Apparatus for determining potential recurrence risk areas of brain tumors and model training method

Abstract

The application discloses a brain tumor potential recurrence risk area determination device and a model training method, relates to the fields of medical technology and artificial intelligence. The device comprises a delineation determination unit configured to determine a brain tumor delineation area based on a target medical image of a target object; a nerve fiber reconstruction unit configured to simulate and reconstruct image information of N target nerve fibers of the target object according to the target medical image; and a risk area prediction unit configured to input the target medical image and the image information of the N target nerve fibers into a target model, perform feature extraction on the received information by the target model, and determine a recurrence risk area of the brain tumor of the target object according to the extracted features by using prior knowledge in a model training stage. The application solves the technical problem of low accuracy in determining the potential recurrence risk area of the brain tumor caused by only observing the risk diffusion of the brain tumor by imaging or judging the risk diffusion of the brain tumor by expert experience in the prior art.

Description

Technical Field

[0001] This application relates to the fields of medical technology and artificial intelligence, and more specifically, to a device and model training method for determining potential recurrence risk areas of brain tumors. Background Technology

[0002] High-grade gliomas are the most common and deadliest brain tumors in humans. Even with standard treatments for these tumors using current medical technology, recurrence is common because glioma cells have a strong invasive ability and can spread to other areas of the brain via white matter fiber bundles.

[0003] White matter fiber tracts provide a pathway for tumor cells to migrate from the primary tumor site, which is one of the reasons why surgery and local treatments often fail to completely remove the tumor. White matter fiber tracts are also a pathway for the spread of malignant gliomas, promoting the widespread diffusion of tumor cells within the brain. Diffusion tensor imaging (DTI) is a highly sensitive imaging technique that can detect subtle interruptions in white matter fiber tracts.

[0004] However, in the current technology, the risk of brain tumor spread is judged based solely on diffusion tensor imaging through imaging observation or expert experience, which is usually difficult to accurately determine the potential recurrence risk area of ​​brain tumor.

[0005] There is currently no effective solution to the above problems. Summary of the Invention

[0006] This application provides a device and model training method for determining potential recurrence risk areas of brain tumors, in order to at least solve the technical problem of low accuracy in determining potential recurrence risk areas of brain tumors caused by relying solely on imaging observation or expert experience to judge the risk of brain tumor spread.

[0007] According to one aspect of this application, an apparatus for determining a potential recurrence risk area of ​​a brain tumor is provided, comprising: a delineation determination unit for determining a delineated area of ​​a brain tumor of a target object based on a target medical image of the target object; a nerve fiber reconstruction unit for simulating and reconstructing image information of N target nerve fibers of the target object based on the target medical image, wherein N is an integer greater than 1; and a risk area prediction unit for inputting the target medical image and the image information of the N target nerve fibers into a target model, extracting features from the received information through the target model, and determining the recurrence risk area of ​​the brain tumor of the target object based on the extracted features using prior knowledge from the model training phase.

[0008] Optionally, the training data for the target model may include at least the following: the original lesion region of the brain tumor of the reference subject, the actual recurrence region of the brain tumor, and the imaging information of the nerve fibers that pass through both the original lesion region and the actual recurrence region.

[0009] Optionally, the device for determining the potential recurrence risk area of ​​brain tumors further includes: an image processing unit for simulating and reconstructing the image information of N nerve fibers of the reference subject based on medical images of the reference subject; a detection unit for detecting, based on the image information of the N nerve fibers of the reference subject, whether each nerve fiber of the reference subject simultaneously passes through the original lesion area of ​​the brain tumor and the actual recurrence area of ​​the brain tumor; a high-risk label setting unit for marking nerve fibers that simultaneously pass through the original lesion area of ​​the brain tumor and the actual recurrence area of ​​the brain tumor as high-risk labels; a low-risk label setting unit for marking nerve fibers that do not simultaneously pass through the original lesion area of ​​the brain tumor and the actual recurrence area of ​​the brain tumor as low-risk labels; and a model training unit for iteratively training the neural network based on the image information of the N nerve fibers of the reference subject, the risk labels marked on each nerve fiber, and the medical images of the reference subject to obtain the target model.

[0010] Optionally, the delineation determination unit includes: a first delineation subunit, used to delineate the brain tumor region based on medical images taken of the reference subject at the time of the first onset of the disease, to obtain first delineation information; a second delineation subunit, used to expand the delineated region in the first delineation information by a preset ratio to obtain second delineation information; and an original disease region determination subunit, used to determine the original disease region of the brain tumor of the reference subject according to the second delineation information.

