Osteochondral visualization device, osteochondral visualization method, and program

Abstract

An osteochondral visualization device (3) comprises: an acquisition unit (31) that acquires a first image obtained by imaging, using a CT examination device (1) which employs a dual-energy method, a region of the human body that includes a joint composed of a bone and cartilage; and a generation unit (331) that generates information derived from the first image, on the basis of the acquired first image. The first image is an image in which the contours of the bone and the cartilage are defined, in which a boundary between the bone and the cartilage is defined, and in which a boundary between the cartilage and soft tissue included in the joint is defined.

Description

Osteochondral delineation device, osteochondral delineation method, and program

[0007] ,

[0001] The present disclosure relates to an osteochondral delineation device and the like.

[0002] In recent years, a dual-energy CT imaging method for imaging CT images by simultaneously irradiating two types of X-rays with different tube voltages has been studied. For example, an image generation device that uses the dual-energy CT imaging method to distinguish bone and uric acid crystals and generates an image in which the bone and uric acid crystals are distinguished has been developed (see Non-Patent Document 1).

[0003] Amir Yashar Tashakkor et al. “Dual-energy computed tomography: a valid tool in the assessment of gout?” International journal of Clinical Rheumatology Volume 7, Issue 1, 2012, Pages 73-79.

[0004] ' Since the image generation device described in Non-Patent Document 1 focuses on bone and uric acid crystals with significantly different CT values, the difficulty for the image generation device described in Non-Patent Document 1 to generate an image in which bone and uric acid crystals are distinguished is relatively low.

[0005] By the way, for a complex human body structure such as a joint composed of bone and cartilage, a technique for generating an image in which bone and cartilage are clearly distinguished is required. However, since the CT values of the soft tissues of the human body (for example, muscles or fat, etc.) and the CT values of cartilage are similar, the boundary between the soft tissue and cartilage cannot be accurately distinguished. Currently, a technique for generating an image in which the contours of bone and cartilage are clarified is not known.

[0006] The present disclosure has been made to solve such problems, and an object thereof is to provide an osteochondral delineation device and the like that can obtain an image in which the contours of bone and cartilage are clarified.

[0007] An osteochondral imaging device according to one aspect of the present disclosure comprises an acquisition unit that acquires a first image obtained by imaging with a CT scanner using the dual-energy method on a part of a human body that includes a joint composed of bone and cartilage, and a generation unit that generates information derived from the first image based on the acquired first image, wherein the first image is an image in which the contours of the bone and the cartilage are clearly defined, the boundary between the bone and the cartilage is clearly defined, and the boundary between the cartilage and the soft tissue included in the joint is clearly defined.

[0008] A method for depicting osteochondral tissue according to one aspect of the present disclosure is a computer-based method for depicting osteochondral tissue, comprising the steps of: acquiring a first image obtained by imaging a part of a person that includes a joint composed of bone and cartilage using a dual-energy CT scanner; and generating information derived from the first image based on the acquired first image, wherein the first image is an image in which the contours of the bone and cartilage are clearly defined, the boundary between the bone and the cartilage is clearly defined, and the boundary between the cartilage and the soft tissue included in the joint is clearly defined.

[0009] A program according to one aspect of this disclosure is a program for causing a computer to execute the osteochondral imaging method.

[0010] According to this disclosure, an osteochondral imaging device, etc., is provided that can acquire images in which the contours of bone and cartilage are clearly defined.

[0011] Figure 1 is a block diagram showing the configuration of a system including an osteochondral imaging device according to an embodiment. Figure 2 is a diagram illustrating the principle of the dual-energy method. Figure 3 is a schematic diagram showing the specific processing flow when the generation unit of the osteochondral imaging device generates a three-dimensional osteochondral model from a plurality of first images. Figure 4 is a sequence diagram showing the operations performed by the system including the osteochondral imaging device according to an embodiment when training a machine learning model. Figure 5 is a schematic diagram showing the specific processing flow of step S18 in Figure 4. Figure 6 is a sequence diagram showing the operations performed by the system including the osteochondral imaging device according to an embodiment during operation. Figure 7 is a schematic diagram showing the specific processing flow from step S24 to step S25 in Figure 6. Figure 8 is a CT image showing the contour of the cartilage of a young pig. Figure 9 is a 3D osteochondral model created from a CT image obtained when the material condition setting is brush-enhanced / calcium suppressed. Figure 10 is a 3D data created from the femur of a young pig using optical scanning. Figure 11 is a diagram showing part of the analysis results by EPMA elemental analysis. Figure 12 shows an example of a CT image obtained during verification.

[0012] The embodiments of this disclosure will be described below with reference to the drawings. The embodiments described below are all specific examples of this disclosure. Therefore, the numerical values, shapes, materials, components, arrangement positions of components, and connection configurations shown in the following embodiments are examples and are not intended to limit this disclosure. Accordingly, any components in the following embodiments that are not described in the independent claims of this disclosure will be described as optional components.

[0013] Please note that each figure is a schematic diagram and not necessarily a strictly accurate representation. Therefore, the scale and other aspects may not necessarily be consistent across all figures. In addition, the same reference numerals are used for substantially identical components in each figure, and redundant explanations are omitted or simplified.

[0014] Furthermore, in this specification, the terms "up" and "down" do not necessarily refer to the upward (vertically upward) and downward (vertically downward) directions in absolute spatial perception.

[0015] Furthermore, the osteochondral imaging device according to this embodiment is applicable to humans (hereinafter also referred to as patients).

