Particle dose determination method, device and equipment based on PET signal nuclide decomposition, medium and program product

Abstract

The invention provides a particle dose determination method, device and equipment based on PET signal nuclide decomposition, a medium and a program product, and relates to application of physics and computer fields in the radiation medical field. The method comprises the following steps: simulating a plurality of sample CT images through a Monte Carlo particle transport simulation technology to obtain sample information corresponding to each sample image; constructing a first training data set based on the multiple pieces of sample information, and training the first initial model by using the first training data set to obtain a first target model for predicting the signal ratio of each nuclide; training a second initial model by taking the first training data set and a signal proportion output by the first target model as a second training data set to obtain a second target model for predicting three-dimensional dose distribution information; and sequentially inputting the target CT image and the PET signal of the target time window into the first target model and the second target model to obtain three-dimensional dose distribution information output by the second target model.

Description

Technical Field

[0001] This invention relates to the fields of physics and computer science, specifically to the application of physics and computer technology in the field of radiomedicine, and more specifically to a method, apparatus, device, medium, and program product for determining particle dose based on the decomposition of PET signal nuclides. Background Technology

[0002] Radiotherapy is one of the important means of cancer treatment. Proton and heavy ion therapy (collectively known as particle therapy) can achieve high-dose deposition in the target area and significantly reduce the radiation dose to surrounding normal tissues by utilizing the superior Bragg peak dose distribution characteristics. This results in higher tumor control rates and lower adverse reactions. The accuracy of dose delivery in particle therapy directly affects the effectiveness of particle therapy.

[0003] Currently, commonly used clinical dose verification methods mainly include offline physical dosimetry, CT-based range reconstruction, and radiochemical tracing. However, these methods have limitations, such as being unable to reflect the real-time location of the particle beam within the body and the actual dose distribution, and being significantly affected by factors such as tissue heterogeneity, scattering, and nuclear reactions. Summary of the Invention

[0004] This invention provides a method, apparatus, equipment, medium, and program product for determining particle dose based on PET signal nuclide decomposition, which is used to at least partially solve one of the above-mentioned technical problems.

[0005] According to a first aspect of the present invention, a method for determining particle dose based on PET signal nuclide decomposition is provided, comprising: simulating multiple sample CT images using Monte Carlo particle transport simulation technology to obtain sample information corresponding to each sample image, wherein the sample information includes independent signal components of each nuclide, PET signal, nuclide distribution information, and dose distribution information within a specified time window; constructing a first training dataset based on the multiple sample information, training a first initial model using the first training dataset to obtain a first target model for predicting the signal proportion of each nuclide; training a second initial model using the first training dataset and the signal proportion output by the first target model as a second training dataset to obtain a second target model for predicting three-dimensional dose distribution information; and sequentially inputting a target CT image and a PET signal from a target time window into the first target model and the second target model to obtain three-dimensional dose distribution information output by the second target model.

[0006] According to an embodiment of the present invention, multiple sample CT images are simulated using Monte Carlo particle transport simulation technology to obtain sample information corresponding to each sample image. This includes: simulating the particle transport process in tissue under different energies, energy divergences, and broadening conditions based on Monte Carlo particle transport simulation technology; for each simulation scenario, calculating the types and spatial distribution information of nuclides generated when particles collide with tissue based on a nuclear reaction model; for each nuclide, calculating the photons generated by the decay of the nuclide within a specified time window to form the independent signal component corresponding to that nuclide; and summing the independent signal components of each nuclide within the specified time window to obtain the PET signal of the specified time window under the simulation scenario.

[0007] According to an embodiment of the present invention, a first training dataset is constructed based on multiple sample information, including: calculating the signal proportion of each nuclide based on the independent signal components of each nuclide and the PET signal respectively; for each simulated scenario, using sample CT images, PET signals and beam parameters as input data, and using the signal proportion of each nuclide as output data, to construct a sample pair corresponding to the simulated scenario; the first training dataset includes multiple sample pairs, and the beam parameters include energy, broadening, incident angle and dose rate.

[0008] According to an embodiment of the present invention, a first initial model is trained using a first training dataset to obtain a first target model for predicting the signal proportion of each nuclide. The training includes: using sample CT images, PET signals, and beam parameters as input data, and the signal proportion of each nuclide as output, training the first initial model through supervised learning so that the first initial model learns the energy dependence characteristics and spatial decay law of the PET signal, thereby obtaining the first target model for predicting the signal proportion of each nuclide.

[0009] According to an embodiment of the present invention, the second initial model is trained using the first training dataset and the signal proportions output by the first target model as the second training dataset to obtain a second target model for predicting three-dimensional dose distribution information. This includes: using the first training dataset and the signal proportions of each nuclide output by the first target model as inputs, and the dose distribution information as outputs, the second initial model is trained through supervised learning to learn the mapping relationship between signal combinations and dose distribution, thereby obtaining a second target model for predicting dose distribution.

[0010] According to an embodiment of the present invention, a target CT image and a PET signal of a target time window are sequentially input into a first target model and a second target model to obtain three-dimensional dose distribution information output by the second target model. This includes: inputting the target CT image, the PET signal image of the target time window, and beam information as first input data into the first target model so that the first target model can determine the signal proportion of each nuclide within the target time window based on the first input data; and inputting the signal proportion of each nuclide output by the first target model and the first input data as second input data into the second target model so that the second target model can determine the dose distribution information based on the second input data.

