Patent Application
20250166182
The invention relates to a joint bone and cartilage measurement system and method, explicitly referring to a measurement system and method for personalized bone and articular cartilage morphology that applies artificial intelligence techniques.
Cartilage is a type of connective tissue that plays a crucial role in human joints by helping to reduce impact and friction during joint movement, thereby ensuring smooth joint function. However, cartilage is susceptible to damage due to aging, excessive use of joints, or trauma, which may lead to osteoarthritis (OA). Osteoarthritis, also known as degenerative arthritis, is characterized by damage to joint cartilage or subchondral bone. It commonly affects the distal joints of the fingers, the base of the thumb, the neck, the lower back, the knees, and the hips, with knee osteoarthritis (Knee OA) being the most prevalent. In this condition, joint inflammation and damage result in changes to the bones, degeneration of tendons and ligaments, and cartilage wear, leading to joint pain, swelling, deformity, impaired joint function, and a decline in quality of life. Therefore, early diagnosis and monitoring of treatment effectiveness are vital for targeted therapy, making the detection of cartilage changes a critical component.
Conventional detection of cartilage changes typically relies on traditional clinical assessments and imaging methods such as X-rays and MRI. The most commonly used grading system for osteoarthritis is the Kellgren-Lawrence Grading System. Based on standing knee X-ray images, this system categorizes knee osteoarthritis into five grades according to the degree of joint space narrowing (indicating cartilage wear) and osteophyte formation. However, both X-ray and MRI methods have limitations. For instance, X-rays cannot provide direct information about the cartilage; they only infer the severity of knee osteoarthritis by assessing the space between bones. Furthermore, differences in evaluations among physicians can occur. MRI is constrained by image quality, making it difficult to delineate the damaged cartilage from the healthy cartilage precisely. Additionally, both methods can only capture static information, require high costs and specialized equipment, and are influenced by the operator's expertise.
Therefore, developing a new joint bone and cartilage measurement system and method is necessary to address the aforementioned conventional issues.
In view of this, one scope of the present invention is to provide a measurement system for personalized bone and articular cartilage morphology to solve the aforementioned problems. The measurement system for personalized bone and articular cartilage morphology, applied to construct a personalized joint skeleton and cartilage model of a subject and calculating a cartilage wear value of the subject based on the personalized joint skeleton and cartilage model, the system comprising: a database, configured to store a dataset, wherein the dataset comprises a plurality of joint bone and cartilage images of a plurality of healthy individuals, a 3D bone and cartilage model corresponding to the plurality of joint bone and cartilage images and a plurality of shape parameters; an image capture module, configured to capture a plurality of bone images of the subject, each of the bone images comprising a first bone image and a second bone image; an image analysis module, coupled with the image capture module and the database, and configured to generate the 3D bone and cartilage model of the subject by applying a statistical shape model (SSM) learned from the dataset through the bone images of the subject and, through principal component analysis (PCA), to generate a plurality of first bone and cartilage shape parameters corresponding to the first bone image and a plurality of second bone and cartilage shape parameters corresponding to the second bone image based on the 3D bone and cartilage model and the shape parameters, generate a first bone and cartilage 3D model and a second bone and cartilage 3D model based on the first bone and cartilage shape parameters and the second bone and cartilage shape parameters; a neural network model, coupled with the image analysis module, and configured to input the first bone and cartilage shape parameters and the second bone and cartilage shape parameters into the neural network model and generate a plurality of first bone motion parameters and a plurality of second bone motion parameters corresponding to the first bone and cartilage shape parameters and the second bone and cartilage shape parameters, wherein the neural network model is trained by a machine learning method based on the dataset; a model building module, coupled with the image analysis module and the neural network model, and configured to generate the personalized joint skeleton and cartilage model based on the first bone and cartilage 3D model, the second bone and cartilage 3D model, the first bone motion parameters and the second bone motion parameters; a novel computation module, coupled with the model building module, and configured to simulate the relative positions of the first bone and cartilage 3D model and the second bone and cartilage 3D model of the subject during exercise based on the personalized joint skeleton and cartilage model, and to calculate an actual cartilage shape and the cartilage wear value of the subject by eliminating overlapping cartilage regions in the first bone and cartilage 3D model and the second bone and cartilage 3D model during exercise.
