Automatic extraction of vertebral landmarks from ultrasound images: A pilot study.

Interpreting ultrasound (US) images of the spine is challenging due to the high variability of the contrast during freehand US acquisitions. In this paper, an automatic method to extract vertebral landmarks (spinous process and laminae) from US images acquired in the transverse plane is presented. Prior knowledge about the vertebral shape and the associated hyper-echoic property is incorporated using the horizontal and vertical projections of the image intensities. After detrending, the mean-value crossing of the projections is used to define the concept of mean boundary and locate landmarks without the need for thresholding or parameter adjustment. The method was evaluated using two datasets: a porcine cadaver dataset (PC) with CT data registered to the US data used as a gold standard, and a healthy human subjects dataset (HH) with a silver standard generated from manual landmarks located on the US data acquired with a curvilinear (6C2) and linear (14L5) probe. The mean sum of distances (MSD) of the landmark extraction to the gold and silver standards is respectively MSD=0.90±1.05 mm for PC, MSD=1.14±1.08 mm (6C2) and MSD=3.54±2.69 mm (14L5) for HH. Results are satisfying on PC and HH with 6C2. Variable contrast quality for 14L5 gives satisfying results for the spinous process but not for the laminae. The proposed approach has the potential to be used for different applications in the context of US spine imaging such as scoliosis follow-up and intra-operative surgical guidance.

Frequency-dependent shear properties of annulus fibrosus and nucleus pulposus by magnetic resonance elastography.

Aging and degeneration are associated with changes in mechanical properties in the intervertebral disc, generating interest in the establishment of mechanical properties as early biomarkers for the degenerative cascade. Magnetic resonance elastography (MRE) of the intervertebral disc is usually limited to the nucleus pulposus, as the annulus fibrosus is stiffer and less hydrated. The objective of this work was to adapt high-frequency needle MRE to the characterization of the shear modulus of both the nucleus pulposus and annulus fibrosus. Bovine intervertebral discs were removed from fresh oxtails and characterized by needle MRE. The needle was inserted in the center of the disc and vibrations were generated by an amplified piezoelectric actuator. MRE acquisitions were performed on a 4.7-T small-animal MR scanner using a spin echo sequence with sinusoidal motion encoding gradients. Acquisitions were repeated over a frequency range of 1000-1800 Hz. The local frequency estimation inversion algorithm was used to compute the shear modulus. Stiffness maps allowed the visualization of the soft nucleus pulposus surrounded by the stiffer annulus fibrosus surrounded by the homogeneous gel. A significant difference in shear modulus between the nucleus pulposus and annulus fibrosus, and an increase in the shear modulus with excitation frequency, were observed, in agreement with the literature. This study demonstrates that global characterization of both the nucleus pulposus and annulus fibrosus of the intervertebral disc is possible with needle MRE using a preclinical magnetic resonance imaging (MRI) scanner. MRE can be a powerful method for the mapping of the complex properties of the intervertebral disc. The developed method could be adapted for in situ use by preserving adjacent vertebrae and puncturing the side of the intervertebral disc, thereby allowing an assessment of the contribution of osmotic pressure to the mechanical behavior of the intervertebral disc.

A physically based trunk soft tissue modeling for scoliosis surgery planning systems.

One of the major concerns of scoliotic patients undergoing spinal correction surgery is the trunk's external appearance after the surgery. This paper presents a novel incremental approach for simulating postoperative trunk shape in scoliosis surgery. Preoperative and postoperative trunk shapes data were obtained using three-dimensional medical imaging techniques for seven patients with adolescent idiopathic scoliosis. Results of qualitative and quantitative evaluations, based on the comparison of the simulated and actual postoperative trunk surfaces, showed an adequate accuracy of the method. Our approach provides a candidate simulation tool to be used in a clinical environment for the surgery planning process.

Statistical model based 3D shape prediction of postoperative trunks for non-invasive scoliosis surgery planning.

One of the major concerns of scoliosis patients undergoing surgical treatment is the aesthetic aspect of the surgery outcome. It would be useful to predict the postoperative appearance of the patient trunk in the course of a surgery planning process in order to take into account the expectations of the patient. In this paper, we propose to use least squares support vector regression for the prediction of the postoperative trunk 3D shape after spine surgery for adolescent idiopathic scoliosis. Five dimensionality reduction techniques used in conjunction with the support vector machine are compared. The methods are evaluated in terms of their accuracy, based on the leave-one-out cross-validation performed on a database of 141 cases. The results indicate that the 3D shape predictions using a dimensionality reduction obtained by simultaneous decomposition of the predictors and response variables have the best accuracy.