[0011] According to a second aspect of the present invention, a particle dose determination device based on PET signal nuclide decomposition is provided, comprising: a simulation module, configured to simulate multiple sample CT images respectively using Monte Carlo particle transport simulation technology to obtain sample information corresponding to each sample image, the sample information including independent signal components of each nuclide, PET signal, nuclide distribution information and dose distribution information under a specified time window; a first training module, configured to construct a first training dataset based on multiple sample information, and train a first initial model using the first training dataset to obtain a first target model for predicting the signal proportion of each nuclide; a second training module, configured to train a second initial model using the first training dataset and the signal proportion output by the first target model as a second training dataset to obtain a second target model for predicting three-dimensional dose distribution information; and an acquisition module, configured to sequentially input the target CT image and the PET signal of the target time window into the first target model and the second target model to obtain the three-dimensional dose distribution information output by the second target model.

[0012] According to a third aspect of the present invention, an electronic device is provided, comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method described above.

[0013] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0014] According to a fifth aspect of the present invention, a computer program product is provided, comprising a computer program or instructions that, when executed by a processor, implement the steps of the above-described method. Attached Figure Description

[0015] The above and other objects, features and advantages of the present invention will become more apparent from the following description of embodiments of the invention with reference to the accompanying drawings, in which:

[0016] Figure 1 A flowchart illustrating a particle dose determination method based on PET signal nuclide decomposition according to an embodiment of the present invention is shown schematically.

[0017] Figure 2 This schematically illustrates a flowchart of a process for simulating multiple sample CT images using Monte Carlo particle transport simulation technology according to an embodiment of the present invention, thereby obtaining sample information corresponding to each sample image.

[0018] Figure 3 This illustration schematically shows a flowchart of constructing a first training dataset based on information from multiple samples according to an embodiment of the present invention;

[0019] Figure 4 The schematic diagram illustrates the principle of training a first initial model according to an embodiment of the present invention;

[0020] Figure 5 The schematic diagram illustrates the principle of training a second initial model according to an embodiment of the present invention;

[0021] Figure 6 The flowchart illustrates, according to an embodiment of the present invention, the sequential input of a target CT image and a target time window PET signal into a first target model and a second target model to obtain three-dimensional dose distribution information output by the second target model;

[0022] Figure 7 This schematic diagram illustrates the structure of a particle dose determination device based on PET signal nuclide decomposition according to an embodiment of the present invention.

[0023] Figure 8 A block diagram of an electronic device for determining particle dose in PET signal nuclide decomposition according to an embodiment of the present invention is shown schematically. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0025] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. The terms “comprising,” “including,” etc., as used herein indicate the presence of a feature, step, operation, and / or component, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0026] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between them; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0027] In the description of this invention, it should be understood that the terms "longitudinal", "length", "circumferential", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the subsystem or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0028] Throughout the accompanying drawings, identical elements are represented by the same or similar reference numerals. Conventional structures or configurations may be omitted where they might cause confusion in understanding the invention. Furthermore, the shapes, dimensions, and positional relationships of the components in the drawings do not reflect actual size, scale, or actual positional relationships. Additionally, any reference symbols placed within parentheses should not be construed as limiting.

[0029] Similarly, to simplify the invention and aid in understanding one or more of the various disclosed aspects, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together into a single embodiment, figure, or description thereof. The use of terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicates that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0030] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0031] This invention provides a method, apparatus, device, medium, and program product for determining particle dose based on PET signal nuclide decomposition. Before introducing the technical solution provided by this invention, the related technologies involved in this invention will be described first.

[0032] Online or near-real-time particle beam monitoring technology based on positron emission tomography (PET) utilizes β⁺ radionuclides (such as...) generated by the interaction of particles with tissue nuclei. 11 C 15 O、 13 The spatial distribution of these radioactive signals is detected by a PET system, thereby indirectly reflecting the particle beam energy deposition and range information.

[0033] However, traditional PET monitoring still faces several key technical bottlenecks: First, the amount of radionuclides generated is affected by the complex influence of beam energy, tissue composition, and nuclear reaction cross-section, resulting in a non-linear correspondence between signal intensity and actual dose; second, the long acquisition time, high image noise, and limited spatiotemporal resolution make it difficult to achieve high-precision dynamic dose mapping.

[0034] With the rapid development of artificial intelligence technology, deep learning-based methods for radiotherapy dose prediction and monitoring have gradually become a research hotspot in the field of particle therapy. Traditional physical and empirical models require complex Monte Carlo simulations and tissue parameter assumptions in dose calculation, resulting in high computational costs and high sensitivity to input accuracy. In recent years, deep neural networks, with their powerful capabilities in high-dimensional nonlinear mapping modeling, have been widely used to directly learn the particle beam energy deposition patterns from multimodal image data, achieving end-to-end prediction from image to dose distribution.

[0035] In particle therapy, researchers have attempted to use multi-source data, including pre-treatment CT images, stopping power ratio (SPR) distribution, and PET signals acquired during treatment, as network inputs. They have employed structures such as deep convolutional networks (CNNs), 3D U-Nets, or spatiotemporal attention networks to learn the complex correspondence between PET activity distribution and actual dose deposition distribution. While these methods overcome the limitation of traditional models that can only qualitatively estimate range deviation based on static activity maps, their accuracy in predicting 3D dose distribution remains insufficient.

