Proton arc lattice radiotherapy plan planning method and system

Abstract

The invention belongs to the technical field of nuclear magnetic resonance imaging systems, and provides a proton arc lattice radiotherapy plan planning method and system, and the method comprises the steps: designing an initial radiotherapy plan according to a beam spot lattice coverage rate and an extrapolation rate in combination with a delivery characteristic; performing adaptive optimization on the beam spot distribution to obtain height-modulated beam spot distribution; and carrying out energy transfer linear density distribution optimization based on the optimal beam spot distribution. Through beam spot distribution optimization and energy transfer linear density optimization, highly modulated beam spot distribution and energy transfer linear density distribution are obtained, lattice target area irradiation is maximized, excellent peak valley dose distribution and biological dose distribution are achieved, meanwhile, machine delivery characteristics are considered, the transmission efficiency is improved, the single treatment period is shortened, and the treatment cost is reduced. And the clinical application of proton arc lattice radiotherapy is promoted.

Description

Technical Field

[0001] This invention belongs to the technical field of nuclear magnetic resonance imaging systems, specifically relating to a proton arc lattice radiotherapy planning method and system. Background Technology

[0002] Spatial segmentation radiation planning technology is a novel radiation planning technique that breaks away from the traditional theory of uniform dose within the target area. By providing non-uniform radiation dose to the target area, that is, alternating high-dose "peaks" and low-dose "valleys," it shows significant advantages in radiation planning for large-volume and complex target areas. It can optimize the dose distribution within the target area while further reducing the radiation dose to the surrounding areas.

[0003] Lattice radiation scheme is an innovative implementation type of spatial segmentation radiation scheme in a three-dimensional scene. It achieves a highly customized peak-valley dose distribution within the target area by constructing multiple locally high-dose spheres (called vertices) with certain intervals within the target area, while maintaining a low dose level at the periphery of the target area to reduce unnecessary dose effects.

[0004] Due to the unique Bragg peak characteristics of proton beams, applying proton technology to space-fractionated radiation programs can achieve more precise dose distribution. Therefore, proton space-fractionated radiation programs have become a promising direction for radiation planning. Furthermore, proton rotation intensity-modulated radiation therapy (IMRT) is currently recognized as the next-generation technology in the field of proton radiation programs.

[0005] However, compared to photon volume modulated radiography (IMRT) schemes, lattice-based IMRT schemes do not show a significant improvement in peak-valley dose distribution optimization. This limitation may stem from the smaller beam field number in traditional IMRT schemes, while existing research has shown that proton arc IMRT schemes have superior dosimetric performance. To address this issue, current research attempts to introduce proton arc IMRT techniques into lattice-based IMRT schemes to improve peak-valley dose distribution optimization. However, these techniques do not yet consider the influence of beam spot distribution and energy transfer linear density distribution, leaving considerable room for improvement in scheme quality and transfer efficiency. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a proton arc lattice radiotherapy planning method and system to solve the problems in the prior art. The technical solution adopted by this invention is as follows:

[0007] This invention provides a proton arc lattice radiotherapy planning method and system. Through initial optimization of the beam spot energy layer, optimization of beam spot distribution, and optimization of energy transfer linear density distribution, a novel proton arc lattice radiotherapy planning method has been developed, achieving superior beam spot distribution, energy transfer linear density distribution, and higher transmission efficiency, and realizing superior peak-valley dose distribution and biological dose distribution.

[0008] This invention proposes a proton arc lattice radiotherapy planning system, comprising a storage device and a processor; the storage device includes at least one set of instructions; the processor is configured to communicate with the storage device, and when executing the instruction set, the processor is configured to instruct the system to perform the following operations:

[0009] Step S1: Design the initial radiotherapy plan based on the beam spot lattice coverage and fall rate combined with delivery characteristics;

[0010] Step S2: Optimize the beam spot distribution based on the original dual active set computation to determine the optimal beam spot distribution;

[0011] Step S3: Solve the energy transfer linear density distribution optimization model based on the alternating direction multiplier method to obtain the optimal energy transfer linear density distribution.

[0012] This invention proposes a planning method, comprising the following steps:

[0013] Step S1: Design the initial radiotherapy plan based on the beam spot lattice coverage and fall rate combined with delivery characteristics;

[0014] Step S2: Optimize the beam spot distribution based on the original dual active set computation to determine the optimal beam spot distribution;

[0015] Step S3: Solve the energy transfer linear density distribution optimization model based on the alternating direction multiplier method to obtain the optimal energy transfer linear density distribution.