[0016] (Embodiment) [Configuration] First, the configuration of the system including the osteochondral imaging device according to this embodiment will be described. Figure 1 is a block diagram showing the configuration of the system including the osteochondral imaging device 3 according to this embodiment. The system shown in Figure 1 comprises a CT scanner 1, an X-ray scanner 2, and an osteochondral imaging device 3.

[0017] The CT scanner 1 is a diagnostic examination device that uses X-rays to capture detailed cross-sectional images of the inside of a person (i.e., inside the body). In this embodiment, the CT scanner 1 is operated by a medical professional to capture a first image showing parts of the human body that include joints composed of bone and cartilage. In this embodiment, the CT scanner 1 acquires the first image using dual-energy CT scanning (hereinafter also simply referred to as dual-energy scanning). Dual-energy CT scanning is a technique that captures a CT image (referred to as the first image in this specification) by simultaneously irradiating with two types of X-rays with different tube voltages.

[0018] Furthermore, the CT scan device 1 includes a communication unit 11 and an imaging unit 12.

[0019] The communication unit 11 is a processing unit (communication circuit) for communicating with the osteochondral imaging device 3. The communication unit 11 transmits the first image captured by the imaging unit 12 to the osteochondral imaging device 3. The communication performed by the communication unit 11 may be wireless communication or wired communication. Furthermore, the specifications of the communication performed by the communication unit 11 are not particularly limited.

[0020] The imaging unit 12 images a part of the human body that includes a joint composed of bone and cartilage, and acquires it as a first image. In this specification, the first image is an image in which the contours of bone and cartilage are clearly defined. "Clearly defined bone contour" means that the boundary between bone and cartilage is clearly defined, and "clearly defined cartilage contour" means that the boundary between bone and cartilage, and the boundary between cartilage and soft tissue are clearly defined. Soft tissue is the tissue included in a joint composed of bone and cartilage, such as muscle and fat.

[0021] X-ray inspection device 2 is a diagnostic inspection device that uses X-rays to image bones inside a person (i.e., inside the body). X-ray inspection device 2, operated by a medical professional, captures a second image of a person's body, specifically showing areas that include joints composed of bone and cartilage. The second image is an image of the area that includes the parts shown in the first image.

[0022] Furthermore, the X-ray inspection device 2 includes a communication unit 21 and an imaging unit 22.

[0023] The communication unit 21 is a processing unit (communication circuit) for communicating with the osteochondral imaging device 3. The communication unit 21 transmits the second image captured by the imaging unit 22 to the osteochondral imaging device 3. The communication performed by the communication unit 21 may be wireless communication or wired communication. Furthermore, the specifications of the communication performed by the communication unit 21 are not particularly limited.

[0024] The imaging unit 22 images parts of the human body that include joints composed of bone and cartilage, and acquires them as a second image.

[0025] The osteochondral imaging device 3 is a device used to evaluate a patient's articular cartilage, cartilage other than articular cartilage (e.g., costal cartilage and auricular cartilage), joint shape, and joint diseases (e.g., fractures, dislocations, osteoarthritis, and rheumatoid arthritis). The osteochondral imaging device 3 may be a portable information terminal such as a tablet, or a stationary information terminal such as a personal computer. The osteochondral imaging device 3 comprises an acquisition unit 31, a storage device 32, an information processing unit 33, and a display unit 34.

[0026] The acquisition unit 31 is a processing unit (communication circuit) for communicating with the CT scanner 1 and the X-ray scanner 2. The acquisition unit 31 acquires a first image from the CT scanner 1 and a second image from the X-ray scanner 2. The communication performed by the acquisition unit 31 may be wireless communication or wired communication. Furthermore, the communication standard performed by the acquisition unit 31 is not particularly limited.

[0027] The storage device 32 is a memory that holds the first image and the second image acquired by the acquisition unit 31, as well as control programs and the like executed by the information processing unit 33.

[0028] The information processing unit 33 controls the entire osteochondral imaging device 3. The information processing unit 33 is implemented by a microcomputer or processor. That is, the functions of the information processing unit 33 are realized by the microcomputer or processor executing a control program stored in the storage device 32. The information processing unit 33 also includes a generation unit 331.

[0029] The generation unit 331 generates information derived from the first image based on the first image acquired by the acquisition unit 31. The generation unit 331 generates a three-dimensional osteochondral model based on multiple first images of the same person but with different imaging areas within the same body part (specifically, multiple first images acquired by the acquisition unit 13). The generation unit 331 also has a trained machine learning model 331a, and generates the three-dimensional osteochondral model by inputting the second image held in the storage device 32 into the machine learning model 331a. The three-dimensional osteochondral model is a three-dimensional image showing the area containing the joint (bone and cartilage, etc.) captured as the first or second image.

[0030] The machine learning model 331a is a model based on a TL embedded network consisting of an autoencoder and a convolutional neural network (CNN). The machine learning model 331a is a model that has been pre-trained before generating a 3D osteochondral model based on the second image. Specifically, the generation unit 331 trains the machine learning model 331a using the acquired (i.e., stored in the memory device 32) first and second images as input data, and the 3D osteochondral model generated from the first image as output data. The specific training of the machine learning model 331a will be described later.

[0031] The display unit 34 displays, for example, a three-dimensional osteochondral model generated by the generation unit 331. The display unit 34 is implemented by a display panel such as a liquid crystal panel or an organic EL panel. The display unit 34 may also display a first image, a second image, etc., as needed.