[0011] Optionally, the delineation determination unit includes: a third delineation subunit, used to delineate the brain tumor region based on medical images taken of the reference subject during the recurrence of the disease, and obtain third delineation information; and a recurrence region determination subunit, used to determine the actual recurrence region of the brain tumor of the reference subject based on the third delineation information.

[0012] Optionally, the risk region prediction unit includes: a label determination subunit, used to determine the risk label carried by each target nerve fiber of the target object based on the extracted features by utilizing prior knowledge from the model training phase of the target model, and to identify the target nerve fibers carrying high-risk labels as high-risk nerve fibers; an image reconstruction subunit, used to reconstruct the high-risk nerve fibers into 3D images based on the coordinate system corresponding to the target object; an image combination subunit, used to combine all the 3D images reconstructed based on the high-risk nerve fibers into a target 3D image; and a risk region determination subunit, used to determine the recurrence risk region of the brain tumor of the target object based on the target 3D image.

[0013] Optionally, the risk region determination subunit includes: a detection module for detecting the density of high-risk nerve fibers in each region of the target 3D image; a rendering module for rendering regions with a density of high-risk nerve fibers greater than a preset threshold as highlighted regions; and a determination module for determining the highlighted regions as recurrence risk regions of the target object's brain tumor.

[0014] Optionally, the third delineation subunit includes: a first image processing module, used to use medical images of the reference subject taken before the recurrence of the disease as the first image; a second image processing module, used to use medical images of the reference subject taken during the recurrence of the disease as the second image; a registration module, used to register the first image and the second image to obtain a registration result; and a delineation module, used to delineate the brain tumor region based on the registration result to obtain third delineation information.

[0015] Optionally, the device for determining the potential recurrence risk area of ​​brain tumors further includes: an image acquisition unit for acquiring multimodal medical images of the target object; and a data enhancement unit for performing data enhancement on the multimodal medical images of the target object, and using the data-enhanced multimodal medical images as the target medical images of the target object.

[0016] According to another aspect of this application, a model training method is also provided, wherein the method includes: simulating and reconstructing image information of N nerve fibers of a reference object based on medical images of a reference object; determining, based on the image information of the N nerve fibers of the reference object, nerve fibers that simultaneously pass through the original onset area and the actual recurrence area of ​​the brain tumor of the reference object; and iteratively training the neural network based on the image information of the original onset area, the actual recurrence area, and the nerve fibers that simultaneously pass through the original onset area and the actual recurrence area of ​​the brain tumor of the reference object to obtain a target model.

[0017] In the device for determining the potential recurrence risk area of ​​brain tumors in this application, the delineation determination unit determines the delineation area of ​​the brain tumor of the target object based on the target medical image of the target object; the nerve fiber reconstruction unit simulates and reconstructs the image information of N target nerve fibers of the target object based on the target medical image, where N is an integer greater than 1; the risk area prediction unit inputs the target medical image and the image information of N target nerve fibers into the target model, the target model extracts features from the received information, and uses the prior knowledge of the model training stage to determine the recurrence risk area of ​​the brain tumor of the target object based on the extracted features.

[0018] As described above, the neural fiber reconstruction unit of this application utilizes target medical images of the target subject to simulate and reconstruct the imaging information of the target subject's neural fibers, providing imaging evidence of the complex interactions between brain tumors and neural fiber tissue. Visualization of neural fibers helps reveal possible pathways of brain tumor spread, which are difficult to determine effectively using traditional imaging observation. Finally, this application also achieves the transformation from data to knowledge by inputting the target medical images and the reconstructed N neural fiber imaging information into a neural network model (target model). Using the target model trained with a large amount of historical data, key imaging features affecting recurrence risk can be automatically extracted, thereby predicting which areas have a higher risk of recurrence. This helps doctors more accurately identify brain regions requiring focused attention, providing important reference value for developing treatment plans such as surgical resection extent and radiotherapy target areas.

