FLASH radiotherapy dose correction and compensator collaborative optimization method

Abstract

This invention relates to the field of radiotherapy technology, and particularly to a method for dose correction and compensator co-optimization in FLASH radiotherapy. The method includes constructing an objective function using the corrected target dose and actual dose in the target area; integrating the objective function with a process constraint penalty function for the compensator thickness to generate an optimization function and determining the constraint conditions; calculating the gradient of the updated optimization function with respect to the beam intensity; calculating the compensator thickness based on the updated beam intensity; verifying the compensator thickness according to the constraint conditions; determining whether the dose deviation meets dose convergence requirements based on the optimized beam intensity and compensator thickness; and outputting an optimized scheme based on the result of whether the compensator thickness meets the constraint conditions. This invention can avoid overcompensation or undercompensation of the target area, preserving the normal tissue protection effect of FLASH radiotherapy, and simultaneously solving the problem in traditional technologies where the dose is within the target range but the compensator cannot be manufactured, thus improving the feasibility of FLASH radiotherapy treatment plans.

Description

Technical Field

[0001] This invention relates to the field of radiotherapy technology, and in particular to a method for correcting FLASH radiotherapy dose and co-optimizing compensators. Background Technology

[0002] FLASH radiotherapy, with its ultra-high dose rate and unique protective effect on normal tissues, has become an important direction for the development of precision radiotherapy. Clinically, it is necessary to ensure that the target dose is accurate to trigger the FLASH effect and kill tumor cells, while also ensuring that the compensator can be actually manufactured and adapted to the radiotherapy equipment to realize the treatment plan. However, when dealing with FLASH radiotherapy dose correction and compensator design, the existing technology performs dose correction separately through algorithms and then designs the compensator independently based on the correction results. It does not take into account the difference between the radiation attenuation coefficient under ultra-high dose rates and traditional low dose rates, which can easily lead to overcompensation of the target area (i.e., the actual dose is higher than the treatment threshold, resulting in excessive irradiation of normal tissues) or undercompensation (i.e., the actual dose is lower than the treatment threshold), which directly weakens the protective effect of normal tissues of the FLASH effect. Furthermore, the beam intensity and compensator thickness are not designed in a coordinated manner, resulting in situations where the theoretical dose is met, but the compensator exceeds the manufacturing limitations and cannot be manufactured.

[0003] Therefore, it is necessary to propose a method for co-optimizing FLASH radiotherapy dose correction and compensator to solve the above problems. Summary of the Invention

[0004] The purpose of this invention is to provide a method for correcting FLASH radiotherapy dose and co-optimizing compensators, in order to solve the problem that overcompensation or undercompensation of the target area weakens the normal tissue protection effect of FLASH and that the theoretical dose is achieved, but the compensator exceeds the manufacturing limitations and cannot be manufactured.

[0005] To achieve the above objectives, the present invention provides the following technical solution: The method for correcting FLASH radiotherapy dose and co-optimizing compensators includes the following steps: S1: Initialize the beam intensity and compensator thickness within the radiation field, obtain the target dose in the target area based on the FLASH radiotherapy physical-biological combined dose correction model, calculate the actual dose based on the current beam intensity and current compensator thickness, and construct the dose deviation objective function between the target dose and the actual dose; S2: Determine the constraints, construct the process constraint penalty function for the compensator thickness based on the thickness range constraint and thickness gradient constraint, and integrate the process constraint penalty function with the dose deviation objective function to generate the optimization function; S3: Update the beam intensity according to the optimization function, and solve the compensator thickness according to the updated beam intensity. Verify the compensator thickness according to the constraint conditions. When the compensator thickness does not meet the constraint conditions, feed back the solved compensator thickness to the beam intensity update step, and re-update the beam intensity and solve the compensator thickness. S4: Recalculate the actual dose based on the optimized beam intensity and compensator thickness, determine whether the dose deviation between the target dose and the actual dose meets the dose convergence requirement, and output the optimization scheme based on whether the compensator thickness meets the constraint conditions.

