An automatic selection and reduction planning method and system for total hip joint replacement prosthesis

Abstract

This application relates to the technical field of orthopedic surgical planning, and particularly to a method and system for automatic selection and reduction planning of total hip arthroplasty (THA) prostheses. The method includes: reconstructing a three-dimensional surface model using patient imaging data, establishing a right-handed rectangular coordinate system on the anterior pelvic plane as a unified reference; automatically matching the acetabular cup and planning its placement posture; fitting the long axis of the femoral medullary canal and locating the trochanteric fossa, extracting the geometric features of each femoral stem model; obtaining the optimal placement posture through pattern search, evaluating the stem-medullary canal fit by distinguishing between distal and proximal compression zones, and outputting the optimal femoral stem model according to a non-compensated multi-level safety ranking; simulating the matching of the head and liner, traversing the neck length offset, and calculating the lower limb length difference and femoral eccentricity to complete the reduction optimization; finally, determining the osteotomy plane and outputting all three-dimensional models and planning parameters. This invention achieves fully automatic and high-precision preoperative planning for total hip arthroplasty (THA), effectively avoiding the risk of intraoperative fracture and ensuring postoperative joint stability.

Description

Technical Field

[0001] This application relates to the technical field of orthopedic surgical planning, and in particular to a method and system for automatic selection and reduction planning of total hip arthroplasty prostheses. Background Technology

[0002] The accuracy of preoperative planning in total hip arthroplasty (THA) directly determines the accuracy of prosthesis placement, postoperative joint stability, and long-term survival. Currently, the commonly used clinical method is still mainly based on two-dimensional X-ray template measurement, which has inherent defects such as loss of three-dimensional skeletal morphology information, reliance on physician experience, and inability to simulate three-dimensional fitting conditions. While existing planning software based on CT three-dimensional reconstruction provides a three-dimensional visualization environment, it still has technical bottlenecks such as inconsistent coordinate systems leading to difficulties in standardizing planning parameters, reliance on manual selection of key anatomical points resulting in large errors, single evaluation indicators for femoral stem selection that do not distinguish between different mechanical requirements of distal and proximal ends, and the inability to coordinate the planning and cutting of the acetabular and femoral sides to optimize the lower limb length difference and femoral offset.

[0003] In existing technologies, such as CN121445532A, a reinforcement learning-based method for selecting prostheses in three-dimensional total hip arthroplasty is provided. This method acquires medullary canal morphology and prosthesis 3D data, performs shape matching screening, prosthesis structure segmentation and constraint calculation, and geometric simplification to generate polyhedral data. Finally, it uses reinforcement learning on the Unity platform to train a selection model to output the prosthesis model with the highest matching degree. However, this method has the following drawbacks: the training process relies on a large amount of labeled data and computational resources; the model's generalization ability is limited by the distribution of training data, making it difficult to adapt to patients with significant variations in medullary canal morphology; and the geometric simplification operation reduces the accuracy of the prosthesis model, potentially causing the selection results to deviate from clinical reality. Summary of the Invention

[0004] The purpose of this invention is to provide an automatic selection and reduction planning method and system for total hip arthroplasty prostheses, which partially solves or alleviates the above-mentioned shortcomings in the prior art, enables fully automatic and high-precision preoperative planning for total hip arthroplasty, effectively avoids the risk of intraoperative fracture, and ensures postoperative joint stability.

[0005] To solve the aforementioned technical problems, the present invention specifically adopts the following technical solution: A first aspect of the present invention is to provide an automatic selection and repositioning planning method for total hip arthroplasty prostheses, comprising the following steps: S1: Acquire patient image data and reconstruct a three-dimensional surface model, which includes models of the pelvis, femur, and medullary cavity; acquire a prosthesis CAD model library, which includes models of the acetabular cup, femoral stem, ball head, and liner; use the pelvic model to establish a right-handed rectangular coordinate system on the anterior plane of the pelvis as a unified spatial reference. S2: Extract the point cloud of the acetabular region of the pelvic model, calculate the opening size to match the acetabular cup model, and place the acetabular cup model in the planned position according to the preset abduction angle and anteversion angle; S3: Extract the medullary canal point cloud from the medullary canal model, obtain the initial medullary canal long axis through layered section processing, and then fit the region of interest again using the lesser trochanter point in the femur model as the anchor point to obtain the fitted medullary canal long axis; project the point cloud of the femur model onto the fitted medullary canal long axis, and select the radial minimum point as the trochanteric fossa; at the same time, extract the stem long axis, shoulder point, neck axis and reference ball head center of each femur stem model in the prosthesis CAD model library; S4: For each femoral stem model, align the stem long axis with the fitted medullary cavity long axis, constrain the insertion depth with the trochanteric socket and set the initial forward tilt angle, and use pattern search within the finite degrees of freedom to find the optimal placement pose; S5: Under the optimal placement posture, evaluate the separation of the distal safety zone and the proximal pressing zone, and output the optimal femoral stem model according to the non-compensated multi-level safety ranking. S6: Match the ball head model and liner model according to the matched acetabular cup model, obtain the supported neck length offsets according to the output femoral stem model, calculate the lower limb length difference and femoral offset after reduction, and select the optimal offset to complete the reduction. S7: Determine the osteotomy plane based on the optimal placement posture and the reset result of step S6, and then assemble the prosthesis model after cutting the femoral model.

[0006] Furthermore, in step S1, establishing a right-handed rectangular coordinate system for the anterior pelvic plane using the pelvic model as a unified spatial reference includes: Identify the bilateral anterior superior iliac spines and pubic tuberosities on the pelvic model; The origin is the midpoint of the line connecting the two anterior superior iliac spines. The positive direction of the X-axis is defined as the direction from the origin to the right anterior superior iliac spine of both sides. The anterior pelvic plane is defined by the plane passing through the bilateral anterior superior iliac spines and the pubic tubercle. The positive Y-axis is defined as the direction located in the anterior plane of the pelvis, perpendicular to the X-axis, and pointing forward. The positive Y-axis is obtained by subtracting the projection component of the temporary vector in the X-axis direction from the temporary vector pointing from the origin to the pubic tubercle and then normalizing it. The positive direction of the Z-axis is defined by the cross product of the X-axis and Y-axis, and the positive direction of the Z-axis is perpendicular to the anterior plane of the pelvis and points towards the head of the patient. Establish a right-handed rectangular coordinate system for the anterior pelvic plane, and transform the three-dimensional surface model and each model in the prosthesis CAD model library to the right-handed rectangular coordinate system for the anterior pelvic plane through rigid body transformation.

[0007] Furthermore, step S2 includes: The triangular patch point cloud of the acetabular fossa region is extracted from the pelvic model. The position of the center of the acetabular fossa and the radius of the sphere are calculated using the least squares spherical fitting method. The acetabular opening size is taken as twice the radius of the sphere. It is compared with the nominal outer diameter of each acetabular cup model in the prosthesis CAD model library. The acetabular cup model with the smallest deviation between the nominal outer diameter and the opening size is selected as the matching result. The geometric center of the matched acetabular cup model is aligned with the center of the sphere as the rotation center of the acetabular cup. The unit normal vector direction of the acetabular cup opening plane is taken as the opening direction. The target direction of the opening plane is determined according to the preset abduction angle and preset forward tilt angle. A rigid body transformation matrix is ​​constructed based on the rotation center and the target direction, and the acetabular cup model is transformed to the planned position.

[0008] Further, in step S3, the medullary canal point cloud is extracted from the medullary canal model, and the initial medullary canal long axis is obtained through layered section processing. Then, the region of interest is extracted using the lesser trochanter point in the femur model as an anchor point, and the fitted medullary canal long axis is obtained by fitting the data again. Extract the medullary cavity point cloud from the medullary cavity model, layer it at equal intervals along the longitudinal direction, calculate the geometric center of each cross section, connect the centers of each cross section to construct the initial medullary cavity centerline and obtain the initial medullary cavity long axis; Using the lesser trochanter point in the femoral model as the anchor point, the region of interest is extracted within a first preset distance from the distal end of the lesser trochanter along the long axis of the initial medullary canal. The point cloud of the medullary cavity within the region of interest is further layered to extract the center of the cross section. The continuous straight line segments with stable center changes and minimum residuals are selected for principal axis fitting to obtain the candidate medullary cavity long axis. Calculate the radial distance from each point in the medullary cavity point cloud to the candidate medullary cavity long axis, remove outliers whose radial distance is greater than a first preset threshold, and fit the remaining points again to obtain the fitted medullary cavity long axis.

[0009] Further, in step S3, projecting the point cloud of the femoral model onto the long axis of the fitted medullary canal and selecting the radial minimum point as the trochanteric fossa includes: Project the point cloud of the femoral model onto the long axis of the fitted medullary canal, and calculate the axial height parameter and radial distance of each point; Using the lesser trochanter point as the proximal constraint, candidate points with an axial height greater than the lesser trochanter height and located within the height range adjacent to the center of the femoral head are selected. Points located within the femoral head bulb are eliminated based on the center and radius of the femoral head. The point with the smallest radial distance among the remaining candidate points is selected as the rotor socket.

