Multi-center radiotherapy plan generation method and device, electronic equipment and computer readable storage medium

Abstract

This invention provides a method, apparatus, electronic device, and computer-readable storage medium for generating multi-center radiotherapy plans. The method includes: receiving coordinates, field parameters, positioning methods, and unified dose constraints from multiple treatment centers; receiving medical images of the patient, target area, and contours of organs at risk; establishing a multi-center collaborative optimization model; optimizing the multi-center collaborative optimization model according to the unified dose constraints to generate a total radiotherapy plan containing a three-dimensional dose distribution; dividing the total plan into several sub-plans; performing dose verification on the sub-plans; and outputting the dose-verified sub-plans as executable sub-plans to the treatment equipment. This invention optimizes the multi-center collaborative optimization model based on unified dose constraints to generate a total radiotherapy plan containing a three-dimensional dose distribution. The plan generation process eliminates the need for repeatedly setting dose constraints and performing multiple positioning verifications, thereby shortening radiotherapy time and reducing the risk of dose deviation. Furthermore, unified optimization across multiple treatment centers reduces the risk of dose interference between treatment centers and avoids the formation of hot and cold spots due to uneven dose superposition, which could affect the safety of organs at risk.

Description

Technical Field

[0001] This invention belongs to the field of medical radiotherapy technology, specifically relating to a method, device, electronic equipment, and computer-readable storage medium for generating multi-center radiotherapy plans. It achieves global dose control, automated plan splitting, and efficient execution through multi-center collaborative optimization. Background Technology

[0002] In radiotherapy, multicenter treatment protocols are required for cases with multiple or long target areas. Commonly used multicenter treatment protocols include manual splitting of the treatment plan, step-by-step overlay optimization, and treatment zone segmentation under mechanical limitations. Manually splitting treatment plans requires physicists to design individual plans for each target center, setting target dose constraints and organ-at-risk protection parameters separately. This process necessitates repeated adjustments to field parameters (such as gantry angle and MLC sequence), taking hours or even days. For example, in the treatment of bilateral breast cancer, intensity-modulated radiotherapy (IMRT) plans must be designed independently for the left and right target areas before dose aggregation, resulting in complex and inefficient procedures.

[0003] Step-by-step optimization, using conventional radiotherapy methods, first optimizes a single isocenter plan and then superimposes the dose contributions from other isocenters. For example, Monte Carlo algorithms can be used to optimize the radiation field parameters for different target areas in stages. However, step-by-step optimization is prone to dose interference risks. For instance, uneven dose superposition may create hot and cold spots, affecting the safety of organs at risk.

[0004] Treatment area segmentation under mechanical limitations is constrained by equipment hardware (e.g., gantry rotation angle ≤180°, treatment bed movement accuracy ±1 mm). A single treatment session cannot cover ultra-long target areas or spatially dispersed multiple target areas (e.g., spinal whole-center irradiation requires three centers: intracranial, upper spine, and lower spine). However, multi-planning requires multiple setup verifications, which may lead to setup errors (e.g., cranio-coccygeal direction deviation ±3 mm), significantly affecting dose distribution.

[0005] In summary, existing multicenter treatment techniques suffer from the following technical problems: 1. Inefficiency: Manual splitting requires repeated dose constraint setting, lacks global optimization algorithm support, and requires multiple setup verifications for multiple treatment plan executions, significantly increasing execution time and dose deviation risk. 2. Uneven dose distribution: Step-by-step optimization leads to the superposition of hot and cold spots in the target area. 3. Insufficient protection of endangered organs: When multiple plans are optimized independently, the cumulative dose is prone to exceed the limit. Summary of the Invention

[0006] To address at least one of the aforementioned technical problems in existing multicenter radiotherapy techniques, the first aspect of this invention provides a method for generating multicenter radiotherapy plans, the detailed technical solution of which is as follows: A method for generating a multicenter radiotherapy plan includes: It receives coordinates, field parameters, positioning methods, and uniform dose constraints from multiple treatment centers, as well as medical images, target areas, and organ-at-risk outlines of patients. Establish a multi-center collaborative optimization model; A unified dose-constrained multi-center collaborative optimization model is used to generate a total radiotherapy plan that includes a three-dimensional dose distribution. The overall plan is broken down into several sub-plans; The sub-plans are dose-validated, and the dose-validated sub-plans are output as executable sub-plans to the treatment device.

