Method for determining a screw trajectory for a pedicle bone screw

Abstract

A method for determining a screw trajectory for a pedicle bone screw, comprising: obtaining a CT image of a target bone region intended to receive a pedicle bone screw, establishing a personalized three-dimensional geometric model of the target bone region based on the CT image, accessing a database comprising three-dimensional bone region models; wherein the bone region models comprise a pedicle screw insertion surface (10) and a pedicle transverse surface (20, 20') for each pedicle, deforming the bone region model to the geometric model of the target bone region, and generating a deformed vertebra model (47) with the pedicle screw insertion surface (10) and the pedicle transverse surface (20, 20'), calculating a maximum bone density for a pedicle screw in the deformed vertebra model (47) of the target bone when the bone material is replaced by the pedicle screw, and outputting a spatial vector of the screw trajectory for the pedicle screw in the deformed vertebra model of the target bone and a length and a diameter of the pedicle screw.

Description

Technical Field

[0001] The present invention relates to a method for determining the screw trajectory of a pedicle screw, a data processing system including means for determining the screw trajectory of a pedicle screw, and a computer program product including instructions that, when executed by a computer, cause the computer to plan the screw trajectory of the pedicle screw. Background Technology

[0002] CN 110946652 A discloses a method for determining the trajectory of a bone screw, comprising the following steps: establishing a personalized three-dimensional geometric model of the target bone and a three-dimensional bone density model containing spatial distribution information of bone density; setting bone screw parameters and calculating the spatial vector of the bone screw trajectory; calculating the spatial spiral of the bone thread as the bone screw thread curve; placing the bone screw thread curve in the personalized three-dimensional geometric model of the target bone; extracting the bone density at the contact point with the bone screw thread in the personalized three-dimensional geometric model of the target bone; and calculating the average value of the bone density at the contact position; determining the bone screw trajectory based on the bone screw parameters, bone density, and bone density distribution. According to this method, by obtaining the coordinates of the contact position between the bone and the bone screw on a three-dimensional scale, the bone density around the screw thread is accurately determined, and the screw trajectory of the bone screw can be determined based on the fixation performance of the bone screw.

[0003] US 2017 / 112575A1 discloses a non-transitory computer-readable medium embodying a machine-executable instruction program for performing operations for pedicle screw positioning, the operations including: receiving image data of at least a portion of the spine; segmenting at least one vertebra of interest in the image data; determining two pedicle regions within the segmented vertebra of interest; determining one or more safe regions within the segmented vertebra of interest; generating two optimal insertion paths within the one or more safe regions, wherein the two optimal insertion paths pass through the respective centers of the pedicle regions; and displaying the two optimal insertion paths for pedicle screw positioning. The method uses voxels and the distances of voxels within the segmented vertebra of interest to the nearest vertebral edge to determine a safe region for screw implantation, assigning voxels to safe regions if the distance to such edge is greater than a threshold distance.

[0004] US 2004 / 240715A1 discloses a method for determining the placement of a pedicle screw, comprising determining a trajectory for placing such a pedicle screw in a patient from a set of 2D images, wherein the step of determining the trajectory includes: calculating an optimal implantation trajectory for the pedicle screw by determining the minimum transverse pedicle width of all 2D image slices containing the vertebra under study, and determining the total minimum transverse pedicle width for the vertebra under study based on the minimum transverse pedicle width of all 2D image slices.

[0005] CN 109199604 A discloses a method for locating the optimal entry point of a pedicle screw, which uses the feature vector of a three-dimensional mesh model as input to a decision tree classification model. Summary of the Invention

[0006] Based on this prior art, one object of the present invention is to provide an alternative method for optimizing screw parameters, including size and orientation, based on patient-specific bone characteristics, particularly bone mineral density of the replaced screw volume.

[0007] According to the invention, this objective is achieved by the features of independent claim 1. Furthermore, further advantageous embodiments are derived from the dependent claims and the description.

[0008] Specifically, this objective is achieved by a method for determining the screw trajectory of a pedicle screw, the method comprising the following steps:

[0009] - Obtain CT images of the target bone region intended for receiving the pedicle screw.

[0010] - Create a personalized three-dimensional geometric model of the target bone region based on CT images including bone density information.

[0011] - Access a database containing three-dimensional bone region models; wherein the bone region models include bone screw insertion surfaces and pedicle transverse surfaces.

[0012] - Deform the bone region model into a geometric model of the target bone region, generating a deformed vertebral model with a bone screw insertion surface and a pedicle transverse surface.

[0013] - When bone material is replaced by pedicle screws in a deformed vertebral model of the target bone region, calculate the optimal screw trajectory of the pedicle screws that maximizes bone density, and

[0014] - Outputs the optimal screw trajectory for the pedicle screw in the deformed vertebral model of the target bone, as well as the length and diameter of the bone screw.

[0015] The database to be accessed includes data associated with a three-dimensional bone mineral density model having predetermined bone screw insertion surfaces and predetermined pedicle transverse surfaces. According to the disclosed embodiments, the three-dimensional bone mineral density model is based on and trained upon examined bone samples. A deformation step then allows these bone screw insertion surfaces and pedicle transverse surfaces to be automatically transferred onto a 3D bone image model of the vertebra of interest.

