the global coordinate system and the position of the CoPS1 in the global and local coordinate systems).
Table 1 Spinal and pelvic parameters of the pre- and post-operative patients and asymptomatic controls. Subjects Pre-operative Post-operative Controls
MT(n = 5) RT/LL(n = 4) MT(n = 5) RT/LL(n = 4)
Thoracic Cobb (°)
Lumbar Cobb (°)
Kyphosis (°)
Lordosis (°)
PI (°)
PT (°)
SS (°)
53 SD 33 40 SD 25 26 SD 7 23 SD 18 –
28 SD 22 55 SD 27 15 SD 13 24 SD 20 –
33 SD 25 32 SD 12 29 SD 15 26 SD 13 47 SD 10
44 SD 17 43 SD 13 53 SD 6 50 SD 13 50 SD 7
39 SD 2 42 SD 1 51 SD 6 55 SD 7 44 SD 7
4 SD 2 11 SD 8 12 SD 3 18 SD 9 9 SD 7
34 SD 17 31 SD 5 38 SD 4 36 SD 9 35 SD 4
984
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
Table 2 Frontal and sagittal balances in pre and post-operative subjects and controls (positive measurements are to the left (in frontal plane) and anterior (in sagittal plane). Subjects Pre-operative Post-operative
MT (n = 5) RT/LL (n = 4) MT (n = 5) RT/LL (n = 4)
Controls n = 12
Frontal balance (mm)
Sagittal balance (mm)
−9 SD 17.9 31 SD 19.1 8.8 SD 11.6 15.9 SD 20.3 −1.1 SD 5.3
3 SD 23.7 −5 SD 41.0 −12.0 SD 17.5 −18.7 SD 36.7 −11 SD 6.4
Table 3 The Young modulus of model components in the FEM.
2.4. Statistical analysis All the statistical analyses were performed in PASW statistics 18.0, SPSS Inc., Chicago, IL. A normality test (Shapiro–Wilk test) was performed to determine whether the measured and calculated parameters are normally distributed. The differences between the average of the spinal and pelvic biomechanical and geometrical parameters were determined between the asymptomatic controls and AIS pre- and post-operatively. A Mann Whitney U test was used to compare the clinical parameters before and after the surgery. A two-tailed Pearson's correlation test was applied to determine the relationship between the biomechanical and geometrical parameters of the spine and pelvis in the studied groups before and after operation. Finally a clustering technique (K-mean cluster), with two clusters permitted, was used to divide the distributions of the compressive stress on the S1 in two low and high stress areas in each subject.
Model component
Element
Poisson's ratio
Young's modulus (MPa)
Vertebral body Pedicle Spinal process Sternum Posterior arc Rib Intercostal ligaments Abdominal cavity Intervertebral disks
Linear 3D elastic beam Linear 3D elastic beam Linear 3D elastic beam Linear 3D elastic beam Linear 3D elastic beam Linear 3D elastic beam Non linear spring Linear hexahedral elements Linear 3D elastic beam
0.3 0.3 0.3 0.2 0.3 0.1 – 0.45 0.45
1000 5000 35,000 10,000 1000 5000 35 0.01 3.5–15
3. Results 3.1. Case presentation Fig. 3 presents the bi-planar radiographs and the position of the high and low stress areas on the sacrum endplate in two scoliotic subjects before and after surgery. Spinal and pelvic parameters of these two patients were presented in Table 5. The sacral loading was more contra-laterally asymmetric before operation while it was more equilibrated after operation in both subjects (Fig. 3). 3.2. Comparison between the spinal and pelvic geometrical parameters preand post-operatively
2.5. Sensitivity analysis The normality test suggested that the data is not normally distributed and thus non-parametric statistical analysis was used to compare the mean values. A Mann Whitney U test showed significant decrease in the frontal plane spinal curvature i.e. thoracic and lumbar Cobb angles (p b 0.05), while the spinal parameters in the sagittal plane i.e. kyphosis and lordosis did not change significantly (p N 0.05). Both frontal and sagittal balances were significantly different between pre-operative RT/LL subjects and controls while only frontal balance was significantly different between RT and controls (p b 0.05). No significant difference was observed between post-operative and control subjects when frontal and sagittal balances were compared (p N 0.05). Thoracic Cobb angle was decreased by 51% in RT subjects and 48% in RT/LL group. Lumbar Cobb angles were decreased by 52% in RT subjects
The sensitivity of the FEM was tested for several design parameters. The modulus of elasticity of the intervertebral disk was multiplied by 0.5 and 1.5 to simulate the flexible versus rigid spine respectively (Clin et al., 2010). The effect of the spinal flexibility on the S1 loading and the position of the CoPS1 was studied. Furthermore the sensitivity of the model to the COM position was tested. The COM position in the frontal plane shifted 1 cm to the left and right sides of the vertebral centers at each vertebral level and the sacral loading was calculated for the new position of the COM to evaluate the sensitivity of the FEM to small changes of this parameter. The reaction forces and the bending moment at the T1 vertebra were calculated to determine the required counter balancing forces to stabilize the FEM.