[0016] The present invention has the following beneficial effects:

[0017] This invention optimizes the optimal energy distribution of the beam spot based on the beam spot lattice coverage and delivery characteristics, improves the machine delivery path, and enhances transmission efficiency, thus shortening the single treatment cycle. By combining the radiotherapy target and lattice distribution, beam spot distribution optimization yields a highly modulated beam spot distribution, thereby accurately determining the machine irradiation range and improving treatment precision. Based on the optimal beam spot distribution, energy transfer linear density distribution optimization is performed, improving the bio-dose calculation engine and obtaining a highly modulated energy transfer linear density distribution, thus ensuring the bio-dose distribution and improving the reliability of dose calculation. This maximizes lattice target area irradiation and achieves superior peak-valley dose distribution, potentially clearing obstacles for the clinical application of proton arc lattice radiotherapy technology and promoting the development of proton lattice therapy technology. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the proton arc lattice radiotherapy planning method of the present invention;

[0019] Figure 2 This is a schematic diagram of the optimal energy layer distribution of the beam spot in this invention;

[0020] Figure 3 This is a schematic diagram of the beam spot distribution in this invention (a. initial beam spot distribution; b. beam spot distribution after beam spot energy layer optimization; c. beam spot distribution after beam spot adaptive distribution optimization);

[0021] Figure 4 This is the energy transfer linear density volume histogram and the corresponding dose volume histogram in this invention;

[0022] Figure 5 This is a dose slice diagram from the present invention. Detailed Implementation

[0023] The following will be based on embodiments of the present invention. Figures 1-5 The technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Unless otherwise specified, the technical means used in the embodiments are conventional means well known to those skilled in the art.

[0024] This invention provides a proton arc lattice radiotherapy planning method and system. Through initial optimization of the beam spot energy layer, optimization of beam spot distribution, and optimization of energy transfer linear density distribution, a novel proton arc lattice radiotherapy planning method has been developed, achieving superior beam spot distribution, energy transfer linear density distribution, and higher transmission efficiency, and realizing superior peak-valley dose distribution and biological dose distribution.

[0025] It should be noted that this invention is a planning system and method, not an improvement to the specific radiotherapy treatment process, but rather a comprehensive overview of the operation and parameters of radiotherapy equipment. The weight of a proton therapy gantry typically exceeds 100 tons. Optimizing the arc-shaped delivery path significantly improves delivery efficiency and shortens the single treatment cycle; adaptive optimization of the beam spot lattice distribution precisely measures the machine's irradiation range, reduces mechanical positioning errors, and maximizes irradiation of the lattice target area, thus improving the treatment effect; and improving the bio-dose calculation engine through an energy transfer linear density distribution optimization algorithm further enhances dose delivery accuracy. Through these optimizations, this invention improves the overall operational performance of proton arc-shaped lattice radiotherapy equipment, including delivery efficiency, mechanical motion accuracy, and dose calculation reliability, providing a more efficient planning scheme for clinical implementation.

[0026] This invention proposes a proton arc lattice radiotherapy planning system, comprising a storage device and a processor; the storage device includes at least one set of instructions; the processor is configured to communicate with the storage device, and when executing the instruction set, the processor is configured to instruct the system to perform the following operations:

[0027] Step S1: Design the initial radiotherapy plan based on the beam spot lattice coverage and fall rate combined with delivery characteristics;

[0028] Step S2: Optimize the beam spot distribution based on the original dual active set computation to determine the optimal beam spot distribution;

[0029] Step S3: Solve the energy transfer linear density distribution optimization model based on the alternating direction multiplier method to obtain the optimal energy transfer linear density distribution.

[0030] Both storage devices and processors are existing technologies. Storage devices can be hard drives, memory, etc.; processors can be computer CPUs or computer processors.

[0031] Based on a proton arc lattice radiotherapy planning system, this invention also proposes a planning method, comprising the following steps:

[0032] Step S1: Design the initial radiotherapy plan based on the beam spot lattice coverage and fall rate combined with delivery characteristics.

