A planning and navigation system for lower limb fracture reduction surgery

Abstract

The application discloses a planning navigation system for lower limb fracture reduction surgery. The system comprises a medical image processing module for segmenting a proximal bone and a distal bone from a lower limb fracture image and reconstructing a three-dimensional surface model of a broken bone; a surgery planning module for point cloud sampling on the three-dimensional surface model of the broken bone to obtain a corresponding point cloud model; extracting a fracture section from the point cloud model; obtaining a coordinate transformation matrix of an initial position of the distal bone and a target position of the distal bone after reduction by point cloud registration on the fracture section; performing fracture reduction path planning based on the coordinate transformation matrix, the fracture reduction path planning being used for guiding translation and rotation of the distal bone to obtain the target position of the distal bone under the condition that the proximal bone is fixed; and a navigation module for visually displaying a surgery process based on the fracture reduction path planning. The application can improve reduction accuracy, reduce reduction force and reduce soft tissue damage such as muscle damage at a fracture of a patient.

Description

Technical Field

[0001] This invention relates to the field of computer-aided medical technology, and more specifically, to a planning and navigation system for lower limb fracture reduction surgery. Background Technology

[0002] Medical image-based surgical navigation refers to a technology that utilizes preoperative or intraoperative patient medical images, computer graphics, and other sensor devices to reconstruct a virtual image of the patient, plan the surgery, and track its execution. Thanks to the widespread clinical use of medical images such as CT and MRI, and advancements in modern personal computer storage, computer-aided surgical planning and navigation systems have seen tremendous development and application.

[0003] Developed by Pieper and colleagues at MIT's Artificial Intelligence Laboratory, an open-source surgical navigation software system for clinical medicine integrates functions such as reading and writing medical images in various file formats and manipulating 2D and 3D image systems. This allows the system to be further developed for specific surgeries to provide a range of functions, including automatic registration, 3D reconstruction, and surgical image guidance.

[0004] Currently, lower limb fracture reduction is typically performed manually by experienced surgeons. Without navigation system assistance, obstructions along the reduction path cannot be detected, leading to a time-consuming and laborious "trial and error" approach. To avoid interosseous collision during reduction, surgeons apply axial stretching, which generates forces in the lower limb. Overcoming these forces increases pressure on the surrounding soft tissues and requires more physical exertion from the surgeon, making it even more difficult to move the fractured bone. Existing navigation systems specifically designed for lower limb fracture reduction surgery are scarce, and systems capable of addressing these clinical challenges are entirely lacking.

[0005] Regarding surgical planning and navigation systems, some systems designed for orthopedic surgeries have been put into clinical use, particularly in joint replacement surgeries. These systems use preoperative medical image analysis and simulated surgery to help surgeons plan surgical pathways or select replacement components. However, planning systems for fracture reduction surgeries are still relatively few. Due to the diversity of lower limb fractures and the complexity of reduction procedures, most surgeries still rely on the surgeon's experience and judgment.

[0006] Regarding lower limb fracture reduction planning, some studies have employed mirrored medical images generated from the contralateral bone and registered the fixed end (proximal bone) to predict the target position of the movable end (distal bone), thereby completing the reduction planning. However, this approach is susceptible to special cases such as imperfect left-right symmetry in the human body and fractures occurring on the contralateral side, which can affect the planning results. Another approach uses statistical models derived from multiple sets of healthy bones to establish corresponding template bone models for reduction planning. However, this approach is highly dependent on the quantity and quality of healthy bone data, and the accuracy and stability of the planning results can be significantly affected by specific individual cases. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide a planning and navigation system for lower limb fracture reduction surgery. This system includes a medical image processing module, a surgical planning module, and a navigation module, wherein:

[0008] The medical image processing module is used to segment the proximal and distal bones from the lower limb fracture image and reconstruct a three-dimensional surface model of the fracture bone, including a three-dimensional surface model of the distal bone and a three-dimensional surface model of the proximal bone.

[0009] The surgical planning module is used to: sample the three-dimensional surface model of the fractured bone to obtain a corresponding point cloud model; extract the fracture surface from the point cloud model; obtain the coordinate transformation matrix of the initial position of the distal bone and its target position after reduction by performing point cloud registration on the fracture surface; and plan the fracture reduction path based on the coordinate transformation matrix. The fracture reduction path planning is used to guide the translation and rotation of the distal bone to obtain the target position of the distal bone while keeping the proximal bone fixed.

