Particle source distribution method and system applied to particle implantation therapy
By optimizing the particle source location using a hybrid optimization algorithm and a block-based source placement algorithm, the problems of reliance on doctors' experience and high computational costs in existing technologies are solved, achieving efficient and accurate particle source placement optimization while reducing computation time and resource requirements.
Patent Information
- Application Number
- CN202411433335.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-15
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-10-15
AI Technical Summary
Existing radioactive particle implantation therapy systems rely on doctors' clinical experience, have high computational costs and are time-consuming. Monte Carlo methods are computationally time-consuming and require high-performance computing resources, while simulated annealing methods suffer from local optima.
A hybrid optimization algorithm is adopted, combining the block-based source placement algorithm, particle swarm optimization (PSO) algorithm, and simulated annealing (SA) algorithm. The particle source placement position is optimized using standard dose templates and abstract dose templates. Global search and local optimization are performed through 3D reconstruction and particle swarm optimization.
It achieves high-precision and fast particle source optimization, reduces computational cost and time, avoids local optima traps, and improves computational efficiency and result accuracy.
Smart Images

Figure CN118949298A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of tumor radiotherapy dose calculation and optimization, and in particular to a particle source distribution method and system for particle implantation therapy. Background Art
[0002] Brachytherapy is an internal radiation therapy method widely used in tumor treatment. The basic principle of this method is to use high-energy rays (usually gamma rays or X-rays) released by radioactive isotopes (such as iodine-125, palladium-103, etc.) to destroy the mitosis of tumor cells, thereby achieving a therapeutic effect. Due to the limited radiation range of radioactive particles, the radiation dose is mainly concentrated in the tumor area, which can effectively reduce damage to surrounding healthy tissues compared with external radiation therapy.
[0003] Previous radioactive seed implantation methods relied heavily on the doctor's experience in both the calculation of the particle radiation dose and the location of the particle source. After the advent of the radioactive seed implantation treatment system (TPS system), the system can help doctors calculate the radiation dose of the nuclide, but the location of the particle source is still highly dependent on the doctor's clinical experience.
[0004] The Chinese patent "CN113181563A - Method, system and medium for particle implantation in tumor radiotherapy dose planning" uses the Monte Carlo method for particle radiation dose calculation. It is widely accepted by researchers because of its high accuracy in dose calculation and is called the "gold standard" for dose calculation. However, its disadvantages are also very obvious: first, the Monte Carlo method requires a large amount of random sampling and simulation to achieve the required accuracy, and the calculation time is long; second, in order to speed up the calculation, it is often necessary to use high-performance computing resources, such as multi-core processors or GPUs, which may be difficult to achieve in some clinical environments; third, the Monte Carlo method requires accurate medium parameters to simulate the movement between particles in order to achieve high measurement accuracy, and existing commercial software cannot provide detailed medium preset parameters, so in actual applications, even if the calculation cost is high, the calculation accuracy obtained by doctors is discounted.
[0005] The simulated annealing method used in particle dose optimization has certain flexibility and global search capabilities in particle source optimization, but it has limitations such as long calculation time, difficulty in parameter tuning, and the possibility of falling into local optimality. Summary of the invention
[0006] The present invention aims to overcome the problems of the existing TPS system, which is over-reliant on the clinical experience of doctors, high computational cost and long time consumption, and proposes a high-precision, lightweight and easy-to-use automated particle source placement method for radioactive particle implantation. The doctor only needs to provide the system with the medical image of the patient's target area, and the system can automatically give the optimal position of the particle source based on the volume information of the target area, the information of the surrounding tissues of the target area, etc.
[0007] In order to achieve the above object, the technical solution adopted by the present invention is: A particle source distribution method for particle implantation therapy, comprising the steps of: Obtain medical images of tumors and surrounding tissues; Determining size data of the tumor according to the medical image, and calculating the number of particles required to be implanted based on the size data; The optimal source position of particles is determined by a hybrid optimization algorithm based on the standard dose template determined by the radiation dose of a single particle; The hybrid optimization algorithm is configured to initialize the particle positions based on the block source placement algorithm, then perform a global search using a particle swarm algorithm to obtain a preliminary range of source placement positions, and determine the optimal source placement position in the preliminary range of the solution using a simulated annealing algorithm; The block source distribution algorithm is configured to determine the particle initialization position that meets the dose requirement according to the dose value corresponding to the abstract dose template formed by converting the standard dose template.