[0033] Step S1 specifically includes:

[0034] S1.1 Determine the number and distribution position of the beam spot corresponding to the initial energy layer for each rack angle, and determine whether the beam spot falls within the lattice. Based on the lattice position, select an appropriate rack angle range to generate initial beam information. Each rack angle includes multiple selectable energy layers and corresponding beam spots. Based on the Bragg peak characteristics, determine the distribution position of the beam spot corresponding to each energy layer. Based on the lattice position information, determine whether the beam spot falls within the lattice.

[0035] S1.2 Calculate the beam spot lattice coverage and fall-off rate for all selectable energy layers. Based on whether the beam spot falls within the lattice, combined with the lattice volume, the volume of the irradiated object (tumor), and the beam spot volume, calculate the beam spot lattice coverage and fall-off rate for each selectable energy layer.

[0036] S1.3 Determining the Optimal Energy Distribution of the Beam Spot Based on Delivery Characteristics: When optimizing energy layer selection, priority is given to the beam spot's coverage of the lattice and its fall-off rate. However, existing proton arc radiotherapy techniques face a bottleneck of low delivery efficiency, mainly due to energy layer switching time; therefore, delivery characteristics must also be considered during optimization. This invention determines the optimal beam spot energy layer distribution based on the beam spot's coverage of the lattice and its fall-off rate, combined with the delivery characteristics of the cyclotron proton therapy system. This ensures that the beam spot distribution maximizes lattice target area irradiation while incorporating accelerator delivery characteristics, making the treatment plan more efficient. The objective function is denoted as:

[0037]

[0038] Where α, β, and γ are adjustment coefficients used to adjust the proportions of beam spot coverage P, beam spot fallout M, and energy switching time T in the objective function, i represents the rack angle, j represents the energy layer count corresponding to angle i, m represents the total number of rack angles, and N represents the number of energy layers in each angle. i The energy layer included at control point i is denoted as E. ij n ij This indicates whether the energy layer is selected; a value of 1 indicates that the energy layer is selected, and a value of 0 indicates that it is not selected. The corresponding beam lattice coverage is P(E). ij ·n ij The corresponding spot lattice fall-off rate is M(E) ij ·n ij ).

[0039] S1.4 Solving for the beam spot energy layer distribution: The intelligent ant colony optimization algorithm is used to solve the above objective function. Specifically, the algorithm parameters are first initialized: the number of ants is selected as 10*m; the size of the pheromone matrix τ is m*max(N). i The initial values ​​are set to the same small value (τ0 = 0.1); the heuristic information σ ij Reflecting the choice of energy layer E ij The local benefit is defined as (∈ is a small constant to avoid denominator of 0); pheromone evaporation rate ρ (take 0.1); pheromone enhancement coefficient Q (take 1.0); exploration weights λ and μ are used to control the balance between pheromone and heuristic information (take λ = 1, μ = 2).

[0040] Subsequently, the problem is mapped to a graph model: where nodes represent energy layers, including energy layer and beam distribution information; edges represent connecting energy layer nodes within the same rack angle; ants select energy layer nodes sequentially from the initial angle, passing each edge only once, forming a complete path, and the optimal path is the optimal beam energy layer distribution.

[0041] Subsequently, each ant iteratively constructs a path selection based on the joint probability distribution of pheromones and heuristic information. For each ant k, the selection probability is calculated. Record the path with the highest probability S k ={E ij};Calculate the objective function value f(S) k )=α∑M(E ij )-β∑P(E ij )-γ∑T(Δ(E ij The pheromone is used to update the path information and find the iterative optimal solution S. * Until the iteration termination condition is met.

[0042] Finally, the algorithm returns the globally optimal path S. best That is, the optimal beam spot energy layer distribution

[0043] S1.5 Design the initial radiotherapy plan. For each gantry angle, retain only the selected optimal energy layer and beam spot, and regenerate the beam information.

[0044] Figure 2 The optimal energy layer distribution map is shown using a cervical cancer case as an example. Figure 3 A beam pattern distribution map is shown.

[0045] Step S2: Optimize the beam spot distribution based on the original dual active set computation to determine the optimal beam spot distribution.

[0046] Step S2 specifically includes:

[0047] S2.1 Based on the beam information of the initial plan, regenerate the dose influence matrix and beam spot weight matrix.

