Proton dose information processing device and method and electronic equipment

Abstract

The invention discloses a proton dose information processing device and method and electronic equipment, and relates to the field of medical science and technology. The device comprises a matrix processing unit used for determining a compression matrix according to a physical information table; the particle simulation processing unit is used for simulating the transport behavior of target particles in a medium through a plurality of thread blocks and a plurality of particle stacks, each thread block comprises N threads, and the N threads are used for processing N particles in parallel; the plurality of particle stacks comprise a first type of particle stacks and a second type of particle stacks; and the dose processing unit is used for collecting dose deposition information of the target particles according to the conveying behavior of the target particles in the medium, and updating the compression matrix according to the collected dose deposition information. The technical problems that in the prior art, when a Monte Carlo method is used for proton dose calculation, massive particles need to be simulated to obtain accurate dose distribution, so that the calculation time is too long, and the calculation efficiency is low are solved.

Description

Technical Field

[0001] The present application relates to the field of medical technology, and specifically, to a proton dose information processing device, method, and electronic equipment. Background Art

[0002] In the field of radiotherapy, especially proton therapy, the Monte Carlo method is regarded as the "gold standard" for dose calculation because it can accurately simulate the particle transport process in the patient's body, including complex physical effects such as scattering, energy deposition and secondary particle generation.

[0003] However, traditional Monte Carlo dose calculation usually requires simulating a large number of particle trajectories to obtain a dose distribution with sufficient accuracy. This process is extremely time-consuming, which greatly limits the optimization and iteration speed of the treatment plan.

[0004] Therefore, there is an urgent need for an efficient and accurate Monte Carlo dose calculation method to shorten the calculation time, improve the calculation efficiency, and reduce the demand for video memory, so as to provide a faster and more accurate dose distribution calculation solution for clinical applications.

[0005] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0006] The present application provides a proton dose information processing device, method, and electronic device to at least solve the technical problem in the prior art of using the Monte Carlo method for proton dose calculation, which requires simulating a large number of particles to obtain an accurate dose distribution, resulting in excessively long calculation time and low calculation efficiency.

[0007] According to one aspect of the present application, a proton dose information processing device is provided, comprising: a matrix processing unit for determining a compression matrix based on a physical information table, wherein the physical information table includes at least spot energy information and spot size information, and the compression matrix is a matrix obtained by compressing a dose deposition matrix based on the spot data; a particle simulation processing unit for simulating the transport behavior of target particles in a medium through multiple thread blocks and multiple particle stacks, wherein each thread block includes N threads, and the N threads are used to process N particles in parallel, where N is an integer greater than 1; the multiple particle stacks include a first type of particle stack and a second type of particle stack, the first type of particle stack is used to store original particles released by a radioactive source, and the second type of particle stack is used to store newly generated particles during the activity of the original particles; a dose processing unit for collecting dose deposition information of the target particles according to the transport behavior of the target particles in the medium, and updating the compression matrix based on the collected dose deposition information.

[0008] Optionally, the proton dose information processing device also includes: an acquisition unit, used to determine an energy particle number table based on the physical information table, wherein the energy particle number table is used to count the minimum number of particles required for a single spot to reach a noise distribution threshold at different energies; a calculation unit, used to sum and calculate the minimum number of particles required for each spot in the radiotherapy plan based on the energy particle number table, to obtain a total number of particles corresponding to the radiotherapy plan; a first determination unit, used to determine the number of particles that need to be simulated for each spot based on the ratio of the total number of particles to the minimum number of particles required for each spot; and a second determination unit, used to establish a spot particle number table based on the number of particles that need to be simulated for each spot.

[0009] Optionally, the matrix processing unit includes: a first determining subunit, used to determine the equivalent depth information corresponding to each spot in the radiotherapy plan based on the physical information table; a second determining subunit, used to determine the dose deposition matrix corresponding to each spot based on the equivalent depth information corresponding to each spot; a third determining subunit, used to compress the dose deposition matrix into a compressed matrix based on the starting point coordinates corresponding to the row index and column index of the dose deposition matrix, and establish a matrix position mapping relationship between the compressed matrix and the dose deposition matrix.

[0010] Optionally, the particle simulation processing unit includes: a first processing subunit, used to select the target beam and target spot used when simulating the transport behavior of the target particles; an initialization subunit, used to initialize the attribute information of the target particles according to the parameter information of the target beam, the parameter information of the target spot and the parameter information of the radiotherapy machine, wherein the attribute information includes: particle position, flight speed, released energy and flight direction; a second processing subunit, used to fill the initialized target particles into the first type of particle stack; a third processing subunit, used to simulate the transport behavior of the target particles in the first type of particle stack in different media through multiple thread blocks, and store the new particles generated by the target particles in the first type of particle stack during the activity in the second type of particle stack.

[0011] Optionally, the proton dose information processing device also includes: a first processing unit, which is used to collect the position update information of the target particles when simulating an area where no attribute information is set in the target particle path, and determine that other attribute information of the target particles except the position is not updated if it is detected that multiple thread blocks are simulating the transport behavior of target particles in a uniform medium; a second processing unit, which is used to determine the flight step length of the target particles in the uniform medium according to the position, speed, and direction of the target particles when entering the uniform medium and the transport path of the target particles from the current position to the time when they leave the uniform medium if it is detected that multiple thread blocks are simulating the transport behavior of target particles in a non-uniform medium; and a third processing unit, which is used to determine the flight step length of the target particles in the non-uniform medium according to the position, speed, and direction of the target particles when entering the non-uniform medium and the transport path of the target particles from the current position to the current grid point range when they leave the non-uniform medium if it is detected that multiple thread blocks are simulating the transport behavior of target particles in a non-uniform medium.