[0036] This invention proposes a particle dose determination method based on PET signal nuclide decomposition, comprising: simulating multiple sample CT images using Monte Carlo particle transport simulation technology to obtain sample information corresponding to each sample image, the sample information including independent signal components of each nuclide, PET signal, nuclide distribution information, and dose distribution information within a specified time window; constructing a first training dataset based on the multiple sample information, and training a first initial model using the first training dataset to obtain a first target model for predicting the signal proportion of each nuclide; training a second initial model using the first training dataset and the signal proportion output by the first target model as a second training dataset to obtain a second target model for predicting three-dimensional dose distribution information; and sequentially inputting the target CT image and the PET signal of the target time window into the first target model and the second target model to obtain the three-dimensional dose distribution information output by the second target model.

[0037] This invention achieves multi-nucleoside decomposition of PET signals through hierarchical training of the first target model and the second target model, thereby improving the physical accuracy of dose assessment.

[0038] This invention constructs a nuclide signal decomposition network (first target model) and intelligently separates the signal proportion map of each nuclide using early PET signals within a specified time period, combined with CT images and beam parameters. This is in contrast to traditional PET monitoring techniques that only use the total β⁺ radioactive signal for dose or range estimation, failing to distinguish between different nuclides (such as ¹¹C, ... 5 O、¹ 0 C) The contribution of the signal causes the estimation results to be affected by the coupling effect of the nuclide generation rate, decay constant and diffusion characteristics, making it difficult to achieve physically accurate dose correspondence. This invention can significantly reduce the systematic error caused by signal aliasing and provide more physically meaningful input features for subsequent dose prediction.

[0039] The second target model in this invention is decomposed 11 C 15 O、 10 Using the C-signal image as the core input, and combining CT density distribution and beam information, a nonlinear relationship between radionuclide activity and energy deposition distribution was established. This model effectively integrates tissue structure, nuclear reaction characteristics, and time decay patterns, achieving accurate reconstruction from image signals to dose spatial distribution. Compared with traditional models based on empirical fitting or univariate regression, the deep learning method of this invention exhibits significant advantages in spatial consistency, numerical accuracy, and stability.

[0040] Because this invention performs nuclide-level decomposition of PET signals before dose prediction, each input feature has a clear physical meaning and can independently verify the generation and decay patterns of the corresponding nuclide. This makes the model prediction results traceable and interpretable, providing a reliable basis for clinical validation and scientific research analysis. In particular, the model maintains good generalization ability under different energy beams, different tissue compositions, or different scan times.

[0041] The dual-model architecture proposed in this invention enables rapid inference under GPU acceleration, completing radionuclide decomposition and dose prediction within 2 minutes of PET signal processing, with a computation time of only a few seconds to a dozen seconds, fully meeting the needs of clinical online monitoring. Furthermore, based on standardized inputs (CT images, PET signals, and beam parameters), this method can be seamlessly integrated into existing particle therapy planning and imaging systems, exhibiting excellent scalability and engineering feasibility. Moreover, the method provided by this invention allows for model transfer and application across different particle types by adjusting Monte Carlo training data and beam input parameters, providing a unified intelligent dose monitoring solution for multiple types of radiotherapy systems.

[0042] Figure 1 A flowchart illustrating a particle dose determination method based on PET signal nuclide decomposition according to an embodiment of the present invention is shown.

[0043] like Figure 1 As shown, the particle dose determination method based on PET signal nuclide decomposition in this embodiment includes operations S110 to S140.

[0044] In operation S110, multiple sample CT images are simulated using Monte Carlo particle transport simulation technology to obtain sample information corresponding to each sample image. The sample information includes the independent signal components of each nuclide, PET signal, nuclide distribution information and dose distribution information under a specified time window.

[0045] In some embodiments, the sample CT images can be image data obtained through actual scans or simulated images generated using computer software and algorithms. The sample CT images can include images of different types and tissue structures. Each sample CT image can be simulated using a Monte Carlo particle transport simulation algorithm. By simulating the particle (proton or heavy ion) incident process under different energies, divergences, and broadening conditions, sample information corresponding to each simulated scenario can be obtained.

[0046] Sample information can include independent signal components of each nuclide, PET signals, nuclide distribution information, and dose distribution information within a specified time window. Different nuclides, due to their varying decay characteristics and interactions with surrounding matter, will produce unique signals within a specified time window. Independent signal components reflect the individual contribution of each nuclide to the overall physical process within that time window. PET signals are the combined signals generated by the decay and interactions of multiple nuclides in the target tissue or organ. Nuclide distribution information describes the spatial distribution of nuclides within the target space, reflecting differences in nuclide concentration or toxicity at different locations. Dose distribution information refers to the distribution of radiation energy deposition within the target space, used to describe the amount of radiation energy absorbed per unit mass of material at different locations.

[0047] In operation S120, a first training dataset is constructed based on multiple sample information. The first initial model is trained using the first training dataset to obtain a first target model for predicting the signal proportion of each nuclide.

[0048] In some embodiments, a first training dataset is constructed by integrating multiple sample information obtained from simulation. The input data of this dataset may include PET signals, sample CT images, and beam information within a specified time window, and the output data may be the signal proportion of each nuclide. The signal proportion of each nuclide can be calculated from the independent signal components of each nuclide and the PET signal.

[0049] A suitable deep learning model can be selected as the initial model. This initial model could include, for example, a U-Net model, a ResNet model, a 3D convolutional network model, or a temporal attention network model. The initial model is trained using a first training dataset. By continuously adjusting the model's parameters, it learns the mapping relationship between input data such as PET and output data on the proportion of each nuclide signal. During training, supervised learning can be employed to optimize the model by minimizing the error between the predicted and true values. Through multiple iterations of training, a primary target model for predicting the proportion of each nuclide signal is obtained.