[0006] Preferably, the steps of initializing the beam intensity and compensator thickness within the radiation field, obtaining the target dose in the target area based on the FLASH radiotherapy physical-biological combined dose correction model, calculating the actual dose based on the current beam intensity and current compensator thickness, and constructing a dose deviation objective function between the target dose and the actual dose include: The target dose in the target area is obtained by using the FLASH radiotherapy physical-biological combined dose correction model. Based on the current beam intensity and the current compensator thickness, the energy deposition of the radiation penetrating the compensator and the human body is calculated to obtain the actual dose. Minimizing the deviation between the target dose and the actual dose is taken as the optimization objective, and a dose deviation objective function is constructed. The calculation formula is: ; in, Let be the dose deviation objective function value, representing the sum of squared deviations between the target dose and the actual dose in the target area under the current beam intensity S and compensator thickness T. The two-dimensional spatial coordinates of the target area Two-dimensional spatial coordinates of the target area The target dose, Two-dimensional spatial coordinates of the target area The actual dose.

[0007] Preferably, the step of determining the constraints, constructing a process constraint penalty function for the compensator thickness based on the compensator's thickness range constraint and thickness gradient constraint, and integrating the process constraint penalty function with the dose deviation objective function to generate an optimization function, includes: The constraints include the thickness range constraint and the thickness gradient constraint of the compensator; The thickness range constraint includes the minimum machinable thickness of the compensator. and maximum machinable thickness ,but ; The thickness gradient constraint includes determining the maximum permissible variation in thickness between adjacent regions. ,but , For the compensator thickness in the target area gradient at; The constraints are transformed into penalty terms, and a process constraint penalty function for obtaining the compensator thickness is constructed. ; Process constraint penalty function With objective function By integrating, we obtain the optimized function. .

[0008] Preferably, the process constraint penalty function for the compensator thickness is constructed by converting the constraint conditions into penalty terms. The steps include: The penalty term includes a thickness range penalty term and a thickness gradient penalty term; Regarding the thickness range penalty, for thicknesses exceeding the minimum machinable thickness... Or maximum machinable thickness The thickness value is calculated, the deviation is summed, and the deviation is calculated. when At that time, there were no penalties. when or At times, there are penalties, indicating violations or deviations. For the thickness gradient penalty term, for those exceeding... The thickness gradient is calculated, the deviation is summed, and the deviation is determined. when At that time, there were no penalties. when At this time, there is a penalty term; the larger the gradient, the heavier the penalty. By setting weighting coefficients and integrating the thickness range penalty term and the thickness gradient penalty term, a process constraint penalty function is obtained. By setting the weighting coefficients for thickness range constraints and thickness gradient constraints respectively. , Integrating the two penalty terms yields the process constraint penalty function. The calculation formula is: ; in, For thickness constraint penalty term, This is a gradient constraint penalty term.

[0009] Preferably, the constraints further include dose constraints, which include target dose rate constraints, dose uniformity constraints, dose conformity constraints, and critical organ dose constraints.

[0010] Preferably, the steps of updating the beam intensity according to the optimization function, solving for the compensator thickness based on the updated beam intensity, and verifying the compensator thickness according to the constraint conditions include: Based on the optimization function, the partial derivative of the optimization function with respect to the beam intensity is calculated using the gradient calculation formula. The partial derivative is used as the gradient of the optimization function with respect to the beam intensity. Based on the gradient direction and the preset learning rate, the current beam intensity is iteratively updated to obtain the updated beam intensity; Based on the updated beam intensity, the compensator thickness is iteratively solved using a quasi-Newton algorithm. Based on the thickness range constraint and thickness gradient constraint in the constraint conditions, the thickness of the compensator obtained by the solution is verified respectively. When the compensator thickness simultaneously satisfies both the thickness range constraint and the thickness gradient constraint, the parameters of the current beam intensity and the compensator thickness are retained. When the compensator thickness is different and the thickness range constraint and thickness gradient constraint are satisfied, the calculated compensator thickness is fed back to the beam intensity gradient calculation, and the beam intensity is updated and the compensator thickness is recalculated until the verification is passed.

[0011] Preferably, the step of feeding back the solved compensator thickness to the beam intensity update step when the compensator thickness does not meet the constraint conditions, and re-performing the beam intensity update and compensator thickness solution, includes: Based on the updated beam intensity and optimization function, the gradient of the optimization function with respect to the compensator thickness is solved using a numerical approximation method. A quasi-Newton algorithm is used to update the inverse matrix of the current iteration based on the parameters of the previous iteration; The compensator thickness is iteratively updated using the gradient and inverse matrix of the compensator thickness. Set a convergence threshold and a maximum number of iterations. Calculate the difference between the compensator thickness in the current iteration and the previous iteration. If the difference is less than the convergence threshold, or the number of iterations reaches the maximum number of iterations, the iteration is considered converged, and the compensator thickness is output. If the iteration is not converged, return to the optimization function to calculate the gradient of the compensator thickness and continue iterating until the iteration converges.