[0032] [Regarding the Principle of the Dual Energy Method] The inventors have found that by applying a relational expression obtained from the effective atomic number of human bone and cartilage and the linear attenuation coefficient of human bone and cartilage, the CT scanner 1 using the dual energy method may be able to clarify (i.e., differentiate) the boundary between human cartilage and soft tissue. The principle of the dual energy method used by the CT scanner 1 will be explained below. Figure 2 is a diagram illustrating the principle of the dual energy method. Figure 2 is a cross-sectional view showing an example of when the CT scanner 1 images a part of a human body that includes a joint. Note that the X-ray tube voltage shown in Figure 2 is just an example, and other tube voltages may be used.

[0033] As shown in Figure 2, the detector of the CT scanner 1 detects (measures) the count values ​​(Count 80kV and Count 140kV) of two types of X-rays that penetrate human bone and cartilage. The count values ​​of the two types of X-rays are shown by the following formula, for example. L1 is the effective atomic number of human bone, and L2 is the effective atomic number of human cartilage. μ1 and μ'1 are the linear attenuation coefficients of human bone, and μ2 and μ'2 are the linear attenuation coefficients of human cartilage.

[0034] (Math. 1) Count80kV=μ1L1+μ2L2 (Formula 1) (Math. 2) Count140kV=μ′1L1+μ′2L2 (Formula 2)

[0035] The inventors found that by applying the above formulas (1) and (2), the imaging unit 12 of the CT scanner 1 may be able to clarify the boundary between cartilage and soft tissue. Based on this finding, the inventors considered that by using the CT scanner 1 employing the dual-energy method, it would be possible to generate a first image in which the contours of human bone and cartilage are clearly defined, leading to the present invention.

[0036] In the above, an example was described in which the CT scanner 1 generates a first image that depends on the effective atomic number of human bone and cartilage in order to explain the principle of the dual-energy method, but the invention is not limited to this example. For example, a CT scanner 1 that uses a dual-energy method that depends on a physical quantity other than the effective atomic number is also within the scope of this disclosure.

[0037] [Generation of a 3D Osteochondral Model] The following describes a specific example of when the generation unit 331 of the osteochondral imaging device 3 generates a 3D osteochondral model based on a first image held in the storage device 32. Figure 3 is a schematic diagram showing the specific processing flow when the generation unit 331 of the osteochondral imaging device 3 generates a 3D osteochondral model from a plurality of first images. Figure 3(a) shows an example of a plurality of first images held in the storage device 32. Figure 3(b) shows an example of a 3D osteochondral model generated by the generation unit 331 based on the plurality of first images shown in Figure 3(a). The first image shown in Figure 3(a) is an image of the elbow joint among the parts of a human body.

[0038] The multiple first images shown in Figure 3(a) are images of the same person, but of the same body part, with different imaging areas. Furthermore, the areas indicated by hatching in the first image are regions where human cartilage is present. For the sake of clarity, an example is shown where the areas containing human cartilage are indicated by hatching; however, these areas do not necessarily need to be indicated by hatching or other means.

[0039] The three-dimensional osteochondral model shown in (b) of FIG. 3 is an example of the three-dimensional osteochondral model generated by the generation unit 331. The generation unit 331 generates a three-dimensional osteochondral model by extracting and connecting the contours of bone and cartilage from each of a plurality of first images.

[0040] [Learning of Machine Learning Model] FIG. 4 is a sequence diagram showing operations executed by a system including the osteochondral delineation apparatus 3 according to the present embodiment during learning of the machine learning model 331a. The operations shown in FIG. 4 are operations (that is, learning of the machine learning model) executed before the osteochondral delineation apparatus 3 generates a three-dimensional osteochondral model from the second image. Further, hereinafter, operations in the case where conditions (for example, the effective atomic number of cartilage and the effective atomic number of soft tissue) that can clarify the boundary between human cartilage and soft tissue are specified in advance will be described.

[0041] First, the imaging unit 12 of the CT inspection apparatus 1 images a site including a joint composed of bone and cartilage among human sites by the operation of a medical staff (S11). The communication unit 11 of the CT inspection apparatus 1 transmits the first image obtained by the imaging in step S11 to the osteochondral delineation apparatus 3 (S12). The first image is an image in which the contours of bone and cartilage are clarified.

[0042] The acquisition unit 31 of the osteochondral delineation apparatus 3 acquires the first image transmitted in step S12 (S13). The acquisition unit 31 stores the first image acquired in step S13 in the storage device 32.

[0043] The imaging unit 22 of the X-ray inspection apparatus 2 images a site including a joint composed of bone and cartilage among human sites by the operation of a medical staff. (S14). The site shown in the second image imaged in step S14 is the same location as the site shown in the first image imaged in step S11. Also, the person from whom the second image is imaged in step S14 is the same person as the person from whom the first image is imaged in step S11.

[0044] The communication unit 21 of the X-ray inspection apparatus 2 transmits the second image obtained by the imaging in step S14 to the osteochondral delineation apparatus 3 (S15).

[0045] The acquisition unit 31 of the osteochondral imaging device 3 acquires the second image transmitted in step S15 (S16). The acquisition unit 31 stores the second image acquired in step S16 in the storage device 32.

[0046] The generation unit 331 of the osteochondral imaging device 3 generates a three-dimensional osteochondral model based on a plurality of first images held in the storage device 32 (S17). In step S17, the plurality of first images used by the generation unit 331 are images with different imaging regions within the parts of the same person. Also, step S17 corresponds to a step of generating information derived from the acquired first image.