[0050] In operation S130, the first training dataset and the signal proportion output by the first target model are used as the second training dataset to train the second initial model, thereby obtaining the second target model for predicting three-dimensional dose distribution information.

[0051] In some embodiments, the original data (such as sample CT images, PET signals within a specified time window, and beam parameters) in the first training dataset are combined with the signal proportion information output by the first target model to construct a second training dataset. The input data in the second training dataset consists of the original image information and the radionuclide signal proportion information processed by the first target model, and the output data consists of dose distribution information.

[0052] A deep learning model suitable for processing multimodal data and performing nonlinear regression can be selected as the second initial model. This second initial model is trained using a second training dataset, enabling it to learn the mapping relationship between the input multimodal data and the output dose distribution information. Through multimodal feature fusion and nonlinear regression learning, an end-to-end mapping from radionuclide signal combinations to spatial energy deposition distribution is achieved. The second initial model can also employ supervised learning, continuously optimizing model parameters to improve the accuracy of dose distribution prediction. After sufficient training, a second target model for predicting three-dimensional dose distribution information is obtained.

[0053] In operation S140, the target CT image and the PET signal of the target time window are sequentially input into the first target model and the second target model to obtain the three-dimensional dose distribution information output by the second target model.

[0054] In some embodiments, target CT images and PET signals within a target time window are acquired; these data can be collected in actual clinical settings. The target CT images and PET signals within the target time window are input into a first target model and a second target model simultaneously. First, the first target model predicts the signal proportion of each nuclide. Then, this information, along with other input data, is input into the second target model, ultimately yielding three-dimensional dose distribution information output by the second target model.

[0055] This invention provides dose distribution prediction based on nuclide decomposition. By using a first target model to intelligently separate the signal proportion of each nuclide from PET signals, CT effects, and beam parameters, it avoids aliasing problems caused by mutual interference between different nuclide signals, effectively reduces system errors, and provides more accurate and physically meaningful input features for subsequent dose prediction by a second target model, thereby effectively improving the accuracy of dose assessment.

[0056] This invention also employs a phased training method for two target models. The first target model focuses on predicting the signal proportion of each nuclide, while the second target model, based on this, combines other multimodal information for dose prediction. This allows each model to focus on a specific task, avoiding overfitting caused by overly complex models and further improving the accuracy of dose prediction.

[0057] Figure 2The flowchart illustrates a process of simulating multiple sample CT images using Monte Carlo particle transport simulation technology according to an embodiment of the present invention to obtain sample information corresponding to each sample image.

[0058] like Figure 2 As shown, this embodiment uses Monte Carlo particle transport simulation technology to simulate multiple sample CT images to obtain sample information corresponding to each sample image, including operations S210 to S240.

[0059] In operation S210, the transport process of particles in tissues under different energy, energy dissipation and broadening conditions was simulated based on Monte Carlo particle transport simulation technology.

[0060] In some embodiments, the sample CT images can cover samples of different tissue types and structures. Energy divergence reflects the degree of dispersion of particle energy in the particle beam, and broadening describes the spatial diffusion of the particle beam. These parameters can be selected according to common situations in actual treatment to simulate the real particle transport environment.

[0061] Monte Carlo particle transport simulation technology can be used to simulate the transport process of particles in tissue for each sample CT image and a set combination of simulation parameters. The Monte Carlo method tracks the trajectory of each particle through random sampling and considers various interactions between particles and tissue atoms, such as elastic scattering, inelastic scattering, and nuclear reactions.

[0062] When operating S220, for each simulation scenario, the types and spatial distribution of nuclides generated when particles collide with tissues are calculated based on the nuclear reaction model.

[0063] In some embodiments, for each simulation scenario, the types and spatial distribution of nuclides generated when particles collide with tissue can be calculated using a nuclear reaction model based on recorded particle transport information. The nuclear reaction model, based on quantum mechanics and nuclear physics, describes the probability and products of reactions between particles of different energies and tissue nuclei. Based on the collision locations and nuclear reaction probabilities of the particles in the simulation, the spatial distribution of the generated nuclides within the tissue is determined.

[0064] In operation S230, for each nuclide, the photons generated by the decay of the nuclide, calculated within a specified time window, form the independent signal component corresponding to that nuclide.

[0065] In some embodiments, the specified time window can be set according to the decay characteristics of the nuclide and the actual PET imaging time range, and can be 1 minute, 2 minutes, 3 minutes, 5 minutes, etc. In this embodiment of the invention, the target time window is set to 2 minutes.

[0066] For each nuclide, the number and spatial distribution of photons generated by its decay are calculated within a specified time window, forming an independent signal component corresponding to that nuclide. Nuclide decay is a stochastic process, but the Monte Carlo method can simulate the decay behavior of a large number of nuclides and statistically determine the photon information generated within the target time window. For example, for ¹¹C nuclide, given its half-life of 20.39 minutes, the number of positrons generated within the target time window can be calculated based on its decay pattern and initial quantity, thus determining the distribution of photons generated by the annihilation of these positrons in the tissue.

[0067] In operation S240, the independent signal components of each nuclide within the target time window are summed to obtain the PET signal for the specified time window in the simulated scenario.

