Multi-objective Position Optimization System for Low-dose Rate Brachytherapy of Prostate Cancer

Abstract

The present invention provides a multi-objective position optimization system for low-dose-rate radioactive seed implantation for prostate cancer, comprising: a processing module, which receives the contour data of the CTV and the dose constraint conditions of the OAR; generates an initial population of candidate seed positions based on the contour data; and performs iterative optimization on the initial population by using a multi-objective genetic algorithm; an output module, configured to perform the following steps: output a Pareto optimal solution set that satisfies all dose constraint conditions, the solution set containing multiple particle spatial coordinate distribution schemes. It avoids the problem of over-focusing on a certain objective in the optimization of the position of radioactive seeds for prostate cancer, can optimize multiple objectives simultaneously, without presetting weight parameters, and significantly reduces the risk of improper weight setting or optimization failure in traditional methods. Multi-objective optimization can significantly improve the dose coverage rate of the target area, reduce the dose to normal tissues, explore a wider solution space, and provide multiple optimization schemes for clinical practice.

Description

Technical Field

[0001] The present invention relates to the technical field of optimizing the dose and position of endocurietherapy for tumors, and more particularly to a multi-objective position optimization system for low-dose rate radioactive seed implantation for prostate cancer. Background Art

[0002] Low-dose rate radioactive seed implantation (LDRBT) for prostate cancer is a local radiotherapy method commonly used to treat early-stage localized prostate cancer. This technique involves implanting radioactive isotopes (usually iodine-125 or palladium-103) into the prostate gland, which continuously releases low-dose radiation to kill cancer cells. LDRBT has the advantages of high local dose and less damage to surrounding tissues, effectively treating tumors while minimizing harm to healthy tissues. This technique is characterized by strong targeting, with radiation mainly concentrated in the prostate area, reducing damage to surrounding normal tissues. Side effects are fewer, and the quality of life of patients is relatively high. Seed implantation is performed through minimally invasive surgery, and patients recover quickly, usually resuming normal activities within a few days. In treatment planning, how to accurately distribute radioactive seeds to ensure the best dose distribution while avoiding over-irradiation of healthy tissues is an important optimization problem.

[0003] Existing optimization methods, such as HIPO, are inverse optimization algorithms for brachytherapy. Based on three-dimensional anatomical information, HIPO can optimize the dose distribution of pre-implanted catheters and determine reasonable catheter positions. Its core advantage lies in reducing high-dose hot spots and providing a more uniform dwell time distribution, thereby reducing the free dwell time between adjacent dwell positions. This method can not only effectively improve the dose coverage of the target area but also reduce the radiation dose received by normal tissues. However, despite the advantages of HIPO in optimizing dose distribution and reducing dose hot spots in HDR brachytherapy, there are still certain limitations. First, its optimization effect highly depends on the initial arrangement of catheters. If the catheter positions are not ideal, it is difficult to achieve the best dose distribution. Second, the computational complexity of HIPO is relatively high, which may lead to an extended optimization time. In addition, HIPO may fall into a local optimal solution, affecting the overall dose distribution quality. Although the algorithm has an automatic optimization function, in some cases, it still requires manual adjustment by users to meet clinical goals, thus posing higher requirements for user experience and intervention capabilities.

[0004] The record of the foregoing background art knowledge is intended to help those of ordinary skill in the art understand the prior art relatively close to the present invention and facilitate the understanding of the inventive concept and technical solution of the present invention. It should be clear that, in the absence of clear evidence indicating that the above content was publicly available before the filing date of this patent application, the above background art should not be used to evaluate the novelty of the technical solution of this application. Summary of the Invention

[0005] Technical problem

[0006] To solve the above problems, the object of the present invention is to provide a multi-objective position optimization system for low-dose rate radioactive seed implantation in prostate cancer. The solution avoids the problem of over-focusing on a single target in the optimization of the position of radioactive seeds in prostate cancer, can optimize multiple targets simultaneously without presetting weight parameters, thus significantly reducing the risk of improper weight setting or optimization failure in traditional methods. Multi-objective optimization can significantly improve the dose coverage rate of the target area, reduce the dose to normal tissues at the same time, and can explore a wider solution space to provide multiple optimization schemes for clinical practice.