[0012] Optionally, the proton dose information processing device also includes: a probability determination unit, which is used to determine the probability distribution information of M physical effects occurring on the target particles based on the property change information of the target particles during the transport behavior after simulating the transport behavior of the target particles in the medium through multiple thread blocks and multiple particle stacks, where M is an integer greater than 1; a physical effect determination unit, which is used to determine the target physical effect from the M physical effects based on the probability distribution information; a particle detection unit, which is used to detect whether new particles are generated based on the target physical effect; and a particle storage unit, which is used to store the generated new particles in the second type of particle stack when detecting the generation of new particles.

[0013] Optionally, the proton dose information processing apparatus further includes: a thread block strategy setting unit configured to set at least one of the following strategies for the plurality of thread blocks during the process of simulating the transport behavior of target particles in a medium using the plurality of thread blocks and the plurality of particle stacks:

[0014] A thread block synchronization strategy is used to input N particles into N threads of a thread block respectively, and after detecting that the N threads of the thread block have completely processed the N particles, determine that the thread block is in a task completion state;

[0015] Thread redistribution strategy, used to assign particles belonging to the same beam or the same spot to the same thread block for processing;

[0016] The thread block update strategy is used to traverse all threads after the number of simulated particle transport reaches a preset number. If it is detected during the traversal process that the particles in the i-th thread have been processed, the unprocessed particles are searched in the particle stack and added to the i-th thread for processing, where the i-th thread is any thread.

[0017] Optionally, the first processing subunit includes: a beam selection module, which is used to use the cumulative value of the beam weight or the total weight value of all beams as a beam selection constraint, and select the target beam used to simulate the transport behavior of the target particle according to the numerical interval of the random number; and a spot selection module, which is used to use the cumulative value of the spot weight or the total weight value of all spots as a spot selection constraint, and select the target spot used to simulate the transport behavior of the target particle according to the numerical interval of the random number.

[0018] According to another aspect of the present application, a proton dose information processing method is also provided, which includes: determining a compression matrix based on a physical information table, wherein the physical information table includes at least spot energy information and spot size information, and the compression matrix is a matrix obtained by compressing the dose deposition matrix based on the spot data; simulating the transport behavior of target particles in a medium through multiple thread blocks and multiple particle stacks, wherein each thread block includes N threads, and the N threads are used to process N particles in parallel, and N is an integer greater than 1; the multiple particle stacks include a first type of particle stack and a second type of particle stack, the first type of particle stack is used to store original particles released by the radioactive source, and the second type of particle stack is used to store newly generated particles during the activity of the original particles; according to the transport behavior of the target particles in the medium, the dose deposition information of the target particles is collected, and the compression matrix is updated based on the collected dose deposition information.

[0019] According to another aspect of an embodiment of the present application, an electronic device is also provided, wherein the electronic device includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors control the operation of the above-mentioned proton dose information processing device.

[0020] As can be seen from the above, this application introduces a compression matrix to determine the effective area of the dose deposition matrix based on the physical information table (including spot energy and spot size information), thereby achieving compression processing of the full-size dose deposition matrix. This compression process can reduce unnecessary calculations and memory accesses, thereby significantly reducing video memory usage and data read time.

[0021] In addition, the present application adopts a parallel computing architecture to decompose the particle simulation into multiple thread blocks, each thread block contains N threads for parallel processing of the transport behavior of N particles. This parallelization strategy can increase the speed of particle simulation and reduce the overall particle simulation calculation time. In addition, the present application also sets up two types of particle stacks, namely the first type of particle stack and the second type of particle stack. The first type of particle stack stores the original particles released by the initial radioactive source, and the second type of particle stack stores the new particles generated by these original particles during the transport process. This classification management method ensures the orderliness and traceability of particles at each stage, avoids chaotic interactions between particles, and thus can improve the orderliness and accuracy of particle simulation calculations.

[0022] Finally, at each stage of particle transport, the dose processing unit collects the dose deposition information of the particles and updates the compression matrix based on the collected dose deposition information.

[0023] In summary, this application significantly improves the efficiency of proton dose calculation and greatly shortens the time required for proton dose calculation by means of matrix compression, parallel processing of particle simulation process, and dividing particle stack into primary particle stack and secondary particle stack, thus overcoming the computational inefficiency of traditional calculation methods when processing massive particles. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0025] Figure 1 is a schematic diagram of an optional proton dose information processing device according to an embodiment of the present application;

[0026] Figure 2 is a flow chart of quality dose information processing according to an embodiment of the present application;

[0027] Figure 3 This is a flowchart of an optional proton dose information processing method according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] Figure 1 is a schematic diagram of an optional proton dose information processing device according to an embodiment of the present application, such as Figure 1 As shown, the proton dose information processing device includes: a matrix processing unit 101 , a particle simulation processing unit 102 and a dose processing unit 103 .

[0031] Optionally, the matrix processing unit 101 is used to determine a compression matrix according to a physical information table, wherein the physical information table includes at least spot energy information and spot size information, and the compression matrix is a matrix obtained by compressing the dose deposition matrix based on the spot data.

[0032] Optionally, in this embodiment of the present application, the matrix processing unit 101 is configured to determine a compression matrix based on a physical information table, thereby achieving effective compression of the dose deposition matrix. The physical information table is a crucial component of the radiotherapy plan and contains at least spot energy information and spot size information. Spot energy information describes the energy of particles (e.g., protons) released from the radiation source, while spot size information reflects the distribution of the particle beam within the target area.

