A Simulation Method for Primary Proton Dose Distribution Detection Based on Probability Density and Operator Splitting
By decomposing the proton scattering probability density function equation into spatial, angular, and energy components and using the operator splitting concept for parallel computation, the accuracy and efficiency problems of dose calculation in proton and heavy ion therapy in existing technologies have been solved, and efficient and accurate dose distribution simulation in anisotropic media has been achieved.
Patent Information
- Application Number
- CN202311550550.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-11-21
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2043-11-21
AI Technical Summary
Existing dose calculation techniques in proton and heavy ion therapy suffer from low accuracy and high computational cost in anisotropic media. Existing methods cannot effectively address the limitations of random errors and analytical algorithms.
By employing a method based on probability density and operator splitting, the probability density function equation of proton scattering in the medium is simplified into a Fokker-Planck operator term. The operator splitting concept is used to decompose it into three parts: space, angle, and energy, which are solved separately. The dose distribution of primary protons is calculated using a numerical discretization method.
It enables efficient and accurate calculation of proton dose distribution in anisotropic materials, reducing computational complexity and cost, and improving computational parallelism and accuracy.
Smart Images

Figure CN117592344B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a technology in the field of medical devices, specifically a method for simulating primary proton dose distribution based on probability density and operator splitting. Background Technology
[0002] Dosage calculation in proton and heavy ion therapy is a crucial part of proton and heavy ion radiotherapy planning, and how to quickly and accurately calculate the dose distribution is an important research topic in radiotherapy planning. Existing techniques for calculating the dose distribution of particles in radiotherapy planning include: uniform intensity beam algorithms, pencil beam algorithms, and Monte Carlo methods. However, these techniques each have the following limitations: uniform intensity beam algorithms have low accuracy in estimating the final dose distribution in complex anisotropic media; pencil beam algorithms use a single Gaussian model to describe the transverse dose distribution of the pencil beam, making them unsuitable for non-uniform media; and Monte Carlo methods require tracking tens of millions of particle trajectories, resulting in significant computational costs. Summary of the Invention
[0003] This invention addresses the limitations of existing technologies in simulating random errors in stochastic algorithms and the limitations of analytical algorithms for anisotropic materials. It proposes a method for simulating primary proton dose distribution based on probability density and operator splitting. The complex terms of elastic and inelastic collisions in the probability density function equation of primary protons are simplified into Fokker-Planck operator terms. Utilizing the idea of operator splitting, the simplified equation is decomposed into three distinct parts for separate solving, facilitating parallel computation and reducing computational load. Ultimately, accurate results of primary proton dose deposition in the medium are obtained.
[0004] This invention is achieved through the following technical solution:
[0005] This invention relates to a method for simulating the primary proton dose distribution based on probability density and operator splitting. By simplifying the physical model of the interaction between charged particles and matter, a deterministic partial differential equation for the probability density distribution function of primary proton scattering is obtained. This deterministic partial differential equation is then split by an operator and decoupled into three parts: space, angle, and energy. Each part is then discretized using a numerical discretization method. Finally, the final dose distribution of the primary protons is calculated based on the obtained density distribution of the primary protons.
[0006] This invention relates to a system for implementing the above method, comprising: an initialization unit, a time operator splitting unit, a discretization unit, and a region integration unit, wherein: the initialization unit performs simplification processing based on the physical processes occurring in the medium to obtain partial differential equations describing the density distribution of primary protons; the time operator splitting unit performs time operator splitting based on the equation information satisfied by the primary protons provided by the initialization unit to obtain new equations that are decoupled from each other in each dimension; the discretization unit performs complete discretization processing on the new equations obtained by the time operator splitting unit to obtain a system of linear equations and performs numerical solution; the region integration unit integrates the primary proton density distribution obtained by numerical solution in the region of interest to obtain the final dose distribution of primary protons.