[0012] Preferably, the steps of recalculating the actual dose based on the optimized beam intensity and compensator thickness, determining whether the dose deviation between the target dose and the actual dose satisfies dose convergence, and outputting the optimized scheme based on whether the compensator thickness meets the constraint conditions, include: Based on the optimized beam intensity and compensator thickness, the actual dose is calculated, and combined with the target dose in the target area, the dose deviation value is calculated. According to the dose constraint in the constraint conditions, when the actual dose and the dose deviation value meet the dose constraint, it is determined that the dose converges. Based on the thickness range constraint and thickness gradient constraint of the compensator in the constraint conditions, the thickness of the compensator is checked respectively. If all the constraint conditions are met, the thickness of the compensator is determined to be acceptable. An optimized scheme is generated based on the results of dose convergence and compensator thickness verification.

[0013] The technical effects and advantages of the present invention in the above technical solution are as follows: 1. Unlike the existing technology that first performs dose correction and then designs the compensator, this invention performs coordinated optimization and iteration of beam intensity and compensator thickness, and uses the objective function to unify the dose deviation, so that the corrected target dose matches the actual irradiation dose, avoiding overcompensation or undercompensation of the target area, and preserving the normal tissue protection effect of FLASH radiotherapy.

[0014] 2. By incorporating the process constraints of compensator thickness range and thickness gradient into the optimization process, the optimization scheme can optimize the dosage while simultaneously including the compensator manufacturing process in the optimization scope, thus solving the problem in traditional technology where the dosage meets the standard but the compensator cannot be manufactured.

[0015] 3. By verifying the dual constraints of dose convergence and compensator constraints, the optimized scheme is output for different verification results, ensuring that the optimized scheme meets both clinical radiotherapy treatment requirements and equipment hardware compatibility requirements, thereby improving the feasibility of FLASH radiotherapy treatment schemes. Attached Figure Description

[0016] Figure 1 This is a flowchart of the FLASH radiotherapy dose correction and compensator co-optimization method of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] like Figure 1 As shown, this embodiment provides a method for FLASH radiotherapy dose correction and compensator co-optimization, including the following steps: S1: Initialize the beam intensity and compensator thickness within the radiation field, obtain the target dose in the target area based on the FLASH radiotherapy physical-biological combined dose correction model, calculate the actual dose based on the current beam intensity and current compensator thickness, and construct the dose deviation objective function between the target dose and the actual dose; S2: Determine the constraints, construct the process constraint penalty function for the compensator thickness based on the thickness range constraint and thickness gradient constraint, and integrate the process constraint penalty function with the dose deviation objective function to generate the optimization function; S3: Update the beam intensity according to the optimization function, and solve the compensator thickness according to the updated beam intensity. Verify the compensator thickness according to the constraint conditions. When the compensator thickness does not meet the constraint conditions, feed back the solved compensator thickness to the beam intensity update step, and re-update the beam intensity and solve the compensator thickness. S4: Recalculate the actual dose based on the optimized beam intensity and compensator thickness, determine whether the dose deviation between the target dose and the actual dose meets the dose convergence requirement, and output the optimization scheme based on whether the compensator thickness meets the constraint conditions.

[0019] In one embodiment of the present invention, the steps of initializing the beam intensity and compensator thickness within the radiation field, obtaining the target dose in the target area based on the FLASH radiotherapy physical-biological combined dose correction model, calculating the actual dose based on the current beam intensity and the current compensator thickness, and constructing a dose deviation objective function between the target dose and the actual dose include: S11: The target dose in the target area is obtained by using the FLASH radiotherapy physical-biological combined dose correction model. Based on the current beam intensity and the current compensator thickness, the energy deposition of the radiation penetrating the compensator and the human body is calculated to obtain the actual dose. S12: Minimizing the deviation between the target dose and the actual dose is used as the optimization objective. A dose deviation objective function is constructed, wherein the dose deviation, and the objective function... The calculation formula is: ; in, Let be the objective function value, representing the sum of squared deviations between the target dose and the actual dose in the target area under the current beam intensity S and compensator thickness T. The two-dimensional spatial coordinates of the target area Two-dimensional spatial coordinates of the target area The target dose, Two-dimensional spatial coordinates of the target area The actual dose.