[0008] Further, the step of calculating the number of particles required to be implanted includes: Determine a minimum cuboid surrounding the tumor according to the size data of the tumor, wherein the boundary of the cuboid is tangent to the outer surface of the tumor; The number of particles to be implanted is determined according to the size data of the cuboid.
[0009] Furthermore, the number of particles required to be implanted is expressed as: ; Wherein, L, W, H represent the length, width, and height of the cuboid, respectively, in mm; A0 is the activity of the particle source, in mci.
[0010] Furthermore, the radiation dose of a single particle was obtained by a theoretical calculation method based on AAPM Report No. TG-43.
[0011] Furthermore, when there are multiple particle sources in the space, the cumulative dose values of the multiple particle sources at the same observation point are the accumulation of the radiation dose of each particle source to the observation point.
[0012] Furthermore, the step of initializing the particle position based on the block source distribution algorithm includes: The standard dose template is represented as an abstract dose template in the form of a second cuboid embedded in a first cuboid, the second cuboid is represented as a high-dose area, the area in the first cuboid and not belonging to the second cuboid is represented as a low-dose area, and the position where the centers of the two cuboids overlap is represented as the location of the particle source; The particle initialization position that meets the dose requirement is determined according to the dose value corresponding to the abstract dose template.
[0013] Furthermore, the boundary between the high-dose area and the low-dose area is determined based on the normal distribution; according to the normal distribution trajectory, when the percentage of the accumulated value to the total dose of all points on the axis reaches a preset percentage, this point is used as the dividing point between the high-dose area and the low-dose area.
[0014] Furthermore, the dose values of the high-dose area and the low-dose area are determined according to the average dose values in their respective areas.
[0015] Furthermore, the method also includes the steps of: performing three-dimensional reconstruction on the acquired medical image to form a three-dimensional model, and after determining the particle distribution source, superimposing the isodose surface formed by all particle radiations on the three-dimensional model.
[0016] The present invention also provides a particle source distribution system for particle implantation therapy, comprising at least one processor and at least one memory, wherein the memory stores a computer program, and when the program is executed by the processor, the processor is able to execute the above-mentioned particle source distribution method for particle implantation therapy.
[0017] Compared with the prior art, the present invention has at least the following beneficial effects: (1) When designing the particle source distribution algorithm, a method of pre-packaged standard dose template was used to speed up the calculation and reduce the calculation cost; (2) When the particles are initially arranged, the abstract dose template method is used to make the particle arrangement more reasonable and orderly, further reducing the computational cost of the algorithm and improving the algorithm optimization quality; (3) The particle scattering algorithm uses PSO-SA to overcome the shortcomings of PSO and SA, two common algorithms in the field of particle scattering, and designs relevant parameters so that the algorithm has both convergence speed and convergence quality, greatly avoiding the occurrence of pseudo-optimal situations. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0019] Figure 1 is a flow chart of the steps of the particle source distribution method provided by an embodiment of the present invention; Figure 2 is a schematic diagram of the structure of a particle source provided by an embodiment of the present invention; Figure 3 is a schematic diagram of a particle dose calculation model recommended by AAPM provided in an embodiment of the present invention; Figure 4 is a flow chart of the steps of the hybrid optimization algorithm provided by an embodiment of the present invention; Figure 5 is a schematic top view of an abstracted dosage template provided by an embodiment of the present invention; Figure 6 It is a schematic diagram of forming an abstract dose template provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0020] It should be noted that in the present invention, descriptions such as "first", "second", "one", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.
[0021] In addition, the technical solutions between the various embodiments of the present invention can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.
[0022] The following are specific embodiments of the present invention, and the technical solutions of the present invention are further described in conjunction with the accompanying drawings, but the present invention is not limited to these embodiments.
[0023] In this embodiment, if Figure 1 As shown, a particle source distribution method for particle implantation therapy is provided, comprising the steps of: S1. Acquire medical images of the tumor and surrounding tissues, and perform three-dimensional reconstruction on the acquired medical images to form a three-dimensional model; S2. determining the size data of the tumor according to the medical image, and calculating the number of particles required to be implanted based on the size data; S3, using a hybrid optimization algorithm to determine the optimal source position of particles based on a standard dose template determined by a single particle radiation dose; S4. After determining the particle source distribution, the isodose surface formed by all particle radiations is superimposed on the three-dimensional model.