[0007] Technical solution

[0008] That is, the present invention provides the following technical solution.

[0009] A multi-objective position optimization system for low-dose rate radioactive seed implantation in prostate cancer, comprising:

[0010] A processing module configured to perform the following steps: receiving the contour data of the CTV and the dose constraint conditions of the OAR; generating an initial population of candidate seed positions based on the contour data, wherein the candidate positions are distributed at a preset interval within the range of 0.5 cm to 1 cm outside the CTV; iteratively optimizing the initial population using a multi-objective genetic algorithm, the multi-objective genetic algorithm including the NSGA-II framework, and the optimization objectives being the fitness functions for minimizing the fitness function of the dose outside the CTV below the prescription dose, the fitness function of the dose inside the CTV exceeding the high-dose limit, and the fitness function of the dose of the OAR exceeding the limit;

[0011] An output module configured to perform the following steps: outputting a Pareto optimal solution set that satisfies all dose constraint conditions, the solution set including multiple particle spatial coordinate distribution schemes.

[0012] Further, the step of generating the initial population includes: generating candidate position points at intervals of 0.5 cm in a three-dimensional space; randomly selecting a subset of the candidate position points as initial individuals, and each individual is encoded as a binary or real number sequence of seed coordinates.

[0013] Further, the multi-objective genetic algorithm further includes:

[0014] A crossover recombination module configured to generate new solutions by randomly single-point crossover recombination of individual chromosomes;

[0015] A crowding degree calculation module configured to screen Pareto front solutions based on the distribution density of solutions in the objective space.

[0016] Further, the fitness function includes: for the control points on the periphery of the CTV, a penalty score is calculated when the dose is lower than the prescribed dose:

[0017]

[0018] wherein, represents the fitness function of the control points on the periphery of the CTV, and this function calculates the degree of deviation of all control points from the target; represents the number of control points on the periphery of the CTV; represents the point dose at a distance of from the radiation source, represents the CTV dose control point; represents the CTV prescribed dose; represents the implanted dose.

[0019] Further, the fitness function includes: for the control points inside the CTV, a penalty score is calculated when the dose is higher than the preset high-dose threshold:

[0020]

[0021] wherein, represents the fitness function of the control points inside the CTV, and this function calculates the degree of deviation of all control points from the target; represents the number of control points inside the CTV; represents the point dose at a distance of from the radiation source, represents the CTV dose control point; represents the high dose to be controlled inside the CTV; represents the implanted dose.

[0022] Further, the fitness function includes: for the control points of the OAR, a penalty score is calculated when the dose exceeds the limit:

[0023]

[0024] wherein, represents the fitness function of the control points of the organ at risk; represents all dose reference points of the organ at risk; represents the limiting dose of the organ at risk; represents the implanted dose; the dose of the th dose point is calculated according to the above formula , and the objective function is minimized when the dose of each dose reference point of the OAR is less than the limiting dose .

[0025] Further, the high-dose threshold is set to 150% of the prescribed dose, and the OAR dose limits include a maximum urethra dose of less than 150 Gy and a rectal D2cc dose of less than 100 Gy.

[0026] Further, the system further includes a dose calculation module that calculates the dose contribution of a spatial point according to the following formula:

[0027]

[0028] where represents the dose rate at a distance of from the seed; are all constants; is the average dose rate constant; is the radial dose function; is the anisotropy function.

[0029] Further, the processing module is further configured to: normalize the output Pareto optimal solution set to visualize the solution distribution of multiple objectives; screen the feasible solutions that simultaneously satisfy CTV D90 > 90 Gy, V150 < 50%, urethra Dmax < 150 Gy, and rectal D2cc < 100 Gy.

[0030] Further, the system further includes a user interface module configured to receive weight hyperparameters input by the user and dynamically adjust the convergence direction of the optimization process based on the parameters.

[0031] A computer-readable storage medium stores computer program instructions that, when executed by a processor, implement the functions of the modules in the above-mentioned multi-objective position optimization system for low-dose rate radioactive seed implantation for prostate cancer.

[0032] Further, the computer program instructions include: a module for generating a dose calculation lattice based on voxel data, with a lattice spacing of 1 cm; storing data tables of the radial dose function and anisotropy function of the radiation source generated by Monte Carlo simulation.