[0020] In this embodiment of the invention, as shown in steps S11-S13 above, the initialization process for the beam intensity and compensator thickness within the FLASH radiotherapy field is based on the irradiation angle determined through multi-angle fractionated irradiation optimization. Combined with the clinical radiotherapy planning database, the beam intensity distribution within the field is initialized. Furthermore, the compensator thickness distribution is initialized according to the compensator's manufacturing process. For the actual dose after beam intensity and compensator thickness adjustments, the GBBS photon transport dose calculation model is used. By inputting the patient's CT / MRI anatomical data, FLASH ray physical parameters, initial beam intensity, and initial compensator thickness, the energy deposition process of ray penetration through the compensator and human tissue is simulated to obtain the actual dose distribution. For the target dose, the acquisition process is based on FLASH... The FLASH radiotherapy physical-biological combined dose correction model (e.g., using a gridded Boltzmann transport solver GBBS + dose correction factor DMF + oxygen depletion effect) is specifically applied to the target dose generation stage of radiotherapy planning. The model combines clinical prescription dose, patient anatomy, and radiotherapy equipment parameters to correct the physical dose target of radiotherapy to a target dose that adapts to the FLASH effect and treatment effect. By inputting anatomical data, equipment physical parameters, compensator parameters, clinical treatment dose, and FLASH biological effect parameters into the model, the model outputs the corrected target dose, dose distribution, biological effect index, and dose index, which serve as the basis for optimizing beam intensity and compensator thickness. By combining clinical dose, target dose rate requirements, and dose safety thresholds for organs at risk, the target dose is corrected to obtain the target dose distribution; After obtaining the target dose and the actual dose, an objective function is constructed. The objective function aims to minimize the deviation between the target dose and the actual dose as the computational objective of the model. Its purpose is to reflect the degree of dose matching in the current scheme through the deviation value between the target dose and the actual dose, providing a standard for the joint optimization of beam intensity and compensator thickness. Specifically, the formula for calculating the objective function is as follows: ; in, The objective function value represents the sum of squared deviations between the target dose and the actual dose in the target area under the current beam intensity S and compensator thickness T. The smaller the value, the higher the dose matching degree and the better the optimization effect. The two-dimensional spatial coordinates of the target area Two-dimensional spatial coordinates of the target area The target dose, Two-dimensional spatial coordinates of the target area The actual dose.

[0021] In one embodiment of the present invention, the step of determining the constraints, constructing a process constraint penalty function for the compensator thickness based on the compensator thickness range constraint and thickness gradient constraint, and integrating the process constraint penalty function with the dose deviation objective function to generate an optimization function, includes: S21: The constraints include the thickness range constraint and the thickness gradient constraint of the compensator; S22: The thickness range constraint includes the minimum machinable thickness of the compensator. and maximum machinable thickness ,but ; S23: The thickness gradient constraint includes determining the maximum permissible variation in thickness between adjacent regions. ,but , For the compensator thickness in the target area gradient at; S24: Transform the constraints into penalty terms to construct the process constraint penalty function for the compensator thickness. ; S25: Apply process constraint penalty function With objective function By integrating, we obtain the optimized function. .

[0022] The process constraint penalty function for compensator thickness is constructed by transforming the constraint conditions into penalty terms. The steps include: S241: The penalty term includes a thickness range penalty term and a thickness gradient penalty term; S242: Penalty for thickness range, for thicknesses exceeding the minimum machinable thickness. Or maximum machinable thickness The thickness value is calculated, the deviation is summed, and the deviation is calculated. when At that time, there were no penalties. when or At times, there are penalties, indicating violations or deviations. S243: For the thickness gradient penalty term, for exceeding... The thickness gradient is calculated, the deviation is summed, and the deviation is determined. when At that time, there were no penalties. when At this time, there is a penalty term; the larger the gradient, the heavier the penalty. S244: Set weighting coefficients to integrate the thickness range penalty term and the thickness gradient penalty term to obtain the process constraint penalty function. .

[0023] S211: The constraints also include dose constraints, which include target dose rate constraints, dose uniformity constraints, dose conformity constraints, and critical organ dose constraints.