[0024] In the above process, it is necessary to pre-calculate the radiation dose of a single radioactive particle to the surrounding area, and encapsulate the calculation result into a radiation dose unit as a standard dose template. All subsequent calculations only need to be performed around this standard dose template.
[0025] When calculating the radiation dose of a single particle, I 125 -6702 type particle source as an example, Figure 2 As shown, the standard model of its particle source is: I 125 The shell of the particle is a cylinder with a total length of 4.5 mm, a diameter of 0.8 mm, and a thickness of 0.05 mm. 125 The radiation source is attached to the inside of the cylinder and is located on a silver rod with a length of 3 mm and a diameter of 0.5 mm.
[0026] Here, the dose calculation method used by AAPM (American Association of Physicists in Medicine) in Report TG-43 is selected, that is, the theoretical calculation method based on AAPM Report TG-43 to obtain the radiation dose of a single particle, such as Figure 3 The calculation formula is as follows: ; in, is the air kerma intensity, is the dose rate constant. The other variables in the formula are functions related to the spatial position of the particle source and the observation point. is the geometric factor, is the radial dose function, is an anisotropic function.
[0027] The calculation formula is: ; Where r is the distance from the particle source to the observation point, L is the length of the radiation source, θ is the angle between the observation point and the positive direction of the x-axis (the direction of the titanium tube), and β is the angle between the observation point and the two end points of the particle source.
[0028] The multi-order linear fit used by AAPM is the radial dose function as shown below, where: and is the fitting coefficient, r is the radial distance; ; against The AAPM report does not give a specific calculation formula or method, but instead gives the measurement results at different distances and angles in the form of table data. In actual calculations, it is impossible for all coordinates of the particle source and the observation point to be included in the table. In order to accurately calculate the required distance from the particle source at any coordinate to the observation point, Function value, the data in the table needs to be fitted to ensure the universal applicability and accuracy of the model.
[0029] Data fitting is achieved through polynomial fitting method, and the implementation process is as follows: (1) Construct a quadratic polynomial: The independent variables in the function are r and θ, which are the x and y of the polynomial, and the function value is For z in the polynomial: .
[0030] (2) Construct a matrix X: where n is the number of (x, y) value pairs; .
[0031] (3) Solve the polynomial coefficients by the least squares method: ; in, , z is The function value of .
[0032] Since the treatment cycle of particle implantation lasts about five half-lives, the cumulative radiation dose of a single particle source must be calculated over time. The cumulative dose can be derived based on the attenuation dose formula of radioactive elements in nuclear physics.
[0033] Assume that the dose rate at a point in space is , air kerma intensity (k is a constant coefficient, is the activity of an ideal point source, which will produce exactly the same irradiance as the actual sealed source in the air at the same reference point), then: ; According to the law of radioactive decay: ; in is the physical decay constant of the radioactive nuclide corresponding to the particle: ; Therefore, it is deduced that a single particle at point Absorbed dose at for: ; Where t is the time interval from particle implantation to the time when the absorbed dose is calculated.
[0034] If there are multiple particle sources in space, the cumulative dose value of multiple particle sources at the same observation point is the accumulation of the radiation dose of each particle source to the observation point.
[0035] From the above formula, we can see that the radiation dose of the particle is only related to two variables: the relative position of the particle source and the observation point, and time. That is to say, at the same time, as long as the relative position of the particle position and the observation point does not change, the radiation dose received by the observation point will not change.
[0036] Therefore, this embodiment proposes to calculate the radiation dose of a single particle in advance according to the particle model and particle source activity (unit: mci) as a standard dose template, and package and save it. The user can use the preset standard dose template by selecting the corresponding particle source model and activity.
[0037] After obtaining the coordinate information of the tumor region through medical images, it is necessary to calculate the number of particles required for implantation. The steps include: S21, determining a minimum cuboid surrounding the tumor according to the size data of the tumor, wherein the boundary of the cuboid is tangent to the outer surface of the tumor; S22. Determine the number of particles to be implanted according to the size data of the cuboid.
[0038] The number of particles required to be implanted is expressed as: ; Among them, L, W, H represent the length, width, and height of the cuboid respectively, in mm; A0 is the activity of the particle source, in mci.
[0039] Then in step S4, the optimal source position of each particle needs to be determined through a hybrid optimization algorithm (PSO-SA) according to the standard dose template and the number of particles to be implanted.