Anterior
COPS1 Y
X ri Si
Posterior
Z X Y
A
B
C
Stressmin
Stressmax
Fig. 2. A) Posterior and lateral view of the FEM of spine, pelvis and ribcage. B) The biomechanical compressive stress of the spine in the FEM after applying the gravitational force. C) Detailed view of the stress distribution on the superior sacral endplate (Si), location of the COPS1, and the sacrum radius ri.
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987 Table 4 Trunk slices' weight as a percentage of total body weight. Vertebra level
Percentage of total body weight
T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 L1 L2 L3 L4 L5 Total
1.1% + 8% (head) 1.1% 1.3% + 4% (arms) 1.3% + 4% (arms) 1.3% + 4% (arms) 1.3% 1.4% 1.5% 1.6% 2.0% 2.1% 2.5% 2.4% 2.4% 2.3% 2.6% 2.6% 50.8%
and 58% in RT/LL group after operation. Kyphosis was decreased by 8% and 14% in RT and RT/LL groups respectively. Lordosis was increased by 20% in RT and 12% in RT/LL group. PI was increased by 15% after operation while SS was decreased by 5%. However, the Mann Whitney U test did not show any significant difference in the magnitude of the PI and the SS before and after operation (p N 0.05). PI was the only pelvic parameter, which was significantly correlated to the lumbar lordosis both before and after spinal surgical correction (r = 0.59, r = 0.64, p b 0.05 pre- and post-operatively respectively). 3.3. Comparison between the spinal and pelvic biomechanical parameters pre- and post-operatively The average positions of the CoPS1 and COMTrunk were presented in Table 6 for the studied groups. The medio-lateral parameters were
985
measured in the frontal plane and the anterior–posterior parameters were measured in the sagittal plane with respect to the CHVA (global coordinate system). The position of the CoPS1 with respect to the CHVA was significantly different pre- and post-operatively and between the pre-operative subjects and controls (p b 0.05) while no such difference was observed between the post-operative subjects and controls. The biomechanical loading of the sacrum varied between the pre- and post-operative subjects; the stress distribution on the sacrum was more symmetric i.e. no significant difference between the compressive stresses was measured on the left and right sides of the sacrum after operation (p b 0.05). The distance between the projection of the COMTrunk and the CoPS1 on the transverse plane was decreased in both medio-lateral (88% RT, 65% RT/LL) and anterior–posterior (55% RT, 57% RT/LL) directions after operation. A Mann Whitney U test showed that the distance between the projection of the COMTrunk and CoPS1 on the transverse plane decreased significantly after operation in the cohort of scoliotic subjects (p b 0.05). The anterior–posterior distance between the CoPS1 and COMTrunk in the transverse plane was significantly correlated to the sagittal balance (r = 0.47, p b 0.05) and its medio-lateral distance was significantly related to the frontal balance (r = 0.74, p b 0.05) in the cohort of the pre- and post-operative scoliotic subjects. A significant relationship was observed between the SS and the biomechanical parameters i.e. the position of the CoPS1 and the COMTrunk; as the SS increased the anterior–posterior distance between the COMTrunk and the CoPS1 decreased significantly in the postoperative subjects (r = 0.6, p b 0.05). 3.4. Sensitivity analysis The result of the sensitivity analysis did not show any significant change in the positions of the CoPS1 when the spinal flexibility varied from stiff to normal and to flexible spine (p N 0.05, Mann Whitney U test). Small changes in the position of the COMTrunk, as was explained in the method section, did not significantly change the positions of the
High stress High stress Low stress Low stress
A
High stress
High stress Low stress
B Fig. 3. Bi-planar radiographs and the location of the high and low stress areas on the superior sacral endplate before and after surgery in a A) MT (Patient1) and B) RT/LL (Patient2) (Table 6). The dash line separates the anterior and posterior parts of the superior sacral endplate.
