Intra-Operative Analysis of Scoliosis Surgery in 3-D

Scoliosis is a three-dimensional deformity characterized by coronal, sagittal and axial rotation of the spine. Surgical fusion of the spine is required in severe cases. Assessment of the surgical procedure requires enough accuracy and flexibility to allow planning of individual interventions or implant designs. Conventional 2-D radiography and even 3-D CT scanning have limitations for in-depth analysis of scoliosis that limit the ability to see the three-dimensional deformity and expose the patient to considerable doses of radiation, respectively. Our stereophotogrammetric analysis is able to provide accurate, intra-operative measurement of vertebral movement during surgical manuevres. Stereophoto pairs taken at each stage of the operation and robust statistical techniques can be used to determine rotation, translation, goodness of fit, and overall spinal contour before, during, and after the surgical instrumentation. A demonstration of data available from this system is included.     

A simple method for in vivo measurement of implant rod three-dimensional geometry during scoliosis surgery

Scoliosis is defined as a spinal pathology characterized as a three-dimensional deformity of the spine combined with vertebral rotation. Treatment for severe scoliosis is achieved when the scoliotic spine is surgically corrected and fixed using implanted rods and screws. Several studies performed biomechanical modeling and corrective forces measurements of scoliosis correction. These studies were able to predict the clinical outcome and measured the corrective forces acting on screws, however, they were not able to measure the intraoperative three-dimensional geometry of the spinal rod. In effect, the results of biomechanical modeling might not be so realistic and the corrective forces during the surgical correction procedure were intra-operatively difficult to measure. Projective geometry has been shown to be successful in the reconstruction of a three-dimensional structure using a series of images obtained from different views. In this study, we propose a new method to measure the three-dimensional geometry of an implant rod using two cameras. The reconstruction method requires only a few parameters, the included angle θ between the two cameras, the actual length of the rod in mm, and the location of points for curve fitting. The implant rod utilized in spine surgery was used to evaluate the accuracy of the current method. The three-dimensional geometry of the rod was measured from the image obtained by a scanner and compared to the proposed method using two cameras. The mean error in the reconstruction measurements ranged from 0.32 to 0.45 mm. The method presented here demonstrated the possibility of intra-operatively measuring the three-dimensional geometry of spinal rod. The proposed method could be used in surgical procedures to better understand the biomechanics of scoliosis correction through real-time measurement of three-dimensional implant rod geometry in vivo.     

Biomechanical modeling of posterior instrumentation of the scoliotic spine

Scoliosis is a three-dimensional deformation of the spine that can be treated by vertebral fusion using surgical instrumentation. However, the optimal configuration of instrumentation remains controversial. Simulating the surgical maneuvers with personalized biomechanical models may provide an analytical tool to determine instrumentation configuration during the pre-operative planning. Finite element models used in surgical simulations display convergence difficulties as a result of discontinuities and stiffness differences between elements. A kinetic model using flexible mechanisms has been developed to address this problem, and this study presents its use in the simulation of Cotrel-Dubousset Horizon surgical maneuvers. The model of the spine is composed of rigid bodies corresponding to the thoracic and lumbar vertebrae, and flexible elements representing the intervertebral structures. The model was personalized to the geometry of three scoliotic patients (with a thoracic Cobb angle of 45 degrees, 49 degrees and 39 degrees ). Binary joints and kinematic constraints were used to represent the rod-implant-vertebra joints. The correction procedure was simulated using three steps: (1) Translation of hooks and screws on the first rod; (2) 90 degrees rod rotation; (3) Hooks and screws look-up on the rod. After the simulation, slight differences of 0-6 degrees were found for the thoracic spine scoliosis and the kyphosis, and of 1-8 degrees for the axial rotation of the apical vertebra and for the orientation of the plane of maximum deformity, compared to the real post-operative shape of the patient. Reaction loads at the vertebra-implant link were mostly below 1000 N, while reaction loads at the boundary conditions (representing the overall action of the surgeon) were in the range 7-470 N and maximum torque applied to the rod was 1.8 Nm. This kinetic modeling approach using flexible mechanisms provided a realistic representation of the surgical maneuvers. It may offer a tool to predict spinal geometry correction and assist in the pre-operative planning of surgical instrumentation of the scoliotic spine.     

