Stephane Lavallee

Matching 3-D anatomical surfaces with non-rigid deformations using octree-splines

This paper presents a new method for determining the minimal non-rigid deformation between two 3-D surfaces, such as those which describe anatomical structures in 3-D medical images. Although we match surfaces, we represent the deformation as a volumetric transformation. Our method performs a least squares minimization of the distance between the two surfaces of interest. To quickly and accurately compute distances between points on the two surfaces, we use a precomputed distance map represented using an octree spline whose resolution increases near the surface. To quickly and robustly compute the deformation, we use a second octree spline to model the deformation function. The coarsest level of the deformation encodes the global (e.g., affine) transformation between the two surfaces, while finer levels encode smooth local displacements which bring the two surfaces into closer registration. We present experimental results on both synthetic and real 3-D surfaces.

Pedicle screw placement using image guided techniques.

Clinical evaluation of a computer assisted spine surgical system is presented. Eighty pedicle screws were inserted using computer assisted technology in thoracic and lumbar vertebrae for treatment of different types of disorders including fractures, spondylolisthesis, and scoliosis. Fifty-two patients with severe fractures, spondylolisthesis, or pseudoarthrosis of T10 to L5 were treated using a computer assisted technique on ½ the patients and performing the screw insertion manually for the other ½. At the same time, 28 pedicle screws were inserted in T12 to L4 vertebrae for scoliosis with the help of the computer assisted technique. Surgery was followed in all cases (66 vertebrae; 132 pedicle screws) by postoperative radiographs and computed tomographic examination, on which measurements of screw position relative to pedicle position could be done. For fractures, spondylolisthesis, or pseudarthrosis, comparison between the two groups showed that four screws in 52 (8%) vertebrae had incorrect placement with computer assisted technique whereas 22 screws in 52 (42%) vertebrae had incorrect placement with manual insertion. In patients with scoliosis, four screws in 28(14%) vertebrae had incorrect placement. In all of the patients (132 pedicle screws) there were no neurologic complications. These results show that a computer assisted technique is much more accurate and safe than manual insertion.

Computer-Assisted Spine Surgery: A Technique for Accurate Transpedicular Screw Fixation Using CT Data and a 3-D Optical Localizer

The computer-assisted spine surgery system presented in this paper follows the basic ideas which have been developed for computer-assisted medical interventions (CAMI) in our lab since 1985. There are three steps to insert a linear tool inside vertebral pedicles. First, the surgeon defines an optimal trajectory on pre-operative computed tomography. Second, this trajectory is reported in the operating room coordinate system using an intra-operative sensor and a registration algorithm. Third, a guiding system helps the surgeon follow the selected trajectory. In this paper, we present an implementation of this method that uses only a 3-dimensional optical localizer. Results on cadaver specimens and on the first seven patients are presented.

Computer assisted spine surgery.

The aim of this study was to improve the reliability of pedicle screw insertion. Transpedicle screw insertion may cause neurological, vascular, and mechanical complications. Previous studies of surgical procedures have shown a significant rate of incorrect placement of the screw ranging from 10 to 40%. A new technique that combines preoperative computed tomography (CT) imaging with intraoperative passive navigation was used to perform 64 pedicle screw insertions in the thoracolumbar region. At the same time, 64 pedicle screw insertions were performed manually in the same region and on the same vertebral levels. Surgery was followed in all cases by postoperative radiographs and computed tomography examination, which allowed measurements of screw position relative to pedicle position to be performed. A comparison between the two groups showed that six screws in 64 vertebra (9%) had incorrect placement with the computer-assisted technique whereas 28 screws in 64 vertebra (44%) had incorrect placement with manual insertion. The intraoperative accuracy provided by the computer after registration was better than 1 mm. The good results obtained are similar to those reported in the literature. The cortex penetration observed with the computer-assisted technique was not imputed to computer failure. Errors by the surgeon in acquiring data in the pre- and perioperative steps may explain the six incorrect screw placements. This clinical experience confirms that the accuracy and the reliability of this computer-assisted technique are good.

Incorporating a statistically based shape model into a system for computer-assisted anterior cruciate ligament surgery.