[0007] In some embodiments, the field parameters include field angle, radiation energy, field size, field weight, gantry angle, and MLC sequence; Medical images include CT images, MRI images, and PET images; The multi-center collaborative optimization model calculates the total dose distribution based on the following formula: , Where n is the number of field of fire generated by all treatment centers. Let i be the weight of the i-th shooting field. Let be the dose distribution for the i-th radiation field; Optimize the multi-center collaborative optimization model, with the objective function as follows: , Where m is the number of target regions. Let be the weight coefficient of the j-th target region. For the dose of the j-th target region, Let p be the dose limit for the j-th target region, and p be the number of organs at risk. The weighting coefficient for the o-th organ at risk. For the dose of the oth organ at risk, For the dose limit of the j-th organ at risk, Here, n is the number of shooting fields, and n is the L1 regularization term. For group regularization, C is the number of treatment centers. The number of radiation fields for the l-th treatment center.

[0008] In some embodiments, optimizing the multi-center collaborative optimization model includes: The L-BFGS-B optimization engine was used to optimize the multi-center collaborative optimization model. During the optimization process, Pareto front analysis was used to dynamically adjust the field parameters of each treatment center. The three-dimensional dose distribution is updated in real time through GPU-accelerated dose calculation.

[0009] In some embodiments, optimizing the multi-center collaborative optimization model includes: automatically reducing the weight of the treatment center associated with the organ at risk when the cumulative dose to the organ at risk exceeds a preset threshold.

[0010] In some embodiments, dividing the total plan into several sub-plans includes dividing the total plan into several sub-plans based on the mechanical parameters of the equipment.

[0011] In some embodiments, dose verification of several sub-plans includes: verifying the sub-plan dose using a GPU-accelerated Monte Carlo algorithm, comparing the planned dose with the actual irradiation dose in terms of Gamma pass rate, and the executable sub-plan is a sub-plan with a Gamma pass rate (3% / 3mm) exceeding 95%.

[0012] In some embodiments, the medical images are multimodal images, and after receiving the patient's medical images, target area, and organ at risk contours, the multicenter radiotherapy planning method further includes fusing the multimodal images using rigid and / or non-rigid registration techniques.

[0013] A second aspect of the present invention provides a multicenter radiotherapy planning device, comprising: The receiving module is used to receive the coordinates of multiple treatment centers, field parameters, positioning methods and unified dose constraints, as well as the patient's medical images, target area and organ at risk outlines; The modeling module is used to build multi-center collaborative optimization models. The plan generation module is used to optimize a multi-center collaborative optimization model based on unified dose constraints and generate a total radiotherapy plan that includes a three-dimensional dose distribution. The plan splitting module is used to split the overall plan into several sub-plans; The validation module is used to validate the dosage of the sub-plan and output the sub-plan that has passed the dosage validation as an executable sub-plan to the treatment device.

[0014] A third aspect of the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the program, implements the multicenter radiotherapy planning generation method described in any of the preceding claims.

[0015] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the interactive radiotherapy planning generation method described in any of the preceding claims.

[0016] This invention optimizes a multi-center collaborative optimization model based on unified dose constraints to generate a comprehensive radiotherapy plan containing a three-dimensional dose distribution. The plan generation process eliminates the need for repeatedly setting dose constraints and performing multiple setup verifications, thereby shortening radiotherapy time and reducing the risk of dose deviation. Furthermore, unified optimization across multiple treatment centers reduces the risk of dose interference between centers, preventing hotspots and colds caused by uneven dose superposition, which could compromise organ safety. Attached Figure Description

[0017] Figure 1 A schematic diagram of a single or multiple target areas with multiple treatment centers; Figure 2 This is an execution flowchart of the multicenter radiotherapy plan generation method in an embodiment of the present invention; Figure 3 This is an execution flowchart of the multicenter radiotherapy plan generation method in an embodiment of the present invention; Figure 4 This is a block diagram of a multicenter radiotherapy planning and production apparatus according to an embodiment of the present invention. Detailed Implementation

[0018] It should be noted that the following detailed descriptions are exemplary and intended to provide indicative explanations of the invention. It should be observed that all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which the invention pertains.

[0019] The system architecture of the present invention and the solutions in the prior art will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. It should be noted that the described embodiments are only for explanation and illustration of the present invention, and not all of the contents. Based on the embodiments provided by the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of the present invention.