[0010] Furthermore, the positioning method of the small rotor point is as follows: A local coordinate system for the femur is established based on the point cloud of the femur model and the fitted medullary canal long axis. Candidate bony prominences located on the posteromedial aspect of the proximal femur were selected based on local curvature, surface normal, and axial height. Among the candidate bony prominences, select the continuous cluster of points with the largest projection along the medial-lateral protrusion direction, project the geometric center of the cluster onto the femoral surface, and define the resulting point as the lesser trochanter point.

[0011] Furthermore, step S4 includes: Align the long axis of each bone stem model with the long axis of the fitted medullary cavity, and use the trochanteric fossa to constrain the insertion depth and set the initial anteversion angle as the initial pose. Set axial rotation, lateral translation, and inward / outward tilting as finite degrees of freedom, and set the search range for each degree of freedom; The Hooke-Jeeves pattern search algorithm is adopted, with the stem-medullary cavity matching degree as the objective function, and the pose is updated iteratively by alternately performing probe movement and pattern movement; The iteration terminates when the step size of each degree of freedom is less than the second preset threshold or the preset maximum number of evaluations is reached, and the pose that makes the objective function optimal is output as the optimal placement pose. And / or, The objective function includes: The penetration penalty is determined when the radial distance between a point on the surface of the femoral stem and the point cloud on the inner wall of the medullary canal is less than zero or exceeds a preset safety threshold. The maximum penetration depth is then calculated and a penalty is applied. The gap penalty item calculates the average gap between the surface of the femoral stem and the inner wall of the medullary canal, and applies a penalty to areas with excessive gaps. The tilt penalty item applies a secondary penalty to constrain the prosthesis posture when the inward or outward tilt angle exceeds the preset range.

[0012] Furthermore, in step S5: The evaluation of the distal safety zone includes: taking the area at the second preset distance upward from the distal end of the femoral stem model as the distal safety zone, calculating the radial distance from the point cloud of the medullary canal wall in this area to the surface of the femoral stem model, and counting the maximum overhang distance and the number of overhang layers. The evaluation of the proximal pressing zone includes: taking the area of ​​a predetermined proportional length of the proximal end of the femoral stem model as the proximal pressing zone, calculating the contact rate within this area, and calculating the filling rate of the four quadrants of the medial, lateral, anterior, and posterior sides respectively. The non-compensated multi-level safe sorting is performed according to the following three-level lexicographical order: The first level requires that the maximum overhang distance at the far end is less than the third preset threshold and the number of overhang layers is zero, the contact rate at the near end is greater than the fourth preset threshold and the fill rate in all four quadrants is greater than the fifth preset threshold. The second requirement is that the maximum overhang distance at the far end is less than the sixth preset threshold and the number of overhang layers does not exceed the preset number of layers, and the contact rate at the near end is greater than the seventh preset threshold. The third level selects the femoral stem model with the smallest maximum overhang distance and the highest proximal contact rate.

[0013] Furthermore, step S6 includes: According to the size of the selected acetabular cup model, the corresponding ball head model and liner model are matched from the prosthesis CAD model library. The inner diameter of the liner model is adapted to the diameter of the ball head model, and the outer diameter of the liner model is adapted to the inner diameter of the selected acetabular cup model. Based on the selected femoral stem model, obtain each neck length offset supported by the femoral stem model from the prosthesis CAD model library, traverse each neck length offset and calculate the corresponding ball head center position respectively. The spatial translation vector from the center of the ball head to the rotation center of the acetabular cup is applied to the femoral model, femoral stem model, and ball head model to complete the reduction. The lower limb length difference and femoral offset corresponding to each neck length deviation are calculated respectively. The liner model maintains its relative pose to the acetabular cup model. The difference in lower limb length was used as an indicator to evaluate the equal length of the two lower limbs after surgery, and the femoral offset was used as an indicator to evaluate the stability of the hip joint after surgery. The optimal offset was selected as the neck length offset that made the femoral offset fall within the preset safe range and was closest to the value of the healthy side. If none of the neck length offsets meet the preset safe range of femoral eccentricity, a comprehensive objective function is constructed based on the difference in lower limb length and the deviation between femoral eccentricity and the healthy side value. The neck length offset that minimizes the value of the comprehensive objective function is selected as the optimal offset, and a prompt message is generated.

[0014] Secondly, this application also discloses an automatic selection and reduction planning system for total hip arthroplasty prostheses, the system comprising: A unified coordinate system construction module is configured to acquire patient image data and reconstruct a three-dimensional surface model, which includes models of the pelvis, femur, and medullary cavity; acquire a prosthesis CAD model library, which includes models of the acetabular cup, femoral stem, ball head, and liner; and use the pelvic model to establish a right-handed rectangular coordinate system in the anterior pelvic plane as a unified spatial reference. The acetabular cup matching module is configured to extract point cloud of the acetabular region of the pelvic model, calculate the opening size to match the acetabular cup model, and place the acetabular cup model in the planned position according to the preset abduction angle and anteversion angle. The femoral lateral anatomical feature extraction module is configured to extract medullary canal point cloud from the medullary canal model, obtain the initial medullary canal long axis through layered section processing, and then refit the region of interest by using the lesser trochanter point in the femoral model as the anchor point to obtain the fitted medullary canal long axis; project the point cloud of the femoral model onto the fitted medullary canal long axis, and select the radial minimum point as the trochanteric fossa; at the same time, extract the stem long axis, shoulder point, neck axis and reference ball head center of each femoral stem model in the prosthesis CAD model library; The pose search module is configured to align the long axis of the stem with the fitted medullary cavity long axis of each femoral stem model, constrain the insertion depth with the trochanteric socket and set the initial anteversion angle, and use pattern search to find the optimal placement pose within a limited degree of freedom. The femoral stem screening module is configured to evaluate the distal safety zone and proximal press-fit zone under the optimal placement posture, and output the optimal femoral stem model according to the non-compensated multi-level safety ranking. The reduction module is configured to match the ball head model and the liner model according to the matched acetabular cup model, obtain the supported neck length offsets according to the output femoral stem model, calculate the lower limb length difference and femoral offset after reduction, and select the optimal offset to complete the reduction. The prosthesis model assembly module is configured to determine the osteotomy plane based on the optimal placement posture and the reset result of the reset module, and to assemble the prosthesis model after cutting the femoral model.

[0015] Beneficial technical effects: 1. This application establishes a unified right-hand rectangular coordinate system for the anterior pelvic plane as a spatial reference, and transforms all patient skeletal models and prosthesis models into this coordinate system. This completely eliminates the problem of coordinate inconsistencies caused by differences in patient scanning posture and scanning at different times, making the planning results across cases and across time comparable and reproducible, and providing a standardized spatial reference framework for accurate calculations in subsequent steps.

[0016] 2. This application adopts a fully automatic anatomical feature extraction method based on geometric point cloud processing. By projecting along the long axis of the medullary canal and calculating the axial height and radial distance, combined with multiple constraint screening strategies such as proximal trochanter restriction and femoral head spherical removal, the automatic and accurate positioning of the trochanteric fossa is achieved. At the same time, the automatic fitting of the long axis of the medullary canal is achieved through layered cross-section center fitting and outlier removal. No manual interaction is required throughout the process, eliminating operator subjective error. Unilateral planning can be completed within a few seconds.

[0017] 3. This application addresses the clinical biomechanical needs of metaphyseal press-fit femoral stems by pioneering a zonal evaluation model of "distal anti-penetration and proximal press-fit". The stem-medullary canal fit evaluation area is decoupled into a distal safety zone and a proximal press-fit zone, which are evaluated independently. The distal safety zone strictly prevents the prosthesis from penetrating the medullary canal cortical bone, while the proximal press-fit zone pursues high contact rate and balanced filling in all four quadrants. This precisely matches the design concept of metaphyseal press-fit prostheses and effectively avoids the risk of intraoperative split fracture and postoperative thigh pain caused by excessive pursuit of global filling rate.

[0018] 4. This application employs a non-compensatory multi-level safety sorting strategy, ranking all candidate femoral stem models according to a three-level lexicographical order: the first level requires no distal penetration and optimal proximal compression fitting; the second level allows slight penetration but strictly controls its extent; the third level selects the best model while ensuring minimal penetration. This strategy, at the algorithmic level, strictly prohibits compensating for distal penetration risk with high proximal filling rates, ensuring that no safety hazards can be masked, significantly improving the safety and robustness of the selection decision.