Wherein, the measurement system for personalized bone and articular cartilage morphology further comprises an image registration module. The image registration module is coupled with the image analysis module and the model building module, and configured to adjust and project the angle and position of the first bone and cartilage 3D model and the second bone and cartilage 3D model, and compare the joint bone and cartilage images from the healthy individuals in the dataset by a single-plane dynamic image registration method to calculate the first bone motion parameters and the second bone motion parameters of the subject.
Wherein, each of the bone images of the subject further comprises an X-ray image.
Wherein, the first bone motion parameters and the second bone motion parameters further comprise a translation parameter and a rotation parameter along the three directions of the XYZ axes.
Wherein, the personalized joint skeleton and cartilage model is further visualized through a heatmap and a perspective visualization method.
Another scope of the present invention is to provide a measurement method for personalized bone and articular cartilage morphology. The method comprises the following steps: storing a dataset to a database, wherein the dataset comprises a plurality of joint bone and cartilage images of a plurality of healthy individuals, a 3D bone and cartilage model corresponding to the plurality of joint bone and cartilage images and a plurality of shape parameters; an image capture module capturing a plurality of bone images of the subject, each of the bone images comprising a first bone image and a second bone image; an image analysis module generating the 3D bone and cartilage model of the subject by applying a statistical shape model (SSM) learned from the dataset through the bone images of the subject and, through principal component analysis (PCA), generating a plurality of first bone and cartilage shape parameters corresponding to the first bone image and a plurality of second bone and cartilage shape parameters corresponding to the second bone image based on the 3D bone and cartilage model and the shape parameters, generating a first bone and cartilage 3D model and a second bone and cartilage 3D model based on the first bone and cartilage shape parameters and the second bone and cartilage shape parameters; inputting the first bone and cartilage shape parameters and the second bone and cartilage shape parameters into a neural network model and generate a plurality of first bone motion parameters and a plurality of second bone motion parameters corresponding to the first bone and cartilage shape parameters and the second bone and cartilage shape parameters, wherein the neural network model is trained by a machine learning method based on the dataset; a model building module generating the personalized joint skeleton and cartilage model based on the first bone and cartilage 3D model, the second bone and cartilage 3D model, the first bone motion parameters and the second bone motion parameters; a novel computation module simulating the relative positions of the first bone and cartilage 3D model and the second bone and cartilage 3D model of the subject during exercise based on the personalized joint skeleton and cartilage model, and to calculate an actual cartilage shape and the cartilage wear value of the subject by eliminating overlapping cartilage regions in the first bone and cartilage 3D model and the second bone and cartilage 3D model during exercise.
Wherein, the measurement method for personalized bone and articular cartilage morphology further comprises the following steps: an image registration module adjusting and projecting the angle and position of the first bone and cartilage 3D model and the second bone and cartilage 3D model, and compare the joint bone and cartilage images from the healthy individuals in the dataset by a single-plane dynamic image registration method to calculate the first bone motion parameters and the second bone motion parameters of the subject.
Wherein, each of the bone images of the subject further comprises an X-ray image.
Wherein, the first bone motion parameters and the second bone motion parameters further comprise a translation parameter and a rotation parameter along the three directions of the XYZ axes.
Wherein, the personalized joint skeleton and cartilage model is further visualized through a heatmap and a perspective visualization method.