986
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
Table 5 Pre- and post-operative spinal and pelvic parameters in a patient with a) RT deformity (patient1) and b) RT/LL curve (patient2).
a) Patient1 (RT) b) Patient2 (RT/LL)
Pre-operative Post-operative (16 months) Pre-operative Post-operative (12 months)
Thoracic Cobb (°)
Lumbar Cobb (°)
Kyphosis (°)
Lordosis (°)
PI (°)
PT (°)
SS (°)
43 24
25 8.5
44 32
57 55
50 40
6 13
44 27
68 48
97 54
19 38
61 54
71 67
40 24
31 42
CoPS1. The maximum reaction force at the level of T1 was calculated at 18 N. The maximum bending moment about the Y-axis (My) did not exceed 135 N-mm. 4. Discussion An osseo-ligamentous FEM of the spine and pelvis was used to study the differences between the pre- and post-operative biomechanical loading of the superior endplate of the sacrum in AIS. The effect of the scoliotic surgical correction of the spine on the compressive stress distribution on the S1 endplate was shown in post-operative AIS subgroups. The compressive stress distribution on the S1 superior endplate after spinal operation was more equilibrated contra-laterally. The position of the CoPS1 was related to the sagittal and frontal balance of the pre- and post-operative patients. This result was in line with our hypothesis that the biomechanical loading of the sacral endplate is more similar between controls and post-operative patients despite the fact that the sacrum and pelvis remained unfused. The biomechanical analysis in this study provided a quantitative measure to evaluate the sacral loading and consequently the transferred load between the spine and pelvis in pre- and post-operative AIS patients. Moreover we related the biomechanical loading of the sacrum to the overall postural balance of the AIS. Although the sagittal spino-pelvic alignment in AIS has been studied before, this study is the first to analyze the relationship between the sacral compressive stress and postural balances which can affect the long-term spino-pelvic alignment in AIS. A decrease in the compressive stress asymmetry on the vertebra after surgery can decrease the bone remodeling and subsequent curve progression in scoliosis. The asymmetrical compressive stress of the vertebra has been associated with disk degeneration and curve progression in scoliosis, (Adams et al., 2000) which suggests the importance of considering the vertebral mechanical loading at the unfused spinal sections below the fused spine in AIS particularly at the base of the spine i.e. sacrum to prevent post-surgical curve progression. The correlation between the sagittal and frontal balances and the anterior–posterior distance between the CoPS1 and COMTrunk projections on the transverse plane was shown. An increase in the anterior–posterior distance between the COMTrunk and CoPS1 in the pre-operative patients was accompanied by the pelvic retroversion (decreased SS) and increased distance between C7 and S1 in sagittal plane. In post-operative subjects a posterior shift in the COMTrunk position, in comparison to its pre-operative position, was accompanied by increased SS and an anterior shift in the CoPS1 position. These
changes decreased the anterior–posterior CoPS1–COMTrunk distance in the sagittal plane. The results also showed that a posterior shift in the position of the CoPS1 on the sacral superior endplate in the sagittal plane is linked to the decreased SS and increased distance between the C7 and S1 in the sagittal plane. This phenomenon was observed previously by Roussouly et al. (2013), where it was explained that pelvic retroversion should be considered as a compensatory mechanism in subjects with poor sagittal balance which helps with improving the postural equilibrium during quiet stance. The aforementioned pelvic compensatory mechanism is an attempt to bring back the spine to an upright position however it has been associated with pain, fatigue, and disability particularly in adult patients (Ames et al., 2012; Lafage et