[0019] 5. This application achieves the linkage matching of the acetabular cup, femoral stem, ball head, and liner under the same optimization framework. The selection of the ball head neck length offset is included in the same optimization closed loop with the difference in lower limb length and femoral offset after reduction: For each candidate neck length offset, the difference in lower limb length and femoral offset after reduction are calculated separately. The scheme that makes the femoral offset fall within the preset safe range and is closest to the value of the healthy side is selected first. If all schemes cannot meet the requirements, a prompt message is generated. This completely avoids the local optimal trap caused by independent decision-making of each link in the traditional step-by-step planning, and ensures the equal length of the two lower limbs and the stability of the hip joint after surgery. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. The elements or parts in the drawings are not necessarily drawn to scale. Obviously, the drawings described below are some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0021] Figure 1 This is the overall flowchart of the automatic selection and repositioning planning method for total hip arthroplasty prosthesis proposed in this application.

[0022] Figure 2 This is a schematic diagram of the calibration and definition of the pelvic coordinate system for an APP in this application.

[0023] Figure 3This is a schematic diagram of the results of fitting the long axis of the femoral medullary cavity and locating the trochanteric fossa in this application.

[0024] Figure 4 This is a schematic diagram of a non-compensated multi-level security ranking decision tree in this application.

[0025] Figure 5 This is a schematic diagram of a final planning output result of this application.

[0026] Figure 6 This is a schematic diagram of a module of an automatic selection and repositioning planning system for total hip arthroplasty prosthesis according to this application. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0028] In this document, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" may be used interchangeably.

[0029] In this document, the terms "upper," "lower," "inner," "outer," "front," "rear," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the present invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0030] In this document, unless otherwise explicitly specified and limited, the terms "installed," "equipped with," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0031] In this document, "and / or" includes any and all combinations of one or more of the listed related items.

[0032] In this article, "multiple" means two or more, that is, it includes two, three, four, five, etc.

[0033] Figure 1 This invention presents a flowchart of an automatic selection and reduction planning method for total hip arthroplasty prostheses. (Refer to...) Figure 1 The method specifically includes the following steps: S1: Acquire patient image data and reconstruct a three-dimensional surface model, which includes models of the pelvis, femur, and medullary cavity; acquire a prosthesis CAD model library, which includes models of the acetabular cup, femoral stem, ball head, and liner; use the pelvic model to establish a right-handed rectangular coordinate system on the anterior plane of the pelvis as a unified spatial reference. S2: Extract the point cloud of the acetabular region of the pelvic model, calculate the opening size to match the acetabular cup model, and place the acetabular cup model in the planned position according to the preset abduction angle and anteversion angle; S3: Extract the medullary canal point cloud from the medullary canal model, obtain the initial medullary canal long axis through layered section processing, and then fit the region of interest again using the lesser trochanter point in the femur model as the anchor point to obtain the fitted medullary canal long axis; project the point cloud of the femur model onto the fitted medullary canal long axis, and select the radial minimum point as the trochanteric fossa; at the same time, extract the stem long axis, shoulder point, neck axis and reference ball head center of each femur stem model in the prosthesis CAD model library; S4: For each femoral stem model, align the stem long axis with the fitted medullary cavity long axis, constrain the insertion depth with the trochanteric socket and set the initial forward tilt angle, and use pattern search within the finite degrees of freedom to find the optimal placement pose; S5: Under the optimal placement posture, evaluate the separation of the distal safety zone and the proximal pressing zone, and output the optimal femoral stem model according to the non-compensated multi-level safety ranking. S6: Match the ball head model and liner model according to the matched acetabular cup model, obtain the supported neck length offsets according to the output femoral stem model, calculate the lower limb length difference and femoral offset after reduction, and select the optimal offset to complete the reduction. S7: Determine the osteotomy plane based on the optimal placement posture and the reset result of step S6, and then assemble the prosthesis model after cutting the femoral model.

[0034] In one specific embodiment, the following discloses a specific embodiment of the automatic selection and reduction planning method for total hip arthroplasty prostheses applicable to this application, which specifically includes the following steps: Step 1: Data Loading and Establishment of a Unified Coordinate System This step aims to construct an individualized 3D planning scene and establish a unified anatomical spatial benchmark to eliminate the impact of differences in skeletal spatial posture under different patients and scanning positions on the accuracy of subsequent planning.

[0035] The process involves acquiring patient imaging data, which can be CT scan image sequences (DICOM format) containing the pelvic and femoral regions. Bone boundaries are extracted from the image data using an image segmentation algorithm, generating NIfTI format segmentation labels. A Marching Cubes algorithm is then used to reconstruct the segmentation results into a 3D surface model (STL format). This model includes a pelvic surface model, a femoral surface model, and a medullary canal surface model. The number of triangular facets is controlled within a preset range to ensure subsequent computational efficiency. Simultaneously, a prosthesis CAD model library is acquired, including acetabular cup models, femoral stem models, ball head models, and liner models. Each femoral stem model is associated with a corresponding neck length offset parameter. A right-handed Cartesian coordinate system on the anterior pelvic plane is established on the pelvic surface model as a unified spatial reference.

[0036] On the STL model of the pelvis, key anatomical points of the bilateral anterior superior iliac spines (ASIS) and pubic tubercle (PT) are automatically identified by calculating local curvature and shape indices, and the anterior pelvic plane (APP) is defined accordingly. A right-handed rectangular coordinate system is established based on this plane, specifically defined as follows: Origin of coordinate system Take the midpoint of the line connecting the bilateral anterior superior iliac spines, that is: in and These are the three-dimensional coordinate vectors of the right and left anterior superior iliac spines in the original world coordinate system.

[0037] The X-axis direction is defined as a unit vector pointing from the origin to the right anterior superior iliac spine: The Y-axis is defined as a unit vector located within the APP plane (anterior pelvic plane), perpendicular to the X-axis, and pointing forward. First, a temporary vector pointing from the origin to the pubic tubercle is calculated. , and then from Subtracting its projection component along the X-axis, we obtain a direction vector located in the APP plane and perpendicular to the X-axis, which, after normalization, becomes the positive Y-axis direction. Following the right-hand rule, calculate the cross product of the X and Y axes along the Z-axis to obtain a unit vector perpendicular to the APP plane and pointing towards the head: After defining the coordinate axes, construct the coordinate system from the original world coordinate system to the APP coordinate system. Rigid body transformation matrix This matrix consists of rotation matrices. Translation vector Its homogeneous form is: For any three-dimensional point in the original world coordinate system Its homogeneous coordinates The mapping in the APP coordinate system is as follows: The original coordinate data of the patient's pelvis, femur, medullary canal surface models, and all prosthesis CAD models were transformed using the aforementioned rigid body transformation matrix. The coordinates are uniformly converted to the APP coordinate system. All subsequent planning operations are performed within this unified coordinate system, thereby completely eliminating differences in skeletal posture caused by different scanning positions, and ensuring the comparability and reproducibility of planning results across cases and time periods. For example... Figure 2 As shown, the X, Y, and Z axes of the APP pelvic coordinate system are illustrated.

[0038] Step 2: Acetabular Side Planning (Cup Selection and Placement) This step aims to automatically select the type and spatial placement of the acetabular cup based on the 3D pelvic model in the unified coordinate system of the APP, providing an acetabular side reference for subsequent full-joint linkage planning.

[0039] On the transformed STL model of the pelvis, triangular patch point clouds of the acetabular fossa region are first extracted. The geometric parameters of the acetabular opening are calculated using the least-squares spherical fitting method: Let the acetabular fossa point cloud set be... The center of the acetabular fossa was solved using least-squares spherical fitting. and the radius of the sphere The objective function is: acetabular opening radius Used to match the closest acetabular cup size in the prosthesis library: for each size in the prosthesis library Its nominal outer diameter is Calculate the matching error Select The smallest size is selected as the cup size.

[0040] Then determine the placement pose of the acetabular cup in the APP coordinate system. Let the unit normal vector of the acetabular cup opening plane be... The direction is defined as from the inside of the cup to the outside of the cup opening. For the right hip joint... For the left hip joint , The initial direction is determined by the fitted normal vector of the acetabular fossa opening plane. Abduction angle Defined as The angle between the projection onto the XZ plane (coronal plane) and the X-axis in the same lateral direction; anteversion angle Defined as The angle between the projection onto the XY plane (cross-section) and the X-axis in the same outward direction. The specific calculation formula is: In actual implementation, the abduction angle pass Projection calculation in the XZ plane: Lean angle pass Projection calculation in the XY plane: For the left hip joint, it is necessary to... Invert to maintain left-right symmetry. Determine the target direction of the cup opening plane according to preset adjustable parameters: outward angle. Lean angle Because of individual differences in patients' acetabular fossa angles, the system allows physicians to adjust the aforementioned angle parameters within a preset range. Based on this, the rotation center of the acetabular cup is aligned with the center of the acetabular fossa ball. Align the cup to ensure the inner side of the cup fits against the bone surface of the acetabulum.