Accurate positioning of magnetic microparticles beyond the spatial resolution of clinical MRI scanners using susceptibility artifacts.

Susceptibility-based negative contrast in magnetic resonance imaging (MRI) provides a mean to visualize magnetic microparticles. In the presence of a number of microparticles in the field of view (FOV), the shape of the artifact is affected by the dipole-dipole interaction between the particles. Due to the limited spatial resolution of the clinical MR scanners, the exact positioning of the particles in MR images is not possible. However, the shape of the artifact can shed light on how the particles are distributed within the FOV. In this work, a simulation model and in-vitro experiments were used to study the shape and the amount of the susceptibility artifact for various spacing and angulations between the microparticles. The results showed that for a pair of identical particles with a diameter of D, the signal loss starts to change when particles are separated ~15 × D and they become fully distinguishable when their distance reaches ~ 40 × D.

MRI visualization of a single 15 µm navigable imaging agent and future microrobot.

In magnetic resonance imaging (MRI), the susceptibility-based contrast provides a way to amplify the effects of a magnetic microparticle, whereas its volume is largely inferior to the spatial resolution of the system. This concept presents an approach to visualization by means of susceptibility artifact using ferromagnetic microparticles. In this work, the amount of the susceptibility artifact was investigated using a simulation model and in vitro experiments on stainless steel microspheres measuring 40, 20 and 15 microm in diameter. The results showed that using a clinical MRI system, a single 15 microm microsphere is detectable in gradient-echo scans. The extent of the susceptibility artifact was found to be related to the scan parameters and the particles' sizes. Since the same ferromagnetic microparticle can be used for MRI-based propulsion, these results suggest several potential applications for navigable agents and microrobots involved in therapy, diagnostics, and imaging inside the microvascular network of the human body.

Prediction of the T2-T12 kyphosis in adolescent idiopathic scoliosis using a multivariate regression model.

The paper presents a nonlinear regression model built on the coronal thoracic curvature, the lumbar lordosis and the slope of the first lumbar vertebra in order to estimate the thoracic kyphosis measure between T2 and T12. To train the proposed model, a large database containing scoliotic spines demonstrating several types of scoliotic deformities was used to train the proposed system by a cross-validation method. Validation was performed on patients exhibiting three different types of sagittal thoracic profiles: normal, hypo-kyphotic, and hyper-kyphotic. Results show that a multivariate regression model based on dependent variables is able to predict with a reasonable accuracy the sagittal thoracic kyphosis for the automatic assessment and classification of the spinal curve.

Geometric variability of the scoliotic spine using statistics on articulated shape models.

This paper introduces a method to analyze the variability of the spine shape and of the spine shape deformations using articulated shape models. The spine shape was expressed as a vector of relative poses between local coordinate systems of neighboring vertebrae. Spine shape deformations were then modeled by a vector of rigid transformations that transforms one spine shape into another. Because rigid transformations do not naturally belong to a vector space, conventional mean and covariance could not be applied. The Fréchet mean and a generalized covariance were used instead. The spine shapes of a group of 295 scoliotic patients were quantitatively analyzed as well as the spine shape deformations associated with the Cotrel-Dubousset corrective surgery (33 patients), the Boston brace (39 patients), and the scoliosis progression without treatment (26 patients). The variability of intervertebral poses was found to be inhomogeneous (lumbar vertebrae were more variable than the thoracic ones) and anisotropic (with maximal rotational variability around the coronal axis and maximal translational variability along the axial direction). Finally, brace and surgery were found to have a significant effect on the Fréchet mean and on the generalized covariance in specific spine regions where treatments modified the spine shape.

A 3D visualization tool for the design and customization of spinal braces.

A new tool was developed and validated on an X-ray dummy to allow personalized design and adjustment of spinal braces. The 3D visualization of the external trunk surface registered with the underlying 3D bone structures permits the clinicians to select pressure areas on the trunk surface for proper positioning of correcting pads inside the brace according to the patient's specific trunk deformities. After brace fabrication, the clinicians can evaluate the actual 3D patient-brace interface pressure distribution visualized simultaneously with the 3D model of the trunk in order to customize brace adjustment and validate brace design with respect to the treatment plan.