[0040] In the hybrid optimization algorithm, PSO (particle swarm algorithm) is responsible for global search, quickly exploring the entire search space to find a better solution, and then SA (simulated annealing algorithm) is used for local optimization to further optimize the solution found by PSO, thereby avoiding falling into the local optimum. The process is as follows Figure 4 As shown, the steps include: B1. Initialization parameters; B2. Initialize the population position of particles and randomly initialize the speed of each particle; B3. Calculate the fitness of each particle i, that is, the value of the objective function; B4. Update individual optimal position: If the fitness value of the current particle is better than its historical optimal value, update its historical optimal position; B5. Update the global optimal position: If the fitness value of the current particle is better than the global optimal value, the global optimal position is updated.
[0041] B6. Update speed and position: Each iteration records the position of each (i-th) particle, updates the global optimal position, and the fitness corresponding to the current position and the global optimal fitness.
[0042] If the stopping condition is met (the maximum number of iterations is reached or the particle position no longer changes), stop; otherwise, return to step B3.
[0043] However, the above hybrid optimization algorithm does not solve the problem that the convergence quality of the PSO algorithm itself depends on the initial particle distribution.
[0044] If the initial distribution of the particle source can keep the particle spacing consistent with the scene and the particle arrangement relatively orderly, it can not only greatly reduce the computational burden of the source distribution algorithm, but also improve the convergence quality of the optimization algorithm.
[0045] Therefore, in this embodiment, a method for initializing particle positions based on a block source distribution algorithm is proposed, which includes the following steps: A1. The standard dose template is represented as an abstract dose template in the form of a second cuboid embedded in a first cuboid; A2. Determine the particle initialization position that meets the dose requirement according to the dose value corresponding to the abstract dose template.
[0046] like Figure 5 As shown in the figure, the area inside the second cuboid is represented as a high-dose area, the area inside the first cuboid but not belonging to the second cuboid is represented as a low-dose area, and the center of the cuboid is represented as the location of the particle source. The dose values in the two areas are uniform, that is, the dose in the low-dose area is D min The dose in the high-dose area is D max .
[0047] The high-dose area is defined as a specific area centered on the particle source, where the dose in the area accounts for 95% of the total dose in the entire area. The entire area minus the high-dose area is the low-dose area. When selecting the initial position of the particle, the abstract dose template is placed in the tumor, just like filling an empty house with rectangular boxes, and the center of each box is the location of the particle source.
[0048] In order to find this high-dose area, this embodiment adopts a method of taking the 95% confidence interval under the normal distribution as the high-dose area. The particle radiation dose and distance show an exponential decay trend, and on a certain axis, it will present a symmetrical structure with high in the middle and low on both sides.
[0049] The specific implementation method is: imagine two points moving from the origin to the positive and negative half axes simultaneously along the x-axis, record the coordinate values corresponding to the two points for each movement, and add up these values. When the accumulated value reaches 95% of the total dose of all points on this axis, the specific point on the positive and negative half axes of the x-axis is found, which is called the "endpoint". This endpoint indicates where the high-dose area starts and ends on the x-axis. Using the same method, perform the same operation along the y-axis and z-axis to find the endpoints on the y-axis and z-axis. Figure 6 As shown, Figure 6 The upper left side shows that the particle source gradually decays as the distance from the observation point increases; the right side shows a schematic diagram of the high-dose area formed by the endpoints of the three axes. Then, the average value of the high-dose area and the low-dose area can be taken to determine D max and D min value.
[0050] Therefore, when the particles are initially arranged, the abstract dose template method is used to make the particle arrangement more reasonable and orderly, further reducing the computational cost of the algorithm and improving the quality of algorithm optimization.
[0051] After the particle source is placed, a dose distribution map is drawn on the 3D model to show how different dose levels cover the tumor and surrounding normal tissue. Iso-dose lines in the isodose map are continuous lines that represent areas that receive the same dose, and each line represents a specific dose level. Different dose levels are usually represented by different colors to make them easier to distinguish. High-dose areas may be represented by hot colors (such as red, orange), while low-dose areas may be represented by cold colors (such as blue, green).
[0052] In this embodiment, although the calculation cost of the particle source distribution algorithm itself is increased, the dose calculation process that was previously necessary but could not improve the algorithm quality is reduced. The radiation of the particles to the surroundings is packaged in advance in the system as a standard dose template, which can be used at any time, so there is no need to perform complex calculations, and only an addition calculation that consumes almost no cost is required.