al., 2009; Roussouly et al., 2013). The results of our study are in line with previous observation in sagittal spino-pelvic alignment and pelvic compensatory mechanism and can quantitatively explain this phenomenon from a biomechanical point of view. The pelvic 3D orientation and sacro-pelvic parameters did not vary in the same way i.e. increase or decrease for all the subjects postoperatively, similar to the results in Skalli et al. (2006). The results of the current study showed that the impact of the spinal operation is more prominent on the biomechanical parameters of the pelvis i.e. CoPS1 than on its geometrical parameters e.g. SS and PI which highlights the importance of this biomechanical notion in spino-pelvic evaluation of the scoliotic subjects after surgery. This study had some limitations that should be undertaken in the future; The study was only conveyed on a limited number of subjects. A multicenter study design that evaluates a higher number of subjects with different curve patterns, different fusion levels, surgical techniques, and instrumentation strategies would permit to better investigate the association between the spino-pelvic parameters and the sacral loading post-operatively. The position of the COMTrunk and the spinal material properties were not personalized in the FEM. However the results of the clustering analysis did not change significantly as a function of spinal stiffness or small changes in the position of the COMTrunk. The changes in the muscles' moment arm with respect to spinal vertebrae before and after the surgery that can impact the in vivo vertebral loading, was not included in this FEM. Despite the limitations of the FEM, the small reaction forces at the proximal boundary level suggest that this model is stable and the reaction forces are physiologically realistic. Including the muscle forces, personalized ground reaction forces, and personalized position of the COM would allow not only studying the normalized-scaled stress distribution on the sacrum but also comparing the magnitude of the compressive stress
Table 6 The average position of the pre- and post-operative biomechanical parameters (COM, CoPS1) in the studied groups. CoPS1 (mm) With respect to the CHVA
Pre-operative Post-operative Controls (n = 12)
MT (n = 5) RT/LL (n = 4) MT (n = 5) RT/LL (n = 4)
COM (mm) With respect to the CHVA
Medio-lateral
Postero-anterior
Medio-lateral
Postero-anterior
−8 SD 2 8 SD 4 5 SD 1 4 SD 1 3 SD 4
−6 SD 1 −13 SD 5 −13 SD 2 −21 SD 5 −18 SD 7
−15 SD 11 25 SD 18 6 SD 9 12 SD 15 8 SD 5
1 SD 5 −5 SD 8 −10 SD 6 −16 SD 10 −23 SD 12
S. Pasha et al. / Clinical Biomechanics 30 (2015) 981–987
before and after operation. Another limitation of the model originates from the simulation of the abdominal components; in the FEM the geometry of the abdominal wall was approximated from the geometry of the ribs and vertebrae (Clin et al., 2011) and it was not compared to the 3D surface scans of the trunk. However it was shown that the vertebral loading in the current model when the pressure from the abdominal cavity was included in the FEM is similar to the ones in the literature (0.1–0.2 MPa) (Driscoll et al., 2009 and Stokes, 2007). Although the different components of the intervertebral disks were not simulated in the current FEM this model has proven to provide realistic information on qualitative vertebral compressive loading when it was compared to other models where the growth modulation (Stokes, 2007) and annulus-nucleus (Driscoll et al., 2009) were simulated (Clin et al., 2010, 2011). The FEM simulation was not different between the patients who were treated by the hybrid or all- screw constructs. The focus of the study was on analyzing the impact of the post-surgical spinal geometrical changes on the sacral endplate loading and other surgical parameters such as the surgical technique, implant type and density, fusion length, and the level of fusion were not considered in the analysis. Such analysis