Dynamic pet imaging method, apparatus, electronic device, and storage medium

Abstract

The application provides a dynamic PET imaging method and device, electronic equipment and storage medium, and relates to the technical field of medical PET imaging. The method comprises the following steps: acquiring a first time interval dynamic PET image; performing feature extraction on the first time interval dynamic PET image to obtain a corresponding feature vector; performing mapping on the extracted feature vector to obtain a corresponding pharmacokinetic model parameter; performing parameter conversion on the pharmacokinetic model parameter to obtain a second time interval SUV prediction image; the second time interval is adjacent to and later than the first time interval; and fitting the obtained second time interval SUV prediction image to obtain a parameter Ki image. The embodiment of the application solves the problem that in the related art, a long treatment time cannot stably obtain a high-quality dynamic PET image.

Description

Technical Field

[0001] This application relates to the field of medical PET imaging technology, and more specifically, to a dynamic PET imaging method, apparatus, electronic device, and storage medium. Background Technology

[0002] PET (Positron Emission Computed Tomography) imaging technology is a commonly used tumor detection method in the medical field. It can detect malignant lesions of diseases such as breast cancer, lung cancer, and lymphoma that cannot be detected by magnetic resonance imaging (MRI) and computed tomography (CT), making it a highly effective auxiliary diagnostic method.

[0003] However, this method requires the injection of a tracer into the body, and scanning can only be performed after the tracer has been stably distributed in the body, which greatly reduces patient comfort and prolongs treatment time. Moreover, the dosage of the tracer is also crucial. Since the tracer has radioactive hazards to the human body, excessively high doses cannot be used, but using low doses will significantly reduce the quality of the scanned images and affect the diagnostic results.

[0004] As can be seen from the above, how to reduce treatment time while stably obtaining high-quality dynamic PET images is of great scientific significance and application scenarios in the field of medical diagnosis. Summary of the Invention

[0005] This application provides a dynamic PET imaging method, apparatus, electronic device, and storage medium, which can solve the problems of long treatment times and inability to stably obtain high-quality dynamic PET images in related technologies. The technical solution is as follows:

[0006] According to one aspect of the embodiments of this application, a dynamic PET imaging method is provided, the method comprising: acquiring a dynamic PET image at a first time interval; extracting features from the dynamic PET image at the first time interval to obtain a corresponding feature vector; mapping the extracted feature vector to obtain corresponding pharmacokinetic model parameters; performing parameter transformation on the pharmacokinetic model parameters to obtain a predicted SUV image at a second time interval; the second time interval being adjacent to and later than the first time interval; and fitting the obtained predicted image to obtain a parameter Ki image.

[0007] According to one aspect of the embodiments of this application, a dynamic PET imaging device is provided, the device comprising: an image acquisition module for acquiring a dynamic PET image at a first time interval; the first time interval dynamic PET image is a PET image scanned after a certain time following the injection of a tracer; a feature extraction module for extracting features from the acquired dynamic PET image at the first time interval to obtain a corresponding feature vector; a parameter acquisition module for mapping the extracted feature vector to obtain corresponding pharmacokinetic model parameters; an image prediction module for performing parameter transformation on the pharmacokinetic model parameters to obtain a predicted SUV image at a second time interval; and an image fitting module for fitting the obtained predicted image to obtain a parameter Ki image.

[0008] According to one aspect of the present application, an electronic device includes: at least one processor, at least one memory, and at least one communication bus, wherein a computer program is stored in the memory, and the processor reads the computer program from the memory via the communication bus; when the computer program is executed by the processor, it implements the dynamic PET imaging method as described above.

[0009] According to one aspect of the embodiments of this application, a storage medium is provided on which a computer program is stored, which, when executed by a processor, implements the dynamic PET imaging method as described above.

[0010] According to one aspect of the embodiments of this application, a computer program product includes a computer program stored in a storage medium, a processor of a computer device reads the computer program from the storage medium, and the processor executes the computer program, causing the computer device to implement the dynamic PET imaging method as described above when executed.