[0010] The navigation module is used to visually display the surgical process based on the fracture reduction path planning.

[0011] Compared with existing technologies, the advantages of this invention are as follows: After a lower limb fracture, current clinical practice typically requires surgical reduction of the broken bone to its original position, followed by fixation and healing. The surgical planning and navigation system provided by this invention uses preoperative images of the lower limb fracture patient as input to perform three-dimensional reconstruction of the fracture site, extraction of the fracture surface, and fracture reduction planning. It also considers factors such as potential interosseous collisions and passive stress on soft tissues during reduction to obtain the optimal path for fracture reduction, thereby guiding the surgeon or surgical robot to successfully complete a precise lower limb fracture reduction surgery. Compared to traditional fracture reduction surgery relying on the surgeon's experience, this invention achieves higher reduction accuracy, less damage to nearby muscles and other soft tissues, and significantly reduces the physical and mental burden on the surgeon.

[0012] Other features and advantages of the invention will become clear from the following detailed description of exemplary embodiments of the invention with reference to the accompanying drawings. Attached Figure Description

[0013] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments of the invention and, together with their description, serve to explain the principles of the invention.

[0014] Figure 1 This is a schematic diagram of a planning and navigation system for lower limb fracture reduction surgery according to an embodiment of the present invention;

[0015] Figure 2 This is a schematic diagram illustrating the selection of the long axis of a bone according to an embodiment of the present invention;

[0016] Figure 3 This is a schematic diagram of an initial coordinate system and a target coordinate system according to an embodiment of the present invention;

[0017] Figure 4 This is a flowchart of a reset planning algorithm according to an embodiment of the present invention;

[0018] Figure 5 This is a schematic diagram of a fracture reduction process according to an embodiment of the present invention;

[0019] Figure 6 This is a framework diagram of a fracture reduction surgery navigation system according to an embodiment of the present invention;

[0020] Figure 7 This is a schematic diagram of a repositioning surgery planning module according to an embodiment of the present invention;

[0021] Figure 8 This is a schematic diagram of fracture reduction path planning according to an embodiment of the present invention. Detailed Implementation

[0022] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the invention.

[0023] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.

[0024] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0025] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0026] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0027] This invention relates to a planning and navigation system for lower limb fracture reduction surgery, used to assist surgeons in preoperative surgical planning and intraoperative surgical navigation, thereby improving the surgeon's skill level. See also: Figure 1 As shown, the system as a whole includes a medical image processing module, a surgical planning module, and a navigation module. The medical image processing module is used for segmentation and 3D reconstruction of medical images. The surgical planning module is used to plan the fracture reduction path through processes such as point cloud sampling, fracture surface extraction, and fracture surface registration. The navigation module is used to visually display the surgical execution process.

[0028] I. About the Medical Image Processing Module

[0029] The medical image processing module first performs conventional image segmentation on the input lower limb fracture CT image. For example, it segments the bone into proximal and distal parts using a CT threshold. The segmentation only needs to ensure the fracture site is intact and has a certain axial length; other areas of the fractured bone are not required to be segmented. After obtaining the local medical image, the MarchingCubes algorithm is used to perform 3D reconstruction of the 2D medical tomographic image, resulting in a set of 3D surface models of the fractured bone: one 3D surface model of the distal bone and one 3D surface model of the proximal bone. In one embodiment, the entire 3D reconstruction step can be implemented based on VTK (a visualization tool library).

[0030] II. Regarding the surgical planning module

[0031] 1) Point cloud sampling

[0032] After obtaining the 3D surface model of the fractured bone, which provides the necessary spatial location information, point cloud sampling methods can be used to obtain the original data to be registered. For example, point cloud sampling can be performed using PCL (a cross-platform open-source C++ point cloud processing programming library) to obtain the corresponding point cloud model. To avoid losing cross-sectional feature data and significantly affecting registration accuracy, a uniform sampling method can be used to retain more cross-sectional features. Although this increases the amount of data to some extent, it ensures accuracy in subsequent registration steps.