[0048] S2.2 Construction of the beam spot optimization objective function: Due to the small size of the irradiated sphere, the peak-to-valley dose ratio may be greatly affected by the beam spot distribution. To further provide a highly modulated peak-to-valley spatial dose distribution, an adaptive beam spot allocation strategy is adopted for beam spot distribution constraint optimization. The optimization model is constructed as follows:

[0049]

[0050] Where d represents the dose distribution vector, p represents the required dose, A represents the lattice dose influence matrix, w is the beam spot weight vector (0 indicates the beam spot has been removed), and N is the total number of beam spots. This optimization model includes two objectives: the first is the dose objective term, and the second is the beam spot constraint regularization term. α>0 is the regularization parameter, controlling the balance between beam spot constraint and dose fitting. The constraint w≥0 ensures that the beam spot weights are non-negative.

[0051] S2.3 Solving for the optimal beam spot distribution: Based on the primal dual active set algorithm, the solution algorithm for the above model is designed. First, in order to minimize the beam spot weight vector, the solution function is rewritten as follows, where the dual variable d = A is introduced. T (p-Aw),

[0052]

[0053] Then iterative solutions are performed, with the initial active set... Non-active set Dual variable d (0) =A T (p-Aw (0) The algorithm consists of an outer iteration and an inner loop. The regularization parameter is in the interval α. min ≤α≤α max In the k-th outer iteration, α takes the following values:

[0054]

[0055] In the j-th inner loop, the active set Therefore w j and d j The optimal solution can be found by solving a least-squares problem defined on the dual activity set, as follows:

[0056]

[0057] By adjusting the external stopping criterion in the loop to achieve convergence to the active set stability, solutions with different levels of constraint can be found to balance beam spot distribution and planning quality. This adaptive beam spot allocation strategy significantly reduces beam spots distributed outside the lattice while improving beam spot lattice coverage, thereby further improving the peak-to-valley dose ratio.

[0058] Figure 3 The optimal plaque distribution map is shown using a cervical cancer case as an example.

[0059] Step S3: Solve the energy transfer linear density distribution optimization model based on the alternating direction multiplier method to obtain the optimal energy transfer linear density distribution.

[0060] Step S3 specifically includes:

[0061] S3.1, Adjust the beam spot weight matrix according to the optimal beam spot distribution and generate new beam information and dose influence matrix.

[0062] Section S3.2, Construction of the Energy Transfer Linear Density Distribution Optimization Model: In proton therapy, the relative biological effect (RBE) is typically assumed to be 1.1. However, a constant RBE approximation can introduce significant biological dose errors. Therefore, in this study, due to the monotonic RBE-LET relationship along the proton beam path, we optimize the intralattice energy transfer linear density to further optimize the biological dose distribution in proton arc lattice radiotherapy, considering the following optimization model:

[0063]

[0064] Where A represents the lattice dose influence matrix, w is the optimization variable beam spot weight vector, L is the energy transfer linear density influence matrix, m is the total number of beam spots, and δ let This represents the weighting coefficient of the energy transfer linear density objective term. The optimization model includes two objectives: the first objective function is the classic dose fidelity loss term; the second objective function is a constraint regularization term used to ensure the energy transfer linear density distribution. Based on the non-negativity constraint of the beam spot weight variables, adding a minimum hop count constraint on the weight of each beam spot can improve the planning quality while ensuring a highly modulated energy transfer linear density distribution in the proton arc lattice radiotherapy plan. This constraint is defined as follows: w∈{0}∪[g0,+∞)

[0065] This invention considers both the aforementioned model and the minimum hop count constraint as an optimization model for the energy transfer linear density distribution in proton arc lattice radiotherapy, where p represents the required dose, Ω d For the dose fidelity activity set, Ω l For optimizing the energy transfer linear density term, where l represents the target energy transfer linear density value of the lattice, the model is as follows:

[0066]

[0067] S3.3, Solving for the optimal energy transfer linear density distribution: Based on the alternating direction multiplier method, the algorithm for solving the above model is designed. First, the above model can be rewritten into a standard alternating direction multiplier method framework solution format with two dummy variables, which is the following optimization problem:

[0068]

[0069] Here, y and z are introduced as dummy variables. Specifically, y separates the dose target and the energy transfer linear density target, while z is used to decouple the minimum hop count constraint. The minimum hop count constraint is strengthened using the function m(z), defined as follows:

[0070]

[0071] The optimal energy transfer linear density distribution is obtained by solving the problem using the alternating direction multiplier method. Specifically, the parameter w is initialized first. (0)y (0) , z (0) ,Lagrange multiplier λ (0) μ (0) Define the maximum iteration parameter X, the tolerance ∈, and the penalty parameters ρ1 and ρ2.