[0041] Finally, construct the placement transformation matrix for the acetabular cup. The matrix contains a translation component. and rotating parts ,in The opening normal vector in the coordinate system of the acetabular cup CAD model (usually) Rotate to the target direction .set up and The axis of rotation between them is The rotation angle is ,but The following is given by Rodrigues's rotation formula: in, It is the identity matrix. For vectors The antisymmetric matrix. The selected acetabular cup STL model is transformed using the rigid body transformation matrix. Place it in the planned location, that is, for any vertex on the cup model. Its homogeneous coordinate transformation is .

[0042] Step 3: Automatic extraction of femoral lateral anatomical features 3.1 Automatic Fitting of the Long Axis of the Fibrous Cavity Extracting medullary canal point clouds from segmentation labels or STL models Then, convert it to the unified coordinate system of the APP. First, construct the initial medullary cavity centerline: align the medullary cavity point cloud along the Z-axis at equal intervals. (Usually 2~5mm) is used for layering, and the first layer is... Layered cross-sectional point cloud Calculate its geometric center : ; Connecting the centers of each section sequentially yields the initial medullary cavity centerline. And thereby obtain the initial axial direction. .

[0043] The lesser trochanter point is then used as the proximal anatomical anchor point. In some embodiments, the method for locating the lesser trochanter point is as follows: Based on the femoral surface point cloud and the initial medullary canal axis, a local femoral coordinate system is established with the initial medullary canal axis as the axial reference. In the local femoral coordinate system, the local curvature changes and surface normal directions of each point on the femoral surface are calculated, and points located in the medial posteromedial region of the proximal femur, exhibiting a convex local curvature and with a normal pointing in the medial posteromedial direction are selected as candidate points; then, with the initial medullary canal axis as the reference, a continuous cluster of points with the largest projection value along the medial lateral projection direction is selected from the candidate points, and the point obtained by projecting the geometric center of this cluster onto the femoral surface is defined as the lesser trochanter point. .

[0044] With small rotor point As the proximal anatomical anchor point, along the initial medullary canal axis. A preset distance at the far end of the small rotor (Typically, a point cloud of the region of interest (ROI) within the medullary canal is extracted within a range of 50~80mm.) .

[0045] Multi-layer cross-sectional center extraction was performed on the point cloud of the medullary canal within the ROI. A sliding window method was used to screen continuous straight line segments with stable center changes and small residuals: for continuous... Center of the layer section Calculate its linear fitting residuals : in The starting point position, This is a unit direction vector. (Selection) For the largest continuous interval less than a preset threshold, principal component analysis (PCA) or least squares linear fitting is performed on the center point of the cross section on the straight line segment to obtain the long axis direction of the candidate medullary cavity. and a point on the axis .

[0046] Further calculations were performed on the long axis from each point in the medullary canal point cloud to the candidate medullary canal. Radial distance: Statistics of all The distribution, excluding those with a radial distance greater than Peripheral outliers (i.e.) The mean, (Standard deviation), only the radial core point cloud is retained. .right PCA fitting was performed again to obtain the final long axis direction of the femoral medullary cavity. and a point on its axis (Usually, the intersection of the small rotor plane and the axis is taken): ; in The core point cloud geometry is used. Through the above two-stage fitting process (coarse-to-fine) and the elimination of peripheral outliers, the obtained medullary canal long axis is robust to segmentation noise and irregular medullary canal morphology, and more realistically reflects the biomechanical channel of prosthesis insertion.

[0047] 3.2 Automatic positioning of the trochanteric socket based on the long axis projection of the medullary canal This step uses the long axis of the femoral medullary cavity obtained in step 3.1. and a point on the axis It serves as a spatial reference point for automatically locating the rotor socket reference point.

[0048] Point cloud of the femoral surface Projected onto the long axis of the medullary canal Above, calculate each surface point Compared to Two key parameters: Axial height parameters: Radial distance: With small rotor point For proximal anatomical constraints, the proximal femoral surface region located in the proximal direction of the lesser trochanter and within the height range adjacent to the center of the femoral head was selected as the candidate region for the trochanteric fossa: ; in The height of the small rotor, The height of the center of the femoral head. 10mm is usually used.

[0049] Simultaneously, a spherical region was extracted from the point cloud of the femoral head articular surface obtained from preoperative segmentation, and the center of the femoral head was obtained by least-squares spherical fitting. and radius Points on the femoral head surface are eliminated based on the femoral head center and radius to avoid misidentifying the femoral head as the trochanteric fossa: for any candidate point ,like If it is, then it is determined to be the femoral head point and is excluded: ; In the remaining candidate points In this process, a reasonable height zone is selected based on the height distribution along the long axis of the medullary cavity, eliminating false points such as excessively low distal endpoints of longitudinal grooves and excessively high axial exit points, thus obtaining a candidate point set. Within this reasonable height zone, select the radial distance from the long axis of the medullary cavity. The smallest point on the femoral surface is used as the reference point for the trochanteric fossa. : Geometrically, this point corresponds to the concave region on the posteromedial aspect of the proximal femur closest to the long axis of the medullary canal, which is the trochanteric fossa in the classic anatomical sense. This method uses fully automated logic involving long axis projection, multi-constraint screening, and radial minimum optimization to accurately locate the trochanteric fossa, with a positioning accuracy controllable within 1 mm.

[0050] Figure 3 This diagram illustrates the results of fitting the long axis of the femoral medullary canal and trochanteric fossa localization in this application. The figure uses the three-dimensional surface contour of the femur (gray semi-transparent model) as the background, with the yellow solid portion representing the fitted three-dimensional shape of the medullary canal (i.e., the expected placement path of the femoral stem), the blue dashed line representing the fitted long axis of the medullary canal, and the red dots representing the automatically located trochanteric fossa.

[0051] 3.3 Automatic Extraction of Geometric Features from Femoral Stem CAD Model For each femoral stem CAD model (STL format) in the prosthesis library, the following baseline geometric features are automatically extracted.

[0052] Let the set of point clouds on the surface of the femoral stem be... Long axis of the femoral stem Obtained by PCA fitting of the point cloud of the main body of the handle, taking the direction of the first principal component: ; in The subset of the femoral stem surface point cloud belonging to the stem trunk is obtained by filtering based on the stem geometry prior. The geometry prior includes: removing points in the proximal neck region and the distal tip region along the stem long axis, and retaining the part with a uniformly varied cross-sectional shape in the middle section as the stem trunk region.

[0053] Femoral stem shoulder point Defined in the direction of the major axis The point with the largest projection on the upper part of the implant is the outermost point of the proximal shoulder. in The handle point is the geometric center of the cloud.

[0054] anteroposterior axis of femoral stem The vector perpendicular to the major axis, indicating the anterior-posterior direction of the prosthesis, is obtained through principal direction analysis of the handle section: ; femoral stem neck Obtained by fitting the centerline of the conical surface of the prosthesis neck. Neck brace / cone end point. Defined as a journal shaft The far-end point on the reference ball head. Defined as located on the neck axis, offset inwards along the neck axis by 4.3 mm from the end point of the neck brace: The aforementioned geometric features were extracted and stored as feature vectors for each candidate femoral stem type. This is for use in subsequent initial placement and attitude optimization steps.

[0055] Step 4: Initial placement and attitude optimization of femoral stem candidate models This step aims to establish an initial placement pose for each candidate femoral stem model based on the anatomical and prosthesis geometric features obtained in step 3, and to find the optimal spatial pose that matches the medullary canal for each candidate model through numerical optimization search with finite degrees of freedom, providing accurate pose input for subsequent zonal evaluation and model selection.

[0056] 4.1 Initial anatomical placement of the femoral stem For each candidate femoral stem model Its geometric eigenvectors are The initial placement pose is determined according to the following rules: Orientation Alignment: Constructing Rotation Transformations The long axis of the femoral stem Rotate to the long axis of the medullary canal that is fitted in step 3.1. Orientation alignment. Let the axis of rotation be... Rotation angle ,but The following is given by Rodrigues's rotation formula: Depth constraints: Constructing translation transformations This makes the shoulder point of the implant... The Z-coordinate in the APP coordinate system is equal to the rotor socket located in step 3.2. The Z-coordinate is used to implement the insertion depth constraint of the handle: Initial rotation (forward tilt angle setting): Constructing the long axis around the medullary canal rotation Based on the anteroposterior axis of the femoral stem The initial anteversion angle of the prosthesis is determined based on its relationship with the proximal posterior condyle plane or the femoral neck axis. (can be found) (Preset adjustment within the range). Let the rotation axis be... The rotation angle is ,but for: Finally, the initial transformation matrix Combine them in the following order (rotate first, then translate): For any vertex on the femoral stem prosthesis Its initial placement location is .