[0033] 2) Extract the fracture surface

[0034] To achieve point cloud registration of fracture fracture surfaces, the registration region of the point cloud must first be determined. Since the fracture surfaces of distal and proximal bones are actually planes or curved surfaces before fracture, this characteristic can be used to determine the registration region, i.e., the two planes or curved surfaces where the fracture occurs. These two fracture fracture surfaces can be extracted using various methods, such as the curvature threshold method or the bone long axis normal filtering method.

[0035] Specifically, for extracting cross-sectional point clouds using the curvature thresholding method, the bone surface undergoing a fracture will have a region of abrupt curvature change around the fracture surface. This strongly curved fracture region can be distinguished from the smooth surface of the bone to obtain the characteristic fracture line. Changes in curvature can be reflected in changes in the angle between adjacent normal vectors. For example, the normal vector of each point can be defined as the local plane normal vector fitted to the neighborhood of that point. The gentler the curvature change, the smaller the angle between adjacent normal vectors; the more drastic the curvature change, the larger the angle between adjacent normal vectors. For instance, performing a kd-tree neighbor search on the acquired point cloud data, by selecting the kd-tree search radius r, the set of neighboring points {p1, p2, p3... p...} of point p can be obtained. k Let the arithmetic mean of the angles between a point and the normal vectors of its neighbors be defined as the angles between a point and the normal vectors of its neighbors in the point cloud. i The characteristic quantity f of the degree of change of the normal vector i The calculation formula is as follows:

[0036]

[0037] Where, θ ij Point p i The normal vector of its neighboring point p j The angle between the normal vectors, where k represents point p. i The number of neighboring points.

[0038] This feature can be used to measure the curvature of the plane fitted to each point; a larger feature indicates a greater change in curvature. By setting a threshold ε, point cloud portions resembling flat bone surfaces are filtered out, retaining f. i For points greater than ε, the entire bone model point cloud is traversed using a kd-tree search method to obtain the feature point cloud at the fracture site.

[0039] For the bone long axis normal filtering method, due to the complexity of fractures, in cases where the fracture surface point cloud has gaps or discontinuities, the fractured bone may not have obvious fracture line characteristics. In other words, in some cases, it is difficult to find a suitable threshold ε to separate the outer region of the fracture surface using curvature changes, thus affecting the registration accuracy of the curvature thresholding method. Therefore, in one embodiment, the method of extracting the cross-sectional point cloud using the angle between the bone long axis and the normal is used. Compared with the curvature thresholding method, this method extracts the entire cross-sectional feature, including points on the entire point cloud cross-section. Specifically, it includes two parts: determining the direction of the bone long axis and extracting the cross-sectional point cloud. The "bone long axis" is a vector representing the axial direction of the lower limb bones, and its position is not explicitly required. Therefore, when determining the bone long axis, a method of manually selecting two points on the side of the bone to generate the bone long axis vector can also be used. Figure 2 As shown, the long axis l of the bone is obtained by manually selecting two points on the side of the bone.

[0040] Because the lower limb bones are roughly cylindrical, the normals of most points on the bone lateral wall point cloud are approximately perpendicular to the long axis of the bone, while the angle between the normals of points on the fracture surface point cloud and the long axis of the bone is smaller. Utilizing this characteristic, a suitable empirical threshold can be set to classify the point cloud; points with an angle greater than the threshold are considered to be located on the lateral wall, otherwise on the fracture surface. The method for obtaining the point normal vector is similar to the curvature threshold method mentioned above: a kd-tree search is used to traverse the point cloud and obtain the neighbor set of each point. The point set is then used to fit a plane using the least squares method to obtain p. i The normal vector v of the point i The angle between the normal and the major axis of the point cloud of the fractured bone model can be obtained by calculating the cosine of the angle:

[0041]

[0042] Among them, v i It is p i The normal vector of a point, v ix It is v i The component on the x-axis, v iy It is v i The component on the y-axis, v iz It is v i In the z-axis component, l represents the direction vector of the bone's long axis.

[0043] It should be noted that in practical applications, the curvature threshold method or the bone long axis normal filtering method can be selected to extract the fracture surface, or a combination of these two methods or other methods can be used to extract the fracture surface.