[0072] Next, iterative updates are performed. In the k-th iteration, w, y, and z are updated as follows:

[0073]

[0074] Update the Lagrange multipliers:

[0075]

[0076] The residuals and iteration count are then checked to determine if convergence has terminated. This optimizes the lattice energy transfer linear density distribution while maintaining the dosimetric objectives, thus ensuring the optimal biological dose distribution. Figure 4 The optimal energy transfer linear density volume histogram and the corresponding dose volume histogram for a cervical cancer case are shown.

[0077] Figure 5 A dose-slice diagram of a cervical cancer case is shown.

[0078] This invention simultaneously optimizes the beam spot energy distribution, beam spot distribution, and energy transfer linear density distribution in proton arc lattice radiotherapy, constructing a complete treatment planning framework. This achieves superior peak-valley dose distribution and highly modulated energy transfer linear density distribution, providing a feasible and efficient proton arc lattice radiotherapy planning method. In this invention's method, energy distribution is optimized based on the patient's lattice beam spot coverage and delivery characteristics, and the arc delivery path is optimized, significantly improving delivery efficiency and shortening the single treatment cycle. An adaptive beam spot allocation strategy optimizes the beam spot distribution, accurately refining the machine's irradiation range, reducing mechanical positioning errors, and maximizing irradiation of the lattice target area. Optimizing the energy transfer linear density distribution improves the bio-dose calculation engine, ensuring bio-dose distribution and further improving dose delivery accuracy. This generates a highly efficient proton arc lattice radiotherapy plan, enhancing radiotherapy delivery efficiency and providing superior peak-valley dose distribution and highly modulated energy transfer linear density distribution. This is expected to clear obstacles for the clinical application of proton arc lattice radiotherapy and promote the development of spatially fractionated radiotherapy.

[0079] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Any modifications, alterations, substitutions, or variations made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention shall fall within the protection scope defined by the claims of the present invention.

Claims

1. A proton arc lattice radiotherapy treatment planning system, characterized by, The system comprises a storage device and a processor; the storage device comprises at least one set of instruction sets; the processor is configured to communicate with the storage device, and when the instruction sets are executed, the processor is configured to instruct the system to perform the following operations: Step S1, designing an initial radiotherapy plan according to beam spot lattice coverage and spillage rate combined with delivery characteristics; Step S2, optimizing the beam spot distribution based on the original dual active set algorithm, and determining the optimal beam spot distribution; Step S3, solving the fluence density distribution optimization model based on the alternating direction multiplier method, and obtaining the optimal fluence density distribution.

2. The proton arc lattice radiotherapy treatment planning system of claim 1, wherein, When the processor executes step S1, the following steps are included: S1.1, determining the number and distribution position of beam spots corresponding to each gantry angle initial energy layer, and judging whether the beam spot falls within the lattice: according to the lattice position, selecting the gantry angle range, generating the initial beam information, wherein each gantry angle range includes multiple selectable energy layers and corresponding beam spots, determining the beam spot distribution position corresponding to each energy layer, and judging whether the beam spot falls into the lattice according to the lattice position information; S1.2, calculating the beam spot lattice coverage and spillage rate corresponding to all selectable energy layers: according to whether the beam spot falls into the lattice, combining the lattice volume, the irradiated object volume and the beam spot volume, calculating the beam spot lattice coverage and spillage rate corresponding to the selectable energy layers; S1.3, determining the optimal beam spot energy layer distribution based on the coverage and spillage rate of the beam spot on the lattice combined with the delivery characteristics of the cyclotron proton therapy system, and ensuring that the beam spot distribution maximizes the irradiation of the lattice target area while combining the accelerator delivery characteristics, the objective function is denoted as: Wherein, α, β, γ are adjustment coefficients, used to adjust the proportion of beam spot-lattice coverage rate P, beam spot-lattice out-fall rate M and energy switching time T in the objective function, i represents the gantry angle, j represents the corresponding energy layer count in angle i, m represents the total number of gantry angles, and the number of energy layers in each angle is N i ; the energy layer included in the control point i is recorded as E ij , n ij indicates whether the energy layer is selected, and if it is 1, the energy layer is selected, and if it is 0, the energy layer is not selected, the corresponding beam spot-lattice coverage rate is P(E ij ·n ij ), and the corresponding beam spot-lattice out-fall rate is M(E ij ·n ij ). S1.4, solving the beam spot energy layer distribution: using the intelligent ant colony optimization algorithm to solve the objective function.

