Dynamic tumor tracking method and device, computer equipment and storage medium

Abstract

The invention relates to a dynamic tumor tracking method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring 4DCT image data of a patient, and performing mixed data enhancement on the 4DCT image data to generate a 3DCT image representing each respiratory stage of the lung; based on the 3D CT image, generating an X-ray image in a respiratory cycle of a simulated patient by adopting a mathematical simulation method; registering the X-ray image and the 4DCT image data to generate a tumor tracking model of the patient; and inputting an X-ray image of a patient in a radiotherapy process into the tumor tracking model, and outputting a tumor tracking result through the tumor tracking model. According to the invention, real-time three-dimensional tracking of the tumor can be realized by using a single low-dose X-ray image under any geometric condition, high-precision tumor positioning can be provided on the premise of ensuring low radiation dose, the method is suitable for any projection angle and source-object distance change conditions, and various geometric relationship changes in the treatment process can be flexibly coped with.

Description

Technical Field

[0001] The present application belongs to the field of medical image processing technology, and in particular, relates to a dynamic tumor tracking method, device, computer equipment and storage medium. Background Art

[0002] During radiotherapy, the target position of tumor patients will change due to physiological factors such as respiratory movement and cardiac movement. In order to ensure that the treatment dose is accurately delivered to the target area and avoid damage to the patient's normal tissue, it is necessary to track the tumor dynamically to monitor the target position in real time. In traditional technologies, image-guided technologies such as CBCT (Cone Beam Computed Tomography) are used for tumor tracking, which has the problems of time delay and high radiation dose. It is difficult to cope with the position changes of tumors caused by physiological movements such as breathing in real time, which limits its wide application in clinical practice. With the application of deep learning in medical image processing, the accuracy of tumor tracking has been significantly improved. However, existing deep learning models usually assume that geometric conditions are fixed and cannot cope with changes in the distance between the source and the object, the projection angle, etc. In actual clinical applications, especially in dynamic image-guided radiotherapy, the distance between the source and the object will fluctuate due to changes in patient position, respiratory movement or other physiological changes. Existing deep learning models are difficult to adapt to these geometric changes.

[0003] At present, the mainstream tumor tracking methods in radiotherapy mainly include tumor tracking based on CBCT and CT (Computed Tomography) image guidance and tumor tracking based on 2D / 3D image registration. Among them, the tumor tracking method based on CBCT and CT relies on obtaining multi-angle cone beam projections during radiotherapy to reconstruct 3D images, and then registers with the planned CT to locate the tumor. However, CBCT imaging requires the acquisition of hundreds of projection images, which greatly increases the patient's radiation dose. At the same time, the imaging process is time-consuming and cannot meet the needs of real-time dynamic monitoring, especially when the position of the target area changes due to breathing or heartbeat. The tumor tracking scheme based on 2D / 3D image registration uses 2D ray projections in radiotherapy and planned CT images to achieve tumor tracking, which can effectively reduce the radiation dose and shorten the processing time. In order to achieve the matching of 2D images and 3D targets, linear dimensionality reduction methods such as PCA (Principal Component Analysis) are usually used. However, PCA is difficult to capture complex nonlinear motions. Therefore, when dealing with irregular physiological movements such as breathing and heartbeat, the registration accuracy is limited and it is difficult to meet clinical requirements. Summary of the invention

[0004] The present application provides a dynamic tumor tracking method, apparatus, computer device and storage medium, aiming to solve at least one of the above-mentioned technical problems in the prior art to a certain extent.

[0005] In order to solve the above problems, this application provides the following technical solutions:

[0006] A dynamic tumor tracking method, comprising:

[0007] Acquire 4DCT image data of the patient, and perform hybrid data enhancement on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs;

[0008] Based on the 3DCT image, a mathematical simulation method is used to generate an X-ray image simulating a patient's respiratory cycle;

[0009] registering the X-ray image and the 4DCT image data to generate a tumor tracking model of the patient;

[0010] The X-ray images of the patient during radiotherapy are input into the tumor tracking model, and the tumor tracking results are output through the tumor tracking model.