A novel system for thE 3-D reconstruction of the human spine and rib cage from biplanar X-ray images.

The main objective of this study was to develop a 3-D X-ray reconstruction system of the spine and rib cage for an accurate 3-D clinical assessment of spinal deformities. The system currently used at Sainte-Justine Hospital in Montreal is based on an implicit calibration technique based on a direct linear transform (DLT), using a sufficiently large rigid object incorporated in the positioning apparatus to locate any anatomical structure to be reconstructed within its bounds. During the time lapse between the two successive X-ray acquisitions required for the 3-D reconstruction, involuntary patient motion introduce errors due to the incorrect epipolar geometry inferred from the stationary object. An approach using a new calibration jacket and explicit calibration algorithm is proposed in this paper. This approach yields accurate results and compensates for involuntary motion occurring between X-ray exposures.

Semi-automatic detection of scoliotic rib borders using chest radiographs.

Stereoradiography is a well known technique to obtain 3D reconstructions of the rib cage. However, clinical applications are limited by the associated 2D rib detection method. Either this detection is widely supervised and time-consuming for the user, or it is fully automatic and not accurate enough for proper 3D reconstruction or clinical indices extraction. To address these issues, we propose a novel, semi-automated technique for detecting scoliotic rib borders in PA-0 degrees and PA-20 degrees chest X-ray images, using a modified edge-following approach. The novelty consists in following multiple promising edges simultaneously. Detections are initiated from starting points (input by the user) along the upper and lower rib edges and the final rib border is obtained by finding the most parallel pair among the detected edges. Promising results show the superiority of this approach over classical rib detection in terms of accuracy. Moreover, the proposed method is of great relevancy in the scoliotic context since scoliotic ribs present very few shape priors, due to their irregularities, and hence, standard rib detection techniques become unsuitable.

Accuracy assessment of human trunk surface 3D reconstructions from an optical digitising system.

The lack of reliable techniques to follow up scoliotic deformity from the external asymmetry of the trunk leads to a general use of X-rays and indices of spinal deformity. Young adolescents with idiopathic scoliosis need intensive follow-ups for many years and, consequently, they are repeatedly exposed to ionising radiation, which is hazardous to their long-term health. Furthermore, treatments attempt to improve both spinal and surface deformities, but internal indices do not describe the external asymmetry. The purpose of this study was to assess a commercial, optical 3D digitising system for the 3D reconstruction of the entire trunk for clinical assessment of external asymmetry. The resulting surface is a textured, high-density polygonal mesh. The accuracy assessment was based on repeated reconstructions of a manikin with markers fixed on it. The average normal distance between the reconstructed surfaces and the reference data (markers measured with CMM) was 1.1 +/- 0.9 mm.

A new X-ray calibration/reconstruction system for 3D clinical assessment of spinal deformities.

The main objective of this study was to develop a 3D X-ray reconstruction system of the spine and rib cage for an accurate clinical assessment of spinal deformities. The proposed system uses an explicit calibration technique and a new calibration object composed of: (1) a set of radiopaque markers embedded in a jacket worn by the patient during the X-ray exposures; (2) six control markers to define a reference vertical plane. Computer simulations were performed to evaluate the accuracy of the 3D reconstruction procedure when different kind of displacements were applied on a reference model. Clinical indices computed from the 3D X-ray reconstruction of the spine for 24 scoliotic subjects were compared to those obtained with the DLT method. The results of the evaluation study showed that the new system allows the patient to adopt a normal attitude without any constraint, compensating for its displacement between exposures.

Self-calibration of biplanar radiographs for a retrospective comparative study of the 3D correction of adolescent idiopathic scoliosis.

A novel technique for the 3D reconstruction of the spine from X-ray images is presented. The algorithm is based on the self-calibration of biplanar radiographs. It allows the 3D reconstruction of spines from old uncalibrated preoperative and postoperative radiographs. The reliability of the new self-calibration technique was investigated by validating its results against those of the Direct Linear Transform (DLT) on real images. An accuracy experiment was also performed using a dry spine specimen under controlled conditions. The results indicate that self-calibration is a viable technique, accurate enough to extract meaningful 3D clinical data for retrospective studies.

Investigation of muscle recruitment patterns in scoliosis using a biomechanical finite element model.