[0016] According to the disclosed embodiments, deformation of the bone region model includes providing a 3D pedicle transverse surface within the geometry of the target bone. This is used to identify possible bone screw trajectories. According to the disclosed embodiments, the 2D pedicle transverse surface for the pedicle transverse surface is based on the 3D pedicle transverse surface and is determined as the plane with the minimum transverse pedicle width in the pedicle. The plane with the minimum transverse pedicle width corresponds to the plane (transverse plane) of the 3D pedicle transverse surface, where the pedicle has a minimum width in the transverse direction.

[0017] The 3D and 2D lateral surfaces are modeled within the geometry of the target bone region in the database and deformed along with other elements of the geometry on the target bone region. This allows computation to begin without a direct evaluation of the vertebral region of interest.

[0018] The deformation step of the bone region model may alternatively include providing a sagittal plane within the geometrically deformed vertebral model to determine the transverse surface of the pedicle. According to another embodiment, the deformation of the bone region model may include determining the vertebral foramen within the geometrically deformed vertebral model to define the transverse surface of the pedicle as the minimum diameter of bone material on the lateral surface of the vertebral foramen.

[0019] According to the disclosed embodiments, in the process of determining the 2D transverse surface of the pedicle, a first threshold is provided as a contour safety distance. This safety distance can be, for example, 1 or 2.5 mm, and is generated as a 3D curve within the outer edge of the pedicle, delimiting the voxels used in the 3D and subsequent 2D transverse surfaces of the pedicle in this method. Furthermore, a second threshold as a length safety distance can be provided during the determination of screw length based on the body side surface of the body of the bone region, so that solutions that would extend beyond the other side of the body region of the vertebra (i.e., screw trajectory and the length and diameter of the pedicle bone screw) are not calculated.

[0020] Furthermore, the sagittal plane of the body of the predetermined bone region provides a third threshold for the profile and tip of the screw, defining a prohibited area for the screw, which does not have to pass through the prohibited area, so that either bone screw can be introduced into the same body of the pedicle, or only one of two bone screws can be introduced into the same body in such a way that the two bone screws do not occupy the same position.

[0021] The initial conditions for the calculation step may include the value of the starting screw, wherein on one side, the central axis of the starting screw passes through the center point of the 2D pedicle transverse surface, and on the other side, the bone screw insertion surface of the enveloping cylinder of the starting screw is within the corresponding insertion surface.

[0022] In an alternative approach, a starting point can be selected on the pedicle field surface adjacent to the insertion point to include the central portion of the screw trajectory within the 3D transverse surface, for example, using a least-squares method with the selected insertion point / surface connecting the screw cylinder and the center of the 3D pedicle transverse surface thus defined. According to the disclosed embodiments, in the geometric model of the target bone, the calculation step is provided with screw length, screw diameter with boundary conditions of a predetermined maximum length, and initial parameters of a predetermined maximum diameter.

[0023] The steps of this invention can be performed on a data processing system including means for performing the steps of the method claims. The data processing system includes a computer storage system for a database having a statistical shape model (SSM) and a data storage device for inputting CT images. These images can be prepared in advance or in real-time for calculating screw trajectories. The data processing system typically also includes input units such as a keyboard, mouse, or touchscreen / tablet for selecting, for example, the vertebra of interest, and output devices such as displaying and storing the calculated data for the screw.

[0024] This document further discloses a computer program product including instructions that, when executed by a computer, cause the computer to perform steps of a method according to any embodiment disclosed herein.

[0025] Further embodiments of the invention are set forth in the dependent claims.

[0026] The purpose of the embodiments disclosed herein is to provide an efficient method for determining the optimal screw trajectory of the pedicle screw to maximize bone density when bone material is replaced by a pedicle screw in a deformed vertebral model of a target bone region. This objective is achieved by the features of claim 10. In particular, a particularly efficient method for determining the optimal screw trajectory includes:

[0027] - Use the bone screw insertion surface and the transverse surface of the pedicle to identify the space of possible projection planes of the deformed vertebral model;

[0028] - Scan the space of possible projection planes in order to determine a set of intersection density projections;

[0029] - Scan the density projection of the cross region of this group and calculate the corresponding projected bone mineral density score;

[0030] - The optimal screw trajectory is determined to have a direction perpendicular to the projection plane, which corresponds to the highest score of the projected bone mineral density score.

[0031] According to a specific embodiment, the length of the bone screw is determined using a three-dimensional geometric model of the target bone region and an optimal screw trajectory, as the distance between the bone screw insertion surface and the transverse surface of the pedicle in the direction of the optimal screw trajectory.

[0032] According to a specific embodiment, the step of identifying the space of possible projection planes of a deformed vertebral model includes: defining a sphere in the deformed vertebral model around the center of the target bone region; defining the space of possible projection planes as a set of planes perpendicular to the surface of the sphere; and limiting the space of possible projection planes to a set of projection planes that intersect between the bone screw insertion surface and the transverse surface of the pedicle.