[0047] The generation unit 331 of the osteochondral imaging device 3 uses the first image acquired in step S13 and the second image acquired in step S16 as input data, and inputs the three-dimensional osteochondral model generated in step S17 as output data into the machine learning model 331a, thereby training the machine learning model 331a (S18).

[0048] Hereinafter, the processing flow in the training of the machine learning model 331a described in FIG. 4 will be described using a specific example. FIG. 5 is a schematic diagram showing the specific processing flow of step S18 in FIG. 4. The first image and the second image shown in FIG. 5 are images of the elbow joint among the parts of a person.

[0049] The training of the machine learning model 331a is performed in three main steps.

[0050] First, let's explain the first step ((i) AutoEncoder). In the first step, the generation unit 331 inputs the first image as input data and the 3D osteochondral model as output data to the machine learning model 331a. Specifically, the machine learning model 331a performs dimensionality reduction by extracting the features (Latent vector) from one first image obtained by imaging with the CT scanner 1 in order to extract the features (Latent vector) necessary to generate the 3D osteochondral model (Encoder). Then, the machine learning model 331a learns the process of reconstructing the 3D image from the extracted features using the 3D osteochondral model of the output data as the ground truth (Decoder).

[0051] Next, we will explain the second step ((ii) Regression Network). In the second step, the generation unit 331 takes a second image obtained by imaging with the X-ray inspection device 2 as input data and inputs a three-dimensional osteochondral model as output data to the machine learning model 331a. Specifically, the machine learning model 331a extracts features (Latent vectors) from the input second image (Predictor). Then, using the three-dimensional osteochondral model of the output data as the ground truth, the machine learning model 331a learns the process of generating a three-dimensional osteochondral model from the second image by coordinating with the process of reconstructing a three-dimensional image from the extracted features (Decoder in the first step).

[0052] Finally, let's explain the third step (optimization). In the third step, the generation unit 331 inputs the second image as input data to the machine learning model 331a and optimizes the machine learning model 331a based on the goodness of fit between the features extracted from the second image and the 3D osteochondral model generated based on the second image. The generation unit 331 performs optimization until the loss term, which represents the goodness of fit between the features extracted from the second image and the 3D osteochondral model generated based on the second image, falls below a certain level.

[0053] Thus, before generating a three-dimensional osteochondral model based on a second image obtained by imaging with the X-ray inspection device 2, the generation unit 331 of the osteochondral imaging device 3 trains a machine learning model 331a using a first image obtained by imaging with the CT inspection device 1 and a second image obtained by imaging with the X-ray inspection device 2 as input data, and a three-dimensional osteochondral model as training data as output data. As a result, the generation unit 331 of the osteochondral imaging device 3 can generate a three-dimensional osteochondral model by inputting only the second image obtained by imaging with the X-ray inspection device 2 into the machine learning model 331a.

[0054] [Operation of the Machine Learning Model] Figure 6 is a sequence diagram showing the operations performed by the system including the osteochondral imaging device 3 according to the embodiment during operation. Note that the operations shown in Figure 6 are performed after the operations shown in Figure 4 (i.e., training of the machine learning model 331a) are completed.

[0055] First, the imaging unit 22 of the X-ray inspection device 2 is operated by a medical professional to image a part of the human body that includes joints composed of bone and cartilage (S21). In step S21, the imaging unit 22 images the same parts as those learned by the machine learning model 331a (i.e., parts that include the parts imaged by the imaging unit 12 of the CT inspection device 1 in step S11, or the parts imaged by the imaging unit 22 of the X-ray inspection device 2 in step S14).

[0056] The communication unit 21 of the X-ray inspection device 2 transmits the second image obtained in step S21 to the osteochondral imaging device 3 (S22).

[0057] The acquisition unit 31 of the osteochondral imaging device 3 acquires the second image transmitted in step S22 (S23). The acquisition unit 31 then stores the second image acquired in step S23 in the storage device 32.

[0058] The generation unit 331 of the osteochondral imaging device 3 inputs the second image held in the storage device 32 to the machine learning model 331a (S24) and generates a three-dimensional osteochondral model (S25).

[0059] The display unit 34 of the osteochondral imaging device 3 displays the three-dimensional osteochondral model generated in step S25 (S26).

[0060] The following section will explain the processing flow for generating the three-dimensional osteochondral model described in Figure 6, using a specific example. Figure 7 is a schematic diagram showing the specific processing flow from step S24 to step S25 in Figure 6. The second image shown in Figure 7 is an image of the elbow joint, one of the body parts of a human.

[0061] As shown in Figure 7, the generation unit 331 inputs a second image obtained by imaging by the X-ray inspection device 2 (the image shown above the machine learning model 331a) into the machine learning model 331a (S24), and generates a three-dimensional osteochondral model (the image shown to the right of the machine learning model 331a) (S25).

[0062] Furthermore, if the osteochondral imaging device 3 generates a three-dimensional osteochondral model for a site other than the site shown in Figure 7, the system including the osteochondral imaging device 3 according to this embodiment can perform the operation shown in Figure 4 again for that other site to pre-train the machine learning model 331a.

[0063] The osteochondral imaging device 3 according to this embodiment has the following advantages.

[0064] Generally, MRI scans are used to evaluate a patient's articular cartilage. However, MRI scans require restraining the patient's movement, which presents challenges, especially for pediatric patients. Furthermore, sedatives or anesthetics may be administered to pediatric patients to acquire images using the MRI scanner, which places a significant burden on them.