[0033] Optionally, the matrix processing unit 101 can traverse each spot parameter according to the physical information table, calculate the equivalent depth and influence range of the spot, and thus determine the contribution area of the spot to the dose deposition matrix. This process can avoid redundant calculations of non-contributing areas.

[0034] Subsequently, the matrix processing unit 101 may further determine a compression matrix based on the physical information table, wherein the size of the compression matrix is limited to the dose deposition effective area (e.g., the contribution area of the dose deposition matrix determined above). It should be noted that by avoiding processing the entire dose deposition matrix and instead performing subsequent processing on the compression matrix, memory requirements and computation time can be reduced.

[0035] After creating the compressed matrix, the matrix processing unit 101 initializes the video memory to ensure that the size of the compressed matrix matches the dose deposition effective area. In addition, an index mapping table can be established to ensure that each element of the compressed matrix can accurately correspond to the corresponding element in the original dose deposition matrix.

[0036] As can be seen from the above, compared to using a full-size dose deposition matrix for proton dose calculation, using a compressed matrix consumes less memory resources to complete the calculation. This is particularly important for accelerated calculations using GPUs, as it frees up more video memory space for parallel processing of more particles. Furthermore, despite the compression of the matrix, the precise mapping relationship (i.e., the index mapping table described above) ensures that the accuracy of the dose calculation is not affected, thereby ensuring the quality of the radiotherapy plan.

[0037] In an optional embodiment, the particle simulation processing unit 102 is used to simulate the transport behavior of target particles in a medium through multiple thread blocks and multiple particle stacks, wherein each thread block includes N threads, and the N threads are used to process N particles in parallel, where N is an integer greater than 1; the multiple particle stacks include a first-type particle stack and a second-type particle stack, the first-type particle stack is used to store original particles released by the radioactive source, and the second-type particle stack is used to store newly generated particles during the activity of the original particles.

[0038] Optionally, the particle simulation processing unit 102 can fully utilize the parallel computing capability of the GPU to efficiently simulate the transport behavior of particles in a medium (such as human tissue) through multiple thread blocks and particle stacks.

[0039] Optionally, particle simulation processing unit 102 can assign particle simulation tasks to multiple thread blocks, each containing N threads. Particle simulation processing unit 102 allows each thread to independently simulate the entire transport process of a particle, including complex physical effects such as particle energy loss, scattering behavior, and the possible generation of secondary particles. Furthermore, by leveraging the parallel nature of the GPU, multiple thread blocks can simultaneously process different particles, thereby achieving the goal of simulating the transport behavior of a large number of particles in parallel and reducing the time required for particle simulation calculations.

[0040] Optionally, the particle simulation processing unit 102 also employs a dual-type particle stack design, comprising a first-type particle stack and a second-type particle stack. The first-type particle stack (primary particle stack) is used to store primary particles released directly from the radiation source. At the start of the simulation, these primary particles form the basis of the calculation, with their initial position, energy, and direction precisely set. As the simulation progresses, the primary particles interact with other matter during transport, potentially generating new particles. These new particles, referred to as secondary particles, are stored in the second-type particle stack (secondary particle stack). Storing newly generated particles in the secondary particle stack ensures that all new particles generated by physical effects are effectively tracked and simulated, thereby ensuring that the ultimately calculated dose distribution information is complete and accurate.

[0041] When particles in a thread block (provided by the primary particle stack) generate new particles during the transport process, these newly generated particles will be transferred to the secondary particle stack and wait for subsequent simulation calculations.

[0042] It should be noted that the particle simulation processing unit 102 achieves efficient simulation of the particle transport process by employing particle stack classification management and thread block parallel computing. This approach not only enables the processing of large numbers of particles, but also ensures computational accuracy while significantly reducing the total simulation time. This efficient particle simulation strategy is particularly important in high-precision applications such as proton therapy, helping to accelerate dose distribution calculations and provide timely and accurate support for clinical decision-making.

[0043] In an optional embodiment, the dose processing unit 103 is configured to collect dose deposition information of the target particles according to the transport behavior of the target particles in the medium, and update the compression matrix according to the collected dose deposition information.

[0044] Optionally, the dose processing unit 103 can continuously monitor the behavior of target particles (protons or other radioactive particles) as they are transported through the medium. For example, if a collision or energy deposition between the target particles and the medium is detected, the dose processing unit 103 will collect relevant dose information, including but not limited to the amount of energy released by the particles, the particle's most recent position, and the particle's interaction with surrounding matter.

[0045] It should be noted that the dose deposition information collected based on particle transport behavior is not directly applied to the original dose deposition matrix. Instead, it is mapped to the compressed matrix previously generated by the matrix processing unit 101 based on the location of particle activity. Based on the collected dose deposition information, the dose processing unit 103 updates the corresponding elements of the compressed matrix to reflect the actual state of particle energy deposition. Because the compressed matrix only retains data from the effective dose deposition region, the dose processing unit 103 can avoid accessing a large number of invalid regions when updating the compressed matrix. This not only saves computation time but also significantly reduces memory consumption. This optimization is particularly important when processing high-precision, large-scale dose distributions.

[0046] In an optional embodiment, the proton dose information processing device further includes: an acquisition unit, configured to determine an energy particle number table based on the physical information table, wherein the energy particle number table is used to count the minimum number of particles required for a single light spot to reach a noise distribution threshold at different energies; a calculation unit, configured to sum and calculate the minimum number of particles required for each light spot in the radiotherapy plan based on the energy particle number table to obtain a total number of particles corresponding to the radiotherapy plan; a first determination unit, configured to determine the number of particles required to simulate each light spot based on a ratio of the total number of particles to the minimum number of particles required for each light spot; and a second determination unit, configured to establish a spot particle number table based on the number of particles required to simulate each light spot.