[0033] According to a specific embodiment, the step of scanning the space of possible projection planes to determine a set of cross-region density projections includes: selecting a plurality of discretely distributed projection planes within the space of possible projection planes; determining the cross region of the bone screw insertion surface and the transverse surface of the pedicle for each of the plurality of discretely distributed projection planes; and summing all voxels (representing bone density) of the deformed vertebral model within the cross region and perpendicular to the projection plane to obtain the corresponding cross-region density projection.

[0034] According to a particular embodiment, scanning the set of cross-region density projections and calculating the corresponding projected bone mineral density score includes summing voxels of the cross-region density projections within a region demarcated by one or more possible bone screw diameters for each possible location of the bone screw within the corresponding cross-projection, wherein the optimal screw trajectory has an axis passing through the center of the screw diameter, which corresponds to the highest score of the projected bone mineral density score.

[0035] According to a specific embodiment, in the preparation step, voxels outside the target bone region are removed / deleted from the deformed vertebral model. The remaining voxels of the deformed vertebral model include bone mineral density information related to the target bone region.

[0036] According to a further embodiment, to improve the algorithm's performance, the scanning of the space of possible projection planes can be performed relatively coarsely in the first iteration. A new set of cross-region density projections is then defined in a narrower space near the possible projection plane with the highest projected bone density score, providing finer resolution at locations where high scores were found in previous iterations. This iterative process can be repeated and optionally combined with an optimization function, resulting in a highly efficient method for determining the optimal screw trajectory. Attached Figure Description

[0037] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings, which are for illustrative purposes and not for limiting the scope of the invention. In the drawings,

[0038] Figure 1 A flowchart of a method according to an embodiment of the present invention is shown;

[0039] Figure 2 shows two different CT views and 3D model images. The CT view has the vertebra of interest masked, and the 3D model image represents the CT data associated with the actual bone represented in the model image.

[0040] Figure 3A A schematic diagram is shown, viewed from above the vertebra of interest, illustrating two embodiments on different sides of the sagittal plane;

[0041] Figure 3B It shows from Figure 3A A side view viewed from the left, showing the insertion and pedicle surfaces according to an embodiment of the present invention;

[0042] Figure 3C It shows from Figure 3A A side view viewed from the right side, showing the insertion and pedicle surfaces according to an embodiment of the present invention;

[0043] Figure 4A A schematic perspective view is shown, taken from above the vertebra of interest, with one screw inserted into a configuration provided according to the invention;

[0044] Figure 4B It shows from Figure 4A A schematic perspective view of the vertebrae as shown;

[0045] Figure 5 A flowchart illustrating the calculation of the optimal screw trajectory for the pedicle screw according to a specific embodiment is shown;

[0046] Figure 6 A flowchart illustrating the sub-steps for identifying a space that may project planes, according to a particular embodiment, is shown;

[0047] Figure 7 A flowchart illustrating the sub-steps of scanning a possible projection plane according to a particular embodiment is shown;

[0048] Figure 8 A flowchart of the density projection of the scanning cross region according to a specific embodiment is shown;

[0049] Figure 9 A schematic perspective view of voxels of a deformed vertebra model is shown;

[0050] Figure 10 A perspective view of nodes defined on the surface of a sphere (centered on the target bone region) is shown, located at the intersection of the longitude and latitude lines of the sphere;

[0051] Figure 11A-11CA series of schematic perspective views are shown of possible projection planes of a deformed vertebral model perpendicular to a sphere centered on the target bone region;

[0052] Figure 12A The projection plane of the deformed vertebra model is shown;

[0053] Figure 12B It shows Figure 12A The projection plane of the deformed vertebral model, in which the bone screw insertion surface and the transverse surface of the pedicle are marked;

[0054] Figure 12C It shows Figure 12A and 12B The projection plane of the deformed vertebral model marks the intersection area between the bone screw insertion surface and the transverse surface of the pedicle; and

[0055] Figure 12D It shows Figure 12A , 12B The projection plane of the deformed vertebral model of 12C illustrates the possible positioning of pedicle screws with possible diameters within the density projection of the cross region. Detailed Implementation

[0056] Figure 1 A flowchart of a method according to an embodiment of the present invention is shown, describing a method for determining the optimal pedicle screw size and orientation for a patient requiring spinal fusion. This optimization method requires a data collection step 1, which involves medical imaging data obtained as 3D computed tomography (CT) composed of discrete voxels. The CT images include information that allows for the determination of bone density information extracted and stored for the discrete voxels. The step of obtaining the 3D computed tomography data can also be achieved by accessing a database in which pre-recorded images are stored. Furthermore, the use of the 3D computed tomography data serves a dual purpose. On the one hand, as will be explained, the data is used to create a data model 45 of the vertebra of interest 43, which is the vertebra(s) to which the pedicle screw(s) 60 will be inserted. On the other hand, the image data includes information about bone properties, which can be spatially correlated with the discrete voxels. This step of attaching data to a specific voxel can be performed at any time between the end of step 1 and step 7 (see below), i.e., before the optimization calculation begins.

[0057] Pedicle screw loosening is a clinical complication, often associated with low bone mineral density. Therefore, optimizing pedicle screws based on patient-specific bone characteristics can support surgeons in determining spinal fusion.

[0058] The following are Figure 1The explanation of the flowchart is supplemented by referring to the figure references used in the representation of the vertebrae, related data model 45 and deformed vertebrae model 47, and other figures.

