CN-115546450-B - Planning navigation system for lower limb fracture reduction operation

CN115546450BCN 115546450 BCN115546450 BCN 115546450BCN-115546450-B

Abstract

The invention discloses a planning navigation system for a lower limb fracture reduction operation. The system comprises a medical image processing module, an operation planning module, a navigation module and a navigation module, wherein the medical image processing module is used for dividing a proximal bone and a distal bone from a lower limb fracture image and reconstructing a fractured bone three-dimensional surface model, the operation planning module is used for carrying out point cloud sampling on the fractured bone three-dimensional surface model to obtain a corresponding point cloud model, extracting a fracture section from the point cloud model, carrying out point cloud registration on the fracture section to obtain a coordinate transformation matrix of an initial position of the distal bone and a target position after the distal bone is reset, carrying out fracture reduction path planning based on the coordinate transformation matrix, and the fracture reduction path planning is used for guiding the distal bone to carry out translational rotation under the condition of keeping the proximal bone fixed to obtain a distal bone target position, and the navigation module is used for visually displaying a process of carrying out operation based on the fracture reduction path planning. The invention can improve the resetting precision, reduce the resetting force and reduce the damage of soft tissues such as muscles at the fracture of a patient.

Inventors

HE YUCHENG
WU GEYANG
HU YING
ZHANG PENG
QI XIAOZHI
ZHAO BAOLIANG

Assignees

中国科学院深圳先进技术研究院

Dates

Publication Date: 20260508
Application Date: 20221021

Claims (9)

1. A planning navigation system for lower limb fracture reduction surgery, comprising a medical image processing module, a surgery planning module and a navigation module, wherein: The medical image processing module is used for dividing a proximal bone and a distal bone from a lower limb fracture image and reconstructing a broken bone three-dimensional surface model, and comprises a distal bone three-dimensional surface model and a proximal bone three-dimensional surface model; The operation planning module is used for carrying out point cloud sampling on the fractured bone three-dimensional surface model to obtain a corresponding point cloud model, extracting a fracture section from the point cloud model, carrying out point cloud registration on the fracture section to obtain a coordinate transformation matrix of a distal bone initial position and a target position after the fracture section is reset, carrying out fracture reduction path planning based on the coordinate transformation matrix, and carrying out translational rotation on the distal bone under the condition of keeping the fixation of the proximal bone to obtain a distal bone target position; The navigation module is used for visually displaying the operation process based on the fracture reduction path planning; Wherein extracting the fracture section comprises: Performing kd-tree adjacent point search on the obtained point cloud model, and obtaining an adjacent point set of each point by selecting the kd-tree search radius; Calculating a certain point in the point cloud model Characteristic quantity of degree of variation of normal vector Expressed as: Wherein, the Representation points Normal vector of (c) and its vicinity point The included angle of the normal vector is defined, Representation points Adjacent point count of (a); Setting a threshold value Reserved, reserve And (5) taking the points at the broken bones as characteristic point clouds, and further obtaining the fracture section.
2. The system of claim 1, wherein extracting the fracture section comprises: Determining the long axis direction of the bone; calculating the midpoint of the point cloud model The angle between the normal to the long axis of the bone and the direction of the long axis of the bone is expressed as: Wherein, the Is that The normal vector of the point(s), Is that The component in the x-axis is, Is that The component in the y-axis is, Is that The component in the z-axis is, Representing a bone long axis direction vector; and taking the point with the angle smaller than the set threshold value as the characteristic point cloud of the broken bone, and further obtaining the fracture section.
3. The system of claim 1, wherein the fracture reduction path plan is obtained according to the following procedure: establishing a target coordinate system based on the coordinate transformation matrix Obtaining fracture reduction parameters, wherein the target coordinate system comprises an L axis, an S axis and a T axis, the L axis is the long axis direction of the bone passing through the origin, the S axis and the T axis are coordinate axes pointing to the outer side of the body, and the fracture reduction parameters comprise rotation angle parameters around the S axis Rotation angle parameter about T-axis Parameters of rotation angle about the L-axis And a parameter t, wherein Is a direction vector pointing to the external coordinate axis 1; selecting a stretching step value according to the occurrence range and degree of fracture, and setting a stretching distance parameter ; Performing a resetting process by taking the target position of the distal bone as a starting point and taking the initial position of the distal bone as an ending point, wherein the resetting process is respectively finished in sequence 、、 Resetting parameters and compensating for the stretch distance parameters And then finish Resetting parameters, collecting the current coordinate transformation matrix after each parameter is adjusted, adding the current coordinate transformation matrix into a verification queue, wherein l is a bone long axis direction vector, Is a direction vector pointing to the external coordinate axis 2; performing collision detection on the far-end bone position of each step in the verification queue, removing the path of collision between bones, and updating the iteration times Increasing the stretching distance, emptying the verification queue, and reentering the resetting process until collision detection passes; the position of each step of the far-end bone recorded in the verification queue is used as the planned optimal reset path.
4. A system according to claim 3, wherein the collision detection comprises: And traversing the points in the point cloud model, judging whether the points are inside the closed polyhedron of the broken bones by using an injection line method, if the points are not inside, continuing traversing until finishing, and if the points are inside, stopping traversing to indicate that the near-end bones and the far-end bones collide.
5. A system according to claim 3, wherein the target coordinate system The method is established according to the following process: selecting the point cloud gravity center of the fracture section as a target coordinate system Origin O of (2); selecting a distal bone passing through the origin is taken as the long axis direction L of the bone A Z axis of the coordinate system; Bone long axis direction through proximal bone Selecting a second coordinate axis as a T axis as a reference, wherein the second coordinate axis points to an external coordinate axis 1 and the direction vector thereof Is that ; Determining a third coordinate axis as an S axis according to the Cartesian coordinate system rule, wherein the third coordinate axis points to an external coordinate axis 2 and is a direction vector Is that ; The complete target coordinate system is finally obtained Expressed as: Wherein, the Is a direction vector pointing to the external coordinate axis 2, Is a direction vector pointing to the external coordinate axis 1.
6. The system of claim 5, wherein the bone long axis direction vector Obtained according to the following procedure: Calculating a point cloud normal vector after the fracture section is removed from the point cloud model, and obtaining a point cloud normal vector set; Randomly selecting and equally dividing the point cloud normal vector set into a first point cloud normal vector set and a second point cloud normal vector set, and further solving a bone long axis direction vector Expressed as: Wherein, the And The elements in the first point cloud normal vector set and the second point cloud normal vector set are respectively, and the number of the elements in the first point cloud normal vector set and the second point cloud normal vector set is 。
7. The system of claim 5, wherein the fracture reduction parameters are obtained according to the following procedure: Based on the coordinate transformation matrix And the target coordinate system Calculating the initial coordinate system position of the distal bone : In the target coordinate system For reference, calculate its relative position transformation matrix : Transforming the matrix according to the relative position Determining the values required to be adjusted for each parameter in the reverse fracture reduction process, including displacement parameters 、、 And the angle parameter is 、、 Wherein Is a displacement parameter pointing to the external coordinate axis 2, Is directed to the displacement parameter in the direction of the external coordinate axis 1, Is a displacement parameter in the long axis direction, Is a parameter of the rotation angle around the S axis, Is a rotation angle parameter around the T-axis.
8. The system of claim 1, wherein the fracture section is obtained by cleaning outliers, the outlier cleaning comprising: When the point cloud model is subjected to kd-tree search, each point is recorded simultaneously Distance to its neighboring point; and calculating the average distance between each point and the adjacent point, obtaining a standard deviation according to the average distance and each distance, and filtering out the points with the average distance larger than the standard deviation threshold value by setting a threshold value parameter.
9. A system according to claim 3, wherein the step value is selected to be 1 mm-2 mm, and the stretch distance parameter is set to be a step value and a number of iterations Is a product of (a) and (b).