[0047] Optionally, Figure 2 is a flow chart of quality dose information processing according to an embodiment of the present application, such as Figure 2 As shown in Figure 1, the data preparation phase involves first constructing an energy particle count table based on the physical information table. This table records the minimum number of particles required for a single spot to achieve the preset dose calculation uncertainty requirements at different energy levels. Then, by traversing all the spots in the entire radiotherapy plan, the energy particle count table constructed in the previous step is used to determine the number of particles required to simulate each spot. This process involves accumulating the minimum particle counts for each spot and setting a specific number of simulated particles for each spot based on the ratio of the total target particle count to the accumulated minimum particle count, thereby establishing the spot particle count table.

[0048] Optionally, the matrix processing unit 101 also includes: a first determining subunit, used to determine the equivalent depth information corresponding to each spot in the radiotherapy plan according to the physical information table; a second determining subunit, used to determine the dose deposition matrix corresponding to each spot based on the equivalent depth information corresponding to each spot; a third determining subunit, used to compress the dose deposition matrix into a compressed matrix based on the starting point coordinates corresponding to the row index and column index of the dose deposition matrix, and establish a matrix position mapping relationship between the compressed matrix and the dose deposition matrix.

[0049] Alternatively, as Figure 2 As shown in the figure, based on the physical information table, we can conduct in-depth analysis of each spot parameter in the radiotherapy plan. For example, by calculating the equivalent depth corresponding to each spot, we can determine the effective dose deposition area corresponding to each spot. The effective dose deposition area is the focus of dose calculation. By calibrating the effective dose deposition area, we can avoid redundant calculations in ineffective areas, thereby improving dose calculation efficiency.

[0050] Based on the calibrated effective dose deposition area, a compressed matrix can be constructed. This compression matrix is generated by determining the starting point coordinates of the row and column indices, then retaining only the dose information within the effective dose deposition area, thereby reducing the size of the dose deposition matrix (that is, based on the starting point coordinates corresponding to the row and column indices of the dose deposition matrix, the full-size dose deposition matrix corresponding to each spot is compressed into a compressed matrix).

[0051] In an optional embodiment, the particle simulation processing unit 102 includes: a first processing subunit, used to select the target beam and target spot used when simulating the transport behavior of the target particles; an initialization subunit, used to initialize the attribute information of the target particles according to the parameter information of the target beam, the parameter information of the target spot and the parameter information of the radiotherapy machine, wherein the attribute information includes: particle position, flight speed, released energy and flight direction; a second processing subunit, used to fill the initialized target particles into the first type of particle stack; a third processing subunit, used to simulate the transport behavior of the target particles in the first type of particle stack in different media through multiple thread blocks, and store the new particles generated by the target particles in the first type of particle stack during the activity process in the second type of particle stack.

[0052] Optionally, before the particle simulation starts, the first processing subunit is responsible for selecting a specific target beam and target spot from the radiotherapy plan.

[0053] Alternatively, as Figure 2 As shown, the initialization subunit can initialize the target particle's attribute information based on the selected target beam parameters, target spot parameters, and specific parameters of the radiotherapy machine (such as beam energy, divergence angle, machine structure, etc.). This attribute information includes but is not limited to: particle position, flight speed, released energy, and flight direction.

[0054] The initialized target particles are then added to the first particle stack (the main particle stack) by the second processing subunit, preparing to begin simulating the target particle's transport behavior. The first particle stack is the starting point for particle simulation and stores all the original particles to be simulated.

[0055] The third processing subunit uses multiple thread blocks to simulate the transport behavior of target particles in the first particle stack in parallel through different media. This process involves particle energy loss, direction changes, and the possible generation of secondary particles (i.e., new particles). All new particles resulting from physical effects are stored in the second particle stack (secondary particle stack) for subsequent simulations.

[0056] In an optional embodiment, the first processing subunit includes: a beam selection module, which is used to use the cumulative value of the beam weight or the total weight value of all beams as a beam selection constraint condition, and select the target beam used when simulating the transport behavior of the target particle according to the numerical interval of the random number; a spot selection module, which is used to use the cumulative value of the spot weight or the total weight value of all spots as a spot selection constraint condition, and select the target spot used when simulating the transport behavior of the target particle according to the numerical interval of the random number.

[0057] Alternatively, as Figure 2 As shown, this application introduces a weight-based random selection mechanism to select the target beam and target spot for the current simulation, including the following steps:

[0058] In step 1, at the start of the particle simulation, a random number between 0 and 1 is generated. This random number is then used to select a beam from multiple beams as the target beam for the current simulation. The position of this random number is determined based on the beam's weight ratio within the treatment plan. Specifically, the weights of each beam are accumulated and then compared with the total weight of the plan, forming a series of weight ratio intervals. The interval in which the random number falls determines the selection of the corresponding beam as the target beam for simulation. Through this mechanism, the probability of beam selection is proportional to its importance (i.e., weight) in the treatment plan, ensuring that the final particle simulation results are highly representative.

[0059] In step 2, after the target beam is determined, a second random number between 0 and 1 is generated, and a spot in the target beam is selected as the target spot using this random number. Similar to the selection logic of the target beam, the selection of the target spot is also based on the weight ratio. For example, the weight of each spot can be accumulated and compared with the total weight of all spots in the target beam to form a weight ratio interval corresponding to each spot. Based on the weight ratio interval in which the random number falls, the spot corresponding to the weight ratio interval is selected as the target spot. This process ensures the randomness and fairness of the target spot selection, while ensuring the granularity of the simulation, which can be refined to the level of each spot, further enhancing the accuracy of the dose calculation.