[0011] The beneficial effects of the technical solution provided in this application are:

[0012] In the above technical solution, a dynamic PET image at a first time interval is acquired; features are extracted from the dynamic PET image at the first time interval to obtain corresponding feature vectors; the extracted feature vectors are mapped to obtain corresponding pharmacokinetic model parameters; the pharmacokinetic model parameters are transformed to obtain a predicted SUV image at a second time interval; the second time interval is adjacent to and later than the first time interval; and the predicted image is fitted to obtain a parameter Ki image. This effectively solves the problems of long treatment times and the inability to stably obtain high-quality dynamic PET images in related technologies. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below.

[0014] Figure 1 This is a flowchart illustrating a dynamic PET imaging method according to an exemplary embodiment;

[0015] Figure 2 This is a network structure diagram of a pharmacokinetic network model illustrated according to an exemplary embodiment;

[0016] Figure 3 This is a reconstruction result shown according to an exemplary embodiment;

[0017] Figure 4 This is a flowchart illustrating the training process of a pharmacokinetic network model according to an exemplary embodiment;

[0018] Figure 5 This is a structural block diagram of a dynamic PET imaging device according to an exemplary embodiment;

[0019] Figure 6 This is a hardware structure diagram of an electronic device according to an exemplary embodiment;

[0020] Figure 7 This is a structural block diagram of an electronic device according to an exemplary embodiment. Detailed Implementation

[0021] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0022] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0023] As mentioned earlier, during PET imaging, a tracer needs to be injected into the patient's body. Scanning is performed after the tracer has been distributed and stabilized in the body. Common tracers such as 2-fluoro-2-deoxy-D-glucose (18F-FDG) have a stability time of up to one hour in the human body, which greatly reduces patient comfort and prolongs treatment time. At the same time, due to the radioactive hazards of tracers to the human body, excessively high doses cannot be used, but using low doses will reduce image quality.

[0024] To address these issues, existing technologies have proposed several methods, including directly using deep learning to generate parametric images or SUV (standard uptake value) images. However, these methods cannot simultaneously generate SUV and parametric images, nor can they simultaneously generate continuous dynamic PET images at the same time frame. Even if they can be generated directly, the changing trends of adjacent time frames do not conform to the actual time-activity curve.

[0025] It is evident that the challenge of achieving stable acquisition of high-quality dynamic PET images while reducing treatment time remains to be solved.

[0026] Therefore, the dynamic PET imaging method provided in this application can effectively reduce treatment time while stably obtaining high-quality winter PET images. Accordingly, the dynamic PET imaging method is applicable to PET imaging devices, which can be deployed in medical equipment, such as PET scanners.

[0027] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0028] like Figure 1 As shown in the figure, this application provides a dynamic PET imaging method, which may include the following steps:

[0029] Step 310: Obtain the dynamic PET image of the first time interval.

[0030] As mentioned earlier, common tracers such as 2-fluoro-2-deoxy-D-glucose (18F-FDG) remain stable in the human body for up to an hour, which greatly reduces patient comfort and prolongs treatment time.

[0031] In one possible implementation, this application obtains a dynamic PET image of the first time interval after the tracer has been present in the human body for 30 minutes, and uses a deep learning model to predict the image for the next 30 minutes.

[0032] Step 330: Extract features from the dynamic PET image of the first time interval to obtain the corresponding feature vector.

[0033] In one possible implementation, feature extraction of dynamic PET images for the first time interval is based on a feature extraction network.

[0034] In one exemplary embodiment, the feature extraction network employs a residual encoder-decoder structure, wherein the encoder comprises a DoubleConv block and four DownConv blocks; and the decoder comprises an UpConv block.

[0035] The DoubleConv block, which serves as the basic block, consists of two convolutional layers connected to a GroupNorm layer and an activation layer. The activation function in the activation layer is the ReLU function, and the number of channels in each group of the GroupNorm layer is set to 16.