This paper addresses the problem of extrapolating very sparse three-dimensional (3-D) data to obtain a complete surface representation. A new method that uses statistical shape models is proposed and its application to computer-assisted anterior cruciate ligament (ACL) reconstruction is detailed. The rupture of the ACL has become one of the most common knee injuries. One problem during reconstruction is to find the optimal attachment points for the graft. Therefore a system for computer-assisted reconstruction of the ACL has been proposed by TIMC laboratory. During surgery the surgeon collects several data points on the tibial and femoral joint surface with a 3-D localizer system. These 3-D data are used to find those attachment points resulting in a low anisometry of the graft, while preventing impingement between the graft and the femoral notch. As the collected data points only cover a small surface patch of the femur, it is desirable to extrapolate these data to also have a visualization in those areas where no data points are available. A sufficiently good approximation of the actual femur by the model would further allow us to better deal with the notch impingement problem of the graft. The chosen approach is to fit a deformable model to the data points, it can be subdivided into two steps, constructing the model and fitting this model to the data. To incorporate a priori knowledge into the model, the allowed deformations are determined by the statistics of the shape variation of a set of training objects. Matching the training objects together is obtained by elastic registration of surface points using octree splines. The fitting process of the sparse intra-operative data with the statistical model results in a non-linear multi-dimensional function minimization. A hybrid search strategy combining local and global methods is used to avoid local minima. First experimental results with a model generated from 10 femurs are presented, including fitting of the model with both simulated and real intra-operative data.

Image guided operating robot: a clinical application in stereotactic neurosurgery

Describes a system based on the combined use of medical imaging and robot positioning used in stereotactic neurosurgery. The types of interventions performed with the help of this system require a high precision (less than 1 millimeter) of positioning with respect to the patients brain with no direct visibility of the target; they are executed through a hole in the skull of 2.3 millimeters. A typical example is the placement of a stimulating electrode in a particular nucleus of the thalamus for patients suffering from Parkinsons disease. These interventions rely on the intensive use of medical imaging to define the operative strategy. The target is defined with respect to the images. A six-axis robot whose end-effector is a linear guide automatically reaches a configuration generated such that if the surgical instrument (needle, electrode, etc.) is introduced in the guide, the surgeon reaches the target using a linear displacement of the instrument. This system is operational and is in routine clinical use.<<ETX>>

The mesh-matching algorithm: an automatic 3D mesh generator for finite element structures.

Several authors have employed finite element analysis for stress and strain analysis in orthopaedic biomechanics. Unfortunately, the definition of three-dimensional models is time consuming (mainly because of the manual 3D meshing process) and consequently the number of analyses to be performed is limited. The authors have investigated a new patient-specific method allowing automatically 3D mesh generation for structures as complex as bone for example. This method, called the mesh-matching (M-M) algorithm, generated automatically customized 3D meshes of anatomical structures from an already existing model. The M-M algorithm has been used to generate FE models of 10 proximal human femora from an initial one which had been experimentally validated. The automatically generated meshes seemed to demonstrate satisfying results.

Accurate calibration of cameras and range imaging sensor: the NPBS method

A method is presented for modeling and calibration of sensors, to reach an accuracy similar in magnitude to the resolution of the sensors. The authors focus on camera and range image calibration. They show the limitations of existing methods, and present an approach and camera calibration using mathematical B-Spline functions. This method is abbreviated as NPBS (N-Planes B-Spline). It is an extension of the two-planes method, but relies on pure mathematical modeling based on spline approximation theory. The mathematical modeling is combined with a calibration device that enables accurate data acquisition to calibrate a camera and a light plane independently. Experimental results are presented showing accuracy on the determination of a scattered plane.<<ETX>>

Computer-Assisted Knee Anterior Cruciate Ligament Reconstruction: First Clinical Tests

Anterior cruciate ligament reconstruction is a delicate task. The procedure of choice is the patellar tendon bone autograft, but an anisometric position of this tendon often leads to failure. We propose a method that allows positioning of the central part of the ligament graft at the least anisometric sites. The system uses a workstation and a three-dimensional optical localizer to create images that represent knee kinematics. The surgeon uses these images to guide the surgery. This technique has been validated on eight cadavers and 12 patients.

Nonrigid 3-D/2-D Registration of Images Using Statistical Models

This paper presents a new algorithm for reconstruction of 3D shapes using a few x-ray views and a statistical model. In many applications of surgery such as orthopedics, it is desirable to define a surgical planning on 3-D images and then to execute the plan using standard registration techniques and image-guided surgery systems. But the cost, time and x-ray dose associated with standard pre-operative Computed Tomography makes it difficult to use this methodology for rather standard interventions. Instead, we propose to use a few x-ray images generated from a C-Arm and to build the 3-D shape of the patient bones or organs intra-operatively, by deforming a statistical 3-D model to the contours segmented on the x-ray views. In this paper, we concentrate on the application of our method to bone reconstruction. The algorithm starts from segmented contours of the bone on the x-ray images and an initial estimate of the pose of the 3-D model in the common coordinate system of the set of x-ray projections. The statistical model is made of a few principal modes that are sufficient to represent the normal anatomy. Those modes are built by using a generalization of the Cootes and Taylor method to 3-D surface models, previously published in MICCAI’98 by the authors. Fitting the model to the contours is achieved by using a generalization of the Iterative Closest Point Algorithm to nonrigid 3D/2D registration. For pathological shapes, the statistical model is not valid and subsequent local refinement is necessary. First results are presented for a 3-D statistical model of the distal part of the femur.