[0060] In step three, after the target spot is selected, the system can count the number of particles in the spot. The simulated number of particles in the target spot is compared with the spot particle count table, which is pre-set based on the uncertainty requirements. If the simulated particle count reaches or exceeds the minimum number of particles specified in the spot particle count table, the simulation task for the target spot is considered complete, and the system will skip the subsequent particle simulation steps and return directly to step one to select the next beam or spot for simulation. Conversely, if the simulated particle count of the spot does not reach the minimum number of particles specified for that spot, the system will continue to perform particle simulation as planned until the minimum number of particles specified for that spot is met.

[0061] In an optional embodiment, the proton dose information processing device further includes: a first processing unit for, if it is detected that multiple thread blocks are simulating the transport behavior of target particles in an air medium, collecting the position update information of the target particles when simulating an area where no attribute information is set for the target particle path, and determining that other attribute information of the target particles except the position is not updated; a second processing unit for, if it is detected that multiple thread blocks are simulating the transport behavior of target particles in a uniform medium, determining the flight step length of the target particles in the uniform medium according to the position, speed, and direction of the target particles when entering the uniform medium and the transport path of the target particles from the current position to the time when they leave the uniform medium; and a third processing unit for, if it is detected that multiple thread blocks are simulating the transport behavior of target particles in an inhomogeneous medium, determining the flight step length of the target particles in the inhomogeneous medium according to the position, speed, and direction of the target particles when entering the inhomogeneous medium and the transport path of the target particles from the current position to the current grid point range when they leave the inhomogeneous medium.

[0062] Optionally, the present application proposes a refined particle property processing method for particle transport simulation in different medium environments, ensuring the physical accuracy and computational efficiency of the simulation process.

[0063] Alternatively, as Figure 2 As shown in the figure, regarding the simulation of particle transport in air, including the process of transporting particles from the light source to the upper surface of the range shifter and from the range shifter to the outer surface of the CT, since the area through which the particles pass is an air medium, and the influence of air on the particles is mainly reflected in resistance, while the resistance effect is negligible on the path of high-energy protons. Therefore, in the air medium, the system assumes that the transport of particles is only manifested as a movement of position, while their other physical properties such as energy, velocity, and direction remain unchanged. This simplifies the calculation process, avoids the simulation of unnecessary complex physical effects, and thus improves the efficiency of the simulation calculation.

[0064] Alternatively, as Figure 2As shown, regarding the simulation of the particle transport process in a uniform medium, the process includes: when the particle enters a uniform medium (such as water or a solid of a specific density), its transport process will be affected by the properties of the medium, such as scattering and energy loss. In order to accurately simulate these physical processes, this application adopts a step selection strategy based on the current position and velocity direction of the particle. Specifically, the system will use the position of the particle when it enters the uniform medium as the starting point, the velocity direction of the particle as a reference, and the path from the current position to the edge of the medium before leaving as the step length of the particle transport. This strategy ensures that the energy deposition and physical effects of the particle in the uniform medium can be accurately simulated. At the same time, by reasonably selecting the step length, the problem of increased computational burden caused by overly fine step division is avoided.

[0065] Alternatively, as Figure 2 As shown, regarding the simulation of particle transport in inhomogeneous media, including: when particles pass through inhomogeneous media (such as organisms with different tissue densities), the difference in medium properties will significantly affect the transport behavior of the particles, including scattering angle, energy loss rate, etc. Therefore, the present application adopts a more dynamic and sophisticated step selection mechanism for the simulation of particle transport in inhomogeneous media. The system calculates the transport path of the particles within the current medium grid as the step size based on the real-time position and velocity direction of the particles. In this way, the simulation of each step takes into account the influence of local medium properties, achieving the purpose of accurately simulating particle transport behavior in complex medium structures. At the same time, by dynamically adjusting the step size, the efficiency of computing resource utilization is improved, avoiding the problem of over-computation in the inhomogeneous medium area.

[0066] Optionally, the proton dose information processing device also includes: a probability determination unit, which is used to determine the probability distribution information of M physical effects occurring on the target particles based on the property change information of the target particles during the transport behavior after simulating the transport behavior of the target particles in the medium through multiple thread blocks and multiple particle stacks, where M is an integer greater than 1; a physical effect determination unit, which is used to determine the target physical effect from the M physical effects based on the probability distribution information; a particle detection unit, which is used to detect whether new particles are generated based on the target physical effect; and a particle storage unit, which is used to store the generated new particles in the second type of particle stack when detecting the generation of new particles.

[0067] Alternatively, as Figure 2As shown, in the Monte Carlo dose calculation method, the interaction between particles and the medium is one of the key factors that determine the dose distribution. In order to accurately reflect this complex physical process, this application has designed a detailed particle effect evaluation and status update mechanism, including: by analyzing the energy, velocity and other key properties of the particles, combined with the physical properties of the medium they pass through (such as electron density, atomic number, etc.), calculating the probability of different physical effects. These physical effects include but are not limited to scattering, energy deposition, secondary particle generation and other situations. Based on the probability distribution information obtained from the evaluation, the system can determine the specific physical effects of the particles within the current step size, ensuring that the simulation process can truly reflect the physical behavior of the particles in the medium.