Characterization of the age-dependent shape of the pediatric thoracic spine and vertebrae using generalized procrustes analysis

Generalized Procrustes Analysis (GPA) is a superimposition method used to generate size-invariant distributions of homologous landmark points. Several studies have used GPA to assess the three-dimensional (3D) shapes of or to evaluate sex-related differences in the human brain, skull, rib cage, pelvis and lower limbs. Previous studies of the pediatric thoracic vertebrae suggest that they may undergo changes in shape as a result of normative growth. This study uses GPA and second order polynomial equations to model growth and age- and sex-related changes in shape of the pediatric thoracic spine. We present a thorough analysis of the normative 3D shape, size, and orientation of the pediatric thoracic spine and vertebrae as well as equations which can be used to generate models of the thoracic spine and vertebrae for any age between 1 and 19 years. Such models could be used to create more accurate 3D reconstructions of the thoracic spine, generate improved age-specific geometries for finite element models (FEMs) and used to assist clinicians with patient-specific planning and surgical interventions for spine deformity.     

Analysis of the interaction between vertebral lateral deviation and axial rotation in scoliosis

There is a lack of clear biomechanical analyses to explain the interaction of the lateral and axial deformity of the spine in idiopathic scoliosis. A finite element model which represented an isolated ligamentous spine with realistic elastic properties and idealized geometry was used to analyse this interaction. Three variations of this model were used to investigate two different hypotheses about the etiology of scoliosis and to define the forces required to produce a scoliosis deformity. The first hypothesis is that coupling within a motion segment produces the interaction between lateral deviation and axial rotation. The second hypothesis is that posterior tethering by soft tissues in the growing spine produces the observed interaction. Modeling of both hypotheses failed to produce the clinically observed pattern of interaction. Therefore, to find which biomechanical forces were required to produce an idealized scoliosis, prescribed displacements were applied to the model. Production of a double curve scoliosis of 10 degrees Cobb angles required lateral forces on the order of 20 N acting 40 mm anterior to the vertebral body centers. There do not appear to be any anatomic structures capable of producing such forces. Therefore, it seems unlikely that scoliosis deformity can be explained in terms of forces acting on the spine, and understanding of its origins may come from examination of other mechanisms such as asymmetric thoracic growth, or asymmetric vertebral development.     

Comparison of erector spinae and hamstring muscle activities and lumbar motion during standing knee flexion in subjects with and without lumbar extension rotation syndrome

The aim of this study was to compare the activity of the erector spinae (ES) and hamstring muscles and the amount and onset of lumbar motion during standing knee flexion between individuals with and without lumbar extension rotation syndrome. Sixteen subjects with lumbar extension rotation syndrome (10 males, 6 females) and 14 healthy subjects (8 males, 6 females) participated in this study. During the standing knee flexion, surface electromyography (EMG) was used to measure muscle activity, and surface EMG electrodes were attached to both the ES and hamstring (medial and lateral) muscles. A three-dimensional motion analysis system was used to measure kinematic data of the lumbar spine. An independent-t test was conducted for the statistical analysis. The group suffering from lumbar extension rotation syndrome exhibited asymmetric muscle activation of the ES and decreased hamstring activity. Additionally, the group with lumbar extension rotation syndrome showed greater and earlier lumbar extension and rotation during standing knee flexion compared to the control group. These data suggest that asymmetric ES muscle activation and a greater amount of and earlier lumbar motion in the sagittal and transverse plane during standing knee flexion may be an important factor contributing to low back pain.     