[0057] 4.2 Finite Degrees of Freedom Attitude Optimization Search For each candidate femoral stem model, in the initial pose A finite-degree-of-freedom pose search is performed in the vicinity to find the pose that optimizes the stem-canal fit. The optimization variables are defined as follows: 1) Axial rotation angle about the long axis of the medullary canal Search scope ; 2) Vertical translation along the long axis of the medullary canal (X direction), range ; 3) Vertical translation along the long axis of the medullary canal (Y direction), range ; 4) Inward and outward tilt angles (Rotation around the X-axis), range If the limit is exceeded, a penalty will be imposed. 5) Flexion-extension tilt angle (Rotation around the Y-axis), range If the limit is exceeded, a penalty will be imposed.

[0058] For candidate handle models Let its current transformation matrix be The new transformation matrix during the search process It is generated by the incremental combination of each degree of freedom. Let the length of the medullary canal be considered. Rotation The matrix is Rotate around the X-axis The matrix is Rotate around the Y-axis The matrix is Translate along the X direction Translation along the Y direction The translation vector is ,but: The Hooke-Jeeves algorithm (a derivative-free pattern search algorithm) is used to iteratively find the optimal pose. This algorithm searches for the descent direction of the objective function along the coordinate axes by alternately performing exploratory moves and pattern moves, without needing to calculate gradients.

[0059] For each candidate handle model, its objective function Defined as a comprehensive cost function based on the degree of stem-medullary cavity matching, with minimization as the optimization direction. The basic form of the objective function is: ; , , The preset weighting coefficients for each penalty item are defined as follows: Penalty item Calculate the radial distance from the point cloud of the medullary canal wall to the prosthesis surface, and statistically analyze the maximum overhang distance. and number of outer suspension layers : in The preset safety threshold is (usually 1mm). and These are preset weighting coefficients.

[0060] Gap penalty item Calculate the distance from the prosthesis surface to the inner wall of the medullary cavity, and statistically analyze the contact rate and average gap. ; in and The preset weighting coefficients are: ContactRate (contact rate, where a distance less than 0.5mm is considered contact) and MeanGap (average gap between all non-contact points).

[0061] Inclined penalty items : Inward / outward tilting exceeds the preset range or Apply secondary penalties in the following situations: in, These are preset weighting coefficients.

[0062] The specific iterative process of the Hooke-Jeeves algorithm is as follows: Detection movement: sequentially along the five degrees of freedom in step sizes. (Axial rotation step size) lateral translation step size Incline step size ) to conduct detection. For each degree of freedom Calculate separately and If there exists a direction that causes the objective function to decrease, then move the current point to the optimal point in that direction; otherwise, change the step size. Halve.

[0063] Pattern Movement: After completing a round of exploration movement across all degrees of freedom, if the current point is an improvement over the previous exploration starting point, then the pattern movement proceeds along the overall improvement direction. ; in This is the transformation matrix for the starting point of the previous detection round. If the mode shift fails to produce further descent, the system returns to the detection shift phase.

[0064] Termination condition: When the step size of all degrees of freedom is... All are less than the preset threshold (axial rotation) lateral translation ,tilt When the maximum number of evaluations is reached (the number of evaluations is limited for each handle model to ensure clinical response speed), the iteration terminates, and the current optimal transformation matrix is ​​output. .

[0065] To accelerate convergence and avoid getting trapped in local optima, a large initial step size is used in the early stages of the search to broadly explore the parameter space. As iterations progress, the step size is gradually reduced for a refined local search. Simultaneously, axial rotation and lateral translation around the long axis of the medullary canal are allowed to be decoupled and adjusted independently within their respective ranges, while a penalty term is applied when inversion / exversion tilt exceeds a preset range to ensure that the search results strictly meet clinical safety constraints.

[0066] This step involves examining each candidate femoral stem model in the prosthesis bank. Perform the initial placement and attitude optimization process described above, and output the optimal transformation matrix for each model. and its corresponding objective function value This is used for partition evaluation and security sorting in subsequent step 5.

[0067] Step 5: Zonal Evaluation and Security Ranking Selection Based on the optimal pose of each candidate femoral stem model obtained in step 4, this step performs a clinical biomechanical zoning evaluation of the stem-medullary canal fit, and selects the optimal model through a non-compensating multi-level safety sorting strategy. This step also includes a left-right mirror reuse mechanism to ensure algorithm consistency and code reusability.

[0068] 5.1 Evaluation of stem-medullary canal coordination zones The optimal transformation matrix obtained in step 4 Below, regarding the candidate femoral stem type The stem-medullary canal fit was evaluated in zones. Based on the clinical biomechanical requirements of the metaphyseal press-fit prosthesis, the evaluation area was decoupled into two functionally independent regions: the distal safety zone and the proximal press-fit zone.

[0069] Remote security zone evaluation: Defined from the distal end of the prosthesis upwards The region is designated as the remote security zone. Let the set of remote endpoint clouds of the prosthesis be... The point cloud of the medullary cavity wall in this region is .right Each point in Calculate its point cloud to the inner wall of the medullary cavity. The closest distance: ; like (i.e., the prosthesis point is located inside the medullary canal point cloud), or Exceeding the preset positive threshold (i.e., if the gap between the prosthesis and the inner wall of the medullary cavity is too large or if penetration occurs), it is statistically classified as an overhang.

[0070] The remote security zone statistics include the following two core metrics: 1) Maximum overhang distance That is, the maximum penetration depth; 2) Number of outer suspension layers The distal safe zone is divided along the long axis of the medullary canal into: For each layer with equal thickness, count the number of consecutive layers with external overhangs.

[0071] Evaluation of the proximal compression zone: Define proximal end of prosthesis The length is the proximal compression region. Let the proximal endpoint cloud set be... The following four core indicators are used to statistically analyze the near-end compression area: 1) Contact rate: Defined by distance threshold Calculate the contact rate : ; 2) Four-quadrant fill rate: The prosthesis section is divided into four quadrants: inner, outer, anterior, and posterior. The fill rate in each quadrant is calculated separately. For each quadrant... Fill rate It is the ratio of the contact area between the prosthesis surface and the medullary cavity wall in this quadrant to the total area of ​​this quadrant, expressed as a percentage.

[0072] 3) Weakest support angle Calculate the support distribution of the prosthesis within the cross-section and identify the weakest direction angle of the support: ; in For angle Support strength function in the direction.

[0073] 4) Wedged support strength The initial stability of the stem is characterized by measuring the axial self-locking force acquired by the proximal wedge structure of the prosthesis within the medullary cavity, and by calculating the sum of the normal contact forces between the proximal conical surface of the prosthesis and the inner wall of the medullary cavity. in The angle between the normal to the prosthesis surface and the long axis of the medullary cavity. The axial component reflecting the wedging force.

[0074] 5.2 Selection of Non-compensated Multi-level Security Sequencing Reference Figure 4 This application sorts all candidate femoral stem models lexicographically according to the following three-level safety partitioning. The core principle of this sorting strategy is to prohibit the use of high filler ratios to compensate for any penetration risk; that is, the sorting priority strictly follows the clinical principle of safety first and fit second.

[0075] Level 1 (PERFECT level) meets all of the following conditions: Maximum overhang distance of the remote safety zone ; Number of outer suspension layers ; Proximal compression zone contact rate ; Fill rate of all quadrants , .

[0076] Level 2 (REVIEW level) must meet all of the following conditions: Maximum overhang distance of the remote safety zone ; Number of outer suspension layers ; Proximal compression zone contact rate .

[0077] Level 3 (FALLBACK): Candidate models that do not meet the criteria of Level 1 or Level 2 above are classified into Level 3. In this level, models with the smallest global penetration distance / maximum overhang distance and the highest contact rate are selected, with the specific sorting rules as follows: 1. Prioritize based on maximum overhang distance Sort by size from smallest to largest; 2. If Same, according to the number of outer suspension layers Sort by size from smallest to largest; 3. If Same, based on contact rate Sort by size from largest to smallest.

[0078] Within the PERFECT and REVIEW levels, the same sorting rules are applied for priority ranking. The overall three-level lexicographical sorting decision logic is as follows: ; Output the femoral stem model that is ranked first. and its final placement position : ; The aforementioned three-level safety ranking directly encodes the clinical safety red line into algorithmic constraints, ensuring that no penetration risk can be masked by a high filling rate, effectively avoiding the risk of intraoperative split fractures.

[0079] 5.3 Normalization and reuse of left and right side mirror images When dealing with the left hip joint, a mirror normalization reuse mechanism is executed, so that the femoral stem planning algorithm (steps 3.1 to 5.2) developed entirely based on the right side can be directly reused on the left side, ensuring consistency between the left and right side algorithms and code reusability.

[0080] For the left hip joint data, first, all femoral data (including medullary canal point cloud) in the current APP coordinate system are collected. Femoral surface dot cloud Small rotor point Femoral head center (and the already placed prosthetic model, etc.) multiplied by the mirror matrix : ; That is, through mirror transformation of the YZ plane (matrix) This maps the data from the left side to the right side of the space. After this mirror transformation, the left femur is geometrically equivalent to the right femur.