[0019] Therefore, the technical solution of this application solves the technical problem of low accuracy in determining the potential recurrence risk area of ​​brain tumors in the prior art by introducing neural fiber reconstruction and neural network models. Attached Figure Description

[0020] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0021] Figure 1 This is a schematic diagram of an optional device for determining potential recurrence risk areas of brain tumors according to an embodiment of this application;

[0022] Figure 2 This is a flowchart of a model training method according to an embodiment of this application. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] It should also be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, and displayed data) collected in this application are information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with the relevant laws, regulations, and standards of the relevant regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation entry points are provided for users to choose to authorize or refuse. For example, this system has interfaces with relevant users or organizations. Before obtaining relevant information, a request to obtain the information needs to be sent to the aforementioned user or organization through the interface, and the relevant information is obtained only after receiving consent from the aforementioned user or organization.

[0026] Figure 1 This is a schematic diagram of an optional device for determining potential recurrence risk areas of brain tumors according to an embodiment of this application, as shown below. Figure 1 As shown, the device for determining the potential recurrence risk area of ​​brain tumors includes: a delineation and determination unit 101, a nerve fiber reconstruction unit 102, and a risk area prediction unit 103.

[0027] Optionally, the delineation determination unit 101 is used to determine the delineation region of the brain tumor of the target object based on the target medical image of the target object.

[0028] Optionally, the target subject may be a patient suffering from a brain tumor, and the target medical image may be an MRI image or other types of images. The target medical image may be an image sequence, including but not limited to T1-weighted imaging (T1WI), enhanced T1-weighted imaging (CE-T1WI), T2-weighted imaging (T2WI), fluid attenuation inversion recovery (FLAIR) imaging, diffusion-weighted imaging (DWI), and diffusion tensor imaging (DTI).

[0029] Optionally, after obtaining the target medical image of the target object, the delineation determination unit 101 can automatically delineate the brain tumor of the target object using a neural network model, thereby obtaining the delineated area of ​​the brain tumor of the target object. The delineation determination unit 101 can also obtain the delineated area of ​​the brain tumor of the target object by manual delineation. This application does not impose any particular limitation on the choice of delineation method.

[0030] Optionally, the nerve fiber reconstruction unit 102 is used to simulate and reconstruct the image information of N target nerve fibers of the target object based on the target medical image, where N is an integer greater than 1.

[0031] Alternatively, there can be various ways to reconstruct nerve fibers, such as reconstructing nerve fibers through reconstruction algorithms or reconstructing nerve fibers through neural network models.

[0032] The following provides an optional neural fiber reconstruction algorithm, including:

[0033] Step 1: Set seed points. Seed points can be automatically generated from the medical image. For example, seed points can be generated in areas of the medical image where the pixel value is greater than 0.7. An upper limit of 0.98 can also be set to remove potential noise, i.e., seed points can be generated in areas of the medical image where the pixel value is greater than 0.7 and less than 0.98.

[0034] Step 2: Construct a fiber response function to describe the diffusion motion of water molecules within biological tissues, and fit the seed point data and fiber response function into a constant solid angle direction distribution function model.

[0035] Step 3: Based on the constant solid angle direction distribution function model, calculate the direction of the sphere with the highest probability mass function. For example, the maximum angle can be set to 30 degrees. Start the tracking and reconstruction of nerve fibers, and finally reconstruct N nerve fibers.

[0036] It should be noted that the above-described neural fiber reconstruction algorithm is only one possible embodiment. In actual operation, other methods can also be used to complete the reconstruction of neural fibers. For example, a neural fiber reconstruction model can be trained using medical images. This neural fiber reconstruction model can reconstruct the image information of neural fibers based on the image information of medical images and using the prior knowledge related to neural fibers learned by the model during the training phase.

[0037] Optionally, the risk area prediction unit 103 is used to input the target medical image and the image information of N target nerve fibers into the target model, extract features from the received information through the target model, and use the prior knowledge from the model training phase to determine the recurrence risk area of ​​the target object's brain tumor based on the extracted features.

[0038] Optionally, the target model can be a pre-trained neural network model, and the training data of the target model includes at least: the original lesion area of ​​the brain tumor of the reference object, the actual recurrence area of ​​the brain tumor, and the image information of the nerve fibers that pass through both the original lesion area and the actual recurrence area.

[0039] In this model, the original lesion area and the actual recurrence area of ​​the brain tumor in the reference subject can be represented by delineated images. The neural network uses these delineated images as training data. If a nerve fiber of the reference subject crosses both the original lesion area and the actual recurrence area, it is identified as a high-risk nerve fiber; if a nerve fiber does not cross both areas, it is identified as a low-risk nerve fiber. The training objective of the neural network is to accurately determine whether a nerve fiber of the reference subject is high-risk based on the image information of the original lesion area, the actual recurrence area, and the nerve fiber. Once the training objective of the neural network is achieved, the target model can be considered trained.