[0068] In some embodiments, the independent signal components of each nuclide within a target time window are summed to obtain the PET signal for a specified time window in the simulated scenario. This synthesized PET signal simulates the combined effect of photons generated by the decay of different nuclides in actual PET imaging. For example, ... 11 C 15 O、 10 The independent signal components of radionuclides such as C within the target time window are added together to obtain a complete PET signal distribution map, which reflects the total radioactivity at different locations in the tissue.

[0069] This invention comprehensively considers the influence of multiple factors such as particle energy, energy divergence, and broadening on particle transport and nuclide generation. Using Monte Carlo simulation, it can more realistically simulate the physical processes in actual treatment, improving the accuracy and reliability of the simulation results. The sample information (particle transport information, nuclide generation information, and PET signals) corresponding to each sample image obtained from the simulation provides rich data support for dose assessment and signal analysis, thereby improving the accuracy of subsequent dose assessment.

[0070] Figure 3 The flowchart illustrating the construction of a first training dataset based on information from multiple samples according to an embodiment of the present invention is shown.

[0071] like Figure 3 As shown, the construction of the first training dataset based on multiple sample information in this embodiment includes operations S310 to S320.

[0072] In operation S310, the signal proportion of each nuclide information is calculated based on the independent signal components of each nuclide and the PET signal.

[0073] In some embodiments, the first training dataset includes multiple sample pairs, and the beam parameters include energy, broadening, incident angle, and dose rate. Independent signal components reflect the distribution of decay patterns produced by individual nuclides, while the PET signal is the result of the combined effect of decaying photons from all nuclides. For each nuclide, the intensity value of its independent signal component at each spatial location is divided by the intensity value of the PET signal at the same location to obtain the signal proportion of that nuclide at each location. This calculation process is repeated for all grid points in the tissue to obtain the signal proportion distribution of each nuclide throughout the entire tissue region.

[0074] When operating the S320, for each simulation scenario, sample CT images, PET signals, and beam parameters are used as input data, and the signal proportion of each nuclide is used as output data to construct sample pairs corresponding to the simulation scenario.

[0075] In some embodiments, for each simulation scenario, relevant sample CT images, PET signals, and beam parameters are collected. Sample CT images provide anatomical information of the tissue, PET signals contain comprehensive information on the decay of all nuclides, and beam parameters (including energy, broadening, incident angle, and dose rate) describe the characteristics of the particle beam, which affect particle transport and nuclide generation within the tissue. The signal percentage of each nuclide calculated based on operation S310 is used as output data.

[0076] Input and output data are combined to form a sample pair. For example, a sample pair can be represented as (sample CT image, PET signal, beam parameters (energy, broadening, incident angle, dose rate);¹¹C signal proportion,¹ 5 O signal ratio, ...). In this way, a corresponding sample pair is constructed for each simulation scenario.

[0077] Sample pairs constructed from multiple simulated scenarios were collected to form the first training dataset. This dataset covers the proportion of nuclide signals under different tissue structures and beam parameters, providing ample data support for subsequent model training.

[0078] The first training dataset in this embodiment of the invention contains multi-dimensional data, which can more comprehensively describe the characteristics of the simulated scene, so that the model can capture more complex physical relationships and patterns during the learning process, thereby improving the accuracy of model training.

[0079] For example, beam parameters can be extended to include energy spectrum width, scan path, dose rate, and target position variation to enhance the physical consistency of the model. By expanding the beam parameter inputs, the model can consider more factors affecting the physical process when dealing with problems, thereby making the model's predictions closer to the real physical situation and improving the model's physical rationality and accuracy.

[0080] Figure 4 The schematic diagram illustrates the principle of training a first initial model according to an embodiment of the present invention.

[0081] According to an embodiment of the present invention, a first initial model is trained using a first training dataset to obtain a first target model for predicting the signal proportion of each nuclide. The training includes: using sample CT images, PET signals, and beam parameters as input data, and the signal proportion of each nuclide as output, training the first initial model through supervised learning so that the first initial model learns the energy dependence characteristics and spatial decay law of the PET signal, thereby obtaining the first target model for predicting the signal proportion of each nuclide.

[0082] like Figure 4 As shown, in some embodiments, the first initial model can be, for example, a U-Net model, a ResNet model, a 3D convolutional network model, or a temporal attention network model. Supervised learning can be used to train the first target model by taking the PET signals from the first training dataset as input and the corresponding true signal proportions of each nuclide as output. During training, the model parameters are continuously adjusted to minimize the error between the model's predicted signal proportions and the true values.

[0083] For example, sample CT images from each simulated scenario, PET signals within a specified time window (e.g., 2-minute cumulative PET signals), and beam parameters are used as inputs to the model. The first initial network performs forward propagation calculations based on the input data to obtain the predicted proportions of each nuclide signal. The loss function value between the predicted result and the true annotation (i.e., the output proportions of each nuclide signal) is calculated. This loss value is then backpropagated back into the model using a backpropagation algorithm, and the model parameters are updated according to the optimization algorithm. The forward and backpropagation processes are repeated until the model's performance on the validation set no longer improves or reaches a preset number of training epochs.

[0084] The first initial network model trained can be evaluated using test set data. Evaluation metrics (such as mean squared error, mean absolute error, structural similarity index, etc.) can be calculated. The difference between the proportion of nuclide signals predicted by the model and the actual labeling can be used to evaluate the generalization ability of the model. Based on the model evaluation results, the model can be optimized to obtain the first target model.