[0068] Once the target physical effect of a particle is determined, the system can dynamically update the particle's state information based on the specific properties of the target physical effect. For example, if it is a scattering effect, the particle's direction may be randomly deflected; if it is an energy deposition process, the particle's energy will be reduced accordingly; if a new particle is generated, in addition to modifying the state of the current particle, the new particle must also be initialized. This update process strictly adheres to the laws of physics to ensure that each simulation accurately reflects the real-world physical scenario of particle transport. At the same time, the system also checks for new particle generation events so that they can be promptly incorporated into subsequent simulation activities.

[0069] In an optional embodiment, the proton dose information processing device further includes: a thread block strategy setting unit, configured to set at least one of the following strategies for multiple thread blocks in the process of simulating the transport behavior of target particles in a medium through multiple thread blocks and multiple particle stacks: a thread block synchronization strategy, configured to input N particles into N threads of a thread block respectively, and after detecting that the N threads of the thread block have completely processed the N particles, determine that the thread block is in a task completion state; a thread redistribution strategy, configured to assign particles belonging to the same beam or the same spot to the same thread block for processing; a thread block update strategy, configured to traverse all threads after the number of simulated particle transport reaches a preset number, and if it is detected during the traversal process that the particles in the i-th thread have been processed, search for unprocessed particles in the particle stack and add them to the i-th thread for processing, wherein the i-th thread is any thread.

[0070] Alternatively, as Figure 2 As shown, particles in the particle stack are first assigned to threads according to specific rules, ensuring that each thread simulates a single particle. For example, every 32 threads (the specific number depends on the GPU design) can be organized into a thread block to leverage the GPU's parallel computing capabilities. This particle allocation strategy ensures efficient use of computing resources and facilitates subsequent thread synchronization and reallocation.

[0071] During particle simulation, threads within each thread block execute synchronously. This means that all threads in a thread block must complete their assigned particle simulation tasks before they can collectively proceed to the next execution phase. Any thread that completes its task early will wait for the other threads, ensuring that the thread block as a whole completes its task. This synchronization mechanism avoids data inconsistencies caused by varying thread execution speeds.

[0072] Furthermore, to further optimize the parallel efficiency of particle simulations, the system reevaluates particle allocation after reaching a certain stage of simulation (for example, after a preset number of simulated particle transports). For example, the system can reallocate particles belonging to the same beam or spot to the same thread block, reducing inter-thread data exchange and synchronization overhead and improving parallel computing efficiency. Through this thread reallocation process, the system dynamically adjusts the correspondence between particles and threads, ensuring the optimal allocation of computing resources.

[0073] Furthermore, the system can periodically check the particle simulation status of thread blocks. When it detects that particles within a thread have completed their transport behavior simulation and met the predetermined exit criteria, the system does not immediately terminate the thread. Instead, it marks its status as "pending update." The system then traverses the particle stack, searching for new particles to fill these completed threads and continue particle simulation. This allows each thread block to remain active until all particles have completed simulation, improving GPU resource utilization while ensuring the consistency and integrity of the simulation process.

[0074] In an optional embodiment, the dose information processing device further includes: a processing priority setting unit for setting the particle processing priority in the first type of particle stack to be higher than the particle processing priority in the second type of particle stack in the process of simulating the transport behavior of particles in a medium through multiple thread blocks and multiple particle stacks.

[0075] Alternatively, as Figure 2 As shown, two particle stacks are first created: a primary particle stack (corresponding to the first type of particle stack described above) and a secondary particle stack (corresponding to the second type of particle stack described above). The primary particle stack is used to store the original particles currently participating in the dose calculation and transport process, which originate from the beam. The secondary particle stack is specifically used to store new particles generated during the transport process due to physical effects (such as scattering and energy deposition).

[0076] Alternatively, as Figure 2As shown in the figure, when a thread completes the transport simulation of its assigned particles, the system checks the status of the secondary particle stack. If there are newly generated particles in the secondary particle stack, these particles are moved to the main particle stack so that they can be extracted by the new thread and the simulation continues. If there are no new particles in the secondary particle stack, the system automatically generates a certain number of particles and replenishes them in the main particle stack to ensure the continuity of the simulation. This method ensures the dynamic update of the main particle stack and avoids the technical problem of simulation termination due to exhaustion of particle resources.

[0077] During system initialization, each thread extracts a particle from the primary particle stack for simulation. Once a thread completes its particle transport simulation, the system automatically assigns a new particle from the primary particle stack to that thread. This process repeats until all particles in the primary particle stack have been simulated. If the primary particle stack runs out of particles, the system calls newly generated particles from the secondary particle stack and replenishes them to the primary particle stack, thus maintaining simulation continuity.

[0078] In addition, whenever new particles are generated, they are immediately added to the secondary particle stack. The update of the secondary particle stack is synchronized with the particle generation, ensuring the timely storage of new particles.

[0079] It should be noted that, in actual application, it is also possible not to set priorities for the first type of particle stack and the second type of particle stack, and this application does not impose any special limitation on the setting of priorities.

[0080] According to another aspect of the embodiment of the present application, a proton dose information processing method is also provided, wherein: Figure 3 is a flow chart of an optional proton dose information processing method according to an embodiment of the present application, such as Figure 3 As shown, the following steps are included:

[0081] Step S301 : determining a compression matrix according to a physical information table, wherein the physical information table includes at least spot energy information and spot size information, and the compression matrix is a matrix obtained by compressing a dose deposition matrix based on the spot data.

[0082] Optionally, each spot parameter can be traversed first according to the physical information table to calculate the equivalent depth and influence range of the spot, thereby determining the contribution area of the spot to the dose deposition matrix. This process can avoid redundant calculations of non-contributing areas. Subsequently, a compression matrix can be determined according to the physical information table, the size of which is limited to the effective area of dose deposition (for example, the contribution area of the dose deposition matrix determined above). The index and value of the compression matrix are mapped to the corresponding positions of the original dose deposition matrix. It should be noted that by avoiding processing the entire dose deposition matrix and performing subsequent processing on the compressed matrix, memory requirements and calculation time can be reduced.