[0040] As described above, the neural fiber reconstruction unit of this application utilizes target medical images of the target subject to simulate and reconstruct the imaging information of the target subject's neural fibers, providing imaging evidence of the complex interactions between brain tumors and neural fiber tissue. Visualization of neural fibers helps reveal possible pathways of brain tumor spread, which are difficult to determine effectively using traditional imaging observation. Finally, this application also achieves the transformation from data to knowledge by inputting the target medical images and the reconstructed N neural fiber imaging information into a neural network model (target model). Using the target model trained with a large amount of historical data, key imaging features affecting recurrence risk can be automatically extracted, thereby predicting which areas have a higher risk of recurrence. This helps doctors more accurately identify brain regions requiring focused attention, providing important reference value for developing treatment plans such as surgical resection extent and radiotherapy target areas.

[0041] Therefore, the technical solution of this application solves the technical problem of low accuracy in determining the potential recurrence risk area of ​​brain tumors in the prior art by introducing neural fiber reconstruction and neural network models.

[0042] In one optional embodiment, the device for determining the potential recurrence risk area of ​​a brain tumor further includes: an image processing unit for simulating and reconstructing image information of N nerve fibers of a reference subject based on medical images of a reference subject; a detection unit for detecting, based on the image information of the N nerve fibers of the reference subject, whether each nerve fiber of the reference subject simultaneously passes through the original onset area of ​​the brain tumor and the actual recurrence area of ​​the brain tumor; a high-risk label setting unit for marking nerve fibers that simultaneously pass through the original onset area of ​​the brain tumor and the actual recurrence area of ​​the brain tumor as high-risk labels; a low-risk label setting unit for marking nerve fibers that do not simultaneously pass through the original onset area of ​​the brain tumor and the actual recurrence area of ​​the brain tumor as low-risk labels; and a model training unit for iteratively training the neural network based on the image information of the N nerve fibers of the reference subject, the risk labels marked on each nerve fiber, and the medical images of the reference subject to obtain a target model.

[0043] Optionally, the image processing unit simulates and reconstructs the image information of N nerve fibers in the brain of the reference object based on various medical image data (e.g., DTI data) of the reference object.

[0044] One possible implementation is as follows: First, extract the DTI sequence from the medical image data of the reference object, and then use parameters such as anisotropic diffusion (AD), fractional anisotropy (FA), and geometric anisotropy (GA) in the DTI sequence, combined with the fiber response function and constant solid angle direction distribution function (ODF) model, to reconstruct the 3D image information of N nerve fibers.

[0045] Secondly, the detection unit checks whether each reconstructed nerve fiber simultaneously crosses both the original lesion area and the actual recurrence area of ​​the brain tumor. Based on the detection results, the high-risk labeling unit marks nerve fibers that simultaneously cross both the original lesion area and the actual recurrence area as "high-risk." The low-risk labeling unit marks nerve fibers that do not simultaneously cross both the original lesion area and the actual recurrence area of ​​the brain tumor as "low-risk."

[0046] Finally, the model training unit uses the image information of labeled nerve fibers and medical images of the reference object to train a neural network model. This model can accurately determine whether each nerve fiber of the patient belongs to the high-risk category based on the image information of the reconstructed nerve fibers, the original lesion area of ​​the brain tumor, and the actual recurrence area of ​​the brain tumor. For example, the model training unit receives data from the image processing unit and the label setting unit, and inputs this data as the training set into the neural network. Through iterative training, the neural network gradually adjusts the weights and parameters of each network layer until the neural network can accurately identify which nerve fibers belong to the high-risk category and which do not. At this point, the neural network is considered to have entered a convergence state, and the training of the neural network is considered complete.

[0047] In one optional embodiment, the delineation determination unit includes: a first delineation subunit, used to delineate the brain tumor region based on medical images taken of the reference subject at the time of the first onset of the disease, to obtain first delineation information; a second delineation subunit, used to expand the delineated region in the first delineation information by a preset ratio to obtain second delineation information; and an original disease region determination subunit, used to determine the original disease region of the brain tumor of the reference subject according to the second delineation information.