[0044] Furthermore, to improve the accuracy of fracture surface extraction, outlier removal can be performed. This is because both the curvature thresholding method and the bone long axis normal filtering method will produce outliers in the point cloud, which need to be removed by a filtering method. When performing a kd-tree search using the above method, each point p can be recorded simultaneously. i The distance between each point and its neighboring points. Assuming that the distance between each point and its neighboring points follows a Gaussian distribution, the average distance is the mean. The standard deviation is calculated based on the mean and each distance. A threshold parameter is set to filter out points whose average distance is greater than the standard deviation threshold, thus completing the filtering of the feature point cloud.

[0045] 3) Point cloud registration of fracture cross-section

[0046] Registration typically refers to matching images acquired from different sensors, at different times, or under different conditions. Similarly, 3D point cloud registration generally involves matching point clouds in different spatial coordinates to ensure a one-to-one correspondence between each point's coordinates. However, after a fracture, the distal and proximal bone models are not the same object, and theoretically, the sampled point clouds do not have a one-to-one correspondence. But in this embodiment of the invention, the characteristic that fracture surfaces are originally on the same plane is utilized to achieve surface matching and alignment using point cloud registration. For example, ICP (Iterative Closest Point) algorithm is used for point cloud registration. This method does not require a correspondence between two sets of point sets as parameter data. Starting from a random or given initial correspondence, iterative calculation of the optimized registration transformation matrix between the two sets of point sets is performed using optimization analysis methods. For example, in one embodiment, the bone cross-section point clouds selected by the curvature threshold method and the bone long axis normal filtering method are registered separately. By comparing the ICP registration errors, the fracture cross-section obtained by the method with the smaller error is selected as the registration point cloud data, resulting in the coordinate transformation matrix T of the distal bone and its target position after reduction. Point cloud registration can be achieved based on PCL (point cloud library).

[0047] 4) Fracture reduction pathway planning

[0048] First, the parameters of the repositioning path planning algorithm are determined. Using the coordinate transformation matrix obtained through registration, the distal bone is translated and rotated while maintaining proximal bone fixation to obtain its target position, i.e., the target coordinate system C where the proximal and distal bones are located after repositioning. t The centroid of the point cloud used for registration is selected as C. t The origin O is represented here by a vector o(o) in the original coordinate system of the 3D point cloud space. x ,o y ,o z Simultaneously, the direction of the long axis of the bone passing through the origin is selected as C. tThe z-axis of the coordinate system is used, but the bone long axis selected during the extraction of the fracture point cloud is not the accurate direction of the bone long axis. Therefore, it is necessary to redetermine the bone long axis l in a strictly defined sense. For example, the bone long axis can be determined using the normal vector data of the point cloud after the fracture surface has been separated, because after removing the fracture surface, most points are located on the outer surface of the bone. The unit vector orthogonal to the normal vectors v1 and v2 of any two points can be used as the bone long axis.

[0049]

[0050] Among them l 12 This only represents the major axis determined by these two normal vectors. The obtained set of point cloud normal vectors is randomly divided into two sets of n elements each. By traversing the normal vectors in the two sets, the optimal direction of the bone major axis l is determined.

[0051]

[0052] The other two coordinate axes in the coordinate system are perpendicular to l. Using the major axis l' of the other bone as a reference, a second coordinate axis is selected, which is the direction vector t perpendicular to the plane formed by the two broken bones. It generally points outwards, and here it is defined as coordinate axis 1 pointing outwards, with direction vector t being l×l'. According to the Cartesian coordinate system rules, the third coordinate axis, i.e., coordinate axis 2 pointing outwards, has a direction vector s of l×t. Finally, the complete target Cartesian coordinate system is obtained:

[0053]

[0054] The coordinate transformation matrix T obtained from registration is the transformation matrix from the initial position of the distal bone to the target position. After inverting it, the initial coordinate position C of the distal bone can be obtained from the coordinate system of the target position. i :

[0055] C i =T -1 ·C t (6)

[0056] See the initial and target coordinate systems for the distal bone. Figure 3 As shown.