3. The proton arc lattice radiotherapy treatment planning system of claim 2, wherein, When the processor executes step S1.4, the following steps are included: First, the algorithm parameter initialization configuration is completed: the number of ants is selected as 10*m; the pheromone matrix τ size is m*max(N i ), and the initial value is set as the same small value; the heuristic information σ ij reflects the local benefit of selecting the energy layer E ij , and is defined as , ∈ is a constant; the pheromone evaporation rate ρ is 0.1; the pheromone enhancement coefficient Q is 1.0; the exploration weight λ and μ are used to control the balance between pheromone and heuristic information; Then, the problem is mapped to a graph model: where the node represents the energy layer, including the energy layer and beam spot distribution information; the edge is characterized as connecting the energy layer nodes within the same gantry angle; the ants start from the initial angle and select the energy layer nodes in turn, each edge can only pass once, forming a complete path, and the optimal path is the optimal beam spot energy layer distribution; Thereafter, each ant iteratively constructs a selection path according to a joint probability distribution of pheromone and heuristic information, and for each ant k, a selection probability is calculated The path S with the maximum probability is recorded k = {E ij}; the objective function value f(S k ) = a∑M(E ij ) - β∑P(E ij ) - γ∑T(Δ(E ij )) is calculated; pheromone is updated to update path information to find an iterative optimal solution S * until an iterative termination condition is reached. Finally, the algorithm returns the globally optimal path S best i.e. the optimal beamlet energy layer distribution.

4. The proton arc lattice radiotherapy treatment planning system of claim 1, wherein, When the processor executes step S2, the following steps are included: S2.1, according to the beam information of the initial plan, re-generating the dose influence matrix and beam spot weight matrix; S2.2, constructing the beam spot optimization objective function, using an adaptive beam spot allocation strategy to optimize the beam spot distribution constraint, and constructing the following optimization model: Where d represents the dose distribution vector, p represents the required dose, A represents the lattice dose influence matrix, w is the beam spot weight vector, and 0 represents that the beam spot has been removed, and N is the total number of beam spots; the optimization model includes two optimization objectives: the first objective is the dose objective term, and the second objective is the beam spot constraint regularization term, and α>0 is a regularization parameter controlling the balance weight of beam spot constraint and dose fitting; S2.3, Solution of optimal beamlet distribution, the solution algorithm design of optimization model is based on the original dual active set algorithm. Firstly, in order to minimize the beamlet weight vector, the solution function is rewritten as follows, wherein the dual variable d=A T (p-Aw); Subsequently, an iterative solution is performed, initial active set Non-active set Dual variable d (0) = A T (p - Aw (0) ); the algorithm includes outer iterations and inner loops, and the regularization parameter is selected in the interval a min ≤ a ≤ a max In the kth outer iteration, a takes the value In the jth inner loop, the active set w j and d j The optimal solution is found by solving a least squares problem defined on the dual active set: By adjusting the external stopping criterion in the loop, the active set can be converged to the steady state, and solutions with different constraint degrees can be found to balance the beam spot distribution and the plan quality.