[0048] Optionally, let's assume that the medical image taken of the reference subject at the time of the first onset of the disease is image A. The first delineation subunit can support manual delineation or automatic delineation by a neural network model to delineate the brain tumor region in image A and obtain the first delineation information (which can be denoted as delineation 1).

[0049] Then, the second outlining subunit can expand outwards from outlining 1 to obtain second outlining information (which can be denoted as outlining 2). The preset proportion of the expansion can be customized, for example, 20% or 15%, etc. Furthermore, the direction and angle of the expansion can also be customized; this embodiment does not impose any particular limitations on these aspects.

[0050] Finally, the original onset area of ​​the brain tumor in the reference subject was determined based on delineation 2.

[0051] It should be noted that the reason for determining the original lesion area of ​​the brain tumor of the reference object based on delineation 2 is that there may be errors in delineation 1. By expanding outward by a certain proportion, cancer cells on the boundary can be included in the original lesion area of ​​the brain tumor as much as possible, thereby improving the accuracy of subsequent judgment of the recurrence risk area.

[0052] Of course, this application embodiment also supports directly using the delineation 1 to determine the original onset area of ​​the brain tumor of the reference object.

[0053] In one optional embodiment, the delineation determination unit includes: a third delineation subunit, used to delineate the brain tumor region based on medical images taken of the reference subject during the recurrence of the disease, and obtain third delineation information; and a recurrence region determination subunit, used to determine the actual recurrence region of the brain tumor of the reference subject based on the third delineation information.

[0054] Optionally, assuming the medical image taken of the reference subject at the time of recurrence is image B, the third delineation subunit can support manual delineation or automatic delineation by a neural network model to delineate the brain tumor region in image B, obtaining the third delineation information (which can be denoted as delineation 3). The recurrence region determination subunit can determine the actual recurrence region of the brain tumor in the reference subject based on delineation 3.

[0055] In one optional embodiment, the risk region prediction unit includes: a label determination subunit, used to determine the risk label carried by each target nerve fiber of the target object based on the extracted features by utilizing prior knowledge from the model training phase of the target model, and to identify the target nerve fibers carrying high-risk labels as high-risk nerve fibers; an image reconstruction subunit, used to reconstruct the high-risk nerve fibers into 3D images based on the coordinate system corresponding to the target object; an image combination subunit, used to combine all the 3D images reconstructed based on the high-risk nerve fibers into a target 3D image; and a risk region determination subunit, used to determine the recurrence risk region of the brain tumor of the target object based on the target 3D image.

[0056] Optionally, in the application stage of the target model, the image information of the N target nerve fibers reconstructed from the target object and the target medical image data of the target object can be input into the target model. Then, based on the prior knowledge learned by the target model in the model training stage, the risk label carried by each target nerve fiber of the target object can be determined according to the relevant image features of the target nerve fibers and the relevant image features of the target medical images. The target nerve fibers carrying high-risk labels are designated as high-risk nerve fibers.

[0057] Then, based on the coordinate system corresponding to the target object, the image reconstruction subunit reconstructs high-risk nerve fibers into 3D images for display. For example, for 100 target nerve fibers, 100 sets of 3D images will be generated. Subsequently, the 100 sets of 3D images are added together to obtain a target 3D image, which can serve as an indicator image of high-risk recurrence areas specific to the target object.

[0058] In one optional embodiment, the risk region determination subunit includes: a detection module for detecting the density of high-risk nerve fibers in each region of the target 3D image; a rendering module for rendering regions with a density of high-risk nerve fibers greater than a preset threshold as highlighted regions; and a determination module for determining the highlighted regions as recurrence risk regions of brain tumors in the target object.

[0059] Optionally, the detection module can detect the density of high-risk nerve fibers within each region of the target 3D image. A higher density of high-risk nerve fibers within a region indicates a greater likelihood of recurrence risk for brain tumors. Based on this, the rendering module can highlight regions with a high-risk nerve fiber density exceeding a preset threshold to alert doctors that these regions are at risk of recurrence of brain tumors in the target patient. The preset threshold can be customized.

[0060] In one optional embodiment, the third delineation subunit includes: a first image processing module, used to use medical images of the reference subject taken before the recurrence of the disease as the first image; a second image processing module, used to use medical images of the reference subject taken during the recurrence of the disease as the second image; a registration module, used to register the first image and the second image to obtain a registration result; and a delineation module, used to delineate the brain tumor region based on the registration result to obtain third delineation information.