[0053] This embodiment also provides a particle source distribution system for particle implantation therapy, comprising at least one processor and at least one memory, wherein the memory stores a computer program, and when the program is executed by the processor, the processor is able to execute the above-mentioned particle source distribution method for particle implantation therapy.
[0054] The specific embodiments described herein are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in similar ways, but they will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.
Claims
1. A particle source distribution method for particle implantation therapy, characterized in that: Includes steps: Obtain medical images of tumors and surrounding tissues; Determining size data of the tumor according to the medical image, and calculating the number of particles required to be implanted based on the size data; The optimal source position of particles is determined by a hybrid optimization algorithm based on the standard dose template determined by the radiation dose of a single particle; The hybrid optimization algorithm is configured to initialize the particle positions based on the block source distribution algorithm, and then perform a global search using a particle swarm algorithm to obtain a preliminary range of source distribution positions, and determine the optimal source distribution position in the preliminary range using a simulated annealing algorithm; The block source distribution algorithm is configured to determine the particle initialization position that meets the dose requirement according to the dose value corresponding to the abstract dose template formed by converting the standard dose template.
2. A particle source distribution method for particle implantation therapy according to claim 1, characterized in that: The steps to calculate the number of implanted particles required include: Determine a minimum cuboid surrounding the tumor according to the size data of the tumor, wherein the boundary of the cuboid is tangent to the outer surface of the tumor; The number of particles to be implanted is determined according to the size data of the cuboid.
3. A particle source distribution method for particle implantation therapy according to claim 2, characterized in that: The number of implanted particles required is expressed as: ; Wherein, L, W, H represent the length, width, and height of the cuboid, respectively, in mm; A0 is the activity of the particle source, in mci.
4. The particle source distribution method for particle implantation therapy according to claim 1, characterized in that: The single particle radiation dose is obtained by a theoretical calculation method based on AAPM TG-43 report.
5. The particle source distribution method for particle implantation therapy according to claim 4, characterized in that: When there are multiple particle sources in space, the cumulative dose value of multiple particle sources at the same observation point is the accumulation of the radiation dose of each particle source to the observation point.
6. The particle source distribution method for particle implantation therapy according to claim 1, characterized in that: The step of initializing the particle position based on the block source distribution algorithm includes: The standard dose template is represented as an abstract dose template in the form of a second cuboid embedded in a first cuboid, the second cuboid is represented as a high-dose area, the area in the first cuboid and not belonging to the second cuboid is represented as a low-dose area, and the position where the centers of the two cuboids overlap is represented as the location of the particle source; The particle initialization position that meets the dose requirement is determined according to the dose value corresponding to the abstract dose template.
7. The particle source distribution method for particle implantation therapy according to claim 6, characterized in that: The boundary between the high-dose area and the low-dose area is determined based on the normal distribution; according to the normal distribution trajectory, when the percentage of the accumulated value to the total dose of all points on the axis reaches a preset percentage, this point is used as the dividing point between the high-dose area and the low-dose area.
8. The particle source distribution method for particle implantation therapy according to claim 7, characterized in that: The dose values of the high-dose area and the low-dose area are determined according to the average dose values in their respective areas.
9. The particle source distribution method for particle implantation therapy according to claim 1, characterized in that: Also includes the steps: The acquired medical image is three-dimensionally reconstructed to form a three-dimensional model. After the particle distribution source is determined, the isodose surface formed by all particle radiations is superimposed on the three-dimensional model.
10. A particle source system for particle implantation therapy, comprising at least one processor and at least one memory, characterized in that: The memory stores a computer program, and when the program is executed by the processor, the processor is enabled to execute the particle source distribution method for particle implantation therapy as described in any one of claims 1 to 9.
Citation Information
Patent Citations
Radiotherapy dose planning method and system for implanting particles into tumor, and medium
CN113181563A
Adaptive radiation dose calculation simulation method for irregular shaped radioactive source
CN108491583A
Method of tumor radioactive particle layout for laser guided positioning
CN109173084A
Method for randomly filling tumor with radioactive particle source
CN110227213A
Automatic planning method and system based on dose prediction and parameter optimization
CN115620870A
Cited By
Simulation-based path planning method and device, surgical robot and electronic equipment
CN122557975A
An emulation-based path planning method and device, a surgical robot, and an electronic device
CN122557975B