although can highlight the impact of the surgical parameters on the biomechanical loading of the sacrum, will require a larger sample size to include a variety of surgical techniques and should be the subject of another study. The biomechanical parameters of the sacral endplate provided coherent information about the biomechanics of the spino-pelvic alignment after spinal surgery and can be accompanied by the patients' geometrical parameters to better evaluate the long-term results of the spinal surgical treatment in scoliosis. This study focused on characterizing the sacral loading in post PSIF AIS patients and showed the relationship between the compensatory spino-pelvic alignment and the changes in the sacral loading. The relationship between the local sacral biomechanical loading and the overall postural balance in postsurgical AIS was highlighted. 5. Conclusion The impact of the surgical spinal correction on the biomechanical loading of the sacrum was highlighted in AIS subgroups. The notion of the CoPS1 was used to characterize the sacral mechanical loading and relate it to pre- and post-operative postural compensatory mechanisms in AIS subgroups. Acknowledgments The project was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC : PCIPJ346145-11) and Fonds de Recheche du Quebec- Nature et Technologies Recherche sur la Nature et les Technologies (FRQNT). References Adams, M.A., Freeman, B.J.C., Morrison, H.P., Nelson, I.W., Dolan, P., 2000. Mechanical initiation of intervertebral disc degeneration. Spine 25 (13), 1625–1636. Ames, C.P., Smith, J.S., Scheer, J.K., Bess, S., Bederman, S.S., Deviren, V., Lafage, V., Schwab, F., Shaffrey, C.I., 2012. Impact of spinopelvic alignment on decision making in deformity surgery in adults: a review. J. Neurosurg. Spine 16 (6), 547–564. Aubin, C.E., Dansereau, J., de Guise, J.A., Labelle, H., 1996. A study of biomechanical coupling between spine and rib cage in the treatment by orthosis of scoliosis. Ann. Chir. 50 (8), 641–650. Bridwell, K.H., Hanson, D.S., Rhee, J.M., Lenke, L.G., Baldus, C., Blanke, K., 2002. Correction of thoracic adolescent idiopathic scoliosis with segmental hooks, rods, and Wisconsin wires posteriorly: it's bad and obsolete, correct? Spine 27 (18), 2059–2066. Cheriet, F., Remaki, L., Bellefleur, C., Koller, A., Labelle, H., Dansereau, J., 2002. A new X-ray calibration/reconstruction system for 3D clinical assessment of spinal deformities. Stud. Health Technol. Inform. 91, 257–261.
987
Cheriet, F., Laporte, C., Kadoury, S., Labelle, H., Dansereau, J., 2007. A novel system for the 3-D reconstruction of the human spine and rib cage from biplanar X-ray images. IEEE Trans. Biomed. Eng. 54 (7), 1356–1358. Clin, J., Aubin, C.-E., Parent, S., Sangole, A., Labelle, H., 2010. Comparison of the biomechanical 3D efficiency of different brace designs for the treatment of scoliosis using a finite element model. Eur. Spine J. 19 (7), 1169–1178. Clin, J., Aubin, C.E., Lalonde, N., Parent, S., Labelle, H., 2011. A new method to include the gravitational forces in a finite element model of the scoliotic spine. Med. Biol. Eng. Comput. 49, 967–977. Delorme, S., Petit, Y., de Guise, J.A., Labelle, H., Aubin, C.E., Dansereau, J., 2003. Assessment of the 3D reconstruction and high-resolution geometrical modeling of the human skeletal trunk from 2D radiographic images. IEEE Trans. Biomed. Eng. 50 (8), 989–998. Driscoll, M., Aubin, C.-E., Moreau, A., Villemure, I., Parent, S., 2009. The role of spinal concave–convex biases in the progression of idiopathic scoliosis. Eur. Spine J. 18 (2), 180–187. Dubousset, J., 1994. Three-dimensional analysis of the scoliotic deformity. In: Weinstein, S.L. (Ed.), Pediatric Spine: Principles and Practice. Raven Press, New York. El-Rich, M., Shirazi-Adl, A., 2005. Effect of load position on muscle forces, internal loads and stability of the human spine in upright postures. Comput. Methods Biomech. Biomed. Engin. 