[0061] Optionally, by registering the first image and the second image, precise alignment of the two images can be achieved, thereby eliminating image differences caused by factors such as head position and equipment calibration. The registration process includes, but is not limited to, various registration operations such as rigid body transformation, non-rigid body deformation, and feature point matching.

[0062] In one optional embodiment, the apparatus for determining the potential recurrence risk area of ​​a brain tumor further includes: an image acquisition unit for acquiring multimodal medical images of the target object; and a data enhancement unit for performing data enhancement on the multimodal medical images of the target object and using the data-enhanced multimodal medical images as the target medical images of the target object.

[0063] Optionally, for various multimodal medical images such as DTI MRI images, data augmentation units can be used to augment these medical images to obtain a multi-sequence image set as the target medical image. The multi-sequence image set may include images of multiple modalities such as AD, FA, GA, linear, MD, planarity, RD, sphericity, and trace.

[0064] Among them, AD (Axial Diffusivity) is used to reflect the rate of water molecule diffusion along the axis of nerve fibers, and is particularly sensitive for detecting nerve fiber damage.

[0065] Fractional Anisotropy (FA) is used to quantify the differences in diffusion of water molecules in different directions. Its value ranges from 0 to 1. A higher FA value indicates that the tissue has obvious diffusion directionality, which is usually associated with healthy protein fibers.

[0066] GA (Geodesic Anisotropy) is similar to FA, but focuses on the curvature and bifurcation of fiber bundles, and is particularly useful for evaluating the geometric structural properties of fiber bundles.

[0067] MD (Mean Diffusivity) provides information about the average rate at which water molecules diffuse in all directions. MD is often associated with the overall health of an organization.

[0068] RD (Radial Diffusivity) is used to measure the rate of water molecule diffusion perpendicular to the nerve fiber axis. It is often used to study myelin sheath integrity and the radial health of nerve fibers.

[0069] It should be noted that data augmentation can expand the quantity and diversity of images, thereby improving the robustness of the model and enhancing the model's accuracy in predicting potential recurrence risk areas of brain tumors.

[0070] According to another aspect of the embodiments of this application, a model training method is also provided, wherein, Figure 2 This is a flowchart of a model training method according to an embodiment of this application, such as... Figure 2 As shown, the method includes the following steps:

[0071] Step S201: Based on the medical images of the reference object, simulate and reconstruct the image information of N nerve fibers of the reference object.

[0072] Step S202: Based on the imaging information of N nerve fibers of the reference subject, identify the nerve fibers that simultaneously pass through the original onset area of ​​the brain tumor and the actual recurrence area of ​​the brain tumor of the reference subject from among the N nerve fibers.

[0073] Step S203: Based on the image information of the original onset area of ​​the brain tumor, the actual recurrence area of ​​the brain tumor, and the nerve fibers that pass through both the original onset area and the actual recurrence area of ​​the reference object, the neural network is iteratively trained to obtain the target model.

[0074] Optionally, the target subject may be a patient suffering from a brain tumor, and the target medical image may be an MRI image or other types of images. The target medical image may be an image sequence, including but not limited to T1-weighted imaging (T1WI), enhanced T1-weighted imaging (CE-T1WI), T2-weighted imaging (T2WI), fluid attenuation inversion recovery (FLAIR) imaging, diffusion-weighted imaging (DWI), and diffusion tensor imaging (DTI).

[0075] Optionally, after obtaining the target medical image of the target object, a neural network model can be used to automatically delineate the brain tumor of the target object, thereby obtaining the delineated area of ​​the brain tumor of the target object. Alternatively, the delineated area of ​​the brain tumor of the target object can be obtained manually. This application does not impose any particular limitation on the choice of delineation method.

[0076] Alternatively, there can be various ways to reconstruct nerve fibers, such as reconstructing nerve fibers through reconstruction algorithms or reconstructing nerve fibers through neural network models.

[0077] The following provides an optional neural fiber reconstruction algorithm, including:

[0078] Step 1: Set seed points. Seed points can be automatically generated from the medical image. For example, seed points can be generated in areas of the medical image where the pixel value is greater than 0.7. An upper limit of 0.98 can also be set to remove potential noise, i.e., seed points can be generated in areas of the medical image where the pixel value is greater than 0.7 and less than 0.98.