[0020] Example 1 like Figure 2 and Figure 3 As shown, a multicenter radiotherapy plan generation method of the present invention includes the following steps: S1 receives coordinates, field parameters, positioning methods, and uniform dose constraints from multiple treatment centers, as well as medical images of the patient, target area, and outlines of organs at risk.

[0021] S2. Establish a multi-center collaborative optimization model.

[0022] S3. Based on a unified dose constraint, optimize the multi-center collaborative optimization model to generate a total radiotherapy plan that includes a three-dimensional dose distribution.

[0023] S4. Divide the overall plan into several sub-plans.

[0024] S5. Perform dose validation on the sub-plan and output the dose-validated sub-plan as an executable sub-plan to the treatment device.

[0025] The location of the treatment center in step S1 may include methods such as bone localization, gold standard localization, and soft tissue localization. Dose constraints are the dose limits for the target volume (PTV) and organs at risk.

[0026] For example, in whole-central spinal irradiation, the dose constraint is set to: PTV D95 ≥60 Gy, spinal cord D max ≤35 Gy. That is, at least 95% of the PTV volume receives a radiation dose of not less than 60 Gy, and the dose received by any point in the spinal cord must not exceed 35 Gy.

[0027] For example, in the treatment of nasopharyngeal carcinoma with primary lesion invading the skull base and bilateral cervical lymph node metastasis, it is necessary to treat both intracranial and cervical target areas simultaneously. Dose limits should be set for the target areas and organs at risk. Dose limit rings should be set for each target area with distances of 0.2 cm, 0.5 cm, and 0.9 cm, respectively, and a uniform inner diameter of 0.2 cm. Gradually decreasing dose limits should be set to control the rate of dose drop.

[0028] As can be seen, the multi-center radiotherapy planning method of the present invention optimizes the multi-center collaborative optimization model based on unified dose constraints to generate a total radiotherapy plan containing a three-dimensional dose distribution. The plan generation process eliminates the need for repeatedly setting dose constraints and performing multiple setup verifications, thereby shortening radiotherapy time and reducing the risk of dose deviation. Furthermore, unified optimization across multiple treatment centers reduces the risk of dose interference between individual centers, avoiding the formation of hot and cold spots due to uneven dose superposition, which could affect the safety of organs at risk.

[0029] Example 2 like Figure 2 and Figure 3 As shown, a multicenter radiotherapy plan generation method of the present invention includes the following steps: S1 receives coordinates, field parameters, positioning methods, and uniform dose constraints from multiple treatment centers, as well as medical images of the patient, target area, and outlines of organs at risk.

[0030] S2. Establish a multi-center collaborative optimization model.

[0031] S3. Based on a unified dose constraint, optimize the multi-center collaborative optimization model to generate a total radiotherapy plan that includes a three-dimensional dose distribution.

[0032] S4. Divide the overall plan into several sub-plans.

[0033] S5. Perform dose validation on the sub-plan and output the dose-validated sub-plan as an executable sub-plan to the treatment device.

[0034] The difference between this embodiment and Embodiment 1 is that: In step S1 above, the radiation field parameters include the radiation field angle, radiation energy, radiation field size (e.g., length and width dimensions), and MLC sequence. The patient's medical images can be one or more medical images acquired by various existing medical imaging devices, such as CT images, MRI images, and PET images. CT images focus on displaying anatomical structures, MRI images focus on displaying the target volume television (PTV) boundary, and PET images focus on displaying metabolically active areas. The fusion of these three images can further improve the accuracy of target volume delineation.

[0035] Therefore, when the received medical images of a patient are multimodal images including two or more of the following: CT images, MRI images, and PET images, the multimodal images can be fused using rigid and / or non-rigid registration techniques.

[0036] Before administering radiation therapy, doctors or physicists typically outline the target area and organs at risk on medical images. In addition, they may draw one or more custom-designed auxiliary radiation therapy structures on the medical images, such as dose drop zones near the target area.

[0037] The multi-center collaborative optimization model in step S2 calculates the total dose distribution based on the following formula: , Where n is the number of field of fire generated by all treatment centers. The dose weight of the target area corresponding to the i-th radiation field. Let be the dose distribution for the i-th radiation field.