Validation of the ‘FeMorph’ software in planning cam osteochondroplasty by incorporating labral morphology

Impingement resulting from a cam deformity may cause pain, limit the hip joint range of motion (RoM) and lead to osteoarthritis. We have previously developed FeMorph software to quantify and plan corrective surgery and predict hip RoM post surgery. This study aimed to validate the software and evaluate the influence of the acetabular labrum on hip RoM. Computed tomography data from 92 femur-pelvis pairs were analysed in conjunction with the inter/intra-observer reliability. Four cadaveric hips were dissected, and the three-dimensional (3D) shape and size of the acetabular labrum for these hips was obtained using laser scan. The influence of the acetabular labrum in the RoM and subsequent planning for corrective surgery were then evaluated in cadavers for models with and without a labrum, and used as a first step towards validation of FeMorph RoM prediction. FeMorph was successfully used to model cam deformities and plan corrective surgery. Three-dimensional alpha angles were reduced to below 50° after virtual surgery without an excessive reduction in femoral neck cross-sectional area, which could increase fracture risk. A mean increase of 8° ± 2° in permitted internal rotation was observed during impingement testing following removal of the labrum. FeMorph provides a reliable and useful method to model and plan cam deformity correction. This study indicates that the presence of the labrum is responsible for a substantial decrease in permitted internal rotation at the hip joint. This has implications for surgical planning models which often only account for bony impingement.     

Three-dimensional-segmentation of the spine with specially adapted tools]

Segmented 3-D data of the spine form the basis for various modern clinical applications. Among these, multisegmental image fusion, image registration and finite element modeling for biomechanical analysis are promising innovative tools capable of facilitating treatment decisions and optimization of individual therapy in the future. However, the complex anatomy of the spine and the often extensive degenerative deformation presenting in clinical practice, generally limit the application of fully automated segmentation. A newly developed software system is presented that meets the complex requirements for image segmentation of the spine through the use of specially adapted interactive tools that take account of its axial skeletal structure. Furthermore, a standardized protocol is introduced that combines the newly developed interactive tools (rotation transformation, warped dissection plane) and standard segmentation tools to permit rapid and accurate segmentation. To date, the software environment presented herein has been applied with success to the segmentation of cervical, thoracic and lumbar spine.     

Automated measurement of neural foramen cross-sectional area during in vivo functional movement

An automated technique to measure neural foramen cross-sectional area during in vivo, multi-planar movements is presented. This method combines three-dimensional (3D) models of each vertebra obtained from CT scans with in vivo movement data collected using high-speed biplane radiography. A novel computer algorithm that automatically traces a path around the bony boundary that defines the neural foramen at every frame of X-ray data is described. After identifying the neural foramen boundary, the cross-sectional area is calculated. The technique is demonstrated using data collected from a patient with cervical radiculopathy who is tested before and after conservative treatment. The technique presented here can be applied when 3D, dynamic, functional movements are performed. Neural foramen cross-sectional area can be quantified at specific angles of intervertebral rotation, allowing for matched comparisons between two trials or two test sessions. The present technique is ideal for longitudinal studies involving subjects who receive conservative or surgical treatments that may affect spine motion.     

Are coupled rotations in the lumbar spine largely due to the osseo-ligamentous anatomy?--a modeling study

Prior studies have found that primary rotations in the lumbar spine are accompanied by coupled out-of-plane rotations. However, it is not clear whether these accompanying rotations are primarily due to passive (discs, ligaments and facet joints) or active (muscles) spinal anatomy. The aim of this study was to use a finite element (FE) model of the lumbar spine to predict three-dimensional coupled rotations between the lumbar vertebrae, due to passive spinal structures alone. The FE model was subjected to physiologically observed whole lumbar spine rotations about in vivo centres of rotation. Model predictions were validated by comparison of intra-discal pressures and primary rotations with in vivo measurements and these showed close agreement. Predicted coupled rotations matched in vivo measurements for all primary motions except lateral bending. We suggest that coupled rotations accompanying primary motions in the sagittal (flexion/extension) and transverse (axial rotation) planes are primarily due to passive spinal structures. For lateral bending the muscles most likely play a key role in the coupled rotation of the spine.     