[0079] Step 2: Construct a fiber response function to describe the diffusion motion of water molecules within biological tissues, and fit the seed point data and fiber response function into a constant solid angle direction distribution function model.

[0080] Step 3: Based on the constant solid angle direction distribution function model, calculate the direction of the sphere with the highest probability mass function, set the maximum angle to 30 degrees, start the tracking and reconstruction of nerve fibers, and finally reconstruct N nerve fibers.

[0081] It should be noted that the above-described neural fiber reconstruction algorithm is only one optional embodiment. In actual operation, other methods can also be used to complete the reconstruction of neural fibers. For example, a neural fiber reconstruction model can be trained using medical images. This neural fiber reconstruction model can reconstruct the image information of neural fibers based on the image information of medical images and using the prior knowledge of neural fibers learned by the model during the training phase.

[0082] Optionally, the target model can be a pre-trained neural network model, and the training data of the target model includes at least: the original lesion area of ​​the brain tumor of the reference object, the actual recurrence area of ​​the brain tumor, and the image information of the nerve fibers that pass through both the original lesion area and the actual recurrence area.

[0083] In this model, the original lesion area and the actual recurrence area of ​​the brain tumor in the reference subject can be represented by delineated images. The neural network uses these delineated images as training data. If a nerve fiber of the reference subject crosses both the original lesion area and the actual recurrence area, it is identified as a high-risk nerve fiber; if a nerve fiber does not cross both areas, it is identified as a low-risk nerve fiber. The training objective of the neural network is to accurately determine whether a nerve fiber of the reference subject is high-risk based on the image information of the original lesion area, the actual recurrence area, and the nerve fiber. Once the training objective of the neural network is achieved, the target model can be considered trained.

[0084] As described above, the neural fiber reconstruction unit of this application utilizes target medical images of the target subject to simulate and reconstruct the imaging information of the target subject's neural fibers, providing imaging evidence of the complex interactions between brain tumors and neural fiber tissue. Visualization of neural fibers helps reveal possible pathways of brain tumor spread, which are difficult to determine effectively using traditional imaging observation. Finally, this application also achieves the transformation from data to knowledge by inputting the target medical images and the reconstructed N neural fiber imaging information into a neural network model (target model). Using the target model trained with a large amount of historical data, key imaging features affecting recurrence risk can be automatically extracted, thereby predicting which areas have a higher risk of recurrence. This helps doctors more accurately identify brain regions requiring focused attention, providing important reference value for developing treatment plans such as surgical resection extent and radiotherapy target areas.

[0085] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein the computer-readable storage medium stores a computer program, wherein when the computer program is executed, the device in which the computer-readable storage medium is located performs the above-described means for determining potential recurrence risk areas of glioma.

[0086] According to another aspect of the embodiments of this application, an electronic device is also provided, wherein the electronic device includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors control the operation of the above-mentioned device for determining potential recurrence risk areas of glioma.

[0087] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0088] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0090] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0092] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0093] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. An apparatus for determining a brain tumor potential recurrence risk area, characterized by, The method comprises the following steps: an outline determination unit is configured to determine a brain tumor outline region of a target object based on a target medical image of the target object; a nerve fiber reconstruction unit is configured to simulate reconstruction of image information of N target nerve fibers of the target object according to the target medical image, wherein N is an integer greater than 1; a risk region prediction unit is configured to input the target medical image and the image information of the N target nerve fibers into a target model, perform feature extraction on the received information by the target model, and determine a recurrence risk region of a brain tumor of the target object according to the extracted features by using prior knowledge in a model training stage. The determination apparatus of the brain tumor potential recurrence risk region further comprises an image processing unit configured to simulate reconstruction of image information of N nerve fibers of a reference object based on a medical image of the reference object; a detection unit configured to detect whether each nerve fiber of the reference object simultaneously passes through a brain tumor original onset region and a brain tumor actual recurrence region of the reference object according to the image information of the N nerve fibers of the reference object; a high-risk label setting unit configured to mark the nerve fiber that simultaneously passes through the brain tumor original onset region and the brain tumor actual recurrence region as a high-risk label; a low-risk label setting unit configured to mark the nerve fiber that does not simultaneously pass through the brain tumor original onset region and the brain tumor actual recurrence region as a low-risk label; and a model training unit configured to perform iterative training on a neural network according to the image information of the N nerve fibers of the reference object, the risk label marked for each nerve fiber, and the medical image of the reference object, to obtain the target model.