The objective of this project is to study the characteristics of trunk muscle recruitment strategies experimentally observed for scoliotic subjects using a finite element model of the trunk. The personalized biomechanical model includes elements representing the osseo-ligamentous structures of the spine, rib cage and pelvis. It also integrates the principal agonistic muscles necessary for trunk movement and a neural control model based on the Equilibrium Point hypothesis (lambda model of Feldman). Muscle recruitment patterns of normal and scoliotic subjects obtained from the simulation of lateral bending movements were qualitatively compared. The generation process of motor control variables was studied by analysing the relationships between central commands and spine segment mobility. Differences in recruitment patterns between normal and scoliotic subjects were observed, especially for paraspinal fascicles crossing the thoracic curve segment. The generation of central commands for normal subjects was strongly correlated with the amplitude of bending, but this relation was weaker for scoliotic subjects and this difference was worst at the apex vertebra. These results show that neuromuscular disorders could occur at a local level. The proposed approach should provide a simulation tool to study the multifactorial origin of scoliosis, and to investigate the implication of muscles and central commands in spinal dysfunctions.

Biomechanical simulation of Colorado instrumentation of the scoliotic spine: a preliminary study.

Few biomechanical models of the scoliotic spine were developed to simulate the Cotrel-Dubousset instrumentation, although none was dedicated to the Colorado system. The objective of this study is to adapt and assess an existing biomechanical model to simulate the effect of the Colorado instrumentation of the scoliotic spine as a function of pre-operative geometry and surgical planning. Fifteen scoliotic patients operated with a Colorado system were analysed using a knowledge extraction technique to simplify surgical procedure and to establish the biomechanical model (boundary conditions, simulation procedures,...). Preliminary results on one patient show that the Colorado surgical technique can be adequately modelled using the preoperative geometric data and limited simulation strategy parameters.

[Perioperative radiographic reconstruction of the scoliotic vertebral column]. / Reconstruction radiographique peropératoire de la colonne vertébrale scoliotique.

We have developed a new per-operative three dimensional (3D) reconstruction technique to evaluate the 3D correction of a scoliotic spine induced by surgery using Cotrel-Dubousset instrumentation. A small object with 15 embedded markers was used to calibrate the radiographic system. During surgery, the calibration object was sterilized and fixed to the patient just before the acquisition of two pairs of posterior-anterior and sagittal radiographs; one pair before the rotation maneuver of the rod and one pair after the maneuver. The markers were digitized on each radiograph and their relative 3D positions were measured to establish the relation between the 3D positions of the anatomical structures and their 2D positions on the radiographs. This relation was used to calculate the 3D position of six anatomical landmarks per vertebra (the centers of the superior and inferior vertebral body endplates and the proximal and distal bodies of both pedicles) from the identification of these landmarks on each radiograph. We made a 3D representation of the thoracic and lumbar spine of three patients with idiopathic scoliosis undergoing corrective surgery by the posterior approach. Clinical indices (Cobb angle, axial rotation and the plane of maximum curvature) computed from the 3D reconstruction of the spine obtained before and after the rotation maneuver of the rod were compared to evaluate the 3D correction performed during the surgery. The new proposed approach allows the surgeon to evaluate the per-operative shape of the spine. This approach is simpler, faster and less risky for the patient than the previous method which employed an electromagnetic digitizer to measure the 3D coordinates of anatomical landmarks located on the posterior part of the spine. Furthermore, the 3D representation of the spine visualized from different points of view gives the surgeon an accurate evaluation of the 3D correction during the surgical procedure.

Self-calibration of a biplane X-ray imaging system for an optimal three dimensional reconstruction.

The purpose of this paper is to elaborate a three dimensional (3D) reconstruction method, using biplane X-ray angiograms acquired daily by the clinician, without any special calibration procedure during the X-ray examination. The absolute geometry of the X-ray imaging system is determined by an iterative procedure based on the minimization of the mean square distance between observed and analytical projections of a set of reference points identified by the clinician on the simultaneous pair of images. Once the geometry of the imaging system is found the 3D structure of interest is retrieved from classical methods of binocular stereovision. This 3D information is a prerequisite for an accurate evaluation of the degree of severity of a vascular structure or motion anomaly and therefore, for establishing an appropriate diagnosis. The proposed 3D reconstruction method is validated on synthetic and real data and is shown to perform robustly and accurately in the presence of noise. The method should be particularly useful in clinical applications as it needs very little intervention from the clinician.