[0057] After establishing the coordinate system, several parameters describing the position of the distal bone are determined for fracture reduction path planning. These parameters, while describing the relative position, are also used to adjust their values ​​to complete the entire reduction process. For example, three positional parameters (or displacement parameters) and three angular parameters relative to the coordinate system are used as the basis for adjustment. First, the target coordinate system C... t Using this as a reference, obtain its relative position transformation matrix T':

[0058]

[0059] In the target coordinate system, the distal bone is moved to its initial position using a position transformation matrix T'. The position parameters in the relative position transformation matrix are the values ​​that need to be adjusted for each parameter during the repositioning reverse process. The translation amount corresponding to T' is the corresponding translation parameter t. s t t t l The angle parameters need to be calculated using the rotation matrix R in T' to obtain the Euler angles. For example, according to the reduction procedure described below, the Euler angles are calculated by rotating around the s, t, and l axes sequentially, with the corresponding angle parameters being α, β, and γ, respectively. All obtained fracture reduction parameters and their descriptions are shown in Table 1.

[0060] Table 1. Fracture Reduction Parameters

[0061]

[0062] Furthermore, in fracture reduction path planning, in order to avoid collision between the two bones during the reduction process, it is preferable to optimize the reduction path at the planned location through collision detection and dynamic analysis of the reduction muscles.

[0063] Specifically, considering the possibility of bone overlap or intersecting during fractures, a collision detection algorithm is introduced to avoid collisions during the reduction process. In a two-dimensional plane, a ray-drawing method can be used to determine if a point is inside a polygon. A ray is drawn from the target point, and the number of intersections between this ray and all edges of the closed polygon is recorded. An odd number indicates the point is inside, and an even number indicates it is outside. For the three-dimensional model of this invention, the fracture bone's three-dimensional data model is actually composed of multiple triangular facets, thus simulating a closed surface model of the fracture bone entity. Point cloud sampling yields the position data of various points on the fracture bone surface. Since the amount of data obtained through uniform sampling is sufficient, the sampled points can approximate any position on the fracture bone. The points in the point cloud model are traversed, and for each point, a ray-drawing method is used to determine if it is inside the closed polyhedron of the fracture bone. If the result is that it is not inside, the traversal continues until the end, indicating no collision has occurred. If the result is that it is inside, the traversal terminates, indicating a collision between the two bones.

[0064] The following section will introduce the fracture reduction path planning algorithm in detail, which includes dynamic analysis of the reduction muscles.

[0065] To avoid collisions between the two bones during repositioning, a stretching distance parameter `dis` is set in the planning algorithm. This parameter, along with the collision detection algorithm, iteratively increases the repositioning plan. A constant step value is set for the parameter `dis`. A smaller step value results in a path closer to the optimal path, while a larger step value reduces the number of algorithm iterations.

[0066] Combination Figure 4 and Figure 5 As shown, the specific reset path planning algorithm is as follows:

[0067] Step S1: First, acquire and store the point cloud model of the fractured bone, the polydata model of the three-dimensional solid surface, and the position transformation matrix T obtained through registration. Then, establish the target coordinate system C based on the determined fracture reduction parameters. t And obtain the corresponding reset parameters;

[0068] Step S2: Select the step value according to the extent and severity of the fracture. For example, it can be selected as 1mm to 2mm. Set the stretching distance parameter dis, which is the product of the step value and the current iteration number n.

[0069] Step S3: Enter the reset process. To ensure that the position after reset is not affected by the calculation process, the entire reset algorithm performs the reset process in reverse. That is, starting from the target position of the distal bone and ending at the initial position of the distal bone, the parameters l+dis, γ, and (s,t) are reset in sequence. Then, the previously set stretching distance parameter dis is compensated, and finally the (α,β) parameter is reset. After each parameter is adjusted, the current coordinate transformation matrix is ​​collected and added to the verification queue.

[0070] Step S4 involves performing collision detection on the distal bone positions of each step in the verification queue. If a bone collision occurs, it indicates that the current path is unusable. The iteration count n is updated, the stretching distance is increased, the verification queue is cleared, and the reset process in step S3 is re-entered. This process continues until the collision detection passes. The position of each distal bone step recorded in the verification queue is the planned optimal reset path. After the initial planning process is completed, interpolation collision detection is added. If the collision detection passes, the planning is completed; otherwise, the iteration count needs to be increased. The purpose of not adding interpolation collision detection in the initial planning step is to reduce the computational load of the planning algorithm. In the initial planning, collision detection only needs to be performed at key points, improving efficiency and reducing the time spent on planning calculations.