5. The proton arc lattice radiotherapy treatment planning system of claim 1, wherein, When the processor executes step S3, the following steps are included: S3.1, adjusting the beam spot weight matrix according to the optimal beam spot distribution and generating new beam information and dose influence matrix; S3.2, Construction of the fluence line density distribution optimization model, the fluence line density in the lattice is optimized to optimize the biological dose distribution in the proton arc lattice radiotherapy, the following optimization model is adopted: where A represents the lattice dose impact matrix, w is the optimization variable beamlet weight vector, L is the fluence line density impact matrix, m is the total number of beamlet spots, δ let represents the fluence line density target term weight coefficient; the optimization model includes two optimization objectives: the first objective function is a dose fidelity loss term; the second objective function is a constraint regularization term for ensuring the fluence line density distribution; on the basis of the non-negative constraint of the beamlet spot weight variable, a minimum jump constraint for each beamlet spot weight is added to ensure that the fluence line density distribution of the proton arc lattice radiotherapy plan is highly modulated while improving the plan quality, and the constraint is defined in the following form: w e {0} U [g0, +∞) The optimization model and the minimum hop constraint are simultaneously considered as a proton arc lattice radiotherapy fluence distribution optimization model, p represents a required dose, Ω d is a dose fidelity term activity set, Ω l is a fluence term optimization activity set, l represents a lattice target fluence value, and the model is as follows: S3.3, Solution of the optimal fluence line density distribution, the algorithm design of the model is based on the alternating direction multiplier method: first, the above model can be rewritten as a standard alternating direction multiplier method framework with two virtual variables, that is, the following optimization problem: Wherein, y and z are introduced as virtual variables, the function m(z) is used to strengthen the minimum hop constraint, and is defined as follows: The optimal energy transmission line density distribution is obtained by solving the alternating direction multiplier method: first, initialize the parameters w (0) ,y (0) ,z (0) , Lagrange multiplier λ (0) ,μ (0) , define the maximum iteration parameter X, tolerance ∈ and penalty parameters ρ1, ρ2; Secondly, the iteration is updated, and when the kth iteration is updated, w, y and z are updated as follows: Update the Lagrange multiplier: Then check the residual and the number of iterations to determine whether to converge and terminate; thereby optimizing the best lattice fluence line density distribution while ensuring the dosimetric target, while ensuring the biological dose distribution.

6. A planning method applied to the proton arc lattice radiotherapy planning system of any one of claims 1-5, comprising the following steps: Step S1, design an initial radiotherapy plan according to the beam spot lattice coverage and the leakage rate combined with the delivery characteristics; Step S2, optimize the beam spot distribution based on the original dual active set algorithm to determine the optimal beam spot distribution; Step S3, solve the fluence line density distribution optimization model based on the alternating direction multiplier method to obtain the optimal fluence line density distribution.

7. The planning method of claim 7, step S1 comprising: S1.1, determine the number and distribution position of beam spots corresponding to each gantry angle initial energy layer, and determine whether the beam spot falls within the lattice: according to the lattice position, select the gantry angle range, generate the initial beam information, wherein each gantry angle range includes multiple selectable energy layers and corresponding beam spots, determine the beam spot distribution position corresponding to each energy layer, and determine whether the beam spot falls into the lattice according to the lattice position information; S1.2, calculate the beam spot lattice coverage and leakage rate corresponding to all selectable energy layers: according to whether the beam spot falls into the lattice, combined with the lattice volume, the irradiated object volume and the beam spot volume, calculate the beam spot lattice coverage and leakage rate corresponding to the selectable energy layers; S1.3, determine the optimal beam spot energy layer distribution based on the beam spot coverage and leakage rate on the lattice combined with the delivery characteristics of the cyclotron proton therapy system, ensure that the beam spot distribution maximizes the irradiation of the lattice target area while combining the accelerator delivery characteristics, and the objective function is denoted as: Wherein, α, β, γ are adjustment coefficients, used to adjust the proportion of beam spot-lattice coverage rate P, beam spot-lattice out-fall rate M and energy switching time T in the objective function, i represents the gantry angle, j represents the corresponding energy layer count in angle i, m represents the total number of gantry angles, and the number of energy layers in each angle is N i ; the energy layer included in the control point i is recorded as E ij , n ij indicates whether the energy layer is selected, 1 if selected, 0 if not selected, and the corresponding beam spot-lattice coverage rate is P(E ij ·n ij ), and the corresponding beam spot-lattice out-fall rate is M(E ij ·n ij ); S1.4, solution of beam spot energy layer distribution: the intelligent ant colony optimization algorithm is adopted to solve the objective function.

8. A planning method according to claim 7, characterized in that, Step S1.4 comprises: First, the algorithm parameter initialization configuration is completed: the number of ants is selected as 10*m; the pheromone matrix τ size is m*max(N i ), and the initial value is set as the same small value; the heuristic information σ ij reflects the local benefit of selecting the energy layer E ij , and is defined as , ∈ is a constant; the pheromone evaporation rate ρ is 0.1; the pheromone enhancement coefficient Q is 1.0; the exploration weight λ and μ are used to control the balance between the pheromone and the heuristic information; Then, the problem is mapped to a graph model: where the node represents the energy layer, including the energy layer and beam spot distribution information; the edge represents the connection between the energy layer nodes in the same gantry angle; the ant selects the energy layer node from the initial angle, each edge can only pass once, forming a complete path, and the optimal path is the optimal beam spot energy layer distribution; Thereafter, each ant iteratively constructs a selection path according to a joint probability distribution of pheromone and heuristic information, and for each ant k, a selection probability is calculated The path S with the maximum probability is recorded k = {E ij}; a target function value f(S k ) = a∑M(E ij ) - β∑P(E ij ) - γ∑T(Δ(E ij )) is calculated; pheromone is updated to update path information to find an iterative optimal solution S * until an iterative termination condition is reached. Finally, the algorithm returns the globally optimal path S best i.e. the optimal beamlet energy layer distribution.

