Deep learning algorithm-based spinal MRI precise synthesis CT image processing system and method for central axis type spinal arthritis

Abstract

The invention discloses a deep learning algorithm-based spinal MRI precise synthesis CT image processing system and method for central axis type spinal arthritis, and the system comprises a data obtaining unit which is used for obtaining a three-dimensional T1 weighted multi-gradient echo sequence, a T2 weighted sequence and a water-fat separation sequence which are obtained through the magnetic resonance imaging of the spinal column of a target user; the spatial registration unit is used for inputting the multi-sequence data into a deformable registration network so as to carry out spatial registration on the multi-modal features; the feature processing unit is used for carrying out feature fusion on the registered multi-modal features and carrying out double-domain splitting based on an anatomical attention gating mechanism to obtain skeleton features and soft tissue features; and the image fusion unit is used for synthesizing a synthesized CT image with an HU value corresponding to the spine of the target user based on the skeleton features and the soft tissue features, so that radiation-free high-precision CT image acquisition is realized, the diagnosis grading requirement on the axis type spine joint in the spine can be met, and the influence of radiation on the target user can be reduced.

Description

Technical Field

[0001] The present disclosure generally relates to the interdisciplinary field of medical image processing and artificial intelligence, and specifically to a system and method for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm. Background Art

[0002] Axial spondyloarthritis (axSpA) is a chronic inflammatory arthritis with axial spondyloarthritis as the main manifestation.

[0003] While the current mainstream Magnetic Resonance Imaging (MRI) has the advantage of being sensitive to fibrous tissue, it is insensitive to cortical bone, making it difficult to clearly identify early signs of osteophytes. Furthermore, signals from osteophytes and surrounding fibers often overlap, making quantitative assessment difficult. While combining CT with computed tomography (CT) can address this issue of clear identification, the risk of ionizing radiation from CT limits repeat examinations, making it difficult to follow up with patients.

[0004] Based on this, how to provide a CT image that can clearly display structural damage such as osteophytes and calcifications without the need for radiation scanning has become an urgent problem to be solved. Summary of the Invention

[0005] In view of the above-mentioned defects or deficiencies in the existing technology, it is hoped to provide a system and method for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm to achieve radiation-free high-precision CT image acquisition, which can not only meet the diagnostic grading requirements of axial spinal joints of the spine, but also reduce the impact of radiation on the target users.

[0006] In a first aspect, the present invention provides a system for accurately synthesizing CT images of axial spondyloarthritis spinal MRI images based on a deep learning algorithm, comprising:

[0007] A data acquisition unit, configured to acquire a three-dimensional T1-weighted multi-gradient echo sequence, a T2-weighted sequence, and a water-fat separation sequence obtained by performing magnetic resonance imaging on the spine of a target user;

[0008] a spatial registration unit, configured to input the three-dimensional T1-weighted multi-gradient echo sequence, the T2-weighted sequence, and the water-fat separation sequence into a deformable registration network to perform spatial registration on multimodal features;

[0009] The feature processing unit is used to fuse the registered multimodal features and perform dual-domain splitting based on the anatomical attention gating mechanism to obtain bone features and soft tissue features;

[0010] An image fusion unit is configured to synthesize a synthetic CT image having an HU value corresponding to the target user's spine based on the bone features and the soft tissue features.

[0011] In some embodiments, the feature processing unit further includes a biomechanical prior guidance subunit:

[0012] The biomechanical priori guidance subunit is used to analyze the biomechanical stress weights between the vertebrae of the target user;

[0013] The feature processing unit is further configured to:

[0014] When performing dual-domain splitting based on the anatomical attention gating mechanism, the bone features are determined in response to the biomechanical stress weights between the vertebrae.

[0015] In some embodiments, the feature processing unit is further configured to:

[0016] The biomechanical stress weights are spliced ​​with the multimodal fusion features and used as input features of the skeleton feature splitting branch in the dual-domain splitting, so as to obtain the skeleton features through the dual-domain splitting.