Adaptation of a clinical fixation device for biomechanical testing of the lumbar spine

In-vitro biomechanical testing is widely performed for characterizing the load-displacement characteristics of intact, injured, degenerated, and surgically repaired osteoligamentous spine specimens. Traditional specimen fixture devices offer an unspecified rigidity of fixation, while varying in the associated amounts and reversibility of damage to and “coverage” of a specimen – factors that can limit surgical access to structures of interest during testing as well as preclude the possibility of testing certain segments of a specimen. Therefore, the objective of this study was to develop a specimen fixture system for spine biomechanical testing that uses components of clinically available spinal fixation hardware and determine whether the new system provides sufficient rigidity for spine biomechanical testing. Custom testing blocks were mounted into a robotic testing system and the angular deflection of the upper fixture was measured indirectly using linear variable differential transformers. The fixture system had an overall stiffness 37.0, 16.7 and 13.3 times greater than a typical human functional spine unit for the flexion/extension, axial rotation and lateral bending directions respectively – sufficient rigidity for biomechanical testing. Fixture motion when mounted to a lumbar spine specimen revealed average motion of 0.6, 0.6, and 1.5° in each direction. This specimen fixture method causes only minimal damage to a specimen, permits testing of all levels of a specimen, and provides for surgical access during testing.     

Biomechanical design of load simulation in multiple spinal segments]

Multisegmental biomechanical studies on the lumbar spine are steadily increasing in importance. Only in this way can we acquire knowledge about the physiological behaviour of the entire lumbar spine. Furthermore, these studies allow us to analyse in vitro the biomechanics of manipulated lumbar spines after various surgical operations on the spine. A load simulator was developed to investigate multisegmental lumbar spine mobility, and its function was investigated in an initial study on 19 fresh--frozen specimens of human lumbar spine. After x-ray examination and determination of the bone mineral density, the specimens were loaded up to 10 Nm in the automatic electromechanical loading system under flexion/extension, lateral bending and axial rotation. An ultrasound-based motion analysis system was used to measure the displacements of the vertebrae involved.     

Automated measurement of neural foramen cross-sectional area during in vivo functional movement

An automated technique to measure neural foramen cross-sectional area during in vivo, multi-planar movements is presented. This method combines three-dimensional (3D) models of each vertebra obtained from CT scans with in vivo movement data collected using high-speed biplane radiography. A novel computer algorithm that automatically traces a path around the bony boundary that defines the neural foramen at every frame of X-ray data is described. After identifying the neural foramen boundary, the cross-sectional area is calculated. The technique is demonstrated using data collected from a patient with cervical radiculopathy who is tested before and after conservative treatment. The technique presented here can be applied when 3D, dynamic, functional movements are performed. Neural foramen cross-sectional area can be quantified at specific angles of intervertebral rotation, allowing for matched comparisons between two trials or two test sessions. The present technique is ideal for longitudinal studies involving subjects who receive conservative or surgical treatments that may affect spine motion.     

A demonstration of validity of 3-D video motion analysis method for measuring finger flexion and extension

This study demonstrates the validity of using 3-D video motion analysis to measure hand motion. Several researchers have devised ingenious methods to study normal and abnormal hand movements. Although very helpful, these earlier studies are static representations of a dynamic phenomenon. Despite the many studies of hand motion using scientifically impeccable techniques, little is known about digital motion, and there are still few researchers investigating dynamic three-dimensional motion of the hand. Results from a three-camera video motion analysis system were compared to those from the "gold standard", 2-D lateral view fluoroscopy. We used these two methods to record hand motion simultaneously during unrestricted flexion and extension of the index finger of the dominant hand in 6 neurologically normal, healthy volunteers. After collection and post-processing, the waveforms of the PIP, DIP and MCP joint angles were compared using the adjusted coefficient of multiple determination (R2(a), or CMD). The mean CMD values for the MCP, PIP and DIP joint angle waveforms were 0.96, 0.98 and 0.94, respectively, suggesting a close similarity between motion of comparable joints analyzed by the 2-D and 3-D methods. This shows that the method of 3-D motion analysis is capable of accurately quantifying digital joint motion. It is anticipated that 3-D motion analysis, in addition to being used as a research tool, will also have clinical applications such as surgical planning in neuromuscular disorders and the documentation of abnormal motion in many other pathological hand conditions.     

A novel method for the three-dimensional (3-D) analysis of orthodontic tooth movement-calculation of rotation about and translation along the finite helical axis.