9. A planning method according to claim 7, characterized in that, Step S2 comprises: S2.1, regenerate the dose influence matrix and beam spot weight matrix according to the beam information of the initial plan; S2.2, the construction of beamlet optimization objective function, a self-adaptive beamlet allocation strategy is used to optimize the beamlet distribution constraint, which is constructed as the following optimization model: Where d represents the dose distribution vector, p represents the required dose, A represents the lattice dose influence matrix, w is the beamlet weight vector, and if it is 0, it means that the beamlet has been removed, and N is the total number of beamlet points; The optimization model includes two optimization objectives: the first objective is the dose objective term, and the second objective is the beamlet constraint regularization term, and a>0 is the regularization parameter, which controls the balance weight of beamlet constraint and dose fitting; S2.3, Solution of optimal beamlet distribution, the solution algorithm design of optimization model is based on the original dual active set algorithm. Firstly, in order to minimize the beamlet weight vector, the solution function is rewritten as follows, wherein the dual variable d=A T (p-Aw); Subsequently, an iterative solution is performed, initial active set Non-active set Dual variable d (0) = A T (p - Aw (0) ); the algorithm includes outer iterations and inner loops, and the regularization parameter is selected in the interval a min ≤ a ≤ a max In the kth outer iteration, a takes the value In the jth inner loop, the active set w j and d j The optimal solution is found by solving a least squares problem defined on the dual active set: By adjusting the external stopping criterion in the loop, the active set can be converged to stability, and solutions with different constraint degrees can be found to balance the beamlet distribution and the plan quality.

10. The proton arc lattice radiotherapy treatment planning system of claim 7, wherein, Step S3 includes: S3.1, adjusting the beamlet weight matrix according to the optimal beamlet distribution and generating new beam information and dose influence matrix; S3.2, the construction of the fluence line density distribution optimization model, the fluence line density in the lattice is optimized to optimize the biological dose distribution in the proton arc lattice radiotherapy, and the following optimization model is used: where A represents the lattice dose impact matrix, w is the optimization variable beamlet weight vector, L is the fluence line density impact matrix, m is the total number of beamlet spots, δ let represents the fluence line density target term weight coefficient; the optimization model includes two optimization objectives: the first objective function is a dose fidelity loss term; the second objective function is a constraint regularization term for ensuring the fluence line density distribution; on the basis of the non-negative constraint of the beamlet spot weight variable, a minimum jump constraint for each beamlet spot weight is added to ensure that the fluence line density distribution of the proton arc lattice radiotherapy plan is highly modulated while improving the plan quality, and the constraint is defined in the following form: w∈{0}∪[g0,+∞) The optimization model and the minimum hop constraint are simultaneously considered as a proton arc lattice radiotherapy fluence distribution optimization model, p represents a required dose, Ω d is a dose fidelity term activity set, Ω l is a fluence term optimization activity set, l represents a lattice target fluence value, and the model is as follows: S3.3, solving the optimal fluence line density distribution, based on the alternating direction multiplier method, the algorithm design of the model is as follows: first, the above model can be rewritten as a standard alternating direction multiplier method framework with two virtual variables, that is, the following optimization problem: Where y and z are introduced as virtual variables, the function m(z) is used to strengthen the minimum hop constraint, and is defined as follows: The optimal energy transmission line density distribution is obtained by solving by the alternate direction multiplier method: first, initializing parameters w (0) , y (0) , z (0) , Lagrange multipliers λ (0) , μ (0) , defining a maximum iteration parameter X, a tolerance ∈ and penalty parameters ρ1, ρ2; Secondly, the iteration is updated, and w, y and z are updated at the kth iteration: Update the Lagrange multiplier: Then check the residual and the number of iterations to determine whether to converge and terminate; Thus, the best lattice fluence line density distribution is obtained while ensuring the dosimetric target, and the biological dose distribution is ensured.

Images