[0036] The DownConv block includes a Maxpool layer and a DoubleConv block, where the Maxpool layer is configured to downsample with a sampling factor of 2.

[0037] The UpConv block consists of a transposed convolutional layer and a DoubleConv block.

[0038] It is worth mentioning that, for each block at the same level, a skip connection is used between the encoder and decoder.

[0039] Step 350: Map the extracted feature vectors to obtain the corresponding pharmacokinetic model parameters.

[0040] In one possible implementation, mapping the extracted feature vectors to obtain the corresponding pharmacokinetic model parameters is based on a pharmacokinetic network model.

[0041] In an exemplary embodiment, the pharmacokinetic network model is a 2-organism, 3-compartment model, such as... Figure 2 As shown, the pharmacokinetic network model includes several convolutional layers, GroupNorm layers, LeakyReLU layers, and Hardsigomid layers.

[0042] Specifically, its network structure is shown in Table 1 below.

[0043] Table 1. Structure of the Pharmacokinetic Network Model

[0044]

[0045]

[0046] Step 370: Perform parameter transformation on the pharmacokinetic model parameters to obtain the SUV prediction image for the second time interval.

[0047] The second time interval is adjacent to the first time interval but later than the first time interval.

[0048] Specifically, the time-activity curves of voxels in human tissues and organs can be viewed as solving the following ordinary differential equation:

[0049]

[0050]

[0051] C T (t)=C1(t)+C2(t) (3)

[0052] Where K1, k2, k3, and k4 are model parameters, which are constants for a given tissue or organ. C0(t) is the tracer concentration-time function in arterial blood, which can be directly obtained from the original image; C1(t) is the tracer concentration-time function in the tissue or organ that is freely or non-specifically bound; C2(t) is the tracer concentration-time function that is specifically bound, which is the sum of the non-displaceable and specific compartment concentrations; C T (t) is a time function of the total tracer concentration in tissues and organs.

[0053] The voxel values ​​in PET images are directly proportional to the tracer concentration in tissues and organs. This can be achieved by solving for C. T (t) can then be used to solve for the dynamic PET prediction image. Solving equations (1), (2), and (3) above yields C. T The equation for (t) is:

[0054]

[0055] In equation (4), a, b, α1, and α2 are four constants determined by K1, k2, k3, and k4, and their specific relationships are as follows:

[0056]

[0057] Considering that each voxel is composed of blood and tissue, it is necessary to introduce a parameter f representing the blood volume fraction. v Therefore, the formula for the final time-activity curve is:

[0058]

[0059] In equation (6), S A λ is the initial activity of the tracer, λ is the decay rate of the radioactive isotope, and C is the initial activity of the tracer. WB (t) is the concentration of the tracer in whole blood, which is replaced by the plasma input function C0(t). Represents all parameters of the model, i.e. Therefore, the above formula can be used to obtain the time-activity curve of a voxel at any point in time.

[0060] Since dynamic PET images require scanning the time activity curves of voxels over a period of time to reconstruct the image, it is necessary to integrate and average the above equation (6), as expressed in the following formula:

[0061]

[0062] Among them, SUV (normalized ingress value) is the predicted image after the original image is reconstructed by the image reconstruction module.

[0063] Step 390: Fit the obtained SUV prediction image for the second time interval to obtain the parameter Ki image.

[0064] For SUV images, the quantification accuracy is often affected by factors such as measurement time and tracer concentration changes during the acquisition process. Therefore, using the net inflow rate (Ki) of dynamic PET imaging as a quantification indicator can more accurately reflect the rate at which the tracer is metabolized by human tissue.

[0065] In one exemplary embodiment, the predicted image is fitted using the Patlak method to obtain a Ki parameter image.