[0065] Furthermore, MRI machines are generally installed in large hospitals and are often not available in smaller hospitals such as clinics in urban areas, which presents a problem in that examinations using MRI machines are not possible.

[0066] Therefore, the osteochondral imaging device 3 generates a three-dimensional osteochondral model using only one second image obtained from imaging by the X-ray examination device 2, which can acquire images in a relatively short time. As a result, the osteochondral imaging device 3 can accurately depict the shape of the articular cartilage and realize an examination that is less burdensome for the patient (especially pediatric patients).

[0067] Furthermore, the osteochondral imaging device 3 generates a three-dimensional osteochondral model from a single X-ray image (i.e., a second image) obtained by imaging with the X-ray inspection device 2. This enables doctors working in small hospitals that do not have an MRI device but do have an X-ray inspection device 2, such as clinics in urban areas, to make accurate diagnoses and provide appropriate treatment.

[0068] [Verification] The inventors conducted several verifications to identify combinations of material values ​​that can visualize the contour of cartilage (hereinafter also referred to as the cartilage region) by imaging with the CT scanner 1 using the dual-energy method. Specifically, the inventors conducted several verifications to identify combinations of effective atomic numbers (hereinafter also referred to as material conditions) of two known substances that can emphasize and visualize the contour of cartilage (particularly the boundary between cartilage and soft tissue). The details of the verifications conducted by the inventors are described below in order.

[0069] [1. Verification using young pigs] First, the inventors used the knee joints of young pigs to explore combinations of material values ​​that could visualize the cartilage region of young pigs. In other words, the inventors explored material conditions that could clarify the contour of the cartilage of young pigs. Young pigs refer to pigs in their early stages, and in this verification, the verification was conducted on pigs that were four weeks old.

[0070] The inventors performed imaging using a CT scanner 1 employing the dual-energy method with various material condition settings. Material condition settings refer to the setting of combinations of effective atomic numbers of known substances (material conditions) that can be set when imaging with the CT scanner 1 employing the dual-energy method. In the CT scanner 1 used by the inventors, there are a total of 45 components that can be set in the material condition settings, and the number of components that can be set simultaneously in the material condition settings is 2. Therefore, the inventors comprehensively searched for combinations of material values ​​from 1980 (= 45 × 44) combinations of material values ​​that could depict the contour of cartilage of young pigs. Figure 8 shows a CT image when the contour of cartilage of young pigs was depicted. Figure 8(a) shows a CT image obtained when the material condition setting was Brushite-weighted / Calcium-suppressed. Figure 8(b) shows a CT image obtained when the material conditions were set to brushite enhancement / hydroxyapatite suppression.

[0071] As shown in Figures 8(a) and 8(b), it was confirmed that the contour of the cartilage of young pigs (the area shown by the dashed line in each figure) can be depicted when the material conditions are set to brushite emphasis / calcium suppression, or when the material conditions are set to brushite emphasis / hydroxyapatite suppression.

[0072] Furthermore, the inventors created a three-dimensional osteochondral model showing the three-dimensional shape of the bones and cartilage of young pigs from CT images in which the contours of the bones and cartilage of young pigs were clarified by setting the material conditions to brush-enhanced / calcium suppression, and confirmed the accuracy of the created three-dimensional osteochondral model. In addition, the inventors used three-dimensional data created from the bones and cartilage constituting the femur excised from young pigs as a comparison target for the three-dimensional osteochondral model. In this verification, the inventors excised femurs from young pigs after imaging with the CT scanner 1 using the dual-energy method, and created three-dimensional data from the excised femurs. The inventors removed the soft tissue from the excised femurs and then created three-dimensional data from the femurs using optical scanning. In other words, the three-dimensional data created from the femurs is data showing the actual three-dimensional shape of the bones and cartilage. Figure 9 shows the three-dimensional osteochondral model created from CT images obtained when the material conditions were set to brush-enhanced / calcium suppression. Figure 9(a) shows a three-dimensional osteochondral model viewed from the positive Z-axis direction. Figure 9(b) shows a three-dimensional osteochondral model viewed from the negative X-axis direction. Figure 10 shows three-dimensional data created from the femur of a young pig using optical scanning. Figure 10(a) shows the three-dimensional data viewed from the positive Z-axis direction. Figure 10(b) shows the three-dimensional data viewed from the negative X-axis direction.

[0073] The inventors confirmed the accuracy of the 3D osteochondral model by comparing the 3D osteochondral model shown in Figure 9 with the 3D data shown in Figure 10. Specifically, the inventors confirmed the accuracy of the 3D osteochondral model by comparing the degree of agreement between the 3D osteochondral model shown in Figure 9 and the 3D data shown in Figure 10. As a result of comparing the 3D osteochondral model shown in Figure 9 with the 3D data shown in Figure 10, the Dice coefficient was 0.87, which was a relatively good result. In other words, a high degree of agreement was obtained between the 3D osteochondral model and the 3D data shown in Figure 10, confirming that the accuracy of the 3D osteochondral model was relatively good.

[0074] [2. Component Analysis of Articular Cartilage of Young Pigs] Next, the inventors performed a component analysis of articular cartilage of young pigs (for example, cartilage in joint areas including the elbow joint). Specifically, similar to the search for material conditions conducted in [1. Verification using young pigs] above, the inventors performed a component analysis of articular cartilage of young pigs in order to identify the effective atomic number (in other words, material conditions for articular cartilage of young pigs) that can emphasize and depict the contour of the articular cartilage of young pigs. The reason the inventors conducted such a search was that there were no existing programs that had material condition settings that could emphasize and display the contour of articular cartilage of young pigs on CT images.