In summary, the present invention provides a measurement system for personalized bone and articular cartilage morphology that integrates dynamic X-ray technology with artificial intelligence (AI) to reconstruct the structure of joint bones and cartilage. The system further reconstructs a model of the patient's knee joint bones and cartilage, accurately calculating the degree and location of cartilage changes. Additionally, the personalized joint skeleton and cartilage model created by the invention uses heatmap and perspective visualization to offer an intuitive understanding of cartilage health, enabling faster and more comprehensive cartilage assessments. This AI-based system for measuring 3D joint bone and cartilage through planar X-rays provides physicians with dynamic and precise diagnostic information, facilitating the development of personalized treatment plans for each patient and evaluating the effectiveness of treatments.
For the sake of the advantages, spirits and features of the present invention can be understood more easily and clearly, the detailed descriptions and discussions will be made later by way of the embodiments and with reference of the diagrams. It is worth noting that these embodiments are merely representative embodiments of the present invention, wherein the specific methods, devices, conditions, materials and the like are not limited to the embodiments of the present invention or corresponding embodiments. Moreover, the devices in the figures are only used to express their corresponding positions and are not drawing according to their actual proportion.
In the description of this specification, the description with reference to the terms “an embodiment”, “another embodiment” or “part of an embodiment” means that a particular feature, structure, material or characteristic described in connection with the embodiment including in at least one embodiment of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment. Furthermore, the particular features, structures, materials or characteristics described may be combined in any suitable manner in one or more embodiments. Furthermore, the indefinite articles “a” and “an” preceding a device or element of the present invention are not limiting on the quantitative requirement (the number of occurrences) of the device or element. Thus, “a” should be read to include one or at least one, and a device or element in the singular also includes the plural unless the number clearly refers to the singular.
Please refer to
Please refer to both
Additionally, the personalized joint skeleton and cartilage model can be generated by adjusting the shape parameters of the model based on joint features, utilizing a statistical shape model (SSM) combined with principal component analysis (PCA). In practical applications, the measurement system for personalized bone and articular cartilage morphology described in this embodiment can be applied to reconstruct a personalized knee joint model. In practice, once the knee joint sample model database has been established, the sample models in the database can be further used as training models to compute an average model and establish shape variation. The shape variation can be derived through principal component analysis (PCA), yielding the eigenvalue λi and the eigenvector ϕi. The following formula can express a new knee joint model:
Where bi represents the shape model parameters corresponding to the model.
In practice, artificial intelligence (AI) can further be used to construct a 3D bone and cartilage model of the joint. Combining the model with dynamic X-ray kinematic calculation methods makes it possible to obtain the bone distance distribution between the femur and tibia during passive flexion and extension movements at each point in time. Furthermore, the bone distance distributions can be compared with the cartilage 3D model previously established using artificial intelligence to assess cartilage thickness and calculate penetration areas, penetration depth, and penetration volume, so as to evaluate the extent of cartilage wear and its severity distribution. The personalized joint skeleton and cartilage model can be further visualized through a heatmap and a perspective visualization method. This method helps in a more comprehensive evaluation of joint cartilage health, and detecting cartilage changes through this process can serve as a novel method for assessing the effectiveness of treatments.
Furthermore, in practice, the database plays a crucial role in machine learning, providing the foundational data for training, evaluating, and improving machine learning models. As a result, the machine learning model can learn from the data within the database and adjust its parameters accordingly. In the embodiment described herein, the data stored in the database can include the plurality of bone images of the plurality of joints, the plurality of bone models corresponding to the bone images, 3D bone and cartilage models, the plurality of position parameters of the joints, and the plurality of 3D cartilage models.
In practice, creating a personalized joint skeleton and cartilage model requires dynamic X-ray images from different planes. Through image registration techniques, the skeletal components of the reconstructed three-dimensional (3D) joint model are compared to dynamic X-ray images, and the optimal model shape parameters are determined using artificial intelligence (AI) and machine learning methods. This enables the reconstruction of a personalized 3D joint model. The image registration process involves geometric transformations of two or more images to align or match them spatially. It can also compare and match 3D objects with 2D images to determine their correct position in real-world space. This process can include translation, rotation, and scaling to align specific features or regions between images. The personalized joint skeleton and cartilage model can further include the distal femur, proximal tibia, and cartilage in the present embodiment. The image registration method is also used to determine the relative positions of the distal femur, proximal tibia, and cartilage to obtain an accurate knee joint model.