[0081] For any three-dimensional point Its mirror transformation is For any transformation matrix (Including rotation and translation), its mirror transformation is: ; because For integral matrix ( The above formula simplifies to: ; In the mirror space (logical right side), the femoral stem planning algorithm developed entirely based on the right side in steps 3.1 to 5.2 is directly reused to obtain the optimal stem model and its placement pose in the mirror space. Then, all the planned poses and models are transformed through inverse mirror transformation. Return to the left side space: For any model vertex in the mirror space Its left spatial coordinates are For placing the transformation matrix in the mirror space The transformation matrix of the left space is: ; Actual testing showed that the difference between directly planning the left side and mirror reusing the planning left side was less than [a certain value]. This proves that the mirror reuse mechanism effectively ensures the consistency and symmetry of the planning results on both sides.

[0082] This step ultimately outputs the femoral stem model that is ranked first. Its final placement pose in the original space (the right side is the APP coordinate system, and the left side is the APP coordinate system after reverse mirroring). The various evaluation indicators for this model (maximum distance of the far end, number of layers of the outer suspension, contact rate of the near end, four-quadrant filling rate, etc.) are used for the linkage reset in step 6.

[0083] Step 6: Ball head / liner linkage matching and femoral reduction Based on the acetabular cup selected in step 2, the femoral stem selected in step 5, and their final placement posture, this step automatically completes the linkage matching between the ball head diameter and the liner model. Through a traversal search of different neck length offsets, the femoral component is repositioned to the anatomical position. Finally, among multiple repositioning schemes, the scheme that achieves the optimal balance between leg length discrepancy (LLD) and femoral offset (FO) is selected.

[0084] 6.1 Ball head and liner linkage matching Based on the acetabular cup model selected in step 2 Determine the appropriate ball head diameter. The matching relationship between the inner diameter of the acetabular cup and the diameter of the acetabular head is determined by a compatibility matrix provided by the prosthesis manufacturer. Predefined: ; For the selected acetabular cup model, there are usually one or more selectable ball head diameters (such as 28mm, 32mm, 36mm). The system defaults to the standard ball head diameter that matches the inner diameter of the cup, while allowing doctors to adjust it in preset options.

[0085] Based on the femoral stem model selected in step 5 and its final placement position Iterate through the set of neck length offsets supported by this model. Each offset This corresponds to the offset of the ball head center along the neck axis. Let the reference ball head center extracted in step 3.3 be... The direction of the neck axis is Then the offset The corresponding center position of the ball head is: ; The liner model is based on the diameter of the ball head. It automatically matches the inner diameter of the acetabular cup. Let the liner model be... Its selection function is: ; in, Indicates the selected acetabular cup model. This represents the liner selection function. The specific implementation of this function is based on the compatibility matrix provided by the prosthesis manufacturer. This matrix predefines the range of liner models that can be matched between different ball head diameters and different acetabular cup models.

[0086] When placing the liner, align the plane of the liner cup opening with the plane of the acetabular cup opening. Let the normal vector of the acetabular cup opening determined in step 2 be... The center of rotation of the cup is The placement transformation matrix of the lining for: ; in Let be the transformation matrix for the placement of the acetabular cup. This is a fixed transformation matrix for the liner model from its own coordinate system to the coordinate system of the acetabular cup model. This transformation is determined by the geometric thickness of the liner and specifically involves translating a predetermined distance along the normal direction of the liner cup opening plane, so that the outer surface of the liner fits against the inner surface of the acetabular cup, and the liner cup opening plane is coplanar with the acetabular cup opening plane. The opening normal vector of the liner is consistent with the opening normal vector of the acetabular cup. .

[0087] 6.2 Femoral reduction and neck length deviation optimization For each available neck length offset Calculate the center of the matched ball head To the center of rotation of the acetabular cup Spatial translation vector: ; The translation vector is applied simultaneously to the femoral model, the already placed femoral stem model, and the femoral head model to complete the reduction. For the femoral model... any vertex on Its position after reset is: ; The same translation transformation is applied to both the femoral stem model and the femoral head model. This is to maintain the relative pose between the femoral components.

[0088] After the reset is complete, for each offset scheme Calculate the following two key clinical indicators: 1) Lower limb length difference (LLD): In the APP coordinate system, LLD is defined as the difference in vertical distance from the "teardrop" to the lesser trochanter between the affected and healthy sides. ; in and The Z coordinates of the teardrop points on the affected and healthy sides in the APP coordinate system are shown (automatically extracted from the pelvic model). The Z-coordinate of the lesser trochanter point on the affected side after repositioning. Here is the Z-coordinate of the lesser trochanter point on the healthy side. The sign convention for LLD is: a positive value indicates lengthening of the affected lower limb, and a negative value indicates shortening of the affected lower limb.

[0089] 2) Femoral eccentricity (FO): Measure the horizontal distance from the center of rotation of the femoral head to the longitudinal axis of the femur on the coronal plane (XZ plane) of the APP coordinate system. Let the direction of the femoral longitudinal axis be the long axis of the medullary canal fitted in step 3.1. In the projection direction on the coronal plane, the center of rotation of the femoral head is the center of the femoral head after reduction. A point on the longitudinal axis of the femur is Then the formula for calculating FO is: ; in is the unit vector of the projection of the long axis of the medullary cavity onto the coronal plane.

[0090] Among several optional neck length offsets, the optimal one is selected using the following objective function: ; in The target offset is (usually the healthy side FO value or the median of the preset normal range). and Weighting coefficients (default) , ).

[0091] The priority design of preference constraints is as follows: 1) Primary constraint: Prioritize selection that enables Falling into the preset seat belt And closest to the healthy side value Offset scheme; 2) Secondary constraints: Among the solutions that satisfy the primary constraints, the one that prioritizes the solution that satisfies the primary constraints is selected. The minimum solution; 3) Final fault tolerance: If there is no bias that fully satisfies the above conditions, then choose to make the comprehensive objective function... The smallest candidate is selected, and a warning message is generated and output to a subsequent JSON report.

[0092] The above-mentioned linkage matching and reduction optimization ensures that the acetabular cup, femoral stem, ball head and liner complete linkage matching under the same optimization framework, avoiding the problem of inconsistency between local and global optimality in step-by-step planning, so that LLD and FO, the two key clinical indicators that determine postoperative gait quality and joint dislocation rate, achieve optimal balance.

[0093] Step 7: Osteotomy and Result Output 7.1 Determining the osteotomy plane and cutting the femoral model Based on the final reset pose determined in step 6, and combined with the small rotor point located in step 3.2... Rotor pit and the direction of the long axis of the medullary canal fitted in step 3.1. Determine the femoral neck osteotomy plane.

[0094] First, determine the point through which the osteotomy plane passes. The osteotomy plane is too small; the preset distance above the trochanter is too small. The midpoint of the line connecting a point at a certain location and the trochanteric fossa. Assume the lesser trochanteric point moves along the long axis of the medullary canal. The following point is: ; osteotomy plane passes through point Defined as: ; Normal direction of the osteotomy plane From the long axis of the medullary cavity relative to the femoral neck axis The average direction is determined as follows: ; The direction of the normal line takes into account both the axial direction of the long axis of the medullary canal (ensuring that the osteotomy surface is coordinated with the insertion direction of the stem) and the natural direction of the femoral neck (ensuring that the osteotomy surface is consistent with the direction of the anatomical neck), so that the proximal femur after osteotomy can well accommodate the prosthesis stem.

[0095] For femoral surface model any vertex on Its signed distance to the osteotomy plane is: ; according to Model trimming based on symbols: retaining those that satisfy... The apex (i.e., the portion proximal to the osteotomy plane) is removed. The distal portion. The trimmed femur model. for: ; After trimming, the triangular facets at the osteotomy plane are filled with boundaries and repaired with meshes to ensure that a closed and smooth osteotomy surface mesh model is generated.

[0096] 7.2 Output of Planning Results Documents Based on the final reduction pose and the trimmed femoral model, the following 3D model file (STL format) is output: `femur_resected.stl`: The trimmed STL model of the reduced femur. This model already includes the reduction translation transformation applied in step 6.2. This means that it has been reduced to an anatomical position that matches the acetabular cup.

[0097] `stem_{model}_{size}_final.stl`: The STL model of the femoral stem after reduction, including the complete stem body and neck structure. The file name includes the stem model selected in step 5. And dimensional parameters. The model has also been subjected to a repositioning translation transformation to maintain the correct relative position with the trimmed femur.

[0098] `cup_{model}_final.stl`: The STL model of the acetabular cup selected in step 2, which contains the complete geometry of the cup and has been placed in the target pose determined in step 2.

[0099] `head_{diameter}_{offset}_final.stl`: The STL model for the spherical head, containing the complete geometry of the spherical head and the adapting cone. The file name includes the spherical head diameter determined in step 6.1. and the neck length offset selected in step 6.2 .

[0100] `liner_{inner diameter}_{outer diameter}_final.stl`: The liner STL model, containing the complete geometry of the liner. The file name includes the inner diameter of the liner (equal to the ball head diameter). ) and outer diameter (equal to the inner diameter of the acetabular cup).