[0071] It should be noted that in the description of the coordinate axes, the bone long axis, C... t The z-axis and l-axis of the coordinate system have the same meaning; the second coordinate axis, external coordinate axis 1, and t-axis have the same meaning; the third coordinate axis, external coordinate axis 2, and s-axis have the same meaning, unless the context otherwise requires.

[0072] In summary, fracture reduction path planning includes collision detection algorithms during fracture reduction, soft tissue stress analysis during reduction, and the introduction of stretching distance parameters. This can solve the problem of bone collision during reduction. It also takes into account the passive stress of muscles during reduction, so that the final reduction path minimizes soft tissue damage near the fracture.

[0073] III. Lower Limb Fracture Reduction Surgical Planning and Navigation System

[0074] By integrating medical image processing, surgical planning, and visualization navigation modules, a planning and navigation system for lower limb fracture reduction surgery can be implemented. This system allows for direct interaction with physicians in clinical applications. For example, the entire system can be developed using C++ code, with the interface framework built on the QT platform, primarily utilizing toolkits such as VTK, ITK, and PCL for construction. The system mainly includes medical image processing and segmentation modules, surgical planning modules, and navigation modules, and works with VTK to achieve image interaction, such as... Figure 6 As shown. It should be understood that the present invention does not limit the way or number of modules are divided, as long as the above functions can be achieved.

[0075] The surgical planning module for fracture reduction includes an interactive area for setting important parameters, facilitating the adjustment of experimental parameters and enabling the clinical use of the fracture reduction navigation system in various fracture conditions. Detailed information about the surgical planning module is as follows: Figure 7 As shown.

[0076] The entire system can complete the entire workflow from acquiring the patient's CT medical images to completing fracture reduction planning and subsequent surgical navigation. The specific operation process is as follows: Figure 1 As shown.

[0077] To further verify the effectiveness of this invention, simulation experiments were conducted. In the experiments, the bone long axis normal filtering method was used to extract the cross-sectional point cloud. Since the selection of the bone long axis is generally done manually, this means that the parameters used for registration of the same bone model are different each time, resulting in different registration results. Furthermore, the point cloud gaps caused by this method may also affect the registration results. Therefore, while ensuring that the point clouds to be registered are identical and other parameters are the same, registration experiments were conducted multiple times using different bone long axis directions to verify the registration error. During registration, the selected angle of the long axis was within a deviation of 30°, a condition that is entirely achievable when clinically selecting the bone long axis. Based on the position transfer matrix obtained from the experiments, the root mean square error in each direction was calculated, and the results are shown in Table 2.

[0078] Table 2 Registration Position Error Data Table

[0079]

[0080] Similarly, the Euler angles of rotation around each axis were calculated based on the position transfer matrix, and the root mean square error of the corresponding rotation angles was calculated. The results are shown in Table 3.

[0081] Table 3 Registration Angle Error Data Table

[0082]

[0083] Experimental results show that manually selecting the major axis does not significantly affect the registration results, and the errors in multiple registration results are all within 0.2 mm, verifying the feasibility and stability of cross-sectional registration. Furthermore, based on the registration results, the provided system is used for repositioning planning, obtaining a sequence of position transformation matrices relative to the initial position of the distal bone in the CT coordinate system, such as... Figure 8 As shown, this sequence represents the planning result and has been interpolated. After spatial registration, it can serve as the target position sequence for the surgical execution mechanism, guiding it to complete the surgery. The planned repositioning path was visualized frame-by-frame through analogy with the surgeon's repositioning method, demonstrating the feasibility of the system's surgical planning.

[0084] In summary, compared with the prior art, the present invention has the following technical effects:

[0085] 1) Currently, lower limb fracture reduction in clinical practice is usually performed manually by surgeons based on experience. This requires a high level of surgical experience from the surgeon, and without visual navigation, reduction is prone to interosseous collision and deviation from the final reduction position. Furthermore, due to muscle traction during reduction, when the fracture has caused a large displacement, the required traction force is significant, greatly consuming the surgeon's physical strength and energy, and also increasing the risk of secondary injury to the patient's muscles and other soft tissues. The surgical planning and navigation system provided by this invention, however, is based on the final target result obtained through registration after reconstruction from CT medical image data, achieving a higher accuracy than the target position found by the surgeon based on experience. Simulations show that the reduction steps designed in this invention can significantly reduce reduction force and protect the patient's muscles and other soft tissues.