[0024] In this embodiment of the invention, as shown in steps S21-S25, S241-S244, and S211 above, the construction of the process constraint penalty function involves converting the set thickness range constraint and thickness gradient constraint into quantifiable penalty terms. During optimization design, additional checks are performed on schemes that violate the constraints on the compensator thickness, i.e., determining whether the compensator thickness violates the thickness range constraint and thickness gradient constraint. The penalty terms include a thickness range penalty term and a thickness gradient penalty term. Based on the compensator thickness, for the thickness range penalty term, for thicknesses exceeding the minimum machinable thickness... Or maximum machinable thickness The thickness value is calculated by summing the deviations, using the following formula: ; when At that time, there were no penalties. when or At that time, there are penalties for violations; the greater the deviation, the heavier the penalty. For the thickness gradient penalty term, for those exceeding... The thickness gradient is calculated, the deviation is calculated and summed, and the calculation formula is: ; when At that time, there were no penalties. when At this time, there is a penalty term; the larger the gradient, the heavier the penalty. After converting the thickness range constraint and thickness gradient constraint into thickness range penalty terms and thickness gradient penalty terms respectively, the thickness range penalty terms and thickness gradient penalty terms need to be integrated to obtain the process constraint penalty function. Specifically, weighting coefficients are set for the thickness range constraint and the thickness gradient constraint, respectively. , Integrating the two penalty terms yields the process constraint penalty function. ,Right now: ; Weighting coefficients for thickness range constraints and thickness gradient constraints , The weighting should be adjusted based on the feasibility of the process: increase the weighting if the focus is on process feasibility, and decrease the weighting if the focus is on dosage accuracy. Process constraint penalty function With objective function By integrating, we obtain the optimized function. Specifically, the calculation formula is as follows: ; in, This represents the global squared deviation between the target dose and the actual dose in the target area, and should be minimized as much as possible. The penalty value for compensator process violations should be minimized as much as possible; The constraints also include dose constraints, which include target dose rate constraints. Dosage uniformity constraints Dosage conformity constraints and critical organ dose constraint This refers to the safe dose threshold for each organ at risk.

[0025] In one embodiment of the present invention, the steps of updating the beam intensity according to the optimization function, solving for the compensator thickness according to the updated beam intensity, and verifying the compensator thickness according to the constraint conditions include: S31: Based on the optimization function, calculate the partial derivative of the optimization function with respect to the beam intensity using the gradient calculation formula. The partial derivative is used as the gradient of the optimization function with respect to the beam intensity. S32: Based on the gradient direction and the preset learning rate, iteratively update the current beam intensity to obtain the updated beam intensity; S33: Based on the updated beam intensity, the compensator thickness is iteratively solved using a quasi-Newton algorithm; S34: Verify the thickness of the compensator obtained by solving according to the thickness range constraint and thickness gradient constraint in the constraint conditions; When the compensator thickness simultaneously satisfies both the thickness range constraint and the thickness gradient constraint, the parameters of the current beam intensity and the compensator thickness are retained. When the compensator thickness is different and the thickness range constraint and thickness gradient constraint are satisfied, the calculated compensator thickness is fed back to the beam intensity gradient calculation, and the beam intensity is updated and the compensator thickness is recalculated until the verification is passed.

[0026] The step of iteratively solving the compensator thickness using a quasi-Newton algorithm based on the updated beam intensity includes: S331: Based on the updated beam intensity and optimization function, the gradient of the optimization function with respect to the compensator thickness is solved using a numerical approximation method. S332: Using a quasi-Newton algorithm, the inverse matrix of the current iteration is updated based on the parameters of the previous iteration; S333: The compensator thickness is iteratively updated using the gradient and inverse matrix of the compensator thickness; S334: Set the convergence threshold and the maximum number of iterations, calculate the difference between the compensator thickness in the current round and the previous round. If the difference is less than the convergence threshold, or the number of iterations reaches the maximum number of iterations, the iteration is determined to be converged, and the compensator thickness is output. If the iteration is determined to be non-converged, return to the optimization function to calculate the gradient of the compensator thickness and continue iterating until the iteration converges.