[0075] The inventors extracted articular cartilage from young pigs and performed a detailed component analysis of the extracted articular cartilage. Specifically, the inventors performed elemental analysis of C, H, and N using a high-sensitivity C, H, N quantitative analyzer (SUMIGRAPH NCH-22F) and elemental analysis of EPMA (Electron Probe Micro Analyzer) using an electron probe microanalyzer (EPMA-1600EPMA). Figure 11 shows a portion of the analysis results from the EPMA elemental analysis. In Figure 11, the vertical axis represents the count (i.e., X-ray intensity), and the horizontal axis represents the wavelength (nm). As shown in Figure 11, the elements C, N, and O were observed. Furthermore, elemental analysis using a high-sensitivity C, H, N quantitative analyzer and EPMA elemental analysis using an electron beam microanalyzer revealed that the mass fraction of the articular cartilage extracted from young pigs was O: 59.0%, C: 23.5%, N: 6.9%, H: 6.7%, Na: 0.6%, Mg: 0.2%, P: 0.6%, S: 0.9%, Cl: 0.2%, K: 0.4%, Ca: 0.5%, and Cu: 0.5%.

[0076] [3. Verification using CT images of the elbow joints of pediatric patients] After confirming through animal experiments using young pigs that the contour of cartilage could be depicted by imaging with the CT scanner 1 using the dual-energy method, the inventors explored material conditions that could depict the contours of bone and cartilage using CT images of the elbow joints of pediatric patients. These CT images were taken during the course of clinical treatment. In this verification, CT images of the elbow joints of five pediatric patients were used. The average age of the five pediatric patients was 8 years, and all were boys. Furthermore, imaging was performed on the left and right elbow joints of each pediatric patient using the CT scanner 1.

[0077] The inventors used the material condition settings obtained from the animal experiments using young pigs described above as candidates for exploration. Furthermore, the inventors explored the optimal material condition settings by referring to the difference in material density between cartilage and soft tissue and the CNR (Contrast Noise Ratio). Figure 12 shows an example of a CT image obtained in the verification. Figure 12(a) shows an example of a simple CT image. Here, a simple CT image is not a CT image obtained by imaging with the CT scanner 1 using the dual-energy method, but a CT image obtained by irradiating with one type of X-ray. In other words, a simple CT image is a CT image obtained by imaging with a normal CT scanner. Also, Figure 12(b) shows a CT image obtained when the material condition setting is iron-weighted / lanthanum-suppressed. Figure 12(c) shows a CT image obtained when the material settings were set to iron-weighted / iodine-suppressed. Figure 12(d) shows a CT image obtained when the material settings were set to calcaneus-weighted / calcium-suppressed. Figure 12(e) shows a CT image obtained when the material settings were brushite-weighted / hydroxyapatite-suppressed. Figure 12(f) shows a CT image obtained when the material settings were set to hydroxyapatite-weighted / calcaneus-suppressed. Note that the CT images shown in Figures 12(b) to 12(f) are CT images obtained by imaging with a CT scanner 1 using the dual-energy method. Also, the CT images shown in Figure 12 are images of the elbow joint of the same pediatric patient.

[0078] As shown in Figure 12(a), in a simple CT image, the CT values ​​of the soft tissues and cartilage of the human body are similar, so the boundary between cartilage and soft tissues of the human body is not clear, and the contour of the cartilage is not depicted. In other words, the CT image shown in Figure 12(a) is not an image in which the contour of the cartilage is clearly defined.

[0079] On the other hand, as shown in Figures 12(b) to 12(f), when the material condition setting is one of the five combinations described above, the CT image obtained by imaging with the CT scanner 1 using the dual-energy method clearly shows the boundary between the cartilage and soft tissue of the human body, and the contour of the cartilage (the area shown by the dashed line in each figure) is depicted. Therefore, when one of the five combinations described above is set as the material condition setting for the CT scanner 1, a three-dimensional osteochondral model showing the joint area of ​​a pediatric patient can be generated by using the CT image obtained by imaging with the CT scanner 1 using the dual-energy method.

[0080] Based on the above verification, the inventors determined that five condition settings—iron-enhanced / lanthanum-suppressed, iron-enhanced / iodine-suppressed, carcaneus-enhanced / calcium-suppressed, brushite-enhanced / hydroxyapatite-suppressed, and hydroxyapatite-enhanced / carcaneus-suppressed—are the optimal material condition settings for clearly defining the contours of human cartilage. As a result, the osteochondral imaging device 3 can generate a three-dimensional osteochondral model showing the joint areas (parts composed of bone and cartilage) of the human body by using CT images obtained from imaging by the CT scanner 1 with one of the five material condition settings applied.

[0081] [Effects] The osteochondral imaging device 3 according to this embodiment includes an acquisition unit 31 that acquires a first image obtained by imaging with a CT scanner 1 using the dual-energy method on a part of the human body that includes a joint composed of bone and cartilage, and a generation unit 331 that generates information derived from the first image based on the acquired first image. The first image is an image in which the contours of bone and cartilage are clearly defined, the boundary between bone and cartilage is clearly defined, and the boundary between cartilage and the soft tissue included in the joint is clearly defined.

[0082] Since this osteochondral imaging device 3 acquires the first image obtained by imaging with the CT scanner 1 using the dual-energy method, it is possible to obtain an image in which the contours of bone and cartilage are clearly defined.