[0011] The technical solution adopted by the embodiment of the present application also includes: after obtaining the 4DCT image data of the patient, it also includes:

[0012] Any phase of the 4DCT image data is selected as a floating image, and the remaining N phases are selected as reference images. The reference image and the floating image are registered using an image registration method to obtain N deformation fields.

[0013] The technical solution adopted by the embodiment of the present application also includes: performing mixed data enhancement on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs, specifically:

[0014] Randomly select two deformation fields from the N deformation fields and superimpose them to obtain phase-to-phase deformation;

[0015] Thin plate spline interpolation is used to simulate random deformation of any phase;

[0016] The phase-to-phase deformation and the random deformation are superimposed with arbitrary weights to obtain a deformation field, and the deformation field is applied to the floating image to obtain a 3D CT image representing each respiratory stage of the lung.

[0017] The technical solution adopted by the embodiment of the present application also includes: based on the 3DCT image, a mathematical simulation method is used to generate an X-ray image simulating the patient's respiratory cycle, specifically:

[0018] The 3DCT image is simulated by using a Monte Carlo simulation method to generate a 2D DRR, and the X-ray image in the patient's respiratory cycle is simulated by using the 2D DRR.

[0019] The technical solution adopted by the embodiment of the present application also includes: the X-ray image and the 4DCT image data are registered to generate a tumor tracking model of the patient, specifically:

[0020] The X-ray image is upgraded to a three-dimensional feature map through a residual block, the three-dimensional feature map is encoded and parsed from different angles using a 3D codec network, and the three-dimensional feature map is aligned with the floating image.

[0021] The technical solution adopted in the embodiment of the present application also includes: the tumor tracking model adopts a deep learning model that combines a 3D codec network and a Unet-KAN mechanism.

[0022] The technical solution adopted by the embodiment of the present application also includes: inputting the X-ray image of the patient during radiotherapy into the tumor tracking model, and outputting the tumor tracking result through the tumor tracking model, specifically:

[0023] A single X-ray image at any angle during the radiotherapy of the patient is obtained, the single X-ray image is input into the patient's tumor tracking model, the three-dimensional position of the tumor and the endangered organ is obtained through the tumor tracking model, and the tumor tracking result is output.

[0024] Another technical solution adopted by the embodiment of the present application is: a dynamic tumor tracking device, comprising:

[0025] Image acquisition module: used to acquire 4DCT image data of the patient, and perform mixed data enhancement on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs;

[0026] Image simulation module: used for generating an X-ray image simulating a patient's respiratory cycle by using a mathematical simulation method based on the 3DCT image;

[0027] Image registration module: used to register the X-ray image and 4DCT image data to generate a tumor tracking model of the patient;

[0028] Tumor tracking module: used to input the X-ray images of the patient during radiotherapy into the tumor tracking model, and output the tumor tracking results through the tumor tracking model.

[0029] Another technical solution adopted by the embodiment of the present application is: a computer device, the computer device includes a processor and a memory coupled to the processor, wherein:

[0030] The memory stores program instructions for implementing the dynamic tumor tracking method;

[0031] The processor is configured to execute the program instructions stored in the memory to control a dynamic tumor tracking method.

[0032] Another technical solution adopted by the embodiment of the present application is: a storage medium storing program instructions executable by a processor, wherein the program instructions are used to execute the dynamic tumor tracking method.