[0027] In this embodiment of the invention, as shown in steps S31-S34 above, the partial derivative of the optimization function with respect to the beam intensity is calculated based on the optimization function, specifically including: according to the optimization function... To optimize the objective, a policy gradient algorithm is employed to calculate the optimization function with respect to beam intensity. The gradient is obtained by taking the partial derivatives. The gradient calculation formula is as follows: ; Based on the gradient direction and the preset learning rate The current beam intensity is iteratively updated, and the calculation formula for the iterative update is: ; The updated beam intensity is calculated based on the iteratively updated calculation formula. Gradient updates are used to align the beam intensity towards the optimization function. Adjust the direction of minimization; Utilizing the updated beam intensity Using this as a fixed input, the BFGS quasi-Newton algorithm is employed to iteratively solve for the compensator thickness. Similarly, towards optimizing the function The direction adjustment is minimized, specifically by using a numerical approximation method to solve for the gradient of the optimization function with respect to the compensator thickness, based on the updated beam intensity and optimization function. The gradient calculation formula is: ; in, This is an approximate gradient value for the compensator thickness. This represents the total optimized function value under the current beam intensity S and compensator thickness T. To keep the beam intensity S constant, the thickness of the compensator is slightly varied from T. Then, the new overall optimization function value, This represents a small change in the thickness of the compensator. The BFGS quasi-Newton algorithm is adopted, based on the inverse of the Hessian matrix from the previous iteration. Iteration step size and gradient difference The inverse of the Hessian matrix for the current round is updated. The updated formula is: ; in, This is the inverse approximation of the Hessian matrix after the (t+1)th iteration following the update. Let be the approximate inverse of the Hessian matrix at the t-th iteration. Let t be the update step size for the compensator thickness in round t. Let be the gradient difference vector in round t. They are , transpose; Based on gradient Inverse of the Hessian matrix The thickness of the compensator is iteratively updated using the following formula: ; After updating the compensator thickness, calculate the difference between the current and previous compensator thicknesses. If the difference is less than the convergence threshold, or the maximum number of iterations is reached, the iteration is considered converged, and the final compensator thickness is output. If convergence has not occurred, return to the optimization function to adjust the compensator thickness. The gradient calculation continues iteratively; After obtaining the compensator thickness through iteration, the calculated compensator thickness is then adjusted according to the thickness range constraint and the thickness gradient constraint. Perform a check to determine if the compensator thickness meets the requirements. If the thickness is less than Correct it to If the thickness is greater than Correct it to ; Calculate the thickness gradient of the compensator Determine whether the condition is met. If the gradient exceeds the limit, a smoothing algorithm can be used to adjust the gradient to within the threshold range. Specifically, the algorithm can use a Gaussian smoothing algorithm. When the thickness gradient exceeds the limit, local Gaussian filtering is performed on the exceeding region. By weighted averaging of the neighborhood thickness value, the rate of change of the thickness is reduced, so that the thickness gradient returns to the allowable range of the process.

[0028] In one embodiment of the present invention, the steps of recalculating the actual dose based on the optimized beam intensity and compensator thickness, determining whether the dose deviation between the target dose and the actual dose satisfies dose convergence, and outputting the optimized scheme based on whether the compensator thickness satisfies the constraint conditions include: S41: Based on the optimized beam intensity and compensator thickness, the actual dose is calculated, and combined with the target dose in the target area, the dose deviation value is calculated. S42: According to the dose constraint in the constraint conditions, when the actual dose and the dose deviation value meet the dose constraint, it is determined that the dose converges; S43: Based on the thickness range constraint and thickness gradient constraint of the compensator in the constraint conditions, the thickness of the compensator is checked respectively. If all the constraint conditions are met, the thickness of the compensator is determined to pass. S44: Generate an optimized scheme based on the determination results of dose convergence and compensator thickness verification.

[0029] In this embodiment of the invention, the final actual dose data of the target area is calculated based on the beam intensity and compensator thickness data obtained by common iterative optimization, and the dose deviation value is calculated based on the target dose data of the target area; dose convergence verification is performed according to the dose constraints in the constraints, and compensator thickness verification is performed according to the thickness range constraints and thickness gradient constraints of the compensator; based on the dose convergence verification and compensator thickness verification, different optimization schemes are output according to different scenarios. Specifically, if both the dose convergence verification and the compensator thickness verification pass, the current optimization parameters are determined to be feasible and can be directly used for clinical radiotherapy planning and compensator fabrication. If the dose convergence verification passes but the compensator thickness verification fails, the FLASH radiotherapy dose is deemed clinically feasible, but the compensator is deemed infeasible. No optimized solution is output, and the compensator's violation is used as feedback data for readjustment. If the dose convergence verification fails but the compensator thickness verification passes, the compensator solution is deemed feasible, but the dose has deviations and cannot be used for clinical treatment. The beam intensity gradient is then updated again. If both the dose convergence verification and the compensator thickness verification fail, the optimized solution is deemed infeasible. The initial parameters of the iteration are reset, and a full-process optimization design is performed to obtain a solution that can pass verification.