[0033] Further, the computer program instructions further include: performing multiple iterative calculations on the randomness of the initialized population and merging all Pareto front solutions to generate a global optimal solution set.

[0034] Further, the processing module is further configured to: dynamically adjust the crossover probability and mutation probability during the optimization process to avoid premature convergence; record the change in fitness value for each iteration, and restart the optimization process when there is no improvement for a continuous preset number of generations.

[0035] Further, the system is communicatively connected to a medical imaging device and receives real-time updates of the contour data of the CTV and OAR.

[0036] Further, the output particle coordinate distribution scheme is displayed through a three-dimensional visualization interface and supports manual adjustment by the user for re-optimization.

[0037] Further, the medium is a solid-state drive, a USB flash drive, or a cloud server storage unit, and the program instructions are encrypted and stored to prevent unauthorized access.

[0038] A computer device, the computer device includes a memory, a processor, a communication interface, and a communication bus; wherein, the memory, the processor, and the communication interface communicate with each other through the communication bus; the memory is used for storing a computer program; the processor is used for executing the computer program stored on the memory, and when the processor executes the computer program, it realizes the functions of each module in the above-mentioned multi-objective position optimization system for low-dose rate radioactive seed implantation for prostate cancer.

[0039] On the basis of conforming to the common knowledge in the art, the above-mentioned preferred conditions can be combined with each other to obtain specific implementation manners.

[0040] Beneficial effects

[0041] According to the present invention, there is provided a multi-objective position optimization system for low-dose rate radioactive seed implantation for prostate cancer, which applies multi-objective optimization to find a balance among multiple conflicting objectives, provides multiple Pareto optimal solutions, and each solution represents the optimization of a certain objective without deteriorating other objectives, avoiding the problem of over-focusing on a certain objective in the optimization of the position of radioactive seeds for prostate cancer, being able to optimize multiple objectives simultaneously without presetting weight parameters, thereby avoiding the risk of improper weight setting or optimization failure in traditional methods. Multi-objective optimization can significantly improve the dose coverage rate of the target area, while reducing the dose to normal tissues, and can explore a wider solution space, providing multiple optimization schemes for clinical use. Verification shows that using multi-objective optimization can obtain multiple schemes that fully meet all dose constraint requirements, providing valuable guidance for the selection of the final scheme. The multi-objective optimization method has achieved the best results in terms of CTV D90, CTV V150, and rectal D2cc, and has a small variance.

[0042] The present invention adopts the above technical solutions to achieve the above object, making up for the deficiencies of the prior art, being reasonably designed and convenient to operate. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] To make the above and / or other objectives, features, and advantages of the present invention more obvious and understandable, the following will briefly introduce the drawings required for the specific implementation of the present invention. Obviously, the drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative efforts.

[0044] Figure 1 Schematic diagrams showing the distribution of the feasible solution spaces for single-objective optimization and multi-objective optimization ((A), (B), and (C) represent the solution space distributions of single-objective optimization under three different constraint conditions; (D), (E), and (F) represent the solution space distributions of multi-objective optimization under the same three constraint conditions));

[0045] Figure 2 Graphs showing the dose-volume histograms of CTV and OARs for single-objective and multi-objective optimization under different weight settings (blue represents the rectum, green represents the urethra, and red represents CTV). Specific Embodiments

[0046] Those skilled in the art can draw on the content of this article and appropriately replace and / or modify process parameters to achieve the same. However, it should be particularly noted that all such similar replacements and / or modifications are obvious to those skilled in the art and are all considered to be included in the present invention. The products and preparation methods described in the present invention have been described through preferred examples. Relevant personnel can obviously make changes or appropriate alterations and combinations to the products and preparation methods described in this article without departing from the content, spirit, and scope of the present invention to implement and apply the technology of the present invention.

[0047] Unless otherwise defined, the technical and scientific terms used in this article have the same meanings as those commonly understood by those of ordinary skill in the art to which the present invention belongs. The present invention uses the methods and materials described in this article; however, other suitable methods and materials known in the art can also be used. The materials, methods, and examples described in this article are only illustrative and are not intended to be limiting. All publications, patent applications, patents, provisional applications, database entries, and other references mentioned in this article are incorporated herein by reference in their entirety. In case of conflicts, the present specification including the definitions shall prevail.