[0083] Furthermore, in the osteochondral imaging device 3 according to this embodiment, the CT scanner 1 generates a first image based on the effective atomic numbers of the bone and cartilage.

[0084] Since the CT scanner 1 generates a first image that clarifies the contours of human bones and cartilage based on the effective atomic numbers of bones and cartilage, the osteochondral imaging device 3 can acquire an image in which the contours of bones and cartilage are clearly defined.

[0085] Furthermore, in the osteochondral imaging device 3 according to this embodiment, the generation unit 331 generates a three-dimensional osteochondral model based on a plurality of first images acquired by the acquisition unit 31, which have different imaging areas within the same region. The three-dimensional osteochondral model is a three-dimensional image that shows the region containing the imaged joint.

[0086] Such an osteochondral imaging device 3 can generate a three-dimensional osteochondral model by extracting and stitching together the contours of bone and cartilage from each of the multiple first images acquired.

[0087] Furthermore, in the osteochondral imaging device 3 according to this embodiment, the acquisition unit 31 further acquires a second image in which the same region as the region included in the first image is captured. The second image is an image obtained by imaging with the X-ray inspection device 2. The generation unit 331 has a machine learning model 331a, and the machine learning model 331a is trained using the acquired first and second images as input data and the generated three-dimensional osteochondral model as output data.

[0088] This osteochondral imaging device 3 pre-trains a machine learning model 331a before generating a three-dimensional osteochondral model based on a single second image obtained by imaging with the X-ray inspection device 2. Therefore, by inputting only the single second image obtained by imaging with the X-ray inspection device 2 into the machine learning model 331a, a three-dimensional osteochondral model can be generated.

[0089] Furthermore, the osteochondral imaging device 3 according to this embodiment further includes a display unit 34. The generation unit 331 generates a three-dimensional osteochondral model by inputting another second image acquired by the acquisition unit 31 into the trained machine learning model 331a without inputting the first image into the trained machine learning model 331a, and displays the generated three-dimensional osteochondral model on the display unit 34.

[0090] This osteochondral imaging device 3 generates a three-dimensional osteochondral model using only one second image obtained from imaging by the X-ray examination device 2, which can acquire images in a relatively short time. As a result, the osteochondral imaging device 3 can accurately depict the shape of the articular cartilage and realize an examination that is less burdensome for the patient (especially pediatric patients).

[0091] Furthermore, the osteochondral imaging device 3 generates a three-dimensional osteochondral model from a single X-ray image (i.e., a second image) obtained by imaging with the X-ray inspection device 2. This enables doctors working in small hospitals that do not have an MRI device but do have an X-ray inspection device 2, such as clinics in urban areas, to make accurate diagnoses and provide appropriate treatment.

[0092] Furthermore, the osteochondral imaging method according to this embodiment is a computer-based osteochondral imaging method that includes the steps of: acquiring a first image (S13) obtained by imaging with a CT scanner 1 using the dual-energy method on a part of a person that includes a joint composed of bone and cartilage; and generating information derived from the first image (S17) based on the acquired first image, wherein the first image is an image in which the contours of bone and cartilage are clearly defined, the boundary between bone and cartilage is clearly defined, and the boundary between cartilage and the soft tissue included in the joint is clearly defined.

[0093] In this method of depicting osteochondral tissue, the first image obtained by imaging with a CT scanner 1 using the dual-energy method is acquired, making it possible to obtain an image in which the contours of bone and cartilage are clearly defined.

[0094] Furthermore, the program according to this embodiment is a program for causing a computer to execute the above-described method of depicting osteochondral tissue.

[0095] Such a program produces the same effects as the osteochondral imaging method according to this embodiment.

[0096] (Other) The osteochondral imaging device, etc. relating to this disclosure has been described above based on the above embodiments, but is not limited to the above embodiments. Without departing from the spirit of this disclosure, various modifications that a person skilled in the art can conceive of may be applied to the above embodiments, and forms constructed by combining components from different embodiments may also be included within the scope of one or more embodiments.

[0097] For example, in the above embodiment, a process executed by a specific processing unit may be executed by another processing unit. Furthermore, the order of multiple processes may be changed, or multiple processes may be executed in parallel.

[0098] Furthermore, the communication method between devices in the above embodiment is not particularly limited. In addition, relay devices (not shown) may be involved in the communication between devices.

[0099] Furthermore, the order of processing described in the flowchart of the above embodiment is just one example. The order of multiple processing steps may be changed, and multiple processing steps may be executed in parallel.

[0100] Furthermore, although the osteochondral imaging device was implemented by a single device in the above embodiment, the components of the osteochondral imaging device may be distributed among multiple devices. In such a case, the components of the osteochondral imaging device may be distributed among the multiple devices in any manner.

[0101] Furthermore, in the above embodiment, each component may be implemented by hardware. For example, each component may be a circuit (or integrated circuit). These circuits may form a single circuit as a whole, or they may be separate circuits. Also, each of these circuits may be a general-purpose circuit or a dedicated circuit.

[0102] Furthermore, some or all of the functions of the osteochondral imaging device according to the above embodiment may be realized by a processor such as a CPU executing a program.

[0103] Some or all of the components constituting each of the above devices may consist of detachable IC cards or standalone modules attached to each device. The IC card or module is a computer system composed of a microprocessor, ROM, RAM, etc. The IC card or module may also include a multi-functional LSI. The microprocessor operates according to a computer program, thereby enabling the IC card or module to achieve its function. The IC card or module may also be tamper-resistant.