[0030] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for the correction and compensation of FLASH radiotherapy doses, characterized in that, Includes the following steps: S1: Initialize the beam intensity and compensator thickness within the radiation field, obtain the target dose in the target area based on the FLASH radiotherapy physical-biological combined dose correction model, calculate the actual dose based on the current beam intensity and current compensator thickness, and construct the dose deviation objective function between the target dose and the actual dose; S2: Determine the constraints, construct the process constraint penalty function for the compensator thickness based on the thickness range constraint and thickness gradient constraint, and integrate the process constraint penalty function with the dose deviation objective function to generate the optimization function; S3: Update the beam intensity according to the optimization function, and solve the compensator thickness according to the updated beam intensity. Verify the compensator thickness according to the constraint conditions. When the compensator thickness does not meet the constraint conditions, feed back the solved compensator thickness to the beam intensity update step, and re-update the beam intensity and solve the compensator thickness. S4: Recalculate the actual dose based on the optimized beam intensity and compensator thickness, determine whether the dose deviation between the target dose and the actual dose meets the dose convergence requirement, and output the optimization scheme based on whether the compensator thickness meets the constraint conditions.

2. The method for correcting FLASH radiotherapy dose and co-optimizing compensators according to claim 1, characterized in that, The steps of initializing the beam intensity and compensator thickness within the radiation field, obtaining the target dose in the target area based on the FLASH radiotherapy physical-biological combined dose correction model, calculating the actual dose based on the current beam intensity and current compensator thickness, and constructing a dose deviation objective function between the target dose and the actual dose include: The target dose in the target area is obtained by using the FLASH radiotherapy physical-biological combined dose correction model. Based on the current beam intensity and the current compensator thickness, the energy deposition of the radiation penetrating the compensator and the human body is calculated to obtain the actual dose. Minimizing the deviation between the target dose and the actual dose is taken as the optimization objective, and a dose deviation objective function is constructed. The calculation formula is: ； in, Let be the target function value for dose deviation, representing the sum of squared deviations between the target dose and the actual dose in the target area under the current beam intensity S and compensator thickness T. The two-dimensional spatial coordinates of the target area Two-dimensional spatial coordinates of the target area The target dose, Two-dimensional spatial coordinates of the target area The actual dose.

3. The method of FLASH radiotherapy dose modification and compensator co-optimization of claim 1, wherein, The step of determining the constraints, constructing a process constraint penalty function for the compensator thickness based on the compensator's thickness range constraint and thickness gradient constraint, and integrating the process constraint penalty function with the dose deviation objective function to generate an optimization function includes: The constraints include the thickness range constraint and the thickness gradient constraint of the compensator; The thickness range constraint includes the minimum machinable thickness of the compensator. and maximum machinable thickness ,but ; The thickness gradient constraint includes determining the maximum permissible variation in thickness between adjacent regions. ,but , For the compensator thickness in the target area gradient at; The constraints are transformed into penalty terms, and a process constraint penalty function for obtaining the compensator thickness is constructed. ; Process constraint penalty function Objective function of dose deviation By integrating the results, we obtain the optimized function. .

4. The method for correcting FLASH radiotherapy dose and co-optimizing compensators according to claim 3, characterized in that, The process constraint penalty function for compensator thickness is constructed by transforming the constraint conditions into penalty terms. The steps include: The penalty term includes a thickness range penalty term and a thickness gradient penalty term; Regarding the thickness range penalty, for thicknesses exceeding the minimum machinable thickness... Or maximum machinable thickness The thickness value is calculated, the deviation is summed, and the deviation is calculated. When there is no penalty term; when or At times, there are penalties, and there are violations or deviations. For the thickness gradient penalty term, a deviation is calculated and summed for thickness gradients that exceed a threshold. When there is no penalty term; when At this time, there is a penalty term; the larger the gradient, the heavier the penalty. By setting weighting coefficients and integrating the thickness range penalty term and the thickness gradient penalty term, a process constraint penalty function is obtained. By setting the weighting coefficients for thickness range constraints and thickness gradient constraints respectively. , Integrating the two penalty terms yields the process constraint penalty function. The calculation formula is: ； in, For thickness constraint penalty term, This is a gradient constraint penalty term.