[0101] It also outputs a structured file `full_pipeline_summary.json`, which records complete planning parameters and evaluation information, such as parameters like LLD=-0.5mm and FO=45mm.

[0102] The above documents can be directly used for: 1. Surgical navigation systems: importing them into navigation software for real-time intraoperative prosthesis placement guidance; 2. 3D printing of individualized surgical guides: generating osteotomy guides or prosthesis placement guides; 3. Preoperative simulation and teaching: reviewing the planning scheme in 3D visualization software; 4. Clinical records and traceability: JSON files can be used as digital archives of surgical records, supporting postoperative review and efficacy analysis.

[0103] Figure 5 This is a schematic diagram of the final planning output of this application, used to demonstrate the femoral stem position, ball head and liner linkage matching, femoral reduction, osteotomy, and planning result output. The diagram uses a semi-transparent pelvis and femur (cut according to the osteotomy plane) as the anatomical background to visually display the complete prosthesis components output by the system linkage. The yellow part is the reduced femoral stem (including the neck structure), the blue part is the acetabular cup and liner placed in the acetabular fossa, and the pink part is the ball head matching the femoral stem neck axis.

[0104] This application establishes a right-handed rectangular coordinate system in the anterior pelvic plane as a unified spatial reference, completely eliminating coordinate inconsistencies caused by differences in patient scanning postures, thus ensuring comparability and reproducibility of planning results across cases and time periods. Through a fully automated anatomical feature extraction method based on geometric point cloud processing, it achieves precise automatic positioning of key anatomical features such as the long axis of the medullary canal, trochanteric fossa, and lesser trochanter, requiring no manual interaction; unilateral planning can be completed within seconds. Furthermore, through a pioneering "distal anti-penetration, proximal compression fitting" zonal evaluation model, the stem-medullary canal fit evaluation is decoupled into a distal safety zone and a proximal compression fitting zone, which are evaluated independently, ensuring precise... The design concept of matching the metaphyseal end compression fitting prosthesis effectively avoids the risk of intraoperative split fracture and postoperative thigh pain caused by excessive pursuit of global filling rate. Through a non-compensatory multi-level safety sorting strategy, the algorithm strictly prohibits compensating for the risk of distal penetration with high proximal filling rate, which significantly improves the safety and robustness of the selection decision. By linking the acetabular cup, femoral stem, ball head, and liner in the same optimization framework, the selection of neck length offset and the difference in lower limb length and femoral offset after reduction are included in the same optimization closed loop, avoiding the local optimal trap of traditional step-by-step planning and ensuring postoperative lower limb isometry and hip joint stability.

[0105] Further reference Figure 6 As an implementation of the above-described method, this application provides an embodiment of an automatic selection and reduction planning system for total hip arthroplasty prostheses. This system embodiment is similar to... Figure 1 Corresponding to the method embodiments shown, the system can be specifically applied to various electronic devices.

[0106] refer to Figure 6 An automated selection and reduction planning system for total hip arthroplasty prostheses, comprising: The unified coordinate system construction module 201 is configured to acquire patient image data and reconstruct a three-dimensional surface model, which includes models of the pelvis, femur, and medullary cavity; acquire a prosthesis CAD model library, which includes models of the acetabular cup, femoral stem, ball head, and liner; and use the pelvic model to establish a right-hand rectangular coordinate system in the anterior plane of the pelvis as a unified spatial reference. The acetabular cup matching module 202 is configured to extract the point cloud of the acetabular region of the pelvic model, calculate the opening size to match the acetabular cup model, and place the acetabular cup model in the planned position according to the preset abduction angle and anteversion angle. The femoral lateral anatomical feature extraction module 203 is configured to extract medullary canal point cloud from the medullary canal model, obtain the initial medullary canal long axis through layered section processing, and then obtain the fitted medullary canal long axis by truncating the region of interest with the lesser trochanter point in the femoral model as the anchor point; project the point cloud of the femoral model onto the fitted medullary canal long axis, and select the radial minimum point as the trochanteric fossa; at the same time, extract the stem long axis, shoulder point, neck axis and reference ball head center of each femoral stem model in the prosthesis CAD model library; The pose search module 204 is configured to align the long axis of the stem with the fitted medullary cavity long axis of each femoral stem model, constrain the insertion depth with the trochanteric socket and set the initial forward tilt angle, and use pattern search to find the optimal placement pose within a limited degree of freedom. The femoral stem screening module 205 is configured to evaluate the distal safety zone and proximal press-fit zone under the optimal placement posture, and output the optimal femoral stem model according to the non-compensated multi-level safety ranking. The reduction module 206 is configured to match the ball head model and the liner model according to the matched acetabular cup model, obtain the supported neck length offsets according to the output femoral stem model, calculate the lower limb length difference and femoral offset after reduction, and select the optimal offset to complete the reduction. The prosthesis model assembly module 207 is configured to determine the osteotomy plane based on the optimal placement posture and the reset result of the reset module, and to assemble the prosthesis model after cutting the femoral model.

[0107] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the following... Figure 1 The method shown.

[0108] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0109] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a computer terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0110] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for automatic selection and repositioning planning of total hip arthroplasty prostheses, characterized in that, Includes the following steps: S1: Acquire patient image data and reconstruct a three-dimensional surface model, which includes models of the pelvis, femur, and medullary cavity; acquire a prosthesis CAD model library, which includes models of the acetabular cup, femoral stem, ball head, and liner; use the pelvic model to establish a right-handed rectangular coordinate system on the anterior plane of the pelvis as a unified spatial reference. S2: Extract the point cloud of the acetabular region of the pelvic model, calculate the opening size to match the acetabular cup model, and place the acetabular cup model in the planned position according to the preset abduction angle and anteversion angle; S3: Extract the medullary canal point cloud from the medullary canal model, obtain the initial medullary canal long axis through layered section processing, and then fit the region of interest again using the lesser trochanter point in the femur model as the anchor point to obtain the fitted medullary canal long axis; project the point cloud of the femur model onto the fitted medullary canal long axis, and select the radial minimum point as the trochanteric fossa; at the same time, extract the stem long axis, shoulder point, neck axis and reference ball head center of each femur stem model in the prosthesis CAD model library; S4: For each femoral stem model, align the stem long axis with the fitted medullary cavity long axis, constrain the insertion depth with the trochanteric socket and set the initial forward tilt angle, and use pattern search within the finite degrees of freedom to find the optimal placement pose; S5: Under the optimal placement posture, evaluate the separation of the distal safety zone and the proximal pressing zone, and output the optimal femoral stem model according to the non-compensated multi-level safety ranking. S6: Match the ball head model and liner model according to the matched acetabular cup model, obtain the supported neck length offsets according to the output femoral stem model, calculate the lower limb length difference and femoral offset after reduction, and select the optimal offset to complete the reduction. S7: Determine the osteotomy plane based on the optimal placement posture and the reset result of step S6, and then assemble the prosthesis model after cutting the femoral model.

2. The method for automatic selection and repositioning planning of total hip arthroplasty prostheses according to claim 1, characterized in that: In step S1, establishing a right-handed rectangular coordinate system for the anterior pelvic plane using the pelvic model as a unified spatial reference includes: Identify the bilateral anterior superior iliac spines and pubic tuberosities on the pelvic model; The origin is the midpoint of the line connecting the two anterior superior iliac spines. The positive direction of the X-axis is defined as the direction from the origin to the right anterior superior iliac spine of both sides. The anterior pelvic plane is defined by the plane passing through the bilateral anterior superior iliac spines and the pubic tubercle. The positive Y-axis is defined as the direction located in the anterior plane of the pelvis, perpendicular to the X-axis, and pointing forward. The positive Y-axis is obtained by subtracting the projection component of the temporary vector in the X-axis direction from the temporary vector pointing from the origin to the pubic tubercle and then normalizing it. The positive direction of the Z-axis is defined by the cross product of the X-axis and Y-axis, and the positive direction of the Z-axis is perpendicular to the anterior plane of the pelvis and points towards the head of the patient. Establish a right-handed rectangular coordinate system for the anterior pelvic plane, and transform the three-dimensional surface model and each model in the prosthesis CAD model library to the right-handed rectangular coordinate system for the anterior pelvic plane through rigid body transformation.

3. The method for automatic selection and repositioning planning of total hip arthroplasty prostheses according to claim 1, characterized in that, Step S2 includes: The triangular patch point cloud of the acetabular fossa region is extracted from the pelvic model. The position of the center of the acetabular fossa and the radius of the sphere are calculated using the least squares spherical fitting method. The acetabular opening size is taken as twice the radius of the sphere. It is compared with the nominal outer diameter of each acetabular cup model in the prosthesis CAD model library. The acetabular cup model with the smallest deviation between the nominal outer diameter and the opening size is selected as the matching result. The geometric center of the matched acetabular cup model is aligned with the center of the sphere as the rotation center of the acetabular cup. The unit normal vector direction of the acetabular cup opening plane is taken as the opening direction. The target direction of the opening plane is determined according to the preset abduction angle and preset forward tilt angle. A rigid body transformation matrix is ​​constructed based on the rotation center and the target direction, and the acetabular cup model is transformed to the planned position.