[0033] Compared with the prior art, the beneficial effects of the embodiments of the present application are as follows: the dynamic tumor tracking method, device, computer equipment and storage medium of the embodiments of the present application combine deep learning with 2D-3D registration technology, use the patient's 4DCT image data to train a tumor tracking model dedicated to the patient, and input a single ultra-low-dose X-ray image at any angle during the patient's radiotherapy into the tumor tracking model to achieve efficient real-time tumor tracking. The tumor tracking model of the embodiment of the present application can achieve real-time three-dimensional tracking of tumors using a single low-dose X-ray image under any geometric conditions, can provide high-precision tumor positioning under the premise of ensuring low radiation dose, and is suitable for any projection angle and source-object distance change conditions, can flexibly respond to various geometric relationship changes during the treatment process, and adapt to more clinical application scenarios. The embodiments of the present application can achieve more accurate radiotherapy dose control in the case of low-dose X-ray images, while improving the treatment effect, it can reduce damage to surrounding normal tissues. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of a dynamic tumor tracking method according to an embodiment of the present application;

[0035] Figure 2 is a schematic diagram of a tumor tracking model according to an embodiment of the present application;

[0036] Figure 3 This is a schematic diagram of the tumor registration results from the end of exhalation to the end of inspiration for lung cancer patient data;

[0037] Figure 4 Schematic diagram of tumor centroid motion error at different angles;

[0038] Figure 5 This is a schematic diagram of the tumor registration results of the digital phantom XCAT from the end of exhalation to the end of inspiration;

[0039] Figure 6 Schematic diagram of the movement trajectory of the tumor center of mass in the digital phantom;

[0040] Figure 7 This is a schematic diagram of the structure of a dynamic tumor tracking device according to an embodiment of the present application;

[0041] Figure 8 A schematic diagram of the computer device structure of an embodiment of the present application;

[0042] Fig. 9 A schematic diagram of the structure of a storage medium according to an embodiment of the present application. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0044] The terms "first", "second", "third" in this application are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined as "first", "second", "third" can expressly or implicitly include at least one of the features. In the description of this application, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. In the embodiments of this application, all directional indications (such as up, down, left, right, front, back...) are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication also changes accordingly. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or computer device that includes a series of steps or units is not limited to the steps or units listed, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or computer devices.

[0045] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0046] Specifically, see Figure 1 , is a flow chart of a dynamic tumor tracking method according to an embodiment of the present application. The dynamic tumor tracking method according to an embodiment of the present application comprises the following steps:

[0047] S100: Acquire 4DCT image data of the patient, and perform hybrid data enhancement processing on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs;

[0048] In this step, 4DCT refers to a low-dose imaging technology that records the trajectory of respiratory motion on CT images in sequence. It records lung CT images within multiple respiratory cycles through rapid imaging technology, and then combines these images with respiratory motion signals to form a three-dimensional image sequence of a part of the respiratory cycle to form a four-dimensional data set, which can dynamically observe the position, shape and size of the tumor during the respiratory movement.

[0049] Furthermore, the 4DCT image data generally includes multiple phases, and the hybrid data enhancement processing for the 4DCT image data specifically includes:

[0050] S101: Select any phase of the 4DCT image data as a floating image, and the remaining N phases as reference images;

[0051] S102: registering the reference image and the floating image using an image registration method to obtain N deformation fields;

[0052] S103: randomly selecting two deformation fields from the N deformation fields for superposition to obtain inter-phase deformation, and using thin plate spline interpolation to simulate random deformation of any phase;

[0053] S104: a large number of deformation fields are obtained by superimposing phase-to-phase deformation and random deformation with arbitrary weights, and the large number of deformation fields are applied to the floating image to obtain a large number of 3DCT images representing various respiratory stages of the lungs.

[0054] It can be understood that the embodiment of the present application generates a large number of 3D CT images covering each expiratory phase of the lungs by performing data enhancement processing on the 4DCT image data, thereby providing rich training data for the model, which is conducive to improving the model's adaptability to respiratory movements.