5. The method of FLASH radiotherapy dose modification and compensator co-optimization of claim 1, wherein: The constraints also include dose constraints, which include target dose rate constraints, dose uniformity constraints, dose conformity constraints, and critical organ dose constraints; the target dose rate constraints... The dose uniformity constraint The dose conformity constraint The critical organ dose constraint This refers to the safe dose threshold for each organ at risk.

6. The method for correcting FLASH radiotherapy dose and co-optimizing compensators according to claim 1, characterized in that, The steps of updating the beam intensity according to the optimization function, solving for the compensator thickness based on the updated beam intensity, and verifying the compensator thickness according to the constraints include: Based on the optimization function, the partial derivative of the optimization function with respect to the beam intensity is calculated using the gradient calculation formula. The partial derivative is used as the gradient of the optimization function with respect to the beam intensity. Based on the gradient direction and the preset learning rate, the current beam intensity is iteratively updated to obtain the updated beam intensity; Based on the updated beam intensity, the compensator thickness is iteratively solved using a quasi-Newton algorithm. Based on the thickness range constraint and thickness gradient constraint in the constraint conditions, the thickness of the compensator obtained by the solution is verified respectively. When the compensator thickness simultaneously satisfies both the thickness range constraint and the thickness gradient constraint, the parameters of the current beam intensity and the compensator thickness are retained. When the compensator thickness is different and the thickness range constraint and thickness gradient constraint are satisfied, the calculated compensator thickness is fed back to the beam intensity gradient calculation, and the beam intensity is updated and the compensator thickness is recalculated until the verification is passed.

7. The method for correcting FLASH radiotherapy dose and co-optimizing compensators according to claim 1, characterized in that, The step of feeding back the calculated compensator thickness to the beam intensity update step when the compensator thickness does not meet the constraint conditions, and then re-performing the beam intensity update and compensator thickness calculation, includes: Based on the updated beam intensity and optimization function, the gradient of the optimization function with respect to the compensator thickness is solved using a numerical approximation method. The calculation formula is as follows: ； in, This is an approximate value for the gradient of the compensator thickness. This represents the total optimized function value under the current beam intensity S and compensator thickness T. To keep the beam intensity S constant, the thickness of the compensator is slightly varied from T. Then, the new overall optimization function value, This represents a small change in the thickness of the compensator. Using a quasi-Newton algorithm, the inverse matrix of the current iteration is updated based on the parameters of the previous iteration. The calculation formula is as follows: ； in, This is the inverse approximation of the Hessian matrix after the (t+1)th iteration following the update. Let be the approximate inverse of the Hessian matrix at the t-th iteration. Let t be the update step size for the compensator thickness in round t. Let be the gradient difference vector in round t. They are , transpose; The compensator thickness is iteratively updated using the gradient and inverse matrix of the compensator thickness. Set a convergence threshold and a maximum number of iterations. Calculate the difference between the compensator thickness in the current iteration and the previous iteration. If the difference is less than the convergence threshold, or the number of iterations reaches the maximum number of iterations, the iteration is considered converged, and the compensator thickness is output. If the iteration is not converged, return to the optimization function to calculate the gradient of the compensator thickness and continue iterating until the iteration converges.

8. The method for correcting FLASH radiotherapy dose and co-optimizing compensators according to claim 1, characterized in that, The steps of recalculating the actual dose based on the optimized beam intensity and compensator thickness, determining whether the dose deviation between the target dose and the actual dose satisfies dose convergence, and outputting the optimized scheme based on whether the compensator thickness meets the constraint conditions include: Based on the optimized beam intensity and compensator thickness, the actual dose is calculated, and combined with the target dose in the target area, the dose deviation value is calculated. According to the dose constraint in the constraint conditions, when the actual dose and the dose deviation value meet the dose constraint, it is determined that the dose converges. Based on the thickness range constraint and thickness gradient constraint of the compensator in the constraint conditions, the thickness of the compensator is checked respectively. If all the constraint conditions are met, the thickness of the compensator is determined to be acceptable. An optimized scheme is generated based on the results of dose convergence and compensator thickness verification.

Images