[0038] The objective function of the optimized multi-center collaborative optimization model in step S3 is: , Where m is the number of target regions. Let be the weight coefficient of the j-th target region. For the dose of the j-th target region, Let p be the dose limit for the j-th target region, and p be the number of organs at risk. The weighting coefficient for the o-th organ at risk. For the dose of the oth organ at risk, For the dose limit of the j-th organ at risk, Here, n is the number of shooting fields, and n is the L1 regularization term. For group regularization, C is the number of treatment centers. The number of radiation fields for the l-th treatment center.

[0039] In other words, the objective function is to minimize the sum of the excess dose to the target area and the excess dose to organs at risk. The L1 regularization term is used to compress the weight of low-contribution fields to zero, thereby automatically eliminating redundant fields. Low-contribution fields are, for example, fields with a dose weight of less than 0.01. The group regularization term is used to reduce the number of treatment centers.

[0040] Step S3, optimizing the multi-center collaborative optimization model, specifically includes: The L-BFGS-B optimization engine was used to optimize the multi-center collaborative optimization model. During the optimization process, Pareto front analysis was used to dynamically adjust the field parameters (including field angle, radiation energy, and field size) of each treatment center.

[0041] The three-dimensional dose distribution is updated in real time through GPU-accelerated computation.

[0042] Specifically, the L-BFGS-B optimization engine combines L-BFGS (Limited-Memory BFGS) and boundary constraint handling techniques. It boasts advantages such as low memory footprint, support for explicit boundary constraints, and fast convergence, reducing the number of iterations and enabling rapid optimization of multi-center collaborative optimization models. Pareto front analysis, a mathematical tool for multi-objective optimization problems, can find the optimal balance between multiple conflicting objectives. In the dynamic adjustment of radiotherapy field parameters (such as field angle, radiation energy, and field size), it can balance the conflicts among multiple objectives in field parameter optimization. Adjustments require balancing multiple clinical goals. For example, tumor control aims to maximize dose coverage and homogeneity of the tumor target area; organ-at-risk protection aims to minimize the radiation dose to organs at risk (such as the heart and spinal cord); and dose distribution conformity aims to optimize the fit between the field shape and tumor geometry. These objectives are often contradictory (e.g., increasing the tumor dose may increase the risk of organ-at-risk damage), and Pareto front analysis can be used to find a better compromise.

[0043] Compared to CPU computation, GPU-accelerated computation can increase the speed of Monte Carlo dose simulation by 50-200 times, thereby supporting real-time dynamic adjustment of radiotherapy plans.

[0044] To further enhance the safety of organs at risk, in this embodiment, the optimized multi-center collaborative optimization model in step S3 further includes: When the cumulative dose to an organ at risk exceeds a preset threshold, the weight of the treatment center associated with that organ at risk is automatically reduced, which means the radiation dose to that organ at risk is reduced.

[0045] Example 3 like Figure 2 and Figure 3 As shown, a multicenter radiotherapy plan generation method of the present invention includes the following steps: S1 receives coordinates, field parameters, positioning methods, and uniform dose constraints from multiple treatment centers, as well as medical images of the patient, target area, and outlines of organs at risk.

[0046] S2. Establish a multi-center collaborative optimization model.

[0047] S3. Based on a unified dose constraint, optimize the multi-center collaborative optimization model to generate a total radiotherapy plan that includes a three-dimensional dose distribution.

[0048] S4. Divide the overall plan into several sub-plans.

[0049] S5. Perform dose validation on the sub-plan and output the dose-validated sub-plan as an executable sub-plan to the treatment device.

[0050] The difference between this embodiment and Embodiment 1 is that: Step S4, which involves breaking down the overall plan into several sub-plans, includes: The overall plan is divided into several sub-plans based on the mechanical parameters of the equipment.

[0051] For example, the overall plan for central spinal irradiation is divided into several three-sub-plans based on the range of motion of the treatment bed (±40 cm), with the three sub-plans covering the cervical, thoracic, and lumbar vertebrae respectively.

[0052] Step S5 involves dose validation of the sub-plan, including: The sub-plan dose was verified using a GPU-accelerated Monte Carlo algorithm. The Gamma pass rate was compared between the planned dose and the actual dose, and sub-plans with a Gamma pass rate of over 95% were selected as executable sub-plans.

[0053] In this embodiment, before outputting the dose-verified sub-plan as an executable sub-plan to the treatment device in step S5, the execution order of the executable sub-plan is optimized.