[0017] In some embodiments, the image fusion unit is further configured to:

[0018] A high stress area between the vertebrae having a biomechanical stress weight greater than or equal to a preset weight is identified, and the HU value corresponding to the high stress area is enhanced.

[0019] In some embodiments, the feature processing unit further includes a osteophyte diffusion prediction subunit:

[0020] The osteophyte diffusion prediction subunit is configured to predict the probability of future osteophyte growth based on the historical osteophyte volume and / or current stress value in the magnetic resonance imaging of the target user;

[0021] The feature processing unit is further configured to:

[0022] When performing dual-domain splitting based on the anatomical attention gating mechanism, the soft tissue feature weight is modified in response to the osteophyte growth probability.

[0023] In some embodiments, the soft tissue feature weight is inversely proportional to the probability of osteophyte growth.

[0024] In some embodiments, the image fusion unit is further configured to:

[0025] In the process of synthesizing a synthetic CT image with HU values ​​corresponding to the target user's spine based on the bone features and the soft tissue features, the synthesis of the synthetic CT is dynamically adjusted based on regional changes between the bone features and the soft tissue features.

[0026] In some embodiments, the image fusion unit is further configured to:

[0027] Based on multimodal fusion features, determine the unevenness of each feature point;

[0028] Determining an image contribution correction coefficient of the soft tissue feature based on the unevenness of each feature point;

[0029] Determining a HU value corresponding to each feature point based on the bone feature, the soft tissue feature, and an image contribution correction coefficient of the soft tissue feature;

[0030] The synthetic CT image is synthesized based on the HU values.

[0031] In some embodiments, the data acquisition unit is further configured to:

[0032] An age correction value of a target user is identified, and a flip angle of the three-dimensional T1-weighted multi-gradient echo sequence is adjusted based on the age correction value.

[0033] In a second aspect, the present invention provides a method for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm, comprising:

[0034] Acquire a three-dimensional T1-weighted multi-gradient echo sequence, a T2-weighted sequence, and a water-fat separation sequence obtained by magnetic resonance imaging of the target user's spine;

[0035] Inputting the three-dimensional T1-weighted multi-gradient echo sequence, the T2-weighted sequence, and the water-fat separation sequence into a deformable registration network to perform spatial registration of multimodal features;

[0036] The registered multimodal features are fused and split into two domains based on the anatomical attention gating mechanism to obtain bone features and soft tissue features.

[0037] A synthetic CT image having an HU value corresponding to the target user's spine is synthesized based on the bone features and the soft tissue features.

[0038] In a third aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the embodiment of the present application when executing the program.

[0039] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium on which a computer program is stored, which, when executed by a processor, implements the method described in the embodiment of the present application.

[0040] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, characterized in that when the computer program is executed by a processor, it implements the method described in the embodiment of the present application.

[0041] The embodiment of the present application provides a system and method for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm. The system and method can perform spatial registration on the three-dimensional T1-weighted multi-gradient echo sequence, T2-weighted sequence and water-fat separation sequence obtained by magnetic resonance imaging through a deformable registration network, thereby improving the spatial consistency of multi-sequence feature data and providing a reliable data basis for the subsequent synthesis of high-precision synthetic CT images. The multimodal fusion features are then subjected to a dual-domain split based on the anatomical attention gating mechanism to obtain bone features and soft tissue features, so that the obtained bone features and soft tissue features are highly consistent with the anatomical features and can be verified with each other. Finally, a synthetic CT image with an HU value is synthesized based on the reliable bone features and soft tissue features, which can make the synthesized synthetic CT image closer to the real CT image corresponding to the target user's spine.

[0042] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Other features, objects and advantages of the present application will become more apparent upon reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0044] Figure 1 A flowchart of a method for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm according to an embodiment of the present application is shown;

[0045] Figure 2 A schematic diagram of the structure of a system for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm according to an embodiment of the present application is shown;

[0046] Figure 3 A structural schematic diagram of a spinal MRI precise synthesis CT image processing system for axial spondyloarthritis based on a deep learning algorithm provided in another embodiment of the present application is shown. DETAILED DESCRIPTION

[0047] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the invention are shown in the accompanying drawings.