The purpose of this study was to establish a novel method for evaluating orthodontic tooth movement in three-dimensional (3-D) space. The present system consisted of the following procedures at a given treatment period: (1) 3-D tooth positions were measured with a 3-D surface-scanning system using a slit laser beam; (2) the 3-D shape data were registered automatically at the maxillary first molars, and the coordinate systems were normalized; (3) the rotation matrix and translation vector were calculated from the automatic registration of the two position data for a given tooth; (4) the finite helical axes of teeth were calculated as the locus of zero rotational displacement; and (5) tooth movement was presented as rotation about and translation along the finite helical axis. To test this system, a male patient (age 22 yr 2 months) with Angle Class III malocclusion and moderate crowding of the anterior teeth, who had been treated using a standard multi-bracket appliance, was used as a model case in this study. Impressions for a dental cast model were taken at five phases; immediately before and after application of the appliance, and 10 days, 1 month and 2 months after beginning treatment. The results demonstrated that the present analytical method can more simply describe the movement of a given tooth by rotation about and translation along the finite helical axis, and provides quantitative visual 3-D information on complicated tooth movement during orthodontic treatment.     

Quantification of the spatial strain distribution of scoliosis using a thin-plate spline method

The objective of this study was to quantify the three-dimensional spatial strain distribution of a scoliotic spine by nonhomogeneous transformation without using a statistically averaged reference spine. The shape of the scoliotic spine was determined from computed tomography images from a female patient with adolescent idiopathic scoliosis. The shape of the scoliotic spine was enclosed in a rectangular grid, and symmetrized using a thin-plate spline method according to the node positions of the grid. The node positions of the grid were determined by numerical optimization to satisfy symmetry. The obtained symmetric spinal shape was enclosed within a new rectangular grid and distorted back to the original scoliotic shape using a thin-plate spline method. The distorted grid was compared to the rectangular grid that surrounded the symmetrical spine. Cobb's angle was reduced from 35° in the scoliotic spine to 7° in the symmetrized spine, and the scoliotic shape was almost fully symmetrized. The scoliotic spine showed a complex Green–Lagrange strain distribution in three dimensions. The vertical and transverse compressive/tensile strains in the frontal plane were consistent with the major scoliotic deformation. The compressive, tensile and shear strains on the convex side of the apical vertebra were opposite to those on the concave side. These results indicate that the proposed method can be used to quantify the three-dimensional spatial strain distribution of a scoliotic spine, and may be useful in quantifying the deformity of scoliosis.     

Prediction of biomechanical parameters in the lumbar spine during static sagittal plane lifting.

A combined approach involving optimization and the finite element technique was used to predict biomechanical parameters in the lumbar spine during static lifting in the sagittal plane. Forces in muscle fascicles of the lumbar region were first predicted using an optimization-based force model including the entire lumbar spine. These muscle forces as well as the distributed upper body weight and the lifted load were then applied to a three-dimensional finite element model of the thoracolumbar spine and rib cage to predict deformation, the intradiskal pressure, strains, stresses, and load transfer paths in the spine. The predicted intradiskal pressures in the L3-4 disk at the most deviated from the in vivo measurements by 8.2 percent for the four lifting cases analyzed. The lumbosacral joint flexed, while the other lumbar joints extended for all of the four lifting cases studied (rotation of a joint is the relative rotation between its two vertebral bodies). High stresses were predicted in the posterolateral regions of the endplates and at the junctions of the pedicles and vertebral bodies. High interlaminar shear stresses were found in the posterolateral regions of the lumbar disks. While the facet joints of the upper two lumbar segments did not transmit any load, the facet joints of the lower two lumbar segments experienced significant loads. The ligaments of all lumbar motion segments except the lumbosacral junction provided only marginal moments. The limitations of the current model and possible improvements are discussed.     

A demonstration of the validity of a 3-D video motion analysis method for measuring finger flexion and extension

This study demonstrates the validity of using 3-D video motion analysis to measure hand motion. Several researchers have devised ingenious methods to study normal and abnormal hand movements. Although very helpful, these earlier studies are static representations of a dynamic phenomenon. Despite the many studies of hand motion using scientifically impeccable techniques, little is known about digital motion, and there are still few researchers investigating dynamic three-dimensional motion of the hand. Results from a three-camera video motion analysis system were compared to those from the “gold standard”, 2-D lateral view fluoroscopy. We used these two methods to record hand motion simultaneously during unrestricted flexion and extension of the index finger of the dominant hand in 6 neurologically normal, healthy volunteers. After collection and post-processing, the waveforms of the PIP, DIP and MCP joint angles were compared using the adjusted coefficient of multiple determination (R²_a, or CMD). The mean CMD values for the MCP, PIP and DIP joint angle waveforms were 0.96, 0.98 and 0.94, respectively, suggesting a close similarity between motion of comparable joints analyzed by the 2-D and 3-D methods. This shows that the method of 3-D motion analysis is capable of accurately quantifying digital joint motion.