[0055] S110: Based on the 3D CT image, a 2D DRR (Digitally Reconstructed Radiograph) is generated by a mathematical simulation method, and the X-ray images at different stages of the patient's respiratory cycle are simulated by the 2D DRR;

[0056] In this step, DRR generates simulated X-ray images from 3DCT images by mathematical simulation, which can not only generate a large number of 2D images at multiple angles and different distances, but also simulate different noise levels and artifact characteristics, which is conducive to improving the generalization ability of the model in the clinical environment and providing the model with data that is closer to the actual clinical situation. Among them, the mathematical simulation method includes but is not limited to Monte Carlo simulation (a mathematical technique based on random numbers) simulation algorithm.

[0057] S120: performing image registration on the X-ray image and the floating image to generate a tumor tracking model of the patient;

[0058] In this step, the tumor tracking model uses a deep learning model that combines the 3D codec network and the Unet-KAN mechanism. The model structure is as follows: Figure 2 As shown. The registration method of the X-ray image and the floating image is specifically as follows: first, the X-ray image is upgraded to a three-dimensional feature map through a residual block, and then the three-dimensional feature map is encoded and parsed from different angles using a 3D codec network, and finally the three-dimensional feature map and the floating image are registered to map the X-ray image to the three-dimensional position of the tumor and the endangered organ in the three-dimensional image. Among them, the tumor tracking model combines the 3D codec network and the Unet-KAN mechanism to improve the ability to capture spatial information, thereby improving the accuracy of tumor tracking. In addition, the embodiment of the present application encodes the three-dimensional feature map from different angles through a 3D codec network, so that the tumor tracking model can cope with changes in source-target distance and angle caused by physiological changes such as patient position, breathing or heart, ensure the tracking accuracy under different geometric changes, and enhance the generalization performance of the tumor tracking model.

[0059] Based on the above, before radiotherapy, the embodiment of the present application obtains a tumor tracking model dedicated to the patient through hybrid data enhancement, Monte Carlo simulation and two-dimensional / three-dimensional registration algorithm training for the patient's 4DCT image data, so that the tumor tracking model can accurately map the two-dimensional ultra-low-dose X-ray projection image with the three-dimensional target area and the position of the organs at risk during radiotherapy, so that the tumor tracking model can accurately generate the conversion relationship from the two-dimensional image to the three-dimensional position, realize efficient and real-time tumor tracking, and ensure high-precision positioning based on the patient's individualized information.

[0060] S130: obtaining a single X-ray image of any angle during the patient's radiotherapy, inputting the single X-ray image into the patient's tumor tracking model, and outputting the three-dimensional position of the tumor and the organ at risk through the tumor tracking model;

[0061] In this step, a single ultra-low-dose X-ray image of any angle during the patient's radiotherapy is obtained and input into the patient's tumor tracking model. The X-ray image is quickly analyzed by the tumor tracking model, and the three-dimensional position of the tumor and the organ at risk is output in real time. The embodiment of the present application can obtain the target area position in a short time with only a single low-dose X-ray image, realizes efficient real-time tumor tracking, and greatly reduces the patient's radiation dose and reduces damage to surrounding normal tissues, solving the clinical limitations of the traditional C image guidance method due to high dose and delay.

[0062] In order to verify the feasibility and effectiveness of the embodiments of the present application, experiments were conducted on a public dataset of lung cancer patients and a digital phantom XCAT. Figure 3 As shown, it is a schematic diagram of the tumor registration results of lung cancer patient data from the end of exhalation to the end of inhalation, wherein the odd-numbered columns are unregistered images, and the even-numbered rows are registered images. It can be seen that both the lungs and the tumors have obtained relatively good registration results, wherein the average DSC accuracy of the lungs is 0.97, and the overall tumor localization time is about 100 milliseconds, indicating that the embodiment of the present application can realize real-time low-dose 3D tumor tracking. Figure 4 This is a schematic diagram of the tumor center of mass motion error at different angles. It can be seen that the low-dose X-ray error at most angles is within 1 mm, achieving better tumor tracking results. Figure 5 This is a schematic diagram of the tumor registration results of the digital phantom XCAT from the end of exhalation to the end of inspiration. Figure 6 Schematic diagram of the motion trajectory of the tumor centroid in the digital phantom. It can be seen that the embodiment of the present application can also well realize tumor tracking on the digital phantom.