[0054] Optional optimization methods include, for example: Mechanical motion optimization: The execution sequence of the planned sub-plans is limited by the gantry angle to avoid the cumulative error caused by frequent movement of the treatment bed (e.g., the error in the craniotail direction is reduced from ±3 mm to ±1 mm). A path optimization algorithm is used to automatically generate the shortest path for the treatment bed to improve the efficiency and accuracy of multi-treatment plan execution.

[0055] After the dose-validated sub-plans are output as executable sub-plans to the treatment equipment, formal radiotherapy can be administered to the patient according to the execution sequence of the executable sub-plans. The radiotherapy process is as follows: Secure the patient to the treatment bed. Position the patient automatically according to the positioning method of Sub-Plan 1, and then execute Sub-Plan 1 using the pre-defined tracking method. After Sub-Plan 1 treatment is completed, the system automatically moves the patient to the positioning position of Sub-Plan 2 and executes Sub-Plan 2 using the pre-defined tracking method. Continue executing all sub-plans according to the above steps.

[0056] To enable those skilled in the art to more clearly understand the execution process of the multicenter radiotherapy planning method of the present invention, the execution process of the multicenter radiotherapy planning method of the present invention will be described in more detail below through Examples 4 and 5.

[0057] Example 4 Case characteristics: Nasopharyngeal carcinoma, primary lesion invading the skull base, bilateral cervical lymph node metastasis, requiring simultaneous treatment of intracranial and cervical target areas.

[0058] During the data acquisition (input) phase, sub-millimeter registration is achieved through multimodal image fusion (such as 1mm slice-thickness CT, T1WI / T2WI MRI and PET-CT). Operators delineate the target area (GTV / CTV / PTV) and the contours of organs at risk on the computer system and evaluate the delineation results. Among them, MRI / PET-CT is used to correct complex boundaries such as the skull base invasion area.

[0059] Skeletal tracking is used for localization of the skull region, while soft tissue tracking is used for localization of lymph node metastases in the neck. Localization centers, localization modes, and different treatment node modes are set for each region.

[0060] Set dose limits for target areas and organs at risk. Set dose limit rings for each target area at distances of 0.2 cm, 0.5 cm and 0.9 cm, respectively, with a uniform inner diameter of 0.2 cm. Set gradually decreasing dose limits to control the rate of dose drop.

[0061] The treatment plan is optimized to obtain a total radiotherapy plan, which is then broken down into several sub-plans.

[0062] After treatment assessment, based on the established positioning center and the remaining radiation fields, the obtained overall radiotherapy plan is divided into three sub-plans. Subsequently, a path optimization algorithm (such as a genetic algorithm) is used to plan the movement path of the treatment bed and determine the execution order of the three sub-plans, namely Sub-plan 1, Sub-plan 2, and Sub-plan 3.

[0063] Sub-plan 1, sub-plan 2, and sub-plan 3 are sent to the treatment device for implementation. The operator only needs to perform the initial positioning of the patient for sub-plan 1 and carry out the treatment for sub-plan 1. Then the system automatically performs the positioning of the patient for sub-plan 2 and carries out the treatment for sub-plan 2. Subsequently, the system automatically performs the positioning of the patient for sub-plan 3 and carries out the treatment for sub-plan 3.

[0064] Example 5 The patient requires full spinal irradiation, with the target area covering the cervical, thoracic, and lumbar vertebrae, exceeding the total length of conventional single-center treatment. Traditional approaches require three treatment centers (intracranial, upper spine, and lower spine), resulting in low treatment efficiency due to multiple treatment plans and separate treatments; sequential treatment with multiple positioning can lead to cumulative errors (±2 mm). The implementation process of this embodiment is as follows.

[0065] Data input: Receive the coordinates and field parameters of 5 treatment centers, some of which are: Initial values ​​for field weights: 0.2 for treatment center 1, 0.2 for treatment center 2, 0.2 for treatment center 3, 0.2 for treatment center 4, and 0.2 for treatment center 5.

[0066] Import the patient's CT / MRI fusion images to delineate the target areas (cervical spine, thoracic spine, lumbar spine) and organs at risk (spinal cord, lungs, kidneys).

[0067] Multi-center collaborative optimization: A global optimization model is established, with the objective function as described in Example 2 above.