It is anticipated that 3-D motion analysis, in addition to being used as a research tool, will also have clinical applications such as surgical planning in neuromuscular disorders and the documentation of abnormal motion in many other pathological hand conditions.     

Spinal shape changes resulting from scoliotic spine surgical instrumentation expressed as intervertebral rotations and centers of rotation

This paper reports the changes in spinal shape resulting from scoliotic spine surgical instrumentation expressed as intervertebral rotations and centers of rotation. The objective is to test the hypothesis that the type of spinal instrumentation system (Cotrel-Dubousset versus Colorado) does not influence these motion parameters. Intervertebral rotations and centers of rotation of the scoliotic spines were computed from the pre- and post-operative radiographs of 82 patients undergoing spinal correction. The three-dimensional (3D) reconstruction of six anatomical landmarks was achieved for each of the thoracic and lumbar vertebrae. A least-squares approach based on singular value decomposition was used to calculate the rigid body transformation parameters. Average centers of rotation for all intervertebral levels are located in the neural canal at the mid-sagittal plane and approximately at the superior endplate level of the inferior vertebra. Intervertebral rotations have components in all planes: 6.7 degrees (frontal), 5.5 degrees (sagittal) and 4.5 degrees (transverse) RMS for all intervertebral levels. Nearly all intervertebral rotations and centers of rotation are not significantly different for the two instrumentation systems. Various intervertebral rotations and 3D reconstruction errors were simulated on a theoretical model of a lumbar functional unit to assess the proposed method. Intervertebral rotation errors were 1.7 degrees when simulating 3D errors of 3mm on the position of the landmarks. Maximum errors for the position of centers of rotation were below 1cm in the case of intervertebral rotations larger than 2.5 degrees (most cases), but were larger (38 mm) for small intervertebral rotations (<1 degrees ). The type of instrumentation system did not influence intervertebral rotations and centers of rotation. These results provide valuable data for the development and validation of simulation models for surgical instrumentation of idiopathic scoliosis.     

Extent of Spine Deformity Predicts Lung Growth and Function in Rabbit Model of Early Onset Scoliosis

Early onset deformity of the spine and chest wall (initiated <8 years of age) is associated with increased morbidity at adulthood relative to adolescent onset deformity of comparable severity. Presumably, inhibition of thoracic growth during late stage alveolarization leads to an irreversible loss of pulmonary growth and thoracic function; however the natural history of this disease from onset to adulthood has not been well characterized. In this study we establish a rabbit model of early onset scoliosis to establish the extent that thoracic deformity affects structural and functional respiratory development. Using a surgical right unilateral rib-tethering procedure, rib fusion with early onset scoliosis was induced in 10 young New Zealand white rabbits (3 weeks old). Progression of spine deformity, functional residual capacity, total lung capacity, and lung mass was tracked through longitudinal breath-hold computed tomography imaging up to skeletal maturity (28 weeks old). Additionally at maturity forced vital capacity and regional specific volume were calculated as functional measurements and histo-morphometry performed with the radial alveolar count as a measure of acinar complexity. Data from tethered rib rabbits were compared to age matched healthy control rabbits (N = 8). Results show unilateral rib-tethering created a progressive spinal deformity ranging from 30° to 120° curvature, the severity of which was strongly associated with pulmonary growth and functional outcomes. At maturity rabbits with deformity greater than the median (55°) had decreased body weight (89%), right (59%) and left (86%) lung mass, right (74%) and left (69%) radial alveolar count, right lung volume at total lung capacity (60%), and forced vital capacity (75%). Early treatment of spinal deformity in children may prevent pulmonary complications in adulthood and these results provide a basis for the prediction of pulmonary development from thoracic structure. This model may also have future use as a platform to evaluate treatment effectiveness.