[0048] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0049] axSpA is a chronic inflammatory rheumatic disease that seriously affects the quality of life of patients, and the spine is a key site for its early diagnosis and efficacy evaluation. MRI mainly uses magnetic fields and radio waves for imaging, emphasizing the contrast and details of soft tissues. MRI is very effective when detecting changes in soft tissues (such as inflammation, edema, etc.). However, MRI has weak imaging capabilities for bone structure because the minerals in the bones have little effect on the MRI signal, resulting in insufficient sensitivity in detecting bone destruction. In MRI images, bone tissue usually has low signal intensity and is difficult to separate from surrounding tissues, making the interpretation of bone destruction more difficult.

[0050] The high spatial resolution and bone density-sensitive imaging characteristics provided by CT make it more accurate in identifying tiny bone destruction and better able to reveal subtle changes in bone, including bone destruction and sclerosis. However, CT carries the risk of ionizing radiation and is not suitable for frequent repeated examinations, especially in young patients. Many patients may have undergone MRI examinations before, and repeating CT scans will not only increase the radiation risk but also lead to wasteful repeated examinations. Therefore, the development of a radiation-free imaging method that can accurately reconstruct bone details has important clinical value. Deep learning has shown great potential in the field of medical image synthesis and reconstruction. By constructing an accurate cross-modal image generation model, it is expected to achieve lossless conversion from MRI to CT, providing innovative solutions for the early diagnosis and treatment monitoring of axSpA.

[0051] In order to further illustrate the technical solutions provided by the embodiments of the present application, this is described in detail below with reference to the accompanying drawings and specific embodiments. Although the embodiments of the present application provide the method operation instruction steps shown in the following embodiments or drawings, more or fewer operation instruction steps may be included in the method based on conventional or no creative labor. In the steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order provided by the embodiments of the present application. The method may be executed in the order of the methods shown in the embodiments or drawings or in parallel during the actual processing process or when the device is executed.

[0052] It should be noted that the acquisition or use of data in the embodiments of this application requires the user's consent. The relevant data can only be obtained after the user's authorization and permission, and the acquisition or use of the data complies with the provisions of relevant laws and regulations.

[0053] Please refer to Figure 1 , Figure 1 The flowchart of the method for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm provided by an embodiment of the present application is shown. Figure 1 As shown, the method includes:

[0054] Step 101: Acquire a three-dimensional T1-weighted multi-gradient echo sequence, a T2-weighted sequence, and a water-fat separation sequence obtained by performing magnetic resonance imaging on the spine of a target user.

[0055] Specifically, magnetic resonance imaging of the target user's spine is performed, and by setting the parameters TR = 5.2ms, TE = 2.5ms, and a layer thickness of 0.8mm, a three-dimensional T1-weighted multi-gradient echo sequence is obtained. By setting the parameters TR 3000ms, TE = 80ms, a T2-weighted sequence is obtained, and a water-fat separation sequence of the target user's spine is obtained.

[0056] The three-dimensional T1-weighted multi-gradient echo sequence in magnetic resonance imaging (MRI) is abbreviated as 3D-T1-MGE. 3D refers to three-dimensional volumetric imaging, which achieves arbitrary plane reconstruction (e.g., sagittal, coronal, and axial) by continuously acquiring isotropic voxels, avoiding the inter-slice gaps of traditional 2D stratification. T1 weighting emphasizes T1 relaxation differences in tissues and displays anatomical structures by adjusting repetition time (TR) and echo time (TE). Multi-gradient echo (MGE) sequences utilize multiple gradient echoes to acquire signals at different TE values ​​after a single radiofrequency excitation, improving scanning efficiency and supporting quantitative analysis, providing the data foundation for subsequent synthesis of synthetic CT with Hounsfield Hnit (HU) values.

[0057] In a preferred embodiment, before performing magnetic resonance imaging on the target user, an age correction value of the target user is further identified, and the inversion angle of the three-dimensional T1-weighted multi-gradient echo sequence is adjusted based on the age correction value.

[0058] In other words, the reversal angle of the 3D T1-weighted gradient echo sequence can be adjusted based on the user's age. Specifically, the target user's age is obtained. If the target user's age is greater than or equal to a preset age, the target user is an adult and no adjustment is required. The corresponding age correction value is 0. If the target user's age is less than the preset age, the target user is a child and adjustment is required. The age correction value is then further obtained and the reversal angle of the 3D T1-weighted gradient echo sequence is adjusted based on the age correction value.

[0059] Optionally, the age correction amount may be a fixed value, such as 5°, or a value linearly related to age, which is not specifically limited in this application.

[0060] For example, the age correction amount is a fixed value of 5°. When the target user is identified as a child, the inversion angle of the three-dimensional T1-weighted gradient echo sequence is adjusted from 15° to 10°.

[0061] Therefore, the present application adjusts the inversion angle of the three-dimensional T1-weighted gradient echo sequence to reduce the specific absorption ratio (SAR) to meet the safety requirements of children's scanning.

[0062] In step 102 , the three-dimensional T1-weighted multi-gradient echo sequence, the T2-weighted sequence, and the water-fat separation sequence are input into a deformable registration network to perform spatial registration of multimodal features.

[0063] It should be noted that since the spine has a large spatial range, especially including the area where the target user's respiratory movement occurs, it is easy to cause spatial deviations between different sequences during magnetic resonance imaging. Based on this, this application proposes to use a deformable registration network to spatially align the features of the three-dimensional T1-weighted multi-gradient echo sequence, T2-weighted sequence, and water-fat separation sequence to eliminate the misalignment between sequences caused by respiratory movement or changes in the target user's body position, ensure the spatial consistency of multimodal features, and enable the aligned multi-sequence feature data to accurately superimpose pathological features.

[0064] For example, the following formula can be used for spatial registration:

[0065]

[0066] Among them, φ is the deformation field, which represents the non-rigid deformation of the sequence feature image, φ id is the unit deformation field, i.e. the initial state without displacement, λ1 is the regularization coefficient used to balance the deformation smoothness, λ2 is the regularization coefficient used to balance the multimodal alignment accuracy, and NCC represents the calculation of the characteristic image I of the two images (the three-dimensional T1 weighted multi-gradient echo sequence). T1 and characteristic images of T2-weighted sequences I T2), is the operator of the deformation field scope image, which represents the image I corresponding to the T2 weighted sequence T2 Perform non-rigid deformation.

[0067] In step 103, the registered multimodal features are fused and dual-domain split is performed based on the anatomical attention gating mechanism to obtain bone features and soft tissue features.

[0068] It should be noted that the registered multimodal features are subjected to feature fusion to obtain multimodal fusion features corresponding to magnetic resonance imaging.

[0069] It should also be noted that the dual-domain splitting of the multimodal fusion features based on the anatomical attention gating mechanism is essentially to split the multimodal fusion features using the skeletal feature splitting branch and the soft tissue feature splitting branch, so that the skeletal features obtained by the skeletal feature splitting branch conform to the biomechanical stress, and the soft tissue features obtained by the soft tissue feature splitting branch meet the prediction results of osteophyte diffusion, so that the skeletal features split by the anatomical attention gating mechanism conform to the biomechanical development trend of osteophyte formation, that is, the pathological mechanism of osteophyte calcification, and at the same time, the soft tissue features can effectively avoid the prediction results of osteophyte diffusion, reduce the probability of overlap between soft tissue features and skeletal features, and reduce feature conflicts during post-synthesis of CT.

[0070] Specifically, in some embodiments, since the cortical bone signal in multiple magnetic resonance imaging sequences is low, based on this, the present application further proposes to dynamically model the mechanical transmission path of the spine based on biomechanical prior guidance, so as to increase the anatomical attention mechanism to the osteophyte formation area based on the biomechanical stress conditions.

[0071] Preferably, the adaptive adjacency matrix is ​​used to calculate the biological stress weights between vertebrae:

[0072]

[0073] Among them, A ij is the biomechanical stress weight between the two vertebrae i and j, Stress(V i , V j ) is the stress value between the two vertebrae, α is the biomechanical weight coefficient, are the attention query vector and key vector, which are used to calculate the feature similarity between vertebrae.

[0074] Furthermore, after obtaining the biomechanical weights between vertebrae, the biomechanical stress weights are used as the anatomical attention mechanism corresponding to the multimodal fusion feature map. Specifically, the biomechanical stress weights are concatenated with the multimodal fusion features and used as input features for the skeletal feature decomposition branch in the dual-domain decomposition process to obtain skeletal features. In other words, the biomechanical stress weights are used as an attention mask to select skeletal features in the skeletal feature decomposition branch, thereby improving the accuracy of skeletal feature recognition.

[0075] As a specific embodiment, based on the pathological mechanism of osteophyte formation, it is known that the osteophyte formation area is subjected to higher stress. Therefore, the high stress area, such as A ij The area with HU value > 0.7 is regarded as the osteophyte area, which is assigned to the bone branch in the dual-domain splitting stage. In the synthetic image, the matching degree between the osteophyte area and the real CT image can be further improved by enhancing the corresponding HU value.

[0076] In another feasible embodiment, the osteophyte diffusion prediction model is used to predict the probability of future osteophyte growth based on the historical osteophyte volume and / or current stress value in the target user's magnetic resonance imaging.

[0077] It should be understood that the osteophyte diffusion prediction model is trained using a large number of expert-annotated MRI images and real CT images.

[0078] Optionally, the osteophyte growth for the target user can be predicted based on the historical osteophyte volume recorded in at least one historical feature image of the target user, or based on the aforementioned vertebral stress value, or can be predicted using both the historical osteophyte volume and the current stress value when the information is relatively abundant. This application does not make specific limitations.

[0079] It should be noted that the predicted probability of osteophyte growth in the future is the predicted probability of osteophyte growth corresponding to the current target user's magnetic resonance imaging moment. It should also be noted that the osteophyte growth probability is also used to represent the probability value of the feature position being an osteophyte.

[0080] That is to say, after obtaining the probability of osteophyte growth corresponding to each feature position point, the soft tissue feature weight can be modified in response to the probability of osteophyte growth when performing dual-domain splitting based on the anatomical attention gating mechanism.

[0081] Specifically, the weight of soft tissue features is inversely proportional to the probability of osteophyte growth.

[0082] For example, when the probability of osteophyte growth is P(ΔV t ), we can use 1-P(ΔV t ) as the weight corresponding to the soft tissue feature.

[0083] In other words, before performing dual-domain splitting, the corrected osteophyte growth probability is used as the weight corresponding to the soft tissue feature and concatenated with the multimodal fusion feature. This concatenated feature is then used as the input feature for the soft tissue feature decomposition branch in the dual-domain splitting process to obtain the soft tissue feature. In other words, the corrected osteophyte growth probability is used as an attention mask to select soft tissue features in the soft tissue feature decomposition branch, thereby improving the accuracy of soft tissue feature recognition.

[0084] Therefore, the embodiment of the present application performs dual-domain splitting based on the anatomical attention gating machine, which can effectively split the bone features based on the biomechanical stress conditions of osteophyte formation and split the soft tissue features based on the probability of osteophyte growth, thereby ensuring the accuracy of the splitting of bone features and soft tissue features and reducing the risk of overlap between the two. At the same time, different weights are used to split the two types of features, verify each other, and improve the accuracy of feature splitting.

[0085] Step 104 : synthesize a synthetic CT image having HU values ​​corresponding to the target user's spine based on the bone features and the soft tissue features.

[0086] Optionally, the HU values ​​corresponding to the bone features and the soft tissue features may be determined separately, and then the HU values ​​may be used to synthesize the synthetic CT image.

[0087] Specifically, the synthesis of the synthetic CT image can be dynamically adjusted based on the regional changes between the bone features and the soft tissue features.

[0088] In a feasible embodiment, the unevenness of each feature point can be determined based on the multimodal fusion features, and the image contribution correction coefficient of the soft tissue feature can be determined based on the unevenness of each feature point; the HU value corresponding to each feature point can be determined based on the bone features, soft tissue features and the image contribution correction coefficient of the soft tissue feature, and a synthetic CT image can be synthesized based on the HU value.

[0089] Specifically, in the embodiment of the present application, the feature variance can be used to determine the unevenness of the multimodal fusion features. Based on the image judgment principle, the high variance area is the inflammation area, and the low variance area is the osteophyte area. Based on this, the dynamic correction weight when generating the synthetic CT image is determined, that is, the image contribution correction coefficient.

[0090] Optionally, the image contribution correction coefficient is positively correlated with the unevenness of the multimodal fusion features, that is, the higher the unevenness of the feature distribution, the larger the image contribution correction coefficient is, and the closer it is to 1, so as to enhance the influence of soft tissue features on image reconstruction; the lower the uniformity of the feature distribution, the smaller the image contribution correction coefficient is, and the closer it is to 0, so as to ensure the influence of bone features on image reconstruction.

[0091] For example, the HU value corresponding to each feature point can be determined using the following formula:

[0092] HU=G bone (F bone-out )+β·tan(G soft (F soft-out ))

[0093] Among them, HU is the HU value corresponding to each feature point, F bone-out is the bone feature output by the bone feature branch, G bone (F bone-out ) is the HU value output by the bone feature branch, F soft-out is the soft tissue feature output by the soft tissue feature branch, G soft (F soft-out ) is the HU value output by the soft tissue feature branch, and β is the image contribution correction coefficient.

[0094] Therefore, the HU value generated by the soft tissue feature is corrected and adjusted during the HU value generation process through the image contribution correction coefficient, so that the conflict between the soft tissue information and the bone information in the process of synthesizing the synthetic CT image can be effectively prevented, thereby effectively ensuring the reliability of the synthesized synthetic CT image.

[0095] Therefore, the embodiment of the present application provides a method for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm. The method can perform spatial registration on the three-dimensional T1-weighted multi-gradient echo sequence, T2-weighted sequence and water-fat separation sequence obtained by magnetic resonance imaging through a deformable registration network, thereby improving the spatial consistency of multi-sequence feature data and providing a reliable data basis for the subsequent synthesis of high-precision synthetic CT images. The multimodal fusion features are then split into two domains based on the anatomical attention gating mechanism to obtain bone features and soft tissue features, so that the obtained bone features and soft tissue features are highly consistent with the anatomical features and can be verified with each other. Finally, a synthetic CT image with an HU value is synthesized based on reliable bone features and soft tissue features, which can make the synthesized synthetic CT image closer to the real CT image corresponding to the target user's spine.

[0096] It should be noted that although the operations of the present method are described in a particular order in the drawings, this does not require or imply that the operations must be performed in this particular order, or that all illustrated operations must be performed to achieve desirable results.

[0097] Figure 2 A structural diagram of a system for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm provided in one embodiment of the present application is shown.

[0098] like Figure 2As shown, the deep learning algorithm-based spinal MRI accurate synthesis CT image processing system 10 for axial spondyloarthritis includes:

[0099] The data acquisition unit 11 is used to acquire a three-dimensional T1-weighted multi-gradient echo sequence, a T2-weighted sequence, and a water-fat separation sequence obtained by performing magnetic resonance imaging on the target user's spine;

[0100] a spatial registration unit 12, configured to input the three-dimensional T1-weighted multi-gradient echo sequence, the T2-weighted sequence, and the water-fat separation sequence into a deformable registration network to perform spatial registration on multimodal features;

[0101] A feature processing unit 13 is used to perform feature fusion on the registered multimodal features and perform dual-domain splitting based on the anatomical attention gating mechanism to obtain bone features and soft tissue features;

[0102] The image fusion unit 14 is configured to synthesize a synthetic CT image having an HU value corresponding to the target user's spine based on the bone features and the soft tissue features.

[0103] In some embodiments, as Figure 3 As shown, the feature processing unit 13 further includes a biomechanical priori guidance subunit 131:

[0104] The biomechanical priori guidance subunit 131 is used to analyze the biomechanical stress weights between the vertebrae of the target user;

[0105] The feature processing unit 13 is further configured to:

[0106] When performing dual-domain splitting based on the anatomical attention gating mechanism, the bone features are determined in response to the biomechanical stress weights between the vertebrae.

[0107] In some embodiments, the feature processing unit 13 is further configured to:

[0108] The biomechanical stress weights are spliced ​​with the multimodal fusion features and used as input features of the skeleton feature splitting branch in the dual-domain splitting, so as to obtain the skeleton features through the dual-domain splitting.

[0109] In some embodiments, the image fusion unit 14 is further configured to:

[0110] A high stress area between the vertebrae having a biomechanical stress weight greater than or equal to a preset weight is identified, and the HU value corresponding to the high stress area is enhanced.

[0111] In some embodiments, as Figure 3 As shown, the feature processing unit 13 further includes a osteophyte diffusion prediction subunit 132:

[0112] The osteophyte diffusion prediction subunit 132 is configured to predict the probability of future osteophyte growth based on the historical osteophyte volume and / or current stress value in the target user's MRI;

[0113] The feature processing unit 13 is further configured to:

[0114] When performing dual-domain splitting based on the anatomical attention gating mechanism, the soft tissue feature weight is modified in response to the osteophyte growth probability.

[0115] In some embodiments, the soft tissue feature weight is inversely proportional to the probability of osteophyte growth.

[0116] In some embodiments, the image fusion unit 14 is further configured to:

[0117] In the process of synthesizing a synthetic CT image with HU values ​​corresponding to the target user's spine based on the bone features and the soft tissue features, the synthesis of the synthetic CT is dynamically adjusted based on regional changes between the bone features and the soft tissue features.

[0118] In some embodiments, the image fusion unit 14 is further configured to:

[0119] Based on multimodal fusion features, determine the unevenness of each feature point;

[0120] Determining an image contribution correction coefficient of the soft tissue feature based on the unevenness of each feature point;

[0121] Determining a HU value corresponding to each feature point based on the bone feature, the soft tissue feature, and an image contribution correction coefficient of the soft tissue feature;

[0122] The synthetic CT image is synthesized based on the HU values.

[0123] In some embodiments, the data acquisition unit 11 is further configured to:

[0124] An age correction value of a target user is identified, and a flip angle of the three-dimensional T1-weighted multi-gradient echo sequence is adjusted based on the age correction value.

[0125] It should be understood that the modules or modules described in the axial spondyloarthritis spinal MRI accurate synthesis CT image processing system 10 based on deep learning algorithm are the same as those in the reference Figure 1The various steps in the described method correspond to each other. Therefore, the operations and features described above for the method are also applicable to the axial spondyloarthritis spinal MRI precise synthesis CT image processing system 10 based on deep learning algorithm and the modules contained therein, and will not be repeated here. The axial spondyloarthritis spinal MRI precise synthesis CT image processing system 10 based on deep learning algorithm can be pre-implemented in the browser or other security applications of the electronic device, and can also be loaded into the browser or its security application of the electronic device by downloading or other means. The corresponding modules in the axial spondyloarthritis spinal MRI precise synthesis CT image processing system 10 based on deep learning algorithm can cooperate with the modules in the electronic device to implement the solution of the embodiment of the present application.

[0126] The several modules or units mentioned in the detailed description above are not necessarily divided into one module or unit. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0127] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operating instructions of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the aforementioned module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than the order marked in the accompanying drawings. For example, the boxes represented by two connections can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operating instruction, or can be implemented using a combination of dedicated hardware and computer instructions.

[0128] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment, or may exist independently without being incorporated into the electronic device. The computer-readable storage medium stores one or more programs, which, when used by one or more processors, execute the method for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm described in the present application.

[0129] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the aforementioned disclosed concepts. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A deep learning algorithm-based spinal MRI accurate synthesis CT image processing system for axial spondyloarthritis, characterized by: include: A data acquisition unit, configured to acquire a three-dimensional T1-weighted multi-gradient echo sequence, a T2-weighted sequence, and a water-fat separation sequence obtained by performing magnetic resonance imaging on the spine of a target user; a spatial registration unit, configured to input the three-dimensional T1-weighted multi-gradient echo sequence, the T2-weighted sequence, and the water-fat separation sequence into a deformable registration network to perform spatial registration on multimodal features; The feature processing unit is used to fuse the registered multimodal features and perform dual-domain splitting based on the anatomical attention gating mechanism to obtain bone features and soft tissue features; An image fusion unit is configured to synthesize a synthetic CT image having an HU value corresponding to the target user's spine based on the bone features and the soft tissue features.

2. The system for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on deep learning algorithm according to claim 1 is characterized in that: The feature processing unit further includes a biomechanical priori guidance subunit: The biomechanical priori guidance subunit is used to analyze the biomechanical stress weights between the vertebrae of the target user; The feature processing unit is further configured to: When performing dual-domain splitting based on the anatomical attention gating mechanism, the bone features are determined in response to the biomechanical stress weights between the vertebrae.

3. The system for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on deep learning algorithm according to claim 2 is characterized in that: The feature processing unit is further configured to: The biomechanical stress weights are spliced ​​with the multimodal fusion features and used as input features of the skeleton feature splitting branch in the dual-domain splitting, so as to obtain the skeleton features through the dual-domain splitting.

4. The system for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on deep learning algorithm according to claim 2 is characterized in that: The image fusion unit is further configured to: A high stress area between the vertebrae having a biomechanical stress weight greater than or equal to a preset weight is identified, and the HU value corresponding to the high stress area is enhanced.

5. The system for accurately synthesizing CT images of axial spondyloarthritis using spinal MRI based on deep learning algorithm according to claim 2, characterized in that: The feature processing unit further includes a osteophyte diffusion prediction subunit: The osteophyte diffusion prediction subunit is configured to predict the probability of future osteophyte growth based on the historical osteophyte volume and / or current stress value in the magnetic resonance imaging of the target user; The feature processing unit is further configured to: When performing dual-domain splitting based on the anatomical attention gating mechanism, the soft tissue feature weight is modified in response to the osteophyte growth probability.

6. The system for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on deep learning algorithm according to claim 5 is characterized in that: The soft tissue feature weight is inversely proportional to the osteophyte growth probability.

7. The system for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on deep learning algorithm according to claim 1 is characterized in that: The image fusion unit is further configured to: In the process of synthesizing a synthetic CT image with HU values ​​corresponding to the target user's spine based on the bone features and the soft tissue features, the synthesis of the synthetic CT is dynamically adjusted based on regional changes between the bone features and the soft tissue features.

8. The system for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on deep learning algorithm according to claim 7 is characterized in that: The image fusion unit is further configured to: Based on multimodal fusion features, determine the unevenness of each feature point; Determining an image contribution correction coefficient of the soft tissue feature based on the unevenness of each feature point; Determining a HU value corresponding to each feature point based on the bone feature, the soft tissue feature, and an image contribution correction coefficient of the soft tissue feature; The synthetic CT image is synthesized based on the HU values.

9. The system for accurately synthesizing CT images of axial spondyloarthritis using spinal MRI based on deep learning algorithm according to claim 1, characterized in that: The data acquisition unit is further configured to: An age correction value of a target user is identified, and a flip angle of the three-dimensional T1-weighted multi-gradient echo sequence is adjusted based on the age correction value.

10. A method for accurately synthesizing CT images of axial spondyloarthritis spinal MRI based on a deep learning algorithm, characterized in that: include: Acquire a three-dimensional T1-weighted multi-gradient echo sequence, a T2-weighted sequence, and a water-fat separation sequence obtained by magnetic resonance imaging of the target user's spine; Inputting the three-dimensional T1-weighted multi-gradient echo sequence, the T2-weighted sequence, and the water-fat separation sequence into a deformable registration network to perform spatial registration of multimodal features; The registered multimodal features are fused and split into two domains based on the anatomical attention gating mechanism to obtain bone features and soft tissue features. A synthetic CT image having an HU value corresponding to the target user's spine is synthesized based on the bone features and the soft tissue features.

Images