[0048] Unless specifically stated, the materials, methods, and examples described in this article are only exemplary and not restrictive. Although methods and materials similar or equivalent to those described in this article can be used for the implementation or testing of the present invention, suitable methods and materials are still described herein.

[0049] The technical solutions in the embodiments of the present invention will be described clearly and completely below. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention. At the same time, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other.

[0050] It should be understood that any technical solution claimed in the present invention does not involve the diagnosis and treatment of diseases.

[0051] To facilitate the understanding of the embodiments of the present invention, the abbreviations and key terms that may be involved in the embodiments of the present invention are first explained or defined. For the abbreviations or key terms that are not defined, they are all commonly understood by those skilled in the art.

[0052] LDRBT: Low Dose Rate Brachytherapy for Prostate Cancer;

[0053] HIPO: Hybrid Inverse Planning and Optimization;

[0054] CTV: Clinical Target Volume for Prostate Cancer;

[0055] OAR: Organs at Risk;

[0056] NSGA-II: Non-dominated Sorting Genetic Algorithm II;

[0057] Pareto front solution: Refers to the set of solutions that achieve the best trade-off among multiple conflicting objectives;

[0058] HDR: High Dose Rate;

[0059] DVHs: Dose-Volume Histograms.

[0060] In addition, unless otherwise specified, the test methods used in the embodiments are all conventional methods; the materials, reagents, etc. used, unless otherwise specified, can all be obtained from commercial channels. For reagents or instruments without indicating the manufacturer, they are all conventional products that can be obtained through market purchase. All the published cases and other reference materials mentioned in this article are incorporated herein by reference in their entirety.

[0061] The following describes the present invention in detail

[0062] Embodiment 1:

[0063] A multi-objective position optimization system for low-dose rate radioactive seed implantation in prostate cancer is provided, including a processing module and an output module. The processing module is configured to perform the following steps: receiving the contour data of the CTV and the dose constraint conditions of the OAR; generating an initial population of candidate seed positions based on the contour data, where the candidate seed positions are distributed at a preset interval within the range of 0.5 cm to 1 cm outside the CTV; iteratively optimizing the initial population using a multi-objective genetic algorithm, the multi-objective genetic algorithm including the NSGA-II framework, and the optimization objectives being the fitness functions of minimizing the fitness function of the dose outside the CTV below the prescription dose, the fitness function of the dose inside the CTV exceeding the high-dose limit, and the fitness function of the dose of the OAR exceeding the limit; and the output module is configured to perform the following steps: outputting a Pareto optimal solution set that satisfies all dose constraint conditions, the solution set including multiple particle spatial coordinate distribution schemes.

[0064] 1. Optimization problem

[0065] A constrained optimization problem with variables and optimization objectives can be defined in the following form:

[0066]

[0067] where, is a set of multi-objective optimization functions, is the problem search space, is the dimension of the problem.

[0068] Points outside the tumor target region need to meet the minimum prescription dose limit requirement. That is, for the control points outside the CTV, when the dose is lower than the prescription dose, a penalty score is calculated:

[0069]

[0070] where, represents the fitness function of the control points outside the CTV, and this function calculates the degree of deviation of all control points from the target; represents the number of control points outside the CTV; represents the dose at a point at a distance of from the radiation source, represents the CTV dose control point; represents the CTV prescription dose; represents the implanted dose.

[0071] To control the volume of the high-dose region inside the tumor target region, for the points inside the tumor target region, the highest dose limit requirement needs to be met. That is, for the control points inside the CTV, when the dose is higher than the preset high-dose threshold, a penalty score is calculated:

[0072]

[0073] Among them, represents the control point fitness function inside the CTV, which calculates the degree of deviation of all control points from the target; represents the number of internal control points of the CTV; represents the point dose at a distance of from the radiation source, represents the CTV dose control point; represents the high dose to be controlled inside the CTV; represents the implant dose.

[0074] The above and respectively represent the control point fitness functions on the periphery and inside of the CTV, which calculate the degree of deviation of all control points from the target. In the optimization, it is necessary to satisfy that the dose on the periphery of the CTV is higher than the prescription dose while the dose inside the CTV is controlled within a reasonable high-dose range. In this embodiment, the volume of the dose higher than 150% of the prescription dose inside the CTV, that is, V150, is controlled. Therefore, . When each control point on the CTV reaches and is less than the maximum control dose , the fitness function is minimized.