[0066] Specifically, such as Figure 3 As shown, where Figure 3 (a) is a reference SUV image. Figure 3 (b) is the SUV image obtained by combining the pharmacokinetic model. Figure 3 (c) is an SUV image obtained without combining a pharmacokinetic model. It can be seen that the predicted image obtained by the dynamic PET reconstruction method provided in the embodiments of this application is very similar to the reference image.

[0067] In one possible implementation, the pharmacokinetic network model is trained based on a PET image dataset; the construction of the PET image dataset includes: using several dynamic PET images X acquired at first time intervals. i The corresponding image parameters K are obtained using the Patlak method. i and V i According to the image parameter K i and V i and sampling time t i The denoised SUV prediction image y for the second time interval is calculated. i y i =K*t i +V.

[0068] This led to the construction of a PET image dataset containing multiple pairs of dynamic PET images with a first time interval and predicted SUV images with a second time interval.

[0069] For the training process of image reconstruction network models, please refer to [link / reference]. Figure 4 This may include the following steps:

[0070] Step 410: Input the dynamic PET image of the first time interval in the PET image dataset into the basic network model for image prediction to obtain the corresponding predicted image.

[0071] Step 430, based on the predicted image and the predicted image of the target SUV y i Determine the loss function, and then obtain the loss value.

[0072] Step 450: If the loss value meets the set convergence condition, the training is complete; otherwise, update the parameters of the base model until the loss value meets the set convergence condition.

[0073] It should be noted that the convergence condition can be flexibly set according to the actual situation. For example, the convergence condition refers to the convergence condition based on the predicted image and the SUV predicted image y after the second time interval after denoising. i The loss value obtained by the loss function between them meets the threshold, but no specific limit is made here.

[0074] It is worth mentioning that the parameters of the basic model can be obtained by minimizing the mean squared error function, as expressed in the following formula:

[0075] Loss=∑||x' i -y i || 2 +||(x' i -x' i-1 )-(y i -y i-1 )|| 2 (8)

[0076] Where X i ={x i1 ,x i2 ,…,x i28} is achieved by scanning 28 frames of dynamic PET images over one hour, for any X i Obtain the parameter image K using the Patlak method. i (Slope) and V i (Intercept) and sampling time t i ={t 23 ,t 24 ,t 25 ,t 26 ,t 27 ,t28} can then obtain y i , where y i =K*t i +V indicates that the Adam optimizer is used for function optimization.

[0077] After the above training process, a deep learning network model based on the pharmacokinetic network model is obtained.

[0078] The following are embodiments of the apparatus described in this application, which can be used to execute the dynamic PET imaging method involved in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the method embodiments of the dynamic PET imaging method involved in this application.

[0079] Please see Figure 5 This application provides a dynamic PET imaging device 500, including but not limited to: an image acquisition module 510, a feature extraction module 520, a parameter acquisition module 530, an image prediction module 540, and an image fitting module 550.

[0080] The image acquisition module 510 is used to acquire dynamic PET images at the first time interval.

[0081] The feature extraction module 520 is used to extract features from the acquired dynamic PET image of the first time interval to obtain the corresponding feature vector.

[0082] The parameter acquisition module 530 is used to map the extracted feature vectors to obtain the corresponding pharmacokinetic model parameters.

[0083] The image prediction module 540 is used to perform parameter transformation on the pharmacokinetic model parameters to obtain the SUV prediction image for the second time interval.

[0084] The image fitting module 550 is used to fit the obtained predicted image to obtain the parameter Ki image.

[0085] It should be noted that the device control device provided in the above embodiments is only illustrated by the division of the above functional modules when controlling the device. In actual applications, the above functions can be assigned to different functional modules as needed. That is, the internal structure of the device control device will be divided into different functional modules to complete all or part of the functions described above.

[0086] Furthermore, the device control apparatus and device control method embodiments provided in the above embodiments belong to the same concept, and the specific way in which each module performs operations has been described in detail in the method embodiments, and will not be repeated here.

[0087] Figure 6A schematic diagram of the structure of an electronic device is shown according to an exemplary embodiment.