[0063] Based on the above, the dynamic tumor tracking method of the embodiment of the present application combines deep learning with 2D-3D registration technology, uses the patient's 4DCT image data to train a tumor tracking model dedicated to the patient, and inputs a single ultra-low-dose X-ray image at any angle during the patient's radiotherapy into the tumor tracking model to achieve efficient real-time tumor tracking. The tumor tracking model of the embodiment of the present application can achieve real-time three-dimensional tracking of tumors under any geometric conditions using a single low-dose X-ray image, can provide high-precision tumor positioning under the premise of ensuring low radiation dose, and is suitable for any projection angle and source-object distance change conditions, can flexibly respond to various geometric relationship changes during the treatment process, and adapt to more clinical application scenarios. The embodiment of the present application can achieve more accurate radiotherapy dose control in the case of low-dose X-ray images, while improving the treatment effect, it can reduce damage to surrounding normal tissues.

[0064] See also Figure 7, is a schematic diagram of the structure of a dynamic tumor tracking device according to an embodiment of the present application. The dynamic tumor tracking method device 40 according to an embodiment of the present application includes:

[0065] Image acquisition module 41: used to acquire 4DCT image data of the patient, and perform mixed data enhancement on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs;

[0066] Image simulation module 42: used for generating an X-ray image simulating a patient's respiratory cycle by using a mathematical simulation method based on the 3DCT image;

[0067] Image registration module 43: used to register the X-ray image and 4DCT image data to generate a tumor tracking model of the patient;

[0068] Tumor tracking module 44: used to input the X-ray images of the patient during radiotherapy into the tumor tracking model, and output the tumor tracking results through the tumor tracking model.

[0069] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0070] The device provided in the embodiment of the present application can be applied in the aforementioned method embodiment. For details, please refer to the description of the aforementioned method embodiment, which will not be repeated here.

[0071] See also Figure 8 , is a schematic diagram of the computer device structure of an embodiment of the present application. The computer device 50 includes:

[0072] A memory 51 storing executable program instructions;

[0073] A processor 52 connected to the memory 51;

[0074] The processor 52 is used to call the executable program instructions stored in the memory 51 and execute the following steps: obtain the patient's 4DCT image data, and perform mixed data enhancement on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs; based on the 3DCT image, use a mathematical simulation method to generate an X-ray image simulating the patient's respiratory cycle; align the X-ray image and the 4DCT image data to generate a tumor tracking model of the patient; input the X-ray image of the patient during radiotherapy into the tumor tracking model, and output the tumor tracking result through the tumor tracking model.

[0075] The processor 52 may also be referred to as a CPU (Central Processing Unit). The processor 52 may be an integrated circuit chip having signal processing capabilities. The processor 52 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0076] See also Fig. 9 , is a schematic diagram of the structure of the storage medium of the embodiment of the present application. The storage medium of the embodiment of the present application stores program instructions 61 that can implement the following steps: obtaining the patient's 4DCT image data, and performing mixed data enhancement on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs; based on the 3DCT image, using a mathematical simulation method to generate an X-ray image simulating the patient's respiratory cycle; registering the X-ray image and the 4DCT image data to generate a tumor tracking model for the patient; inputting the X-ray image during the patient's radiotherapy into the tumor tracking model, and outputting the tumor tracking result through the tumor tracking model. Among them, the program instruction 61 can be stored in the above storage medium in the form of a software product, including a number of instructions for a computer device (which can be a personal computer, a server, or a network computer device, etc.) or a processor (processor) to perform all or part of the steps of the various implementation methods of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program instructions, or terminal computer devices such as computers, servers, mobile phones, tablets, etc. Among them, the server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0077] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0078] In addition, each functional unit in each embodiment of the present 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 above integrated unit can be implemented in the form of hardware or in the form of software functional units. The above is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the description and drawings of this application, or directly or indirectly used in other related technical fields, is also included in the patent protection scope of the present application.