[0104] Furthermore, the general or specific embodiments of this disclosure may be implemented as a system, apparatus, method, integrated circuit, computer program, or recording medium such as a computer-readable CD-ROM. They may also be implemented as any combination of a system, apparatus, method, integrated circuit, computer program, and recording medium.

[0105] For example, the technology relating to this disclosure may be implemented as a method for visualizing osteochondral tissue performed by a computer, or as a program for causing a computer to perform such an osteochondral tissue visualization method. Furthermore, the technology relating to this disclosure may be implemented as a computer-readable non-temporary recording medium on which the above program is recorded, or as a program product including the above program.

[0106] [Note] (Technical 1) An osteochondral imaging device comprising: an acquisition unit that acquires a first image obtained by imaging with a CT scanner using the dual-energy method on a part of the human body that includes a joint composed of bone and cartilage; and a generation unit that generates information derived from the first image based on the acquired first image, wherein the first image is an image in which the contours of the bone and the cartilage are clearly defined, the boundary between the bone and the cartilage is clearly defined, and the boundary between the cartilage and the soft tissue included in the joint is clearly defined.

[0107] (Technology 2) The osteochondral imaging device according to Technology 1, wherein the CT scanner generates the first image based on the effective atomic numbers of the bone and the cartilage.

[0108] (Technology 3) The osteochondral imaging device according to Technology 1 or 2, wherein the generation unit generates a three-dimensional osteochondral model based on a plurality of first images acquired by the acquisition unit, each of which has a different imaging area within the same region, and the three-dimensional osteochondral model is a three-dimensional image showing the region in which the captured joint is located.

[0109] (Technical 4) The osteochondral imaging apparatus according to Technical 3, wherein the acquisition unit further acquires a second image in which the same part as the part included in the first image is captured, the second image is an image obtained by imaging with an X-ray inspection device, and the generation unit has a machine learning model, and the machine learning model is trained using the acquired first image and second image as input data and the generated three-dimensional osteochondral model as output data.

[0110] (Technical 5) The osteochondral imaging device according to Technical 4, further comprising a display unit, wherein the generation unit generates a three-dimensional osteochondral model by inputting another second image acquired by the acquisition unit into the machine learning model without inputting the first image into the machine learning model that has been trained, and displays the generated three-dimensional osteochondral model on the display unit.

[0111] (Technical 6) A method for depicting osteochondral tissue performed by a computer, comprising the steps of: acquiring a first image obtained by imaging a part of a person that includes a joint composed of bone and cartilage using a dual-energy CT scanner; and generating information derived from the first image based on the acquired first image, wherein the first image is an image in which the contours of the bone and cartilage are clearly defined, the boundary between the bone and the cartilage is clearly defined, and the boundary between the cartilage and the soft tissue included in the joint is clearly defined.

[0112] (Technical 7) A program for causing the computer to execute the osteochondral imaging method described in Technical 6.

[0113] The osteochondral imaging device and the like described herein are useful, for example, as an image evaluation support tool for evaluating the condition of a patient's articular cartilage.

[0114] 1 CT scanner 11, 21 Communication unit 12, 22 Imaging unit 2 X-ray scanner 3 Osteochondral imaging device 31 Acquisition unit 32 Storage unit 33 Information processing unit 331 Generation unit 331a Machine learning model 34 Display unit

Claims

1. An osteochondral imaging device comprising: an acquisition unit that acquires a first image obtained by imaging with a CT scanner using the dual-energy method on a part of the human body that includes a joint composed of bone and cartilage; and a generation unit that generates information derived from the first image based on the acquired first image, wherein the first image is an image in which the contours of the bone and the cartilage are clearly defined, the boundary between the bone and the cartilage is clearly defined, and the boundary between the cartilage and the soft tissue included in the joint is clearly defined.

2. The osteochondral imaging apparatus according to claim 1, wherein the CT scanner generates the first image based on the effective atomic numbers of the bone and the cartilage.

3. The osteochondral imaging device according to claim 1 or 2, wherein the generation unit generates a three-dimensional osteochondral model based on a plurality of first images acquired by the acquisition unit, each of which has a different imaging area within the same region, and the three-dimensional osteochondral model is a three-dimensional image showing the region in which the imaged joint is located.

4. The acquisition unit further acquires a second image in which the same part as the part included in the first image is captured, the second image is an image obtained by imaging with an X-ray inspection device, and the generation unit has a machine learning model, and the machine learning model is trained using the acquired first image and second image as input data and the generated three-dimensional osteochondral model as output data, the osteochondral imaging device according to claim 3.

5. The osteochondral imaging device according to claim 4, further comprising a display unit, wherein the generation unit generates a three-dimensional osteochondral model by inputting another second image acquired by the acquisition unit into the machine learning model without inputting the first image into the machine learning model, and displays the generated three-dimensional osteochondral model on the display unit.

6. A method for depicting osteochondral tissue performed by a computer, comprising the steps of: acquiring a first image obtained by imaging a part of a person that includes a joint composed of bone and cartilage using a dual-energy CT scanner; and generating information derived from the first image based on the acquired first image, wherein the first image is an image in which the contours of the bone and cartilage are clearly defined, the boundary between the bone and the cartilage is clearly defined, and the boundary between the cartilage and the soft tissue included in the joint is clearly defined.

7. A program for causing the computer to execute the osteochondral imaging method described in claim 6.

Images