[0085] For example, the first target model can also be other deep learning models such as multi-branch attention networks, spectral decomposition convolutional networks, or physically constrained generative adversarial networks. Different network structures have their own unique feature extraction and data processing methods, which can analyze and process PET signals from different perspectives, thereby better capturing the differences between different nuclides and improving the model's ability to distinguish between different nuclides.

[0086] The PET signals input to the first initial model can also include PET signals at multiple time points (such as 1 minute, 3 minutes, and 5 minutes). PET signals at different time points can reflect the distribution and decay of radionuclides over time. PET signals at different time points can be collected as model inputs, or signals at different time points can be combined as model inputs.

[0087] By inputting PET signals from multiple time points, the initial model can learn the signal variation patterns of different nuclides at different time points, thereby more accurately capturing the dynamic changes of nuclide composition over time and more precisely estimating the relative contribution of each nuclide. It can also improve the model's adaptability to different scan durations to meet the diverse PET signal acquisition scenarios in actual clinical applications. Even if the actual scan time is not precisely 2 minutes, the trained initial target model can still accurately decompose the nuclide signal ratio based on the learned patterns.

[0088] The PET signal input to the first initial model can also be a signal incorporating Time-of-Flight (ToF) PET information. ToF-PET technology measures the time difference between the arrival of gamma photon pairs at the detector, utilizing the speed of light propagation in a medium to more accurately locate radionuclides within the body, thereby improving the spatial resolution of the image. Conventional PET signals can be combined with ToF-PET information, allowing the input of a PET signal incorporating more precise location information obtained through ToF technology into the first initial model for training.

[0089] ToF-PET technology can more accurately locate the position of radionuclides in the patient's body, so that the first target model can more accurately understand the spatial distribution of different nuclides, thereby better distinguishing the signals of different nuclides, reducing misjudgments and interference, and more accurately decomposing the signal proportion map of each nuclide.

[0090] In this embodiment of the invention, a first initial model is trained using first training data, so that the trained first target model can automatically decompose the relative contributions of different nuclides from the total PET signal, providing basic data at the nuclide level for subsequent dose prediction, which helps to more accurately calculate the contribution of different nuclides to the dose and improve the accuracy of dose prediction.

[0091] Figure 5 The schematic diagram illustrates the principle of training a second initial model according to an embodiment of the present invention.

[0092] According to an embodiment of the present invention, a second initial model is trained using a first training dataset and the signal proportions output by a first target model as a second training dataset to obtain a second target model for predicting three-dimensional dose distribution information. This includes: using the first training dataset and the signal proportions of each nuclide output by the first target model as inputs, and dose distribution information as outputs, training the second initial model through supervised learning, so that the second initial model learns the mapping relationship between signal combination and dose distribution, thereby obtaining a second target model for predicting dose distribution.

[0093] like Figure 5 As shown, in some embodiments, a suitable deep learning architecture can be selected to construct a second initial model for predicting three-dimensional dose distribution information. For example, a 3D convolutional neural network, a U-Net architecture, or a Transformer architecture can be used. Exemplarily, the second initial model can employ a multi-branch input structure, processing input data of different dimensions based on different branches, fusing features extracted from different branches through a feature fusion layer, and performing nonlinear regression through fully connected layers or convolutional layers to output the predicted dose distribution.

[0094] The sample CT images from the first training dataset, the signal proportions of each nuclide output by the first target model, and the corresponding beam parameters are used as inputs to the second initial model. The sample CT images provide tissue structure information, the nuclide signal proportions reflect the distribution of radioactive nuclides, and the beam parameters determine the physical properties of the radiation, thus providing information on the influence of dose distribution from different perspectives. The prepared three-dimensional dose distribution information is then used as the output of the second initial model, clarifying the model's learning objective.

[0095] Supervised learning can be used to train the second initial model. Input data (CT images, radionuclide signal proportions, beam parameters) is fed into the second initial model, which performs forward propagation to obtain the predicted three-dimensional dose distribution information. The loss function value between the predicted and actual dose distribution information is calculated; commonly used loss functions include mean squared error (MSE) and mean absolute error (MAE), which measure the difference between the predicted and actual values. The loss value is backpropagated into the model using a backpropagation algorithm. The model parameters are updated according to optimization algorithms (such as Adam, SGD, etc.), adjusting the model's weights and biases to gradually reduce the loss function value. The forward and backpropagation processes are repeated for multiple iterations until the model's performance on the validation set no longer significantly improves or reaches a preset number of training epochs, resulting in the second target model for predicting the dose distribution.

[0096] In some embodiments, a joint loss function can be used to optimize the second initial model, improving the physical reliability of the prediction by jointly optimizing the dose reconstruction error and the nuclide physical constraint. For example, a joint loss function can be constructed based on the dose reconstruction error loss function and the nuclide physical constraint loss function to incorporate the physical characteristics of the nuclide into the training process of the second initial network, thereby improving the accuracy and rationality of the dose prediction of the second target network.

[0097] This invention trains a second initial model using data from multiple dimensions. These data complement each other, providing the model with more comprehensive and richer information, which helps improve the accuracy of dose prediction. This invention first trains a first initial model to decompose nuclide signals, then uses the nuclide signal distribution as input to the second model, achieving hierarchical data processing. This allows each model to focus on a specific task: the first target network focuses on nuclide signal decomposition, and the second target network focuses on dose prediction, improving the model's learning efficiency and performance while reducing data processing complexity.