[0086] 2) In studies on the reduction of other lower limb fractures, some studies have used mirror images of the contralateral bone to determine the target location. However, this approach is susceptible to special cases such as imperfect left-right symmetry or fractures on the contralateral side, rendering it ineffective or inaccurate. Another approach uses statistical models derived from multiple healthy bones to build a corresponding template bone model, predicting the position of the distal bone in healthy individuals to determine the target location. However, this approach is highly dependent on the quantity and quality of healthy bone data. In contrast, this invention determines the target location in the reduction plan based on cross-sectional point clouds, unaffected by contralateral bone or other individual characteristics. Furthermore, it does not require extensive preoperative model training and construction using large amounts of healthy bone data. Therefore, this invention is more practical and stable.

[0087] 3) This invention introduces a method for reconstructing CT images and sampling point clouds. The point cloud is then processed for cross-sectional extraction and registration to determine the location of the distal bone target. This method is unaffected by individual patient factors and focuses solely on the fracture cross-section itself, thus ensuring applicability and stability. It also incorporates reduction collision detection and soft tissue stress analysis during the reduction process to obtain the optimal fracture reduction path, effectively solving problems in clinical reduction surgery.

[0088] This invention can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of the invention.

[0089] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example, but not limited to, electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0090] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0091] The computer program instructions used to perform the operations of this invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, Python, etc., and conventional procedural programming languages ​​such as "C" or similar languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions. This electronic circuitry can execute the computer-readable program instructions to implement various aspects of the invention.

[0092] Various aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0093] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0094] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0095] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions. It will be known to those skilled in the art that implementation in hardware, implementation in software, and implementation using a combination of software and hardware are equivalent.

[0096] The various embodiments of the present invention have been described above. These descriptions are exemplary and not exhaustive, and are not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein. The scope of the invention is defined by the appended claims.

Claims

1. A planning and navigation system for lower limb fracture reduction surgery, comprising a medical image processing module, a surgical planning module, and a navigation module, wherein: The medical image processing module is used to segment the proximal and distal bones from the lower limb fracture image and reconstruct a three-dimensional surface model of the fracture bone, including a three-dimensional surface model of the distal bone and a three-dimensional surface model of the proximal bone. The surgical planning module is used to: sample the three-dimensional surface model of the fractured bone to obtain a corresponding point cloud model; extract the fracture surface from the point cloud model; obtain the coordinate transformation matrix of the initial position of the distal bone and its target position after reduction by performing point cloud registration on the fracture surface; and plan the fracture reduction path based on the coordinate transformation matrix. The fracture reduction path planning is used to guide the translation and rotation of the distal bone to obtain the target position of the distal bone while keeping the proximal bone fixed. The navigation module is used to visually display the surgical process based on the fracture reduction path planning; Extracting the fracture surface includes: Perform a kd-tree neighbor search on the acquired point cloud model, and obtain the set of neighboring points for each point by selecting the kd-tree search radius; Calculate a point in the point cloud model Feature quantities of the degree of change of the normal vector , is represented as: in, Point The normal vector and its neighboring points The angle between the normal vectors, Point The number of neighboring points; Set threshold ,reserve The points are used as feature point clouds at the fracture site to obtain the fracture surface.

2. The system according to claim 1, characterized in that, Extracting the fracture surface includes: Determine the direction of the long axis of the bone; Calculate the points in the point cloud model The angle between the normal to the bone and the direction of the long axis is expressed as: in, yes The normal vector of a point yes Components on the x-axis, yes Components on the y-axis yes The component along the z-axis, Represents the direction vector of the bone's long axis; Points with angles less than a set threshold are used as feature point clouds at the fracture site, thereby obtaining the fracture surface.