[0007] Technical effect
[0008] This invention utilizes operator splitting to decouple equation sets for efficient parallel processing. It presents an operator splitting decoupling algorithm based on a deterministic algorithm for primary proton dose calculation in proton and heavy ion radiotherapy, and another based on a deterministic algorithm for primary particle dose calculation of different energies and types (protons and heavy ions). Compared to existing technologies, this invention fully considers the elastic collisions of primary particles with atomic nuclei and the inelastic collisions with extranuclear electrons in the medium. Starting from fundamental principles, through theoretical derivation, it presents a model form more conducive to the design of numerical simulation algorithms for partial differential equations, conforming to real-world conditions. This allows for the decoupling of spatial, angular, and energy variables in the probability density function partial differential equations using the idea of operator splitting, making the equations easier to parallelize in various dimensions. This results in a highly accurate and parallel algorithm. For specific types and energies of particles, the cross-sectional data during particle collisions heavily depends on the elemental composition of the medium material. Therefore, using scientific and reliable data sources in anisotropic materials has a significant impact on dose calculation. Traditional analytical algorithms can only perform calculations on isotropic materials, while the Monte Carlo method can also give correct results for anisotropic materials. However, it is a particle algorithm, which is completely different from the calculation approach based on partial differential equations in this method. This method also has strong accuracy and parallelism for anisotropic materials. Attached Figure Description
[0009] Figure 1 This is a flowchart of the present invention;
[0010] Figure 2 The image shown is a rendering of an example. Detailed Implementation
[0011] like Figure 1 As shown, this embodiment relates to a method for detecting the dose distribution of primary protons, including:
[0012] Step 1) Simplify the physical model of the interaction between charged particles and matter, specifically as follows:
[0013] The physical model of a proton beam irradiating the human body is as follows: ,in: For angle variables, Let E be the spatial variable and E be the energy variable. Let be the probability density distribution function of the proton. This is the scattering cross section of protons undergoing nuclear reactions in the medium; this term is ignored when calculating the dose of primary protons. This represents the scattering cross section during an elastic collision between a proton and an atomic nucleus, where the proton energy loss is negligible. This is the scattering cross section of a proton in an inelastic collision with an outer electron, where the change in the proton angle is negligible.
[0014] First, for the primary proton, the physical process of the proton undergoing nuclear reactions and producing secondary particles is neglected; that is, the first and second terms on the right-hand side of the physical model are ignored. For the integral term of the elastic collision between the proton and the atomic nucleus, a small-angle Fokker-Planck approximation is used to transform the integral term into a differential term. Simultaneously, for the integral term of the inelastic collision between the proton and the outer electron, a continuously decreasing approximation is used. Finally, the linear steady-state Boltzmann equation of the integral equation is transformed into an easily solvable differential equation, specifically:
[0015] ,in: The stopping power refers to the energy loss of a proton per unit transport distance. The transport cross-section represents the probability of angular scattering.
[0016] Step 2) Decouple the deterministic partial differential equations using the idea of operator splitting. Specifically, by observing the simplified equations and treating the x-direction as the time term, the coupled equations are decoupled using the idea of operator splitting, resulting in the following set of equations:
[0017]
[0018] Step 3) Discretize the decoupled system of equations, specifically including:
[0019] 3.1) Transform the angle into the frequency space and use the fast algorithm FFT to solve it to achieve the discretization of space and angle;
[0020] 3.2) Semi-discretization of energy using the finite volume method: By solving this system of linear algebraic equations, the probability density distribution function of the primary proton can be obtained. .
[0021] Step 4) Calculate the dose distribution: based on the probability density distribution function of the primary protons obtained in Step 3. The dose distribution of the corresponding region of interest is obtained by the following integration:
[0022] Where: ROI represents the region of interest, E min With E max These represent the minimum energy required for a primary proton to stop moving and the maximum energy required for an incident proton, respectively.
[0023] A specific experiment was conducted with a 100 MeV proton source placed at the center of the left end of a water tank measuring 10cm × 4cm × 4cm. The results were as follows: Figure 2 The experimental data shown are as follows: the red dashed line represents the reference solution obtained by the existing Monte Carlo method, and the black solid line represents the result obtained by the method in the patent. The two agree very well.