Claims

1. A dynamic tumor tracking method, characterized in that: include: Acquire 4DCT image data of the patient, and perform hybrid data enhancement on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs; Based on the 3DCT image, a mathematical simulation method is used to generate an X-ray image simulating a patient's respiratory cycle; registering the X-ray image and the 4DCT image data to generate a tumor tracking model of the patient; The X-ray images of the patient during radiotherapy are input into the tumor tracking model, and the tumor tracking results are output through the tumor tracking model.

2. The dynamic tumor tracking method according to claim 1, characterized in that: After obtaining the 4DCT image data of the patient, the method further includes: Any phase of the 4DCT image data is selected as a floating image, and the remaining N phases are selected as reference images. The reference image and the floating image are registered using an image registration method to obtain N deformation fields.

3. The dynamic tumor tracking method according to claim 2, characterized in that: The hybrid data enhancement is performed on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs, specifically: Randomly select two deformation fields from the N deformation fields and superimpose them to obtain phase-to-phase deformation; Thin plate spline interpolation is used to simulate random deformation of any phase; The phase-to-phase deformation and the random deformation are superimposed with arbitrary weights to obtain a deformation field, and the deformation field is applied to the floating image to obtain a 3D CT image representing each respiratory stage of the lung.

4. The dynamic tumor tracking method according to any one of claims 1 to 3, characterized in that: Based on the 3DCT image, a mathematical simulation method is used to generate an X-ray image simulating the patient's respiratory cycle, specifically: The 3DCT image is simulated by using a Monte Carlo simulation method to generate a 2D DRR, and the X-ray image in the patient's respiratory cycle is simulated by using the 2D DRR.

5. The dynamic tumor tracking method according to claim 4, characterized in that: The registering of the X-ray image and the 4DCT image data to generate a tumor tracking model of the patient is specifically as follows: The X-ray image is upgraded to a three-dimensional feature map through a residual block, the three-dimensional feature map is encoded and parsed from different angles using a 3D codec network, and the three-dimensional feature map is aligned with the floating image.

6. The dynamic tumor tracking method according to claim 5, characterized in that: The tumor tracking model adopts a deep learning model that combines a 3D codec network and a Unet-KAN mechanism.

7. The dynamic tumor tracking method according to claim 6, characterized in that: The step of inputting the X-ray image of the patient during radiotherapy into the tumor tracking model and outputting the tumor tracking result through the tumor tracking model is specifically as follows: A single X-ray image at any angle during the radiotherapy of the patient is obtained, the single X-ray image is input into the patient's tumor tracking model, the three-dimensional position of the tumor and the endangered organ is obtained through the tumor tracking model, and the tumor tracking result is output.

8. A dynamic tumor tracking device, characterized in that: include: Image acquisition module: used to acquire 4DCT image data of the patient, and perform mixed data enhancement on the 4DCT image data to generate 3DCT images representing various respiratory stages of the lungs; Image simulation module: used for generating an X-ray image simulating a patient's respiratory cycle by using a mathematical simulation method based on the 3DCT image; Image registration module: used to register the X-ray image and 4DCT image data to generate a tumor tracking model of the patient; Tumor tracking module: used to input the X-ray images of the patient during radiotherapy into the tumor tracking model, and output the tumor tracking results through the tumor tracking model.

9. A computer device, characterized in that: The computer device includes a processor and a memory coupled to the processor, wherein: The memory stores program instructions for implementing the dynamic tumor tracking method according to any one of claims 1 to 7; The processor is used to execute the program instructions stored in the memory to control a dynamic tumor tracking method.

10. A storage medium, characterized in that: The device stores program instructions executable by a processor, wherein the program instructions are used to execute the dynamic tumor tracking method according to any one of claims 1 to 7.

Images