[0088] It should be noted that this electronic device is merely an example adapted to this application and should not be construed as providing any limitation on the scope of use of this application. Furthermore, this electronic device should not be interpreted as requiring or depending on any specific feature. Figure 6 One or more components of the exemplary electronic device 2000 shown.

[0089] The hardware structure of electronic devices 2000 can vary significantly due to differences in configuration or performance, such as... Figure 6 As shown, the electronic device 2000 includes: a power supply 210, an interface 230, at least one memory 250, and at least one central processing unit (CPU) 270.

[0090] Specifically, power supply 210 is used to provide operating voltage for various hardware devices on electronic device 2000.

[0091] Interface 230 includes at least one wired or wireless network interface 231 for interacting with external devices.

[0092] Of course, in other examples adapted in this application, interface 230 may further include at least one serial-to-parallel conversion interface 233, at least one input / output interface 235, and at least one USB interface 237, etc. Figure 6 As shown, this does not constitute a specific limitation.

[0093] The memory 250 serves as a carrier for resource storage and can be a read-only memory, random access memory, disk, or optical disk, etc. The resources stored on it include the operating system 251, application programs 253, and data 255, etc., and the storage method can be temporary storage or permanent storage.

[0094] The operating system 251 is used to manage and control the various hardware devices and application programs 253 on the electronic device 2000, so as to enable the central processing unit 270 to perform calculations and processing on the massive data 255 in the memory 250. It can be Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, etc.

[0095] Application 253 is a computer program that performs at least one specific task based on operating system 251, and may include at least one module ( Figure 6 (Not shown), each module may contain a computer program for the electronic device 2000. For example, the dynamic PET imaging device can be considered as an application program 253 deployed on the electronic device 2000.

[0096] Data 255 can be photos, pictures, etc. stored on a disk.

[0097] The central processing unit 270 may include one or more processors and is configured to communicate with the memory 250 via at least one communication bus to read computer programs stored in the memory 250, thereby performing calculations and processing on massive amounts of data 255 stored in the memory 250. For example, a dynamic PET imaging method may be performed by the central processing unit 270 reading a series of computer programs stored in the memory 250.

[0098] Furthermore, this application can also be implemented through hardware circuits or a combination of hardware circuits and software. Therefore, the implementation of this application is not limited to any specific hardware circuit, software, or combination thereof.

[0099] Please see Figure 7 This application provides an electronic device 4000, which can be used as a PET detector.

[0100] exist Figure 7 The electronic device 4000 includes at least one processor 4001, at least one communication bus 4002, and at least one memory 4003.

[0101] The processor 4001 and memory 4003 are connected, for example, via a communication bus 4002. Optionally, the electronic device 4000 may also include a transceiver 4004, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of this application.

[0102] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 4001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0103] The communication bus 4002 may include a path for transmitting information between the aforementioned components. The communication bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus 4002 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0104] The memory 4003 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0105] The memory 4003 stores a computer program, and the processor 4001 reads the computer program stored in the memory 4003 through the communication bus 4002.

[0106] When the computer program is executed by the processor 4001, it implements the dynamic PET imaging method in the above embodiments.

[0107] Furthermore, this application provides a storage medium storing a computer program, which, when executed by a processor, implements the dynamic PET imaging method described in the above embodiments.

[0108] This application provides a computer program product including a computer program stored in a storage medium. A processor of a computer device reads the computer program from the storage medium and executes the computer program, causing the computer device to perform the dynamic PET imaging method described in the above embodiments.

[0109] Compared with related technologies, this application achieves high-quality dynamic PET images by combining pharmacokinetic models and deep learning. Furthermore, it replaces the original multi-frame prediction and parametric image generation of dynamic PET with a single model, effectively improving the accuracy of image prediction, enhancing the interpretability of the network, reducing the necessary time for dynamic PET scanning, and increasing patient comfort.