[0083] As can be seen from the above, compared to using a full-size dose deposition matrix for proton dose calculation, using a compressed matrix consumes less memory resources to complete the calculation. This is particularly important for accelerated calculations using GPUs, as it frees up more video memory space for parallel processing of more particles. Furthermore, despite the compression of the matrix, the precise mapping relationship (i.e., the index mapping table described above) ensures that the accuracy of the dose calculation is not affected, thereby ensuring the quality of the radiotherapy plan.

[0084] Step S302 : simulating the transport behavior of target particles in a medium by using multiple thread blocks and multiple particle stacks.

[0085] Each thread block includes N threads, and N threads are used to process N particles in parallel, where N is an integer greater than 1. The multiple particle stacks include a first-type particle stack and a second-type particle stack. The first-type particle stack is used to store the original particles released by the radioactive source, and the second-type particle stack is used to store the newly generated particles during the activity of the original particles.

[0086] Optionally, according to the solution of this application, particle simulation tasks can be assigned to multiple thread blocks, each containing N threads. The particle simulation processing unit allows each thread to independently simulate the entire transport process of a particle, including complex physical effects such as particle energy loss, scattering behavior, and the possible generation of secondary particles. Furthermore, by leveraging the parallel nature of the GPU, multiple thread blocks can simultaneously process different particles, thereby achieving the goal of simulating the transport behavior of a large number of particles in parallel and thereby shortening the time required for particle simulation calculations.

[0087] Optionally, the present application also adopts a dual-type particle stack design, including: a first-type particle stack and a second-type particle stack. The first-type particle stack (primary particle stack) is used to store the original particles released directly by the radioactive source. At the beginning of the simulation, these original particles form the basis of the calculation, and their initial position, energy, and direction are precisely set. As the simulation progresses, the original particles interact with other substances during transport, potentially generating new particles. These new particles can be called secondary particles, and the secondary particles are stored in the second-type particle stack (secondary particle stack). Storing the newly generated particles in the secondary particle stack ensures that all new particles generated by physical effects can be effectively tracked and simulated, thereby ensuring that the dose distribution information finally calculated is complete and accurate.

[0088] When particles in a thread block (provided by the primary particle stack) generate new particles during the transport process, these newly generated particles will be transferred to the secondary particle stack and wait for subsequent simulation calculations.

[0089] It is important to note that efficient simulation of the particle transport process is achieved by employing particle stack classification management and thread block parallel computing. This approach not only handles large numbers of particles while ensuring computational accuracy, but also significantly reduces the total simulation time required. This efficient particle simulation strategy is particularly important in high-precision applications such as proton therapy, helping to accelerate dose distribution calculations and provide timely and accurate support for clinical decision-making.

[0090] Step S303 : collecting dose deposition information of the target particles according to the transport behavior of the target particles in the medium, and updating the compression matrix according to the collected dose deposition information.

[0091] Optionally, the dose processing unit can continuously monitor the behavior of target particles (protons or other radioactive particles) as they are transported through the medium. For example, if a collision or energy deposition between the target particle and the medium is detected, the dose processing unit will collect relevant dose information, including but not limited to the amount of energy released by the particle, the particle's most recent position, and the particle's interaction with the surrounding material.

[0092] It should be noted that the dose deposition information collected based on particle transport behavior is not directly applied to the original dose deposition matrix. Instead, it is mapped to a compression matrix previously generated by the matrix processing unit based on the location of particle activity. The dose processing unit updates the corresponding elements of the compression matrix based on the collected dose deposition information to reflect the actual state of particle energy deposition. Because the compression matrix only retains data in the effective dose deposition area, the dose processing unit can avoid accessing a large number of invalid areas when updating the compression matrix. This not only saves computation time but also significantly reduces memory consumption. This optimization becomes particularly important when processing high-precision, large-scale dose distributions.

[0093] According to another aspect of an embodiment of the present application, an electronic device is also provided, wherein the electronic device includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors control the operation of the above-mentioned proton dose information processing device.

[0094] The serial numbers of the above embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0095] In the above embodiments of the present application, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0096] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of the units can be a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0097] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0098] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0099] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the existing technology, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, etc. Various media that can store program code.

[0100] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A proton dose information processing device, characterized in that: include: a matrix processing unit, configured to determine a compression matrix according to a physical information table, wherein the physical information table includes at least spot energy information and spot size information, and the compression matrix is a matrix obtained by compressing a dose deposition matrix based on the spot data; a particle simulation processing unit configured to simulate the transport behavior of target particles in a medium using multiple thread blocks and multiple particle stacks, wherein each thread block includes N threads, and the N threads are configured to process N particles in parallel, where N is an integer greater than 1; and the multiple particle stacks include a first type of particle stack and a second type of particle stack, wherein the first type of particle stack is configured to store original particles released by a radioactive source, and the second type of particle stack is configured to store newly generated particles during the activity of the original particles. The dose processing unit is configured to collect dose deposition information of the target particles according to the transport behavior of the target particles in the medium, and update the compression matrix according to the collected dose deposition information.