[0098] Figure 6 The flowchart illustrates, according to an embodiment of the present invention, the sequential input of a target CT image and a target time window PET signal into a first target model and a second target model to obtain three-dimensional dose distribution information output by the second target model.

[0099] like Figure 6 As shown, in this embodiment, the target CT image and the PET signal of the target time window are sequentially input into the first target model and the second target model to obtain the three-dimensional dose distribution information output by the second target model, including operations S610 to S620.

[0100] In operation S610, the target CT image, the PET signal image of the target time window, and the beam information are input as the first input data to the first target model, so that the first target model can determine the signal ratio of each nuclide within the target time window based on the first input data.

[0101] In operation S620, the signal proportions of each nuclide output by the first target model and the first input data are used as the second input data and input to the second target model so that the second target model can determine the dose distribution information based on the second input data.

[0102] In some embodiments, a CT scanning device can be used to scan the target area to obtain a target CT image. A PET scanning device can be used to scan the target area within a specific time window (e.g., 2 minutes) to acquire radioactive signals within the target area and generate a PET signal image. The acquired target CT image and PET signal image are preprocessed to align them spatially so that they correspond accurately during model processing.

[0103] The preprocessed target CT image and the PET signal image of the target time window are input into the first target model, which decomposes the signal proportion of each nuclide based on the input data. The signal proportions of each nuclide output by the first target model are then input together with the first input data into the second target model, which performs dose prediction to obtain three-dimensional dose distribution information.

[0104] In this embodiment of the invention, the input data is processed in layers using a first target model and a second target model. Based on the signal proportion of each nuclide within the target time window determined by the first target model, more accurate nuclide information is provided to the second target model, enabling the second target model to predict the dose distribution more accurately based on these input data, thereby improving the accuracy of dose distribution prediction.

[0105] Figure 7 A schematic block diagram of a particle dose determination device based on PET signal nuclide decomposition according to an embodiment of the present invention is shown.

[0106] like Figure 7 As shown, the particle dose determination device 700 based on PET signal nuclide decomposition in this embodiment includes a simulation module 710, a first training module 720, a second training module 730, and an acquisition module 740.

[0107] The simulation module 710 is used to simulate multiple sample CT images using Monte Carlo particle transport simulation technology to obtain sample information corresponding to each sample image. The sample information includes the independent signal components of each nuclide, PET signal, nuclide distribution information, and dose distribution information within a specified time window. In one embodiment, the simulation module 710 can be used to perform the operation S110 described above, which will not be repeated here.

[0108] The first training module 720 is used to construct a first training dataset based on multiple sample information, and to train a first initial model using the first training dataset to obtain a first target model for predicting the signal proportion of each nuclide. In one embodiment, the first training module 720 can be used to perform the operation S120 described above, which will not be repeated here.

[0109] The second training module 730 is used to train the second initial model using the first training dataset and the signal proportion output by the first target model as the second training dataset, to obtain a second target model for predicting three-dimensional dose distribution information. In one embodiment, the second training module 730 can be used to perform the operation S130 described above, which will not be repeated here.

[0110] The acquisition module 740 is used to sequentially input the target CT image and the PET signal of the target time window into the first target model and the second target model to obtain the three-dimensional dose distribution information output by the second target model. In one embodiment, the acquisition module 740 can be used to perform the operation S140 described above, which will not be repeated here.

[0111] According to embodiments of the present invention, any plurality of modules among the simulation module 710, the first training module 720, the second training module 730, and the acquisition module 740 may be combined into one module, or any one of these modules may be split into multiple modules. Alternatively, at least a portion of the functionality of one or more of these modules may be combined with at least a portion of the functionality of other modules and implemented in one module. According to embodiments of the present invention, at least one of the simulation module 710, the first training module 720, the second training module 730, and the acquisition module 740 may be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging the circuitry, or implemented in hardware or firmware, or in any one of software, hardware, and firmware implementations, or in a suitable combination of any of these. Alternatively, at least one of the simulation module 710, the first training module 720, the second training module 730, and the acquisition module 740 may be implemented at least partially as a computer program module, which can perform corresponding functions when the computer program module is run.

[0112] Figure 8 A block diagram of an electronic device for determining particle dose in PET signal nuclide decomposition according to an embodiment of the present invention is shown schematically.

[0113] like Figure 8As shown, an electronic device 800 according to an embodiment of the present invention includes a processor 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage portion 808 into a random access memory (RAM) 803. The processor 801 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 801 may also include onboard memory for caching purposes. The processor 801 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0114] RAM 803 stores various programs and data required for the operation of electronic device 800. Processor 801, ROM 802, and RAM 803 are interconnected via bus 804. Processor 801 executes various operations of the method flow according to embodiments of the present invention by executing programs in ROM 802 and / or RAM 803. It should be noted that the programs may also be stored in one or more memories other than ROM 802 and RAM 803. Processor 801 may also execute various operations of the method flow according to embodiments of the present invention by executing programs stored in said one or more memories.

[0115] According to an embodiment of the present invention, the electronic device 800 may further include an input / output (I / O) interface 805, which is also connected to a bus 804. The electronic device 800 may also include one or more of the following components connected to the input / output (I / O) interface 805: an input section 806 including a keyboard, mouse, etc.; an output section 807 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN card, modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 87 is also connected to the input / output (I / O) interface 805 as needed. A removable medium 811, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 87 as needed so that computer programs read from it can be installed into the storage section 808 as needed.

[0116] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of the present invention.