[0075] Similarly, for the organ at risk, the fitness function is defined as the penalty score for exceeding the dose limit, that is, for the OAR control point, when the dose exceeds the limit, the penalty score is calculated:

[0076]

[0077] Among them, represents the control point fitness function of the organ at risk; represents all dose reference points of the organ at risk; represents the limiting dose of the organ at risk; represents the implant dose; the dose of the th dose point is calculated according to the above formula , and when each dose reference point of the OAR is less than the limiting dose , the objective function is minimized.

[0078] This study optimizes the problem of 80 variables with 4 constraint conditions, and the constraint conditions are defined as the dose volume limit conditions of the CTV and the OAR. As shown in Table 1.

[0079] Table 1 - Optimization Objectives

[0080]

[0081] For a single objective function, the individual fitness is expressed as the weighted sum of the CTV and OAR fitness:

[0082]

[0083] The weight hyperparameter is set to adjust the importance of the tumor target area and normal tissues in the overall optimization process. By changing these weights, the genetic algorithm can be guided to find solutions that meet the prescription dose requirements. Usually, higher weight values are given to conditions with high optimization difficulty, which helps the optimization process converge to the most optimal goal.

[0084] 2. Encoding and Crossover

[0085] Encoding is defined as whether the seed occupies the candidate position, and the candidate position is determined according to the volume of the tumor target area. The maximum boundary of the needle is 0.5 cm outside the CTV. Considering the actual seed implantation process, a spacing of 0.5 - 1 cm usually needs to be maintained, and on the implantation path, due to the geometric volume of the seed cladding, the minimum distance between two point sources is 0.5 cm. Therefore, in the initialization stage, individual candidate position points are generated at intervals of 0.5 cm in the x, y, and z directions.

[0086] In the initialization process, n candidate positions are randomly selected and m populations are generated, where m is the population size. An individual chromosome in the GA consists of the (x, y, z) encoding of the seed spatial coordinates written one by one. The number of genes required to code a single chromosome is determined by the minimum spacing of seed implantation.

[0087] Crossover and recombination of individuals can recombine different solutions to accelerate the search for the Pareto front, thus finding approximate solutions faster. (Whether it helps with the distance of the solutions).

[0088] Random crossover is to randomly set only one crossover point in the individual coding string, and then exchange part of the chromosomes of two ligand individuals at this point.

[0089] 3. Sorting and Crowding Degree Calculation

[0090] The goal of multi-objective optimization is to find the Pareto front, which is the set of solutions where no single solution can be further optimized without making other objectives worse among all objectives. Individuals in the population should approach the Pareto set in the decision space during the search process, and their mappings in the objective space should gradually converge to the Pareto front. The constraint conditions change the shape and position of the Pareto front and distribute it over multiple feasible regions, causing the population to be easily trapped in some regions and unable to obtain a uniformly distributed Pareto front.

[0091] 4. Calculation of dose volume index (DVI)

[0092] The PTV is obtained from the DICOM RT file with the contours outlined by the treatment planning system. The normal tissues include: urethra, bladder, and rectum. The point cloud data representing these contours is used to generate dot matrices using a polygon scan conversion program. These dot matrices are generated at equal intervals in the Cartesian space for calculating the doses obtained in the target area and normal tissues. In this study, the dot matrix spacing is set to I = 1 cm. If it is too narrow, there will be too many reference points, affecting the calculation time; conversely, if it is set too large, the calculation results will be inaccurate.

[0093] The dose calculation module calculates the dose contribution of spatial points according to the following formula:

[0094]

[0095] where, represents the dose rate at a distance of from the seed; are all constants, = 0.6 mCi, = 1 cm; is the average dose rate constant, Λ = 0.965 cGy / h / cm 2 ; is the radial dose function; is the anisotropy function.