8, 359–368. Jiang, Y., Nagaski, S., You, M., Zhou, J., 2006. Dynamic studies on human body sway by using a simple model with special concerns on the pelvic and muscle roles. Asian J. Control 8, 297–306. Johnson, R.D., Valore, A., Villaminar, A., Comisso, M., Balsano, M., 2012. Sagittal balance and pelvic parameters—a paradigm shift in spinal surgery. J. Clin. Neurosci. 20 (2), 191–196. Kadoury, S., Cheriet, F., Laporte, C., Labelle, H., 2007. A versatile 3D reconstruction system of the spine and pelvis for clinical assessment of spinal deformities. Med. Biol. Eng. Comput. 45, 591–602. Lafage, V., Schwab, F., Patel, A., Hawkinson, N., Farcy, J.P., 2009. Pelvic tilt and truncal inclination: two key radiographic parameters in the setting of adults with spinal deformity. Spine 34 (17), E599–E606. Legaye, J., Duval-Beaupère, G., Hecquet, J., Marty, C., 1998. Pelvic incidence: a fundamental pelvic parameter for three-dimensional regulation of spinal sagittal curves. Eur. Spine J. 7, 99–103. Liu, YK, Laborde, JM, Van Buskirk, WC, 1971. Inertial properties of a segmented cadaver trunk: their implications in acceleration injuries. Aerosp. Med. 42, 650–657. Masso, P.D., Gorton, G.E., 2000. Quantifying changes in standing body segment alignment following spinal instrumentation and fusion in idiopathic scoliosis using an optoelectronic measurement system. Spine 25, 457–462. Moen, K.Y., Nachemson, A.L., 1999. Treatment of scoliosis: an historical perspective. Spine 24, 2570–2575. Pasha, S., Sangole, A.P., Aubin, C.-E., Parent, S., Mac-Thiong, J.-M., Labelle, H., 2010. Characterizing pelvis dynamics in adolescent with idiopathic scoliosis. Spine 35, E820–E826. Pasha, S., Aubin, C.-E., Sangole, A.P., Labelle, H., Parent, S., Mac-Thiong, J.-M., 2014a. Threedimensional spinopelvic relative alignment in adolescent idiopathic scoliosis. Spine 39 (7), 564–570. Pasha, S., Aubin, C.-E., Labelle, H., Parent, S., Mac-Thiong, J.-M., 2014b. Biomechanical loading of the sacrum in adolescent idiopathic scoliosis. Clin. Biomech. 29 (3), 296–303. Pearsall, DJ, Reid, JG, Livingston, LA, 1996. Segmental inertial parameters of the human trunk as determined from computed tomography. Ann. Biomed. Eng. 24, 198–210. Perie, D., Aubin, C.E., Petit, Y., Labelle, H., Dansereau, J., 2004. Personalized biomechanical simulations of orthotic treatment in idiopathic scoliosis. Clin. Biomech. 19 (2), 190–195. Qiu, X.S., Wang, Z.W., Qiu, Y., Wang, W.J., Mao, S.H., Zhu, Z.Z., Qian, B.P., Wang, S.F., Zhu, F., Qiao, J., 2013. Preoperative pelvic axial rotation: a possible predictor for postoperative coronal decompensation in thoracolumbar/lumbar adolescent idiopathic scoliosis. Eur. Spine J. 22 (6), 1264–1272. Roussouly, P., Labelle, H., Rouissi, J., Bodin, A., 2013. Pre- and post-operative sagittal balance in idiopathic scoliosis: a comparison over the ages of two cohorts of 132 adolescents and 52 adults. Eur. Spine J. Suppl. 2, S203–S215. Skalli, W., Zeller, R.D., Miladi, L., Bourcereau, G., Savidan, M., Lavaste, F., Dubousset, J., 2006. Importance of pelvic compensation in posture and motion after posterior spinal fusion using CD instrumentation for idiopathic scoliosis. Spine 31, E359–E366. Stokes, I.A., 1994. Three-dimensional terminology of spinal deformity: a report presented to the Scoliosis Research Society by The Scoliosis Research Society Working Group on 3-D Terminology of Spinal Deformity. Spine 19 (2), 236–248. Stokes, I.A., 2007. Analysis and simulation of progressive adolescent scoliosis by biomechanical growth modulation. Eur. Spine J. 16 (10), 1621–1628. Tanguay, F., Mac-Thiong, J.-M., de Guise, J.A., Labelle, H., 2007. Relation between the sagittal pelvic and lumbar spine geometries following surgical correction of adolescent idiopathic scoliosis. Eur. Spine J. 16 (4), 531–536. Yang, X., Liu, L., Song, Y., Zhou, C., Zhou, Z., Wang, L., Wang, L., 2015. Pre-and postoperative spino-pelvic sagittal balance in adolescent patients with lenke type 5 idiopathic scoliosis. Spine 40 (2), 102–108.