[0117] According to embodiments of the present invention, a computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of the present invention, a computer-readable storage medium may include ROM 802 and / or RAM 803 and / or one or more memories other than ROM 802 and RAM 803 described above.

[0118] Embodiments of the present invention also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to enable the computer system to implement the particle dose determination method based on PET signal nuclide decomposition provided in the embodiments of the present invention.

[0119] When the computer program is executed by the processor 801, it performs the functions defined in the system / apparatus of this invention. According to embodiments of the invention, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0120] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 809, and / or installed from a removable medium 811. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0121] In such an embodiment, the computer program can be downloaded and installed from a network via communication section 809, and / or installed from removable medium 811. When the computer program is executed by processor 801, it performs the functions defined in the system of this embodiment of the invention. According to embodiments of the invention, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0122] According to embodiments of the present invention, program code for executing the computer programs provided in the embodiments of the present invention can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0123] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0124] Those skilled in the art will understand that the features described in the various embodiments of the present invention can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present invention. In particular, the features described in the various embodiments of the present invention can be combined and / or combined in various ways without departing from the spirit and teachings of the present invention. All such combinations and / or combinations fall within the scope of the present invention.

[0125] The embodiments of the present invention have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of the invention. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of the invention, and all such substitutions and modifications should fall within the scope of the invention.

Claims

1. A method for determining particle dose based on PET signal nuclide decomposition, characterized in that, include: The Monte Carlo particle transport simulation technique was used to simulate multiple sample CT images to obtain sample information corresponding to each sample image. The sample information includes the independent signal components of each nuclide, PET signal, nuclide distribution information and dose distribution information under a specified time window. A first training dataset is constructed based on the information from the multiple samples. The first initial model is trained using the first training dataset to obtain a first target model for predicting the signal proportion of each nuclide. The first training dataset and the proportion of signals output by the first target model are used as the second training dataset to train the second initial model, thereby obtaining the second target model for predicting three-dimensional dose distribution information; The target CT image and the PET signal of the target time window are sequentially input into the first target model and the second target model to obtain the three-dimensional dose distribution information output by the second target model.

2. The method according to claim 1, characterized in that, The method involves simulating multiple sample CT images using Monte Carlo particle transport simulation technology to obtain sample information corresponding to each sample image, including: Monte Carlo particle transport simulation technology was used to simulate the transport process of particles in tissues under different energy, energy divergence and broadening conditions. For each simulation scenario, the types and spatial distribution of nuclides generated when particles collide with tissues are calculated based on the nuclear reaction model. For each nuclide, the photons generated by the decay of the nuclide, calculated within a specified time window, form the independent signal component corresponding to that nuclide. The independent signal components of each nuclide within the specified time window are summed to obtain the PET signal of the specified time window in the simulated scenario.

3. The method according to claim 1, characterized in that, The construction of the first training dataset based on the multiple sample information includes: The signal proportion of each nuclide information was calculated based on the independent signal components of each nuclide and the PET signal, respectively. For each simulation scenario, the sample CT images, PET signals, and beam parameters are used as input data, and the signal proportion of each nuclide is used as output data to construct the sample pair corresponding to the simulation scenario. The first training dataset includes multiple sample pairs, and the beam parameters include energy, broadening, incident angle, and dose rate.

4. The method according to claim 1, characterized in that, The step of training the first initial model using the first training dataset to obtain a first target model for predicting the signal proportion of each nuclide includes: Using sample CT images, PET signals, and beam parameters as input data, and the signal proportion of each nuclide as output, the first initial model is trained through supervised learning so that the first initial model learns the energy dependence characteristics and spatial decay law of the PET signal, thereby obtaining the first target model for predicting the signal proportion of each nuclide.

5. The method according to claim 1, characterized in that, The second initial model is trained using the first training dataset and the signal proportion output by the first target model as the second training dataset to obtain a second target model for predicting three-dimensional dose distribution information, including: Using the signal proportions of each nuclide from the first training dataset and the output of the first target model as input, and the dose distribution information as output, the second initial model is trained through supervised learning, so that the second initial model learns the mapping relationship between signal combinations and dose distribution, thereby obtaining the second target model for predicting dose distribution.

6. The method according to claim 1, characterized in that, The step of sequentially inputting the target CT image and the PET signal of the target time window into the first target model and the second target model to obtain the three-dimensional dose distribution information output by the second target model includes: The target CT image, the PET signal image of the target time window, and the beam information are input as the first input data to the first target model, so that the first target model can determine the signal proportion of each nuclide within the target time window based on the first input data. The signal proportions of each nuclide output by the first target model and the first input data are used as the second input data and input into the second target model so that the second target model can determine the dose distribution information based on the second input data.

7. A particle dose determination device based on PET signal nuclide decomposition, characterized in that, include: The simulation module is used to simulate multiple sample CT images using Monte Carlo particle transport simulation technology to obtain sample information corresponding to each sample image. The sample information includes the independent signal components of each nuclide, PET signal, nuclide distribution information and dose distribution information under a specified time window. The first training module is used to construct a first training dataset based on the multiple sample information, and to train a first initial model using the first training dataset to obtain a first target model for predicting the signal proportion of each nuclide. The second training module is used to train the second initial model using the first training dataset and the signal proportion output by the first target model as the second training dataset, so as to obtain the second target model for predicting three-dimensional dose distribution information. The acquisition module is used to sequentially input the target CT image and the PET signal of the target time window into the first target model and the second target model to obtain the three-dimensional dose distribution information output by the second target model.

8. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 6.

Images