4. The method for automatic selection and reduction planning of total hip arthroplasty prostheses according to claim 1, characterized in that: In step S3, the medullary canal point cloud is extracted from the medullary canal model, and the initial medullary canal long axis is obtained through layered section processing. The region of interest is then truncated using the lesser trochanter point in the femur model as an anchor point, and the fitted medullary canal long axis is obtained by fitting the data again. Extract the medullary cavity point cloud from the medullary cavity model, layer it at equal intervals along the longitudinal direction, calculate the geometric center of each cross section, connect the centers of each cross section to construct the initial medullary cavity centerline and obtain the initial medullary cavity long axis; Using the lesser trochanter point in the femoral model as the anchor point, the region of interest is extracted within a first preset distance from the distal end of the lesser trochanter along the long axis of the initial medullary canal. The point cloud of the medullary cavity within the region of interest is further layered to extract the center of the cross section. The continuous straight line segments with stable center changes and minimum residuals are selected for principal axis fitting to obtain the candidate medullary cavity long axis. Calculate the radial distance from each point in the medullary cavity point cloud to the candidate medullary cavity long axis, remove outliers whose radial distance is greater than a first preset threshold, and fit the remaining points again to obtain the fitted medullary cavity long axis.

5. The method for automatic selection and reduction planning of total hip arthroplasty prostheses according to claim 4, characterized in that: In step S3, projecting the point cloud of the femoral model onto the long axis of the fitted medullary canal and selecting the radial minimum point as the trochanteric fossa includes: Project the point cloud of the femoral model onto the long axis of the fitted medullary canal, and calculate the axial height parameter and radial distance of each point; Using the lesser trochanter point as the proximal constraint, candidate points with an axial height greater than the lesser trochanter height and located within the height range adjacent to the center of the femoral head are selected. Points located within the femoral head bulb are eliminated based on the center and radius of the femoral head. The point with the smallest radial distance among the remaining candidate points is selected as the rotor socket.

6. The method for automatic selection and repositioning planning of total hip arthroplasty prostheses according to claim 5, characterized in that: The positioning method of the small rotor point is as follows: A local coordinate system for the femur is established based on the point cloud of the femur model and the fitted medullary canal long axis. Candidate bony prominences located on the posteromedial aspect of the proximal femur were selected based on local curvature, surface normal, and axial height. Among the candidate bony prominences, select the continuous cluster of points with the largest projection along the medial-lateral protrusion direction, project the geometric center of the cluster onto the femoral surface, and define the resulting point as the lesser trochanter point.

7. The method for automatic selection and repositioning planning of total hip arthroplasty prostheses according to claim 1, characterized in that, Step S4 includes: Align the long axis of each bone stem model with the long axis of the fitted medullary cavity, and use the trochanteric fossa to constrain the insertion depth and set the initial anteversion angle as the initial pose. Set axial rotation, lateral translation, and inward / outward tilting as finite degrees of freedom, and set the search range for each degree of freedom; The Hooke-Jeeves pattern search algorithm is adopted, with the stem-medullary cavity matching degree as the objective function, and the pose is updated iteratively by alternately performing probe movement and pattern movement; The iteration terminates when the step size of each degree of freedom is less than the second preset threshold or the preset maximum number of evaluations is reached, and the pose that makes the objective function optimal is output as the optimal placement pose. And / or, The objective function includes: The penetration penalty is determined when the radial distance between a point on the surface of the femoral stem and the point cloud on the inner wall of the medullary canal is less than zero or exceeds a preset safety threshold. The maximum penetration depth is then calculated and a penalty is applied. The gap penalty item calculates the average gap between the surface of the femoral stem and the inner wall of the medullary canal, and applies a penalty to areas with excessive gaps. The tilt penalty item applies a secondary penalty to constrain the prosthesis posture when the inward or outward tilt angle exceeds the preset range.

8. The method for automatic selection and repositioning planning of total hip arthroplasty prostheses according to claim 1, characterized in that, In step S5: The evaluation of the distal safety zone includes: taking the area at the second preset distance upward from the distal end of the femoral stem model as the distal safety zone, calculating the radial distance from the point cloud of the medullary canal wall in this area to the surface of the femoral stem model, and counting the maximum overhang distance and the number of overhang layers. The evaluation of the proximal pressing zone includes: taking the area of ​​a predetermined proportional length of the proximal end of the femoral stem model as the proximal pressing zone, calculating the contact rate within this area, and calculating the filling rate of the four quadrants of the medial, lateral, anterior, and posterior sides respectively. The non-compensated multi-level safe sorting is performed according to the following three-level lexicographical order: The first level requires that the maximum overhang distance at the far end is less than the third preset threshold and the number of overhang layers is zero, the contact rate at the near end is greater than the fourth preset threshold and the fill rate in all four quadrants is greater than the fifth preset threshold. The second requirement is that the maximum overhang distance at the far end is less than the sixth preset threshold and the number of overhang layers does not exceed the preset number of layers, and the contact rate at the near end is greater than the seventh preset threshold. The third level selects the femoral stem model with the smallest maximum overhang distance and the highest proximal contact rate.

9. The method for automatic selection and repositioning planning of a total hip arthroplasty prosthesis according to claim 3, characterized in that, Step S6 includes: According to the size of the selected acetabular cup model, the corresponding ball head model and liner model are matched from the prosthesis CAD model library. The inner diameter of the liner model is adapted to the diameter of the ball head model, and the outer diameter of the liner model is adapted to the inner diameter of the selected acetabular cup model. Based on the selected femoral stem model, obtain each neck length offset supported by the femoral stem model from the prosthesis CAD model library, traverse each neck length offset and calculate the corresponding ball head center position respectively. The spatial translation vector from the center of the ball head to the rotation center of the acetabular cup is applied to the femoral model, femoral stem model, and ball head model to complete the reduction. The lower limb length difference and femoral offset corresponding to each neck length deviation are calculated respectively. The liner model maintains its relative pose to the acetabular cup model. The difference in lower limb length was used as an indicator to evaluate the equal length of the two lower limbs after surgery, and the femoral offset was used as an indicator to evaluate the stability of the hip joint after surgery. The optimal offset was selected as the neck length offset that made the femoral offset fall within the preset safe range and was closest to the value of the healthy side. If none of the neck length offsets meet the preset safe range of femoral eccentricity, a comprehensive objective function is constructed based on the difference in lower limb length and the deviation between femoral eccentricity and the healthy side value. The neck length offset that minimizes the value of the comprehensive objective function is selected as the optimal offset, and a prompt message is generated.

10. An automatic selection and repositioning planning system for total hip arthroplasty prostheses, characterized in that, The system includes: A unified coordinate system construction module is configured to acquire patient image data and reconstruct a three-dimensional surface model, which includes models of the pelvis, femur, and medullary cavity; acquire a prosthesis CAD model library, which includes models of the acetabular cup, femoral stem, ball head, and liner; and use the pelvic model to establish a right-handed rectangular coordinate system in the anterior pelvic plane as a unified spatial reference. The acetabular cup matching module is configured to extract point cloud of the acetabular region of the pelvic model, calculate the opening size to match the acetabular cup model, and place the acetabular cup model in the planned position according to the preset abduction angle and anteversion angle. The femoral lateral anatomical feature extraction module is configured to extract medullary canal point cloud from the medullary canal model, obtain the initial medullary canal long axis through layered section processing, and then refit the region of interest by using the lesser trochanter point in the femoral model as the anchor point to obtain the fitted medullary canal long axis; project the point cloud of the femoral model onto the fitted medullary canal long axis, and select the radial minimum point as the trochanteric fossa; at the same time, extract the stem long axis, shoulder point, neck axis and reference ball head center of each femoral stem model in the prosthesis CAD model library; The pose search module is configured to align the long axis of the stem with the fitted medullary cavity long axis of each femoral stem model, constrain the insertion depth with the trochanteric socket and set the initial anteversion angle, and use pattern search to find the optimal placement pose within a limited degree of freedom. The femoral stem screening module is configured to evaluate the distal safety zone and proximal press-fit zone under the optimal placement posture, and output the optimal femoral stem model according to the non-compensated multi-level safety ranking. The reduction module is configured to match the ball head model and the liner model according to the matched acetabular cup model, obtain the supported neck length offsets according to the output femoral stem model, calculate the lower limb length difference and femoral offset after reduction, and select the optimal offset to complete the reduction. The prosthesis model assembly module is configured to determine the osteotomy plane based on the optimal placement posture and the reset result of the reset module, and to assemble the prosthesis model after cutting the femoral model.

Images