[0068] Dose constraints: At least 95% of the target volume must receive a radiation dose of not less than 60 Gy, and no point in the spinal cord must receive a dose exceeding 35 Gy.

[0069] L1 regularization term (λ1=0.1): Automatically compresses the weight of most of the treatment field in the treatment center to 0 (dose contribution < 0.01), reducing redundancy and increasing treatment efficiency.

[0070] Group regularization term (λ2=0.5): Reduces the number of treatment centers from 5 to 3, simplifying the execution process.

[0071] Plan breakdown and verification: Based on the range of motion of the treatment bed (±10 cm), the overall plan is divided into 3 sub-plans, namely: Sub-plan 1: Cervical spine (center 2) Sub-plan 2: Thoracic vertebrae (center 4) Sub-plan 3: Lumbar spine (Center 5).

[0072] GPU-accelerated Monte Carlo verification: The Gamma pass rate (3% / 3mm) of sub-project 1 was 97.2%, that of sub-project 2 was 96.8%, and that of sub-project 3 was 98.1%, all of which were greater than 95%.

[0073] Sub-plan 1, sub-plan 2, and sub-plan 3 are sent to the treatment device for implementation. The operator only needs to perform the initial positioning of the patient for sub-plan 1 and carry out the treatment for sub-plan 1. Then the system automatically performs the positioning of the patient for sub-plan 2 and carries out the treatment for sub-plan 2. Subsequently, the system automatically performs the positioning of the patient for sub-plan 3 and carries out the treatment for sub-plan 3.

[0074] Technical benefits: Single-session optimization time was reduced from the traditional 4 hours to 40 minutes (GPU acceleration); the cumulative spinal cord dose decreased from 32 Gy in the traditional protocol to 28.5 Gy (below the limit); and positioning error was reduced by 50% (per treatment session). Treatment plan execution time was reduced from 150 minutes to 90 minutes due to the reduction in treatment centers.

[0075] Example 6 like Figure 4 As shown, a multicenter radiotherapy planning device of the present invention includes: The receiving module 1 is used to receive the coordinates of multiple treatment centers, field parameters, positioning methods and unified dose constraints, as well as the patient's medical images, target area and organ at risk outlines. Modeling module 2 is used to establish a multi-center collaborative optimization model; The plan generation module 3 is used to optimize a multi-center collaborative optimization model based on a unified dose constraint and generate a total radiotherapy plan that includes a three-dimensional dose distribution. The plan splitting module 4 is used to split the overall plan into several sub-plans; The verification module 5 is used to verify the dosage of the sub-plan and output the sub-plan that has passed the dosage verification as an executable sub-plan to the treatment device.

[0076] Since the processing procedures of each functional module of the multicenter radiotherapy planning device in this embodiment are consistent with the processing procedures of the multicenter radiotherapy planning methods provided in the foregoing method embodiments, this embodiment will not repeat the detailed processing procedures of each functional module of the interactive radiotherapy planning device. You can directly refer to the relevant content in the foregoing method embodiments.

[0077] Example 7 An electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor and the memory are connected, for example, via a bus. When the processor executes the program, it implements the multicenter radiotherapy planning method of any of the above embodiments.

[0078] Example 8 A computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the multicenter radiotherapy planning method provided in the preceding method embodiments.

[0079] The present invention has been described above in sufficient detail and with certain specificities. Those skilled in the art should understand that the descriptions in the embodiments are merely exemplary, and all changes made without departing from the true spirit and scope of the invention should fall within the protection scope of the invention. The scope of protection claimed by the present invention is defined by the claims, and not by the above descriptions in the embodiments.