[0059] In segmentation step 3, 3D patient images 41 and 42 depicting the spinal region of interest are acquired and segmented to extract 3D geometric models 45 of all vertebrae 43 that need to be fitted.

[0060] Furthermore, a single statistical shape model, or SSM for short, is provided in advance as step 2. This model has predetermined points on the bone screw insertion surface 10 and the pedicle transverse surface 20 or voxel plane as data input for the next processing step of the segmented vertebral image in step 3. According to the disclosed embodiments, the SSM is trained for different vertebrae using a comprehensive dataset of 3D vertebral models. The statistically deformable model obtained in step 2 is capable of representing a wide range of vertebral geometries and can be deformed on the segmented patient model 45 in step 3. This deformation step 4 involves providing a deformation of the predetermined SSM on the 3D geometric model 45, with the screw to be implanted into the vertebra of interest 43, creating a deformed vertebral model 47.

[0061] On the deformable model from step 2, relevant surfaces 10, 20 and / or 20′, 58 are marked to initialize the optimization process: the circumferential contour 58 of the vertebral body between the pedicle screw insertion surface 10, the 3D pedicle transverse surface 20′, the 2D pedicle transverse surface 20, and the endplates 51, 51′. The screw insertion surface 10 is a surface defined on the SSM (Statistical Shape Model) in which the screw body can enter the vertebral material.

[0062] This method, by applying specific deformations of the template model from step 3 to step 4, allows for the automatic identification of anatomical surfaces 10, 20, 20′, and 58 on vertebral structures. This integration reduces the manual steps required for identification and improves its accuracy and robustness.

[0063] Deformation step 4 identifies anatomical surfaces on the patient vertebral model 45 from step 3. These points are initially used to rigidly place the SSM at the corresponding spinal level. Furthermore, in registration step 5, a set of points is performed for non-rigid image registration, which deforms the SSM on the segmented vertebrae 43 of the patient.

[0064] These steps 4 and 5 can be explained in conjunction with Figure 2, which shows two different CT views 41 and 42 and a 3D model image 45. CT views 41 and 42 have vertebrae of interest 43 masked by corresponding lines 44. The 3D model image 45 represents CT data related to the actual bone represented in the model image of Figure 2, which is actually a dataset of voxels. According to the embodiments disclosed herein, the method uses more than two different CT views 41 and 42, which are shown as examples of two CT planes that intersect at right angles when viewed from the perspective of the vertical body line.

[0065] 3D model image 45 shows the vertebra of interest 43, which has a superior end plate 51, an inferior end plate 51′, pedicles 52, transverse process and coastal facet 53, superior articular facet 54, spinous process 55, vertebral foramen 56, and body 57. The deformed 3D model 47 is placed within the input CT image model 45 through a coordinate transformation between the image and physical coordinate systems. The transformation is defined using voxel properties defined in the information paired with the input CT images 41, 42. 3D model 45 is used to mask CT images 41, 42, preserving voxel intensity information within the vertebra of interest 43, referred to as Hounsfield units (HU). Voxel intensity values ​​are converted into bone material properties (Young's modulus):

[0066] ρ app =47+1.22*HU (1)

[0067] E mod =757*(ρ app ) 1.94 (2)

[0068] After deformation, the outline 59 of the marked endplates 51, 51′ from step 3 allows for automatic identification of the patient's pedicle region. A uniform mesh within the pedicle region is created radially from the center of the endplates 51, 51′ using a region growing method. A polar coordinate system is defined on the upper endplate 51 of the vertebral model. In an alternative, the lower endplate 51′ is used for it. Points defining the endplate boundaries 59 are moved by increasing their radial components 8 around the vertebral endplates 51 or 51′ with a consistent stride. For each radial increase, a uniform mesh is generated in the axial direction of the vertebra 43, resulting in a 2D surface that increases perpendicularly to the radial direction around the endplates 51, 51′. A volumetric mesh is generated around the entire vertebral body 57, and since the pedicle 52 is the only structure extending from its sides, the points in the volumetric mesh within the vertebral 3D model represent points within the pedicle region. Furthermore, points within the pedicle 52 must be selected with a threshold distance perpendicular to the vertebral boundary. Such a threshold can be selected as at least 1 mm or 2.5 mm as the minimum safe distance 21, which avoids pedicle screw 60 (see...). Figure 4A (This refers to the penetration of the pedicle wall.) In the accompanying figures, reference numeral 21 is used to show the distance between the 3D pedicle transverse surface 20 or the 2D pedicle transverse surface 20 and the pedicle wall surface. The grid resolution depends on one of the input CT images to include all voxel information in the optimization.

[0069] The above determination of the 3D pedicle transverse surface 20′ and the 2D pedicle transverse surface 20 is based on the endplates 51, 51′. In an alternative approach, a sagittal plane 35 can be used, which can be easily identified by the spinous process surface 55 and the central line passing through the endplates 51, 51′. In another alternative, the vertebral foramen 56 can be identified and used to determine the pedicle 52, wherein the 2D pedicle transverse surface 20 is the smallest diameter portion of the bone material.