[0024] Compared with existing technologies, this method utilizes the idea of operator splitting to decouple the spatial, angular, and energy variables in the partial differential equation of the probability density function, making the equation easy to parallelize in various dimensions; it also has strong accuracy and parallelism for anisotropic materials.
[0025] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.
Claims
1. A method for simulating the detection of primary proton dose distribution based on probability density and operator splitting, characterized in that, By simplifying the physical model of the interaction between charged particles and matter, a deterministic partial differential equation for the probability density distribution function of primary proton scattering is obtained. This deterministic partial differential equation is split and decoupled into three parts: space, angle, and energy, and then discretized using numerical discretization methods. Finally, the final dose distribution of primary protons is calculated based on the obtained density distribution of primary protons. The physical model described above, that is, the physical model of a proton beam irradiating the human body, is as follows: ,in: For angle variables, Let E be the spatial variable and E be the energy variable. Let be the probability density distribution function of the proton. This is the scattering cross section of protons undergoing nuclear reactions in the medium; this term is ignored when calculating the dose of primary protons. This represents the scattering cross section during an elastic collision between a proton and an atomic nucleus, where the proton energy loss is negligible. This is the scattering cross section of a proton in an inelastic collision with an outer electron, where the change in the proton angle is negligible. The simplification of the physical model refers to the following: First, for the primary proton, the physical process of the proton undergoing nuclear reactions and producing secondary particles is ignored, i.e., the first and second terms on the right-hand side of the physical model are neglected. For the integral term of the elastic collision between the proton and the atomic nucleus, a small-angle Fokker-Planck approximation is used to transform the integral term into a differential term. At the same time, for the integral term of the inelastic collision between the proton and the outer electron, a continuous slowing approximation is used. Finally, the linear steady-state Boltzmann equation of the integral equation is transformed into an easily solvable differential equation, specifically: ,in: The stopping power refers to the energy loss of a proton per unit transport distance. The transport cross-section represents the probability of angular scattering.
2. The primary proton dose distribution detection simulation method based on probability density and operator splitting according to claim 1, characterized in that, The aforementioned operator splitting refers to the decoupling of deterministic partial differential equations using the concept of operator splitting. Specifically, by observing the simplified equations and treating the x-direction as a time term, the coupled equations are decoupled using the concept of operator splitting, resulting in a system of equations: .
3. The primary proton dose distribution detection simulation method based on probability density and operator splitting according to claim 1, characterized in that, The discretization process, namely the discretization of the decoupled system of equations, specifically includes: 3.1) Transform the angle into the frequency space and use the fast algorithm FFT to solve it to achieve spatial and angle discretization; 3.2) Semi-discretization of energy using the finite volume method: By solving this system of linear algebraic equations, the probability density distribution function of the primary proton can be obtained. .
4. The method for simulating primary proton dose distribution detection based on probability density and operator splitting according to claim 1, characterized in that, The calculation of the final dose distribution of primary protons based on the obtained density distribution of primary protons refers to: based on the probability density distribution function of primary protons. The dose distribution of the corresponding region of interest is obtained by the following integration: Where: ROI represents the region of interest, E min With E max These represent the minimum energy at which a primary proton stops moving and the maximum energy of an incident proton, respectively.
5. A primary proton dose distribution detection simulation system based on probability density and operator splitting, implementing the method of any one of claims 1-4, characterized in that, include: The system comprises an initialization unit, a time operator splitting unit, a discretization unit, and a region integration unit. The initialization unit simplifies the physical processes of protons in the medium to obtain partial differential equations describing the primary proton density distribution. The time operator splitting unit splits the time operator based on the equations satisfied by the primary protons provided by the initialization unit, resulting in new equations decoupled across dimensions. The discretization unit fully discretizes the new equations obtained by the time operator splitting unit, obtaining a system of linear equations for numerical solution. The region integration unit integrates the primary proton density distribution obtained from the numerical solution within the region of interest to obtain the final dose distribution of the primary protons.
Citation Information
Patent Citations
Method, device and system for simulating particle transport and calculating body dosage in radiotherapy
CN104750992A
Ion therapy apparatus
WO2009139701A1