[0096] Multi-objective optimization provides more choices for clinicians by generating multiple feasible solutions, enabling them to make the best decision according to specific needs. To verify its effect, we tested a typical prostate cancer plan. Figure 1 shows the distribution of the feasible solution space of single-objective optimization and multi-objective optimization. As can be seen from Figure 1 in single-objective optimization ( Figure 1For A), B), and C), after 10 iterations, we selected the highest-ranked solution from the 40 generated solutions. The results showed that all solutions met the prescription dose requirement for the target volume D90, and the normal tissue doses conformed to the constraint conditions (as shown by the red boxes). However, only 2.5% of the solutions met the prescription dose requirement for the target volume V150, 7.5% of the solutions conformed to the high-dose constraint for the urethra, and none of the solutions met the D2cc dose constraint for the rectum. This indicates that in single-objective optimization, none of the 40 solutions could fully meet all dose constraint requirements.

[0097] In contrast, as can be seen from D), E), and F) of Figure 1 , after 10 iterations of multi-objective optimization, Pareto front points were selected from each iteration, and a total of 755 Pareto front points were obtained. Since it is difficult to fully present the Pareto front of the four objectives, we showed the distribution of the solutions under the same conditions as the single-objective function (i.e., CTV D90 > 100 Gy). The results showed that multi-objective optimization not only identified more solutions but also had a wider distribution range of these solutions. On the premise that the CTV met the prescription dose requirement for D90, 3.7%, 8.7%, and 12.4% of the solutions respectively conformed to the dose constraint criteria. Finally, a total of 6 solutions fully met all dose constraint requirements, providing valuable guidance for the selection of the final solution.

[0098] Figure 2 Further, the dose-volume histograms (DVHs) of the CTV and OARs for single-objective and multi-objective optimizations under different weight settings are shown. A, B, C, and D represent four different weight hyperparameters respectively. In the single-objective optimization results, the maximum dose (Dmax) of the urethra (green) was below 150 Gy, within the clinically acceptable range. Among them, the D2cc of the rectum (blue) was the most sensitive to weight changes, and increasing the weight of the rectum significantly reduced the maximum dose of the rectum. Since the urethra is completely contained within the CTV (red), the constraint on the urethra led to a reduction in the seed distribution around it, causing the CTV D90 dose to drop to approximately 91 Gy. However, after increasing the weight in scenario D, the CTV D90 dose recovered to 95 Gy. In addition to the DVI dose control points, the differences among the four scenarios can also be observed from the DVH graphs. In scenarios A and C, the high dose of the rectum led to significant deviations, with a significant right shift of the high-dose region of the rectum (DVH curve), indicating that the weight did not sufficiently suppress the rectum dose, and the urethra dose might increase due to the preferential optimization of the CTV (the urethra is contained within the CTV). In scenario D, although the urethra Dmax met the limit, due to constraint conflicts, the high-dose volume was significantly higher than that of other scenarios.

[0099] The multi-objective optimization method achieved the best results in terms of CTV D90, CTV V150, and rectal D2cc, with a small variance. Although the multi-objective plan had a higher urethra Dmax among the four plans, it still met the optimization goals.

[0100] Example 2:

[0101] Based on the foregoing embodiments, the dose calculation module is optimized. The dose calculation module calculates the dose contribution of a spatial point according to the following formula:

[0102]

[0103] Where, represents the optimized dose rate at a distance of from the seed; are all constants, = 0.6 mCi, = 1 cm; is the average dose rate constant, Λ = 0.965 cGy / h / cm 2 ; is the radial dose function; is the anisotropy function; is the tissue heterogeneity correction factor, which obtains the electron density information through image, , where, represents the effective attenuation coefficient, that is, the actual linear attenuation coefficient of photons in the tissue (cm -1 ); represents the water phantom attenuation coefficient, such as for iodine-125, is taken as about 0.28 cm -1 , and for iridium-192, is taken as about 0.12 cm -1 .

[0104] The inventors surprisingly found that by optimizing the dose contribution formula of the spatial point in the dose calculation module with the tissue heterogeneity correction factor, the accuracy of dose calculation can be improved, especially the dose calculation accuracy at tissue junctions such as the rectum / prostate can be significantly improved, reducing the dose deviation of the latter by at least 5%, reducing the dose deviation caused by model simplification, and optimizing the clinical results.

[0105] Example 3:

[0106] Based on the foregoing embodiments, the processing module is optimized as a whole to obtain