Please refer to
Please refer to
In practice, the database can include a knee joint sample model database, which stores personalized knee joint models of healthy individuals and patients generated from computed tomography (CT) scans. This data serves as a source for machine learning. The personalized knee joint models can include the distal femur, proximal tibia, and cartilage. First, a CT scan is performed over an appropriate joint scan range (i.e., ensuring the knee joint is positioned at the center of the image). The CT images acquired by the image capture module are imported into visualization software for image segmentation, and the segmented data is stacked using a 3D modeling module to construct a 3D bone and cartilage model. The image registration module aligns the 3D bone and cartilage model with other images and computes 3D kinematic information of the knee joint, including translation and rotation along the X, Y, and Z axes. In practice, the image registration module can use a graphic user interface (GUI) developed in MATLAB, implementing single-plane image matching techniques. The interface allows users to adjust the position of bones in 3D space, visually aligning them with dynamic X-ray images. This alignment serves as the initial guess for determining the six degrees of freedom (6-DOF) for the 3D bone and cartilage model. Using the calibrated parameters from the system and distortion correction parameters from the dynamic X-ray system, digitally reconstructed radiographs (DRRs) are generated. By continuously adjusting the 6-DOF position of the 3D model, new DRR images IDRR are generated and compared with the dynamic X-ray images (Ifl). Optimization algorithms are then used to find the optimal spatial position of the bones. The optimization can employ vectorized genetic algorithms, and the similarity measurement can use the gradient difference (GD) method. This method applies horizontal and vertical Sobel operators to both IDRR and Ifl, resulting in four gradient images: dIDRR/di′dIDRR/dj′dIfl/di′dIfl/dj. The difference between the gradient images yields the vertical and horizontal gradient difference images IdiffV and IdiffH, respectively. The similarity metric based on the Gradient Difference method is expressed as:
The data stored in the database and its collection methods are not limited to those described here and can be selected based on user requirements. It should be noted that the modules and functionalities of the measurement system 2 for personalized bone and articular cartilage morphology in the present embodiment are generally similar to those described in the previous embodiments, so they will not be repeated here. The present embodiment uses joint images and data obtained from CT scans to build a training model. Using dynamic X-ray technology and artificial intelligence, it reconstructs the patient's knee joint bones and cartilage using machine learning. In this way, cartilage changes and positioning can be determined more accurately. Furthermore, this embodiment addresses the limitations of conventional X-ray and MRI methods, which cannot directly acquire cartilage information or precisely determine its location. Additionally, it provides dynamic joint information without requiring specialized equipment or being affected by operator skills, thereby reducing time and labor costs.
In summary, the present invention provides a measurement system for personalized bone and articular cartilage morphology that integrates dynamic X-ray technology with artificial intelligence (AI) to reconstruct the structure of joint bones and cartilage. The system further reconstructs a model of the patient's knee joint bones and cartilage, accurately calculating the degree and location of cartilage changes. Additionally, the personalized joint skeleton and cartilage model created by the invention uses heatmap and perspective visualization to offer an intuitive understanding of cartilage health, enabling faster and more comprehensive cartilage assessments. This AI-based system for measuring 3D joint bone and cartilage through planar X-rays provides physicians with dynamic and precise diagnostic information, facilitating the development of personalized treatment plans for each patient and evaluating the effectiveness of treatments.
With the examples and explanations mentioned above, the features and spirits of the invention are hopefully well described. More importantly, the present invention is not limited to the embodiment described herein. Those skilled in the art will readily observe that numerous modifications and alterations of the device may be made while retaining the teachings of the invention. Accordingly, the above disclosure should be construed as limited only by the metes and bounds of the appended claims.
| Number | Date | Country | |
|---|---|---|---|
| 63601235 | Nov 2023 | US |