[0070] Within a 3D uniform grid depicting the 3D transverse surface 20′ of the pedicle region, the 2D transverse surface 20 of the pedicle can be defined at the total minimum transverse pedicle width. Points on this 2D transverse surface 20, together with points in the bone screw insertion surface 10, define the screw trajectory 61 of the pedicle screw 60. The contour 59 of the body 57 surrounding the endplates 51, 51′ includes a body side surface 58. As a further safety margin, the screw length is primarily limited by a minimum safety distance 22 from this body side surface 58. These safety margins ensure that the pedicle bone screw 60 does not perforate the vertebral wall.

[0071] In addition to the main lateral surface 58, the sagittal plane 35 provides a further boundary plane for the pedicle screw 60 when two pedicle bone screws 60 are typically placed through two pedicles 52, as the two pedicle bone screws 60 should not interfere with the placement of the other. A solution (screw trajectory) that does not take into account this boundary condition can be provided for one pedicle bone screw 60, and if these solutions are compatible, i.e., not calculated to occupy the same space, they can be combined with a solution used for another screw examination. The bone screw insertion surface from step 3 is deformed on the patient's vertebra 43 together with the SSM. On the deformed model 47, the bone screw insertion surface 10 and the 3D and 2D pedicle transverse surfaces 20′ and 20′ located within the pedicle, respectively, are deformed. Figure 3A , 3B And shown as shaded lines in 3C. Note that the representation of the deformed 3D model is incomplete in itself, as it also includes, at least before the calculation steps, information about bone properties, especially bone density.

[0072] The optimization space was defined based on the insert 10 and the transverse surfaces 20, 20′ of the pedicle, as well as the bone edges, particularly the lateral surface 58, from the deformed 3D model in step 4. The combination of points from the insert surface 10 and the transverse surface 20 of the pedicle from the mesh defined the screw trajectory 61.

[0073] The input parameters used for optimization are the insertion point on the bone screw insertion surface 10, the pedicle point (i.e., the screw center in the grid region of the 2D pedicle transverse surface 20), the screw diameter, and the screw length, which, taking into account the body side surface 58, can optionally be subtracted from the surface distance 22. All of these are optimized to avoid perforation using the bone structure from the deformed vertebral model 47 of step 4. According to the disclosed embodiment, the insertion point is selected around the center of the surface demarcated as the bone screw insertion surface 10. The insertion point is not strictly a point, but rather a cross-section of the 3D surface of the generally non-uniform vertebral surface with the cylindrical model surface of the screw body, and is indeed a 3D surface. As mentioned above, the starting point in the 2D pedicle transverse surface 20 can be selected as the center of the 2D pedicle transverse surface 20, thereby defining the screw trajectory 61. In other embodiments, a starting point in the pedicle transverse surfaces 20, 20′ adjacent to the insertion point can be selected to include the central portion of the screw trajectory 61 within the 3D pedicle transverse surface 20′, for example, by connecting the selected insertion point and the “center” of the 3D pedicle transverse surface 20′ defined therefrom using least squares.

[0074] Step 8 includes iterative calculations implemented to compare combinations of input parameters (screw insertion surface 10, pedicle point, i.e., screw center in the grid region of the 2D pedicle transverse surface 20, screw diameter, and screw length) and extract the optimal solution. This can be a parameter-based genetic computation method, where the parameters are, for example, sample size in each iteration, mutation percentage, percentage of best-case scenarios carried over to the next iteration, etc., as known to those skilled in the art. A limited number of optimal or near-optimal solutions may also be provided, taking into account the surgeon's choice. A simplified (approximate) cylinder of the pedicle bone screw 60 is created using the combination of input parameters and placed in a 3D deformed vertebral model 47 along a predetermined screw trajectory 61. First, the calculation excludes cylinders of perforated deformed vertebral models 47 from step 4. If the screw is not perforated, the cylinder is placed in a 3D image mask from step 5 using a coordinate transformation between CT images 41, 42 and the physical coordinate system. Image voxels containing bone material property values ​​are extracted within the screw volume and used to calculate the distribution of bone material within the cylinder.

[0075] To initialize the optimization method, a random population of parameter combinations was created. Bone density calculations were used to test each combination. After testing, random variations and recombinations of the parameters were introduced to create a population for the next iteration. The population in the final iteration contained the best-performing parameter combinations.

[0076] The distribution of bone material properties is constrained to the performance of each parameter combination, thus used to optimize the final solution. The optimal screw trajectory and screw size generated by optimization are one that maximizes the voxel-based bone material properties within the screw volume, which is the output of step 9. As mentioned above, the output may also include multiple optimal or near-optimal solutions, which have different advantages in terms of handling, screw head connection between different screws, etc.

[0077] Turn now Figures 5 to 1 2. A particularly effective embodiment for calculating the optimal screw trajectory 61 (step 8) of the pedicle screw 60 will be described.

[0078] Figure 5 A flowchart of steps 8.1 to 8.6 for calculating the optimal screw trajectory 61 of the pedicle screw is shown.

[0079] In the first preparatory step 8.1, the target bone region 43 is removed / deleted from the deformed vertebral model 47. Figure 9 Voxels other than those shown in dark gray. Remaining voxels of deformed vertebra model 47 (in...) Figure 9 (Shown in light gray) Includes bone density information associated with the target bone region 43.