3. The system according to claim 1, characterized in that, The fracture reduction path planning is obtained according to the following process: Establish the target coordinate system based on the coordinate transformation matrix. The fracture reduction parameters are obtained, wherein the target coordinate system includes an L-axis, an S-axis, and a T-axis. The L-axis is the direction of the long axis of the bone passing through the origin, and the S-axis and T-axis are coordinate axes pointing outward from the body. The fracture reduction parameters include rotation angle parameters about the S-axis. Rotation angle parameters around the T-axis Rotation angle parameters around the L-axis and parameters t ,in It is the direction vector pointing to the outer coordinate axis 1; The step value of the stretching is selected based on the extent and severity of the fracture, and the stretching distance parameter is set. ; The reduction process is performed starting from the target position of the distal bone and ending at the initial position of the distal bone, with the following steps completed sequentially. , , Reset the parameters and compensate for the stretching distance parameter. to complete The parameters are reset, and the current coordinate transformation matrix is ​​collected and added to the verification queue after each parameter adjustment. l It is the vector along the long axis of the bone. It is the direction vector pointing to the outer coordinate axis 2; For each step in the validation queue, collision detection is performed on the distal bone position, and paths where inter-bone collisions occur are removed. The iteration count is then updated. Increase the stretching distance, clear the verification queue, and re-enter the reset process until the collision detection passes; The position of each step of the distal bone recorded in the verification queue is used as the planned optimal repositioning path.

4. The system according to claim 3, characterized in that, The collision detection includes: The points in the point cloud model are traversed. For each point, the ray-drawing method is used to determine whether it is inside the closed polyhedron of the fracture. If the result is that it is not inside, the traversal continues until the end; if the result is that it is inside, the traversal is terminated, indicating that the proximal and distal bones have collided.

5. The system according to claim 3, characterized in that, The target coordinate system Established according to the following procedure: The centroid of the point cloud of the fracture surface is selected as the target coordinate system. The origin O; The long axis direction L of the distal bone passing through the origin is selected as... The Z-axis of the coordinate system; Through the long axis direction of the proximal bone As a reference, a second coordinate axis is selected as the T-axis. This second coordinate axis points outward from the body's outer coordinate axis 1, and its direction vector... for ; According to the Cartesian coordinate system rules, the third coordinate axis is determined as the S-axis. This third coordinate axis points outward from the body, and its direction vector is [missing information]. for ; The final complete target coordinate system Represented as: in, It is the direction vector pointing to the outer coordinate axis 2. It is the direction vector pointing to the outer coordinate axis 1.

6. The system according to claim 5, characterized in that, The direction vector of the long axis of the bone Obtained through the following process: Calculate the point cloud normal vectors after removing the fracture surface from the point cloud model to obtain the point cloud normal vector set; The point cloud normal vector set is randomly divided into a first point cloud normal vector set and a second point cloud normal vector set, and then the bone long axis direction vector is calculated. , is represented as: in, and These are elements from the first point cloud normal vector set and the second point cloud normal vector set, respectively. The number of elements in both the first and second point cloud normal vector sets is [number missing]. .

7. The system according to claim 5, characterized in that, The fracture reduction parameters were obtained according to the following process: Based on the coordinate transformation matrix and the target coordinate system Calculate the initial coordinate system position of the distal bone. : Using the target coordinate system Using [a specific reference] as a reference, calculate its relative position transformation matrix. : Based on the relative position transformation matrix Determine the values ​​that need to be adjusted for each parameter during the reverse reduction process, including displacement parameters. , , and angle parameters are , , ,in It refers to the displacement parameter in the direction of coordinate axis 2, which points to the outside of the body. It refers to the displacement parameter in the direction of coordinate axis 1 on the outer side of the body. These are displacement parameters along the major axis. It is a parameter representing the rotation angle around the S-axis. It is the rotation angle parameter around the T-axis.

8. The system according to claim 1, characterized in that, The fracture surface was obtained by removing outliers, and the outlier removal included: When performing a kd-tree search on the point cloud model, each point is recorded simultaneously. Distance to its neighboring points; Calculate the average distance between each point and its neighboring points. Based on this average distance and each distance, calculate the standard deviation. Filter out points whose average distance is greater than the standard deviation threshold by setting a threshold parameter.

9. The system according to claim 3, characterized in that, The step size is selected as 1mm to 2mm, and the stretching distance parameter is set as the step size and the number of iterations. The product of.

Images