[0110] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0111] The above description is only a partial 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. A dynamic PET imaging method, characterized in that, The method includes: Acquire dynamic PET images at the first time interval; Feature extraction is performed on the dynamic PET image at the first time interval to obtain the corresponding feature vector; The extracted feature vectors are mapped to obtain the corresponding pharmacokinetic model parameters; The corresponding tracer time-activity curve was calculated based on the pharmacokinetic model parameters. The SUV prediction image for the corresponding second time interval is obtained by integrating and averaging the tracer time activity curve. The tracer time-activity curve calculated based on the pharmacokinetic model parameters uses the following formula: in, It is the total tracer concentration in the tissue. It is the initial activity of the tracer. It is the decay rate of the radioactive isotope. It is the concentration of the tracer in whole blood. These are all the parameters of the pharmacokinetic model. It is the blood volume fraction; The method for integrating and averaging the tracer time-activity curve to obtain the SUV prediction image for the corresponding second time interval is as follows: Here, SUV refers to the predicted SUV image in the second time interval; the second time interval is adjacent to the first time interval but later than the first time interval. The obtained SUV prediction image for the second time interval is fitted to obtain the parameter Ki image.

2. The method as described in claim 1, characterized in that, The feature extraction of the acquired first time interval dynamic PET image to obtain the corresponding feature vector is implemented based on a feature extraction network, which includes: Residual encoder-decoder; The encoder includes one DoubleConv block and four DownConv blocks; The decoder includes an UpConv block.

3. The method as described in claim 2, characterized in that, The mapping of the extracted feature vectors to obtain the corresponding pharmacokinetic model parameters is implemented based on a pharmacokinetic network model, which includes: Several convolutional layers, GroupNorm layers, LeakyReLU layers, and Hardsigomid layers.

4. The method as described in claim 3, characterized in that, The training process of the feature extraction network and pharmacokinetics network model includes: The dynamic PET images at the first time interval in the PET image dataset are input into the basic network model for image prediction to obtain the corresponding predicted images. Based on the predicted image and the target SUV predicted image Determine the loss function, and then obtain the loss value; If the loss value meets the set convergence condition, the training is complete; otherwise, the parameters of the basic network model are updated until the loss value meets the set convergence condition.

5. The method as described in claim 4, characterized in that, The predicted image of the target SUV It is obtained through the following steps: Based on the acquired dynamic PET images The Patlak method is used to obtain the corresponding image parameter slope. and intercept ; Based on the slope of the image parameters and intercept and sampling time The denoised target SUV prediction image is calculated. .

6. The method as described in claim 5, characterized in that, The loss function is: 。 7. A dynamic PET imaging device, characterized in that, The device includes: The image acquisition module is used to acquire dynamic PET images at the first time interval; The feature extraction module is used to extract features from the dynamic PET image of the first time interval to obtain the corresponding feature vector; The parameter acquisition module is used to map the extracted feature vectors to obtain the corresponding pharmacokinetic model parameters; The image prediction module is used to calculate the corresponding tracer time-activity curve based on the pharmacokinetic model parameters; Integrating and averaging the tracer time-activity curves yields the SUV prediction image for the corresponding second time interval. The tracer time-activity curve calculated based on the pharmacokinetic model parameters uses the following formula: in, It is the total tracer concentration in the tissue. It is the initial activity of the tracer. It is the decay rate of the radioactive isotope. It is the concentration of the tracer in whole blood; The method for integrating and averaging the tracer time-activity curve to obtain the SUV prediction image for the corresponding second time interval is as follows: Among them, SUV refers to the predicted SUV image in the second time interval; The image fitting module is used to fit the obtained predicted image to obtain the parameter Ki image.

8. An electronic device, characterized in that, It includes at least one processor, at least one memory, and at least one communication bus, wherein, The memory stores a computer program, and the processor reads the computer program from the memory via the communication bus; When the computer program is executed by the processor, it implements the dynamic PET imaging method according to any one of claims 1 to 6.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the dynamic PET imaging method as described in any one of claims 1 to 6.

Images