[0080] According to a further embodiment, the voxel mesh of the deformed vertebra model 47 is downsampled to improve computational performance.

[0081] In step 8.2, the space of possible projection planes of the deformed vertebral model 47 is identified using the bone screw insertion surface 10 and the pedicle transverse surface 20. (Refer to...) Figure 6 Describe the details of sub-step 8.2. Figure 6 A flowchart of sub-steps 8.2.1 to 8.2.3 of the step of identifying the space of a possible projection plane according to a particular embodiment is shown. In sub-step 8.2.1, a sphere S is defined in the deformed vertebral model 47 around the center of the target bone region 43 (see...). Figure 10 In substep 8.2.2, the space of possible projection planes is determined, where each possible projection plane is perpendicular to the surface of sphere S. The space of possible projection planes (e.g.) Figures 11A to 11C As shown, (including infeasible projection planes) in step 8.2.3 are demarcated in such a way that only projection planes with orthogonal intersections are retained between the bone screw insertion surface 10 and the transverse surface 20 of the pedicle, with the projection perpendicular to the projection plane, thereby obtaining the space of possible projection planes.

[0082] In step 8.3, the space of possible projection planes is scanned to determine a set of intersection region density projections Idp1-n. (Refer to...) Figure 7 Describe the details of sub-step 8.3. Figure 7 A flowchart of substeps 8.3.1 to 8.3.3 of step 8.3, which involves scanning possible projection planes, is shown. In the first substep 8.3.1, a subset of discretely distributed projection planes P1-n is obtained from the space of possible projection planes. According to a specific embodiment, such as... Figure 10 As shown in the diagram, Figures 11A to 11C As shown in the sequence diagram, nodes are defined on the surface of sphere P at the intersections of the meridians and parallels of the sphere. The meridians and parallels of the sphere are equidistant according to a defined resolution. Discretely distributed projection planes P1-n are arranged perpendicular to sphere S at the nodes, but within the orthogonal intersection space demarcated in step 8.2.3. In the subsequent sub-step 8.3.2, as in Figures 12A to 12CAs depicted in the sequence diagram, on each of the discretely distributed projection planes P1-n, the intersection region I1-n of the bone screw insertion surface 10 and the transverse surface 20 of the pedicle is determined, and the obtained intersection region I1-n defines the space of possible screw trajectories. Subsequently, in substep 8.3.3, all voxels of the deformed vertebral model 47 (representing bone density) within the intersection region I1-n and perpendicular to the projection plane P1-n are summed to obtain the corresponding intersection region density projection Idp1-n for each discretely distributed projection plane P1-n. The voxels of the intersection region density projection Idp1-n each define the sum of all bone densities projected onto the corresponding projection plane P1-n. According to the embodiments disclosed herein, Hounsfield unit (HU) values ​​of CT images are used directly or after being converted to bone material properties according to Young's modulus to obtain the intersection region density projection Idp1-n.

[0083] In step 8.4 of the fourth step, the density projection Idp1-n of each cross region is scanned using the possible bone screw diameters D1-n and / or possible locations L1.1-xy within the cross projection I1-n. (Refer to...) Figure 8 Describe the details of substep 8.4. In the first iterative substep, for each cross projection I1-n, sum the voxels of the cross region density projection Idp1-n within the region demarcated by the possible bone screw diameters D1-m. This process is repeated for each possible location L1.1-xy of the corresponding bone screw diameter D1-m within each possible bone screw diameter D1-m and the corresponding cross region I1-n to obtain the corresponding projected bone density score Pbds1.1.1.1-Pdbsn.mxy. The scan of the cross region I1 is in... Figure 12D The above diagram shows the possible location P1.1.xy of a specific screw diameter D1-m within the corresponding intersection area I1, indicated by the black arrow.

[0084] Subsequently, in step 8.5, the optimal screw trajectory is determined to have a direction perpendicular to the projection plane P1-n, which is an axis passing through the center of the screw diameter D1-m, which corresponds to the highest projected bone mineral density score PbdsMAX.

[0085] After determining the optimal screw trajectory, in step S8.6, the length of the bone screw is determined using the three-dimensional geometric model 45 of the target bone region 43 and the optimal screw trajectory, specifically as the distance between the bone screw insertion surface 10 and the transverse surface 20 of the pedicle in the direction of the optimal screw trajectory.

[0086] According to further embodiments, such as in Figure 5The process of scanning possible projection planes and scanning cross-density projections (steps 8.3 and 8.4), illustrated by dashed arrows, is an iterative process. To improve the algorithm's performance, the incremental scanning of the projection planes (step 3) can be performed relatively coarsely in the first iteration. In particular, the resolution of the meridians of the nodes defined by the discretely distributed projection planes P1-n is relatively low.

[0087] In subsequent iterations, new projection planes are defined around one or more possible projection plane nearest neighbors (angular nearest neighbors on a sphere spanning the center of the target bone region) that have one or more highest projected bone mineral density scores Pbds (calculated in step 4). Specifically, a new subset of discretely distributed projection planes is determined around one or more possible projection plane nearest neighbors that have one or more highest projected bone mineral density scores Pbds, with an increased resolution compared to one or more previous iterations. Steps 8.3.2, 8.3.3, and 8.4 are performed again using the new subset of discretely distributed projection planes P1-n.