Claims

1. A method for generating multicenter radiotherapy plans, characterized in that, The method for generating multicenter radiotherapy plans includes: It receives coordinates, field parameters, positioning methods, and uniform dose constraints from multiple treatment centers, as well as medical images, target areas, and organ-at-risk outlines of patients. Establish a multi-center collaborative optimization model; A unified dose-constrained multi-center collaborative optimization model is used to generate a total radiotherapy plan that includes a three-dimensional dose distribution. The overall plan is broken down into several sub-plans; Dosage validation is performed on several sub-plans, and the sub-plans that pass the dosage validation are output as executable sub-plans to the treatment device; The field parameters include field angle, radiation energy, field size, field weight, gantry angle, and MLC sequence; The medical images include CT images, MRI images, and PET images; The multi-center collaborative optimization model calculates the total dose distribution based on the following formula: ， Where n is the number of field of fire generated by all treatment centers. The dose weight of the target area corresponding to the i-th radiation field. Let be the dose distribution of the i-th radiation field; The objective function for optimizing the multi-center collaborative optimization model is: ， Where m is the number of target regions. Let be the weight coefficient of the j-th target region. For the dose of the j-th target region, Let p be the dose limit for the j-th target region, and p be the number of organs at risk. The weighting coefficient for the o-th organ at risk. For the dose of the oth organ at risk, For the dose limit of the oth organ at risk, where is the coefficient of the L1 regularization term, which is used to compress the weights of low-contribution fields to zero, and n is the number of fields. Here, represents the coefficient of the group regularization term, used to reduce the number of treatment centers, where C is the number of treatment centers. The number of radiation fields for the l-th treatment center.

2. The method for generating multicenter radiotherapy plans as described in claim 1, characterized in that, The optimized multi-center collaborative optimization model includes: The L-BFGS-B optimization engine was used to optimize the multi-center collaborative optimization model. During the optimization process, Pareto front analysis was used to dynamically adjust the field parameters of each treatment center. The three-dimensional dose distribution is updated in real time through GPU-accelerated dose calculation.

3. The method for generating multicenter radiotherapy plans as described in claim 1, characterized in that, The establishment of the multi-center collaborative optimization model also includes: when the cumulative dose to an organ at risk exceeds a preset threshold, automatically reducing the weight of the treatment center associated with that organ at risk.

4. The method for generating multicenter radiotherapy plans as described in claim 1, characterized in that, The process of breaking down the overall plan into several sub-plans includes: The overall plan is divided into several sub-plans based on the mechanical parameters of the equipment.

5. The method for generating multicenter radiotherapy plans as described in claim 1, characterized in that, The dose validation of several sub-plans includes: The sub-plan dose is verified using a GPU-accelerated Monte Carlo algorithm. The Gamma pass rate is compared between the planned dose and the actual dose. The executable sub-plan is a sub-plan with a Gamma pass rate of over 95%.

6. The method for generating multicenter radiotherapy plans as described in claim 1, characterized in that, The medical image is a multimodal image. After receiving the patient's medical image, target area, and organ at risk contours, the multicenter radiotherapy planning method further includes: Multimodal images are fused using rigid and / or non-rigid registration techniques.

7. A multicenter radiotherapy planning device, characterized in that, The multicenter radiotherapy planning device includes: The receiving module is used to receive the coordinates of multiple treatment centers, field parameters, positioning methods and unified dose constraints, as well as the patient's medical images, target area and organ at risk outlines; The modeling module is used to build multi-center collaborative optimization models. The plan generation module is used to optimize a multi-center collaborative optimization model based on unified dose constraints and generate a total radiotherapy plan that includes a three-dimensional dose distribution. The plan splitting module is used to split the overall plan into several sub-plans; The validation module is used to validate the dosage of the sub-plan and output the sub-plan that has passed the dosage validation as an executable sub-plan to the treatment device. The field parameters include field angle, radiation energy, field size, field weight, gantry angle, and MLC sequence; The medical images include CT images, MRI images, and PET images; The multi-center collaborative optimization model calculates the total dose distribution based on the following formula: ， Where n is the number of field of fire generated by all treatment centers. The dose weight of the target area corresponding to the i-th radiation field. Let be the dose distribution of the i-th radiation field; The objective function for optimizing the multi-center collaborative optimization model is: ， Where m is the number of target regions. Let be the weight coefficient of the j-th target region. For the dose of the j-th target region, Let p be the dose limit for the j-th target region, and p be the number of organs at risk. The weighting coefficient for the o-th organ at risk. For the dose of the oth organ at risk, For the dose limit of the oth organ at risk, where is the coefficient of the L1 regularization term, which is used to compress the weights of low-contribution fields to zero, and n is the number of fields. Here, represents the coefficient of the group regularization term, used to reduce the number of treatment centers, where C is the number of treatment centers. The number of radiation fields for the l-th treatment center.

8. An electronic device, characterized in that, The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the program to implement the multicenter radiotherapy planning method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the multicenter radiotherapy planning method according to any one of claims 1 to 6.

Images