Description

Planning navigation system for lower limb fracture reduction operation Technical Field The invention relates to the technical field of computer-aided medical treatment, in particular to a planning navigation system for a lower limb fracture reduction operation. Background Surgical navigation based on medical images refers to techniques for completing a series of procedures, such as reconstruction of a virtual image of a patient, planning of a surgery, tracking of the execution of the surgery, etc., using a patient medical image obtained before or during the surgery, computer graphics, and other sensor devices. Thanks to the large-scale clinical use of medical images such as CT and MRI and the advancement of modern personal computer storage modes, computer surgery planning navigation systems have been greatly developed and applied at present. A surgical navigation open source software system for clinical medicine developed by Pieper et al in the Massa Medicata Fermentata worker artificial intelligence laboratory integrates the functions of reading and writing medical images in various file formats, operating a 2D image system and a 3D image system, and the like, so that the system can be subjected to corresponding secondary development aiming at surgery to obtain a series of surgical navigation software systems with functions of automatic registration, three-dimensional reconstruction, surgical image guidance and the like. Currently, the reduction of the lower limb fracture is usually clinically finished by a professional doctor through experience in hands, and the doctor usually adopts a repeated try method without assistance of a navigation system because the obstacle on the reduction path cannot be detected, so that time and labor are wasted. To avoid an inter-bone collision during reduction, the physician may perform a certain amount of axial stretching, which may cause the lower limb to exert corresponding forces, overcoming which may add more pressure to the surrounding soft tissue that has been damaged, while also requiring more physical effort from the surgeon, which may make it more difficult to accomplish the required movement of the fractured bone. In the prior art, few navigation systems are used for planning the lower limb fracture reduction operation, and the planning navigation system capable of solving the clinical problems is still in a vacant state. For the operation planning navigation system, a part of operation planning systems facing to the bone surgery are put into clinical use, and especially on some joint replacement surgeries, a doctor is helped to complete operation path planning or replacement part selection through preoperative medical image analysis and simulated surgery effect. However, the planning systems for fracture reduction surgery are still less, and most of the planning systems still rely on experience judgment of doctors to complete the surgery due to the diversity of lower limb fracture conditions and great difficulty in reduction operation. Aiming at the aspect of lower limb fracture reduction planning, a part of researches adopt medical images which generate mirror images by utilizing contralateral bones, and the target position of a mobile end (a far-end bone) is predicted by registering a fixed end (a near-end bone), so that the reduction planning is finished, but the scheme has the effects of special conditions of incomplete bilateral symmetry of a human body, fracture of contralateral bones and the like on planning results. In another scheme, a corresponding template skeleton model is established by adopting a statistical model obtained by a plurality of groups of healthy skeletons, so that the reset planning is finished, but the scheme has strong dependence on the quantity and quality of healthy skeleton data, and the accuracy and stability of a planning result are greatly influenced for special individual conditions. Disclosure of Invention The invention aims to overcome the defects of the prior art and provide a planning navigation system for lower limb fracture reduction surgery. The system comprises a medical image processing module, a surgery planning module and a navigation module, wherein: The medical image processing module is used for dividing a proximal bone and a distal bone from a lower limb fracture image and reconstructing a broken bone three-dimensional surface model, and comprises a distal bone three-dimensional surface model and a proximal bone three-dimensional surface model; The operation planning module is used for carrying out point cloud sampling on the fractured bone three-dimensional surface model to obtain a corresponding point cloud model, extracting a fracture section from the point cloud model, carrying out point cloud registration on the fracture section to obtain a coordinate transformation matrix of a distal bone initial position and a target position after the fracture section is reset, carrying out fracture reduction path planni