2. The proton dose information processing device according to claim 1, characterized in that: The proton dose information processing device further includes: an acquisition unit, configured to determine an energy particle number table according to the physical information table, wherein the energy particle number table is used to count the minimum number of particles required for a single light spot to reach a noise distribution threshold at different energies; a calculation unit, configured to sum and calculate the minimum number of particles required for each spot in the radiotherapy plan according to the energy particle number table, to obtain a total number of particles corresponding to the radiotherapy plan; a first determining unit, configured to determine the number of particles required to be simulated for each light spot according to a ratio of the total number of particles to the minimum number of particles required for each light spot; The second determining unit is configured to establish a light spot particle quantity table according to the quantity of particles required to be simulated for each light spot.

3. The proton dose information processing device according to claim 1, characterized in that: Matrix processing unit, including: A first determining subunit is configured to determine, according to the physical information table, equivalent depth information corresponding to each spot in the radiotherapy plan; A second determining subunit is configured to determine a dose deposition matrix corresponding to each light spot according to equivalent depth information corresponding to each light spot; The third determining subunit is used to compress the dose deposition matrix into the compressed matrix based on the starting point coordinates corresponding to the row index and column index of the dose deposition matrix, and establish a matrix position mapping relationship between the compressed matrix and the dose deposition matrix.

4. The proton dose information processing device according to claim 1, characterized in that: Particle simulation processing unit, including: A first processing subunit is used to select a target beam and a target spot used to simulate the transport behavior of the target particles; an initialization subunit, configured to initialize attribute information of the target particles according to parameter information of the target beam, parameter information of the target spot, and parameter information of the radiotherapy machine, wherein the attribute information includes: particle position, flight speed, released energy, and flight direction; a second processing subunit, configured to fill the initialized target particles into the first type particle stack; The third processing subunit is configured to simulate the transport behavior of the target particles in the first type of particle stack in different media through the multiple thread blocks, and store new particles generated during the activity of the target particles in the first type of particle stack in the second type of particle stack.

5. The proton dose information processing device according to claim 4, characterized in that: The proton dose information processing device further includes: a first processing unit configured to, if it is detected that the plurality of thread blocks are simulating the transport behavior of the target particle in the air medium, collect position update information of the target particle when simulating an area where no attribute information is set for the target particle path, and determine that attribute information other than the position of the target particle is not updated; a second processing unit, configured to, if detecting that the plurality of thread blocks are simulating the transport behavior of the target particle in the homogeneous medium, determine a flight step length of the target particle in the homogeneous medium based on a position, speed, and direction of the target particle when entering the homogeneous medium and a transport path of the target particle from a current position to when leaving the homogeneous medium; a third processing unit, configured to, if detecting that the plurality of thread blocks are simulating the transport behavior of the target particle in the inhomogeneous medium, determine a flight step length of the target particle in the inhomogeneous medium based on the position, velocity, and direction of the target particle when entering the inhomogeneous medium and the transport path of the target particle from the current position to the current grid point range when leaving the inhomogeneous medium.

6. The proton dose information processing device according to claim 1, characterized in that: The proton dose information processing device further includes: a probability determination unit, configured to, after simulating the transport behavior of a target particle in a medium using multiple thread blocks and multiple particle stacks, determine probability distribution information of M physical effects occurring on the target particle based on property change information of the target particle during the transport behavior, where M is an integer greater than 1; a physical effect determination unit, configured to determine a target physical effect from the M physical effects according to the probability distribution information; a particle detection unit, configured to detect whether new particles are generated according to the target physical effect; The particle storage unit is configured to store the generated new particle into the second type particle stack when detecting the generation of the new particle.

7. The proton dose information processing device according to claim 1, characterized in that: The proton dose information processing device further includes: The thread block strategy setting unit is configured to set at least one of the following strategies for the plurality of thread blocks during a process of simulating the transport behavior of target particles in a medium using the plurality of thread blocks and the plurality of particle stacks: A thread block synchronization strategy is used to input N particles into N threads of a thread block respectively, and after detecting that the N threads of the thread block have completely processed the N particles, determine that the thread block is in a task completion state; Thread redistribution strategy, used to assign particles belonging to the same beam or the same spot to the same thread block for processing; A thread block update strategy is used to traverse all threads after the number of simulated particle transport reaches a preset number. If it is detected during the traversal process that the particles in the i-th thread have been processed, the unprocessed particles are searched in the particle stack and added to the i-th thread for processing, where the i-th thread is any thread.

8. The proton dose information processing device according to claim 4, characterized in that: The first processing subunit includes: A beam selection module is used to use the accumulated value of the beam weight or the total weight value of all beams as a beam selection constraint condition, and select a target beam used to simulate the transport behavior of the target particle according to the numerical interval of the random number; The light spot selection module is used to use the cumulative value of the light spot weight or the total weight value of all light spots as the light spot selection constraint condition, and select the target light spot used to simulate the transport behavior of the target particle according to the numerical interval of the random number.

9. A proton dose information processing method, characterized in that: include: Determining a compression matrix according to a physical information table, wherein the physical information table includes at least spot energy information and spot size information, and the compression matrix is a matrix obtained by compressing a dose deposition matrix based on the spot data; The transport behavior of target particles in a medium is simulated using multiple thread blocks and multiple particle stacks, wherein each thread block includes N threads, and the N threads are used to process N particles in parallel, where N is an integer greater than 1; the multiple particle stacks include a first type of particle stack and a second type of particle stack, wherein the first type of particle stack is used to store original particles released by a radioactive source, and the second type of particle stack is used to store newly generated particles during the activity of the original particles; According to the transport behavior of the target particles in the medium, the dose deposition information of the target particles is collected, and the compression matrix is updated according to the collected dose deposition information.

10. An electronic device, characterized in that: The device comprises one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors control the operation of the proton dose information processing device according to any one of claims 1 to 8.

Images