[0088] This iterative process (steps 8.3 and 8.4) is repeated a limited number of times. Alternatively or additionally, the iterative process is repeated until the highest projected bone mineral density score PbdsMAX of the current iteration is equal to the highest projected bone mineral density score PbdsMAX of the previous iteration, or exceeds the highest projected bone mineral density score PbdsMAX of the previous iteration by no more than a threshold improvement margin. Optionally, the iterative process is combined with an optimization function to produce a highly efficient method for determining the optimal screw trajectory.

[0089] Finally, after completing the iterative process (steps 8.3 and 8.4), in step 8.5, the optimal screw trajectory is determined to have a direction perpendicular to the projection plane P1.n, which is the axis passing through the center of the screw diameter D1-m, which corresponds to the highest projected bone mineral density score PbdsMAX of any iteration.

[0090] Reference Symbol List

[0091] 1. Data Collection Steps

[0092] 2. Segmentation Steps

[0093] 3. Provide statistical shape models

[0094] 4. Deformation Steps

[0095] 5. Masking and Registration Steps

[0096] 6. Determining parameters from the deformation model

[0097] 7. Determination of Boundary Conditions

[0098] 8. Optimize calculations

[0099] 9 Output Steps

[0100] 10 Bone screw insertion surface

[0101] 20 2D pedicle transverse surface

[0102] 20′ 3D pedicle transverse surface

[0103] 21. Minimum safe distance (pedicle)

[0104] 22 Minimum safe distance (main body)

[0105] 35. Sagittal plane

[0106] 41 First CT View

[0107] 42 Second CT View

[0108] 43 vertebrae of interest

[0109] 44. Boundary Line

[0110] 45. 3D geometric model, i.e., the data model representation of the vertebra of interest, abbreviated as: vertebra model

[0111] 47 Deformed Spine Model

[0112] 51. Top end board

[0113] 51′ Lower endplate

[0114] 52 pedicles

[0115] 53. Transverse processes and costal articular surfaces

[0116] 54. Upper articular surface

[0117] 55. Spinous process

[0118] 56 vertebral foramen

[0119] 57 Main Body

[0120] 58 Main body side view

[0121] 59. Final board outline

[0122] 60 pedicle screws

[0123] 61 Screw trajectory

[0124] P1-n projection plane

[0125] N1-n perpendicular projection plane P1-n

[0126] S-shaped sphere (centered on the target bone region)

[0127] I1-n intersection region

[0128] Idp1-n Cross Region Density Projection

[0129] D1-m Possible screw diameter

[0130] Pbds1.1.1.1-nmxy projection bone mineral density score

[0131] PbdsMAX highest projected bone mineral density score

Claims

1. A method for determining the optimal screw trajectory (61) of a pedicle screw (60), comprising the following steps: - Obtain CT images (41, 42) of the target bone region intended for receiving the pedicle screw (60). - A personalized three-dimensional geometric model of the target bone region is established based on CT images (41, 42) including bone density information (45). - Access a database containing three-dimensional bone region models; wherein the three-dimensional bone region models include bone screw insertion surfaces (10) and pedicle transverse surfaces. - The three-dimensional bone region model is deformed into a personalized three-dimensional geometric model (45) of the target bone region, generating a deformed vertebral model (47), which includes the bone screw insertion surface (10) and the pedicle transverse surface of the three-dimensional bone region model, as well as the bone density information of the personalized three-dimensional geometric model (45), wherein the deformation includes transferring the bone screw insertion surface (10) and the pedicle transverse surface to the personalized three-dimensional geometric model (45). - When the bone material is replaced by pedicle screws (60) in a deformed vertebral model (47) of the target bone region, the optimal screw trajectory (61) of the pedicle screws (60) that maximizes bone density is calculated, and - The optimal screw trajectory (61) of the pedicle screw (60) is calculated by iteratively determining the maximum bone density of the bone material that will be replaced by the volume of the pedicle screw (60) when placed in a deformed vertebral model (47) of the target bone region along a trajectory defined by a point on the screw insertion surface (10) and a point on the pedicle cross-sectional surface (20); and - Output the optimal screw trajectory (61) for the pedicle screw (60) in the deformed vertebral model (47) of the target bone region.

2. The method according to claim 1 further includes the length and diameter of the output pedicle screw (60).

3. The method according to claim 1 or 2, wherein, Transforming a three-dimensional bone region model into a personalized three-dimensional geometric model (45) includes: A 3D pedicle transverse surface is provided within a deformed vertebral model (47) of the target bone, which is modeled in a three-dimensional bone region model within a database and deformed with a personalized three-dimensional geometric model (45).

4. The method according to claim 3, wherein the deformation of the three-dimensional bone region model further includes providing a 2D pedicle transverse surface, wherein the 2D pedicle transverse surface as the pedicle transverse surface is based on the 3D pedicle transverse surface and corresponds to the plane of the minimum transverse pedicle width in the pedicle (52).

5. The method of claim 4, wherein a first threshold (21) is provided as a contour safety distance during the determination of the 2D pedicle transverse surface, the first threshold (21) being generated as a 3D curve within the outer edge of the pedicle (52), delimiting the voxels used in the 3D pedicle transverse surface and the subsequent 2D pedicle transverse surface.

6. The method of claim 3, wherein the starting condition for the step of calculating the optimal screw trajectory (61) includes the value of the starting screw, wherein the central axis of the starting screw is selected as the central portion of the axis within the transverse surface of the 3D pedicle.

7. The method according to claim 4, wherein, The initial conditions for calculating the optimal screw trajectory (61) of the pedicle bone screw (60) include the value of the starting screw, wherein the central axis of the starting screw passes through the center point of the 2D pedicle transverse surface, and wherein the bone screw insertion surface of the enveloping cylinder of the starting screw is within the bone screw insertion surface (10).

8. The method of claim 1, wherein the step of calculating the optimal screw trajectory (61) comprises the following steps: - Create a cylinder that approximates the pedicle bone screw (60) and place it in the deformed vertebral model (47) along the initial screw axis calculated based on the bone screw insertion surface (10) and the transverse surface of the pedicle. - Calculate the bone material density within the cylinder of the approximate pedicle screw (60) in the deformed vertebral model (47).

9. The method of claim 8, further comprising the following steps: - A cylindrical approximation of the pedicle screw (60) in a vertebral model (47) excluding perforation and deformation; and - The bone material properties in the cylinder of the approximate pedicle screw (60) are extracted from image voxels in the deformed vertebral model (47).

10. The method of claim 1, wherein the step of calculating the optimal screw trajectory (61) of the pedicle screw (60) comprises: - Use the bone screw insertion surface (10) and the transverse surface of the pedicle to identify the space of possible projection planes of deformed vertebral models (47); - Scan the space of possible projection planes to determine a set of intersecting region density projections. ; - Scan the density projection of the cross region And calculate the corresponding projected bone mineral density score. ;and - Determine the optimal screw trajectory as having a direction perpendicular to the projection plane. The direction of the projection plane corresponds to the projected bone mineral density score. The highest score in the projected bone mineral density score .

11. The method of claim 10 further comprises using a personalized three-dimensional geometric model (45) of the target bone region and an optimal screw trajectory to determine the length of the bone screw as the distance between the bone screw insertion surface (10) and the transverse surface of the pedicle in the direction of the optimal screw trajectory.

12. The method of claim 10, wherein the step of identifying the space of possible projection planes of the deformed vertebral model (47) comprises: - A sphere (S) is defined in the deformed vertebral model (47) around the center of the target bone region; - Define the space of possible projection planes as a set of planes perpendicular to the surface of the sphere (S); and - The space of possible projection planes is restricted to a group of projection planes that intersect between the bone screw insertion surface (10) and the transverse surface of the pedicle.

13. The method of claim 10, wherein the space of possible projection planes is scanned to determine a set of intersection region density projections. The steps include: - Select multiple discretely distributed projection planes within the space of possible projection planes. ; - For multiple discretely distributed projection planes The intersection area of ​​each defined bone screw insertion surface (10) and the transverse surface of the pedicle. ;and - For the intersection area Inner and perpendicular to the projection plane Summing all voxels of the deformed vertebral model (47) to obtain the corresponding cross-region density projection. The voxels represent bone mineral density.

14. The method of claim 10, wherein, Scan the density projection of the cross region And calculate the corresponding projected bone mineral density score. Including the diameter of one or more possible bone screws Density projection of the intersecting areas within the demarcated area The voxels are summed for use in the corresponding cross projection. For each possible location of the bone screw inside, the optimal screw trajectory has a diameter across the screw. The central axis, the screw diameter corresponds to the projected bone mineral density score. The highest score of projected bone mineral density. .

15. The method according to claim 1, wherein, In the personalized three-dimensional geometric model (45) of the target bone region, the screw length, screw diameter with a predetermined maximum length boundary condition, and initial parameters of the predetermined maximum diameter are provided for the step of calculating the optimal screw trajectory (61) of the pedicle screw (60).

16. The method according to claim 2, wherein, In determining the screw length based on the body side surface (58) of the target bone region, a second threshold (22) is provided as a length safety distance.

17. The method according to claim 1, wherein, The deformation of the three-dimensional bone region model includes providing a sagittal plane (35) within the deformed vertebral model (47) to determine the transverse surface of the pedicle.

18. The method according to claim 1, wherein, The deformation of the three-dimensional bone region model includes determining the vertebral foramen (56) within the deformed vertebral model (47) to determine the transverse surface of the pedicle as the minimum bone material diameter on both sides of the vertebral foramen (56).

19. The method of claim 17, wherein, The sagittal plane (35) of the body of the target bone region provides a third threshold, the outline and tip of the pedicle screw (60) need not pass through the third threshold, so that either pedicle screw (60) can be introduced into the same body (57), or only one of the two pedicle screws (60) can be introduced into the same body (57) in such a way that the two pedicle screws (60) do not occupy the same position.

20. A data processing system comprising means for performing the method steps according to any one of claims 1 to 19.

21. A computer program product comprising instructions that, when executed by a computer, cause the computer to perform the steps of the method according to any one of claims 1 to 19.

Images