2. The apparatus for determining a risk area of potential recurrence of a brain tumor according to claim 1, wherein The training data of the target model at least comprises the brain tumor original onset region of the reference object, the brain tumor actual recurrence region, and the image information of the nerve fiber that simultaneously passes through the original onset region and the actual recurrence region.

3. The apparatus for determining a risk area of potential recurrence of a brain tumor according to claim 1, wherein The outline determination unit comprises: a first outline subunit configured to perform image outlining of a brain tumor region based on a medical image taken when a reference object first falls ill, to obtain first outlining information; a second outline subunit configured to perform outward expansion of the outlined region in the first outlining information by a preset proportion, to obtain second outlining information; an original onset region determination subunit configured to determine a brain tumor original onset region of the reference object according to the second outlining information.

4. The apparatus for determining a risk area of potential recurrence of a brain tumor according to claim 1, wherein The outline determination unit comprises: a third outline subunit configured to perform image outlining of a brain tumor region based on a medical image taken when a reference object falls ill again, to obtain third outlining information; a recurrence region determination subunit configured to determine a brain tumor actual recurrence region of the reference object according to the third outlining information.

5. The apparatus for determining a risk area of potential recurrence of a brain tumor according to claim 1, wherein The risk region prediction unit comprises: a label determination subunit configured to determine a risk label carried by each target nerve fiber of the target object according to the extracted features by the target model using prior knowledge in the model training stage, and to take the target nerve fiber carrying a high-risk label as a high-risk nerve fiber. an image reconstruction subunit, configured to reconstruct the high-risk nerve fiber into a 3D image based on a coordinate system corresponding to the target object; an image combination subunit, configured to combine all the 3D images based on the high-risk nerve fiber into a target 3D image; a risk area determination subunit, configured to determine a recurrence risk area of a brain tumor of the target object according to the target 3D image.

6. The apparatus for determining a risk area of potential recurrence of a brain tumor according to claim 5, wherein The risk area determination subunit comprises: a detection module, configured to detect the density of the high-risk nerve fiber in each area in the target 3D image; a rendering module, configured to render an area with a density of the high-risk nerve fiber greater than a preset threshold as a highlighted area; a determination module, configured to determine the highlighted area as the recurrence risk area of the brain tumor of the target object.

7. The apparatus for determining a risk area of potential recurrence of a brain tumor according to claim 4, wherein The third delineation subunit comprises: a first image processing module, configured to take a medical image of the reference object before the recurrence as a first image; a second image processing module, configured to take a medical image of the reference object when the recurrence occurs as a second image; a registration module, configured to register the first image and the second image to obtain a registration result; a delineation module, configured to perform image delineation of the brain tumor area according to the registration result to obtain third delineation information.

8. The apparatus for determining a risk area of potential recurrence of a brain tumor according to claim 1, wherein, The determination apparatus of the brain tumor potential recurrence risk area further comprises: an image acquisition unit, configured to acquire multi-modal medical images of the target object; a data enhancement unit, configured to perform data enhancement on the multi-modal medical images of the target object, and take the data-enhanced multi-modal medical images as target medical images of the target object.

9. A model training method, comprising: comprising: simulating and reconstructing image information of N nerve fibers of the reference object based on medical images of the reference object; determining nerve fibers that simultaneously pass through an original brain tumor onset area and an actual brain tumor recurrence area of the reference object from the N nerve fibers according to the image information of the N nerve fibers of the reference object; iteratively training a neural network according to the original brain tumor onset area, the actual brain tumor recurrence area of the reference object, and the image information of the nerve fibers that simultaneously pass through the original onset area and the actual recurrence area, to obtain the target model in any one of claims 1 to 7. The training process of the target model comprises: simulating reconstruction of image information of N nerve fibers of a reference object based on a medical image of the reference object; detecting whether each nerve fiber of the reference object simultaneously passes through a brain tumor original onset area and a brain tumor actual recurrence area of the reference object according to the image information of the N nerve fibers of the reference object; marking a nerve fiber that simultaneously passes through the brain tumor original onset area and the brain tumor actual recurrence area as a high-risk label; marking a nerve fiber that does not simultaneously pass through the brain tumor original onset area and the brain tumor actual recurrence area as a low-risk label; and performing iterative training on a neural network according to the image information of the N nerve fibers of the reference object, the risk label of each nerve fiber and the medical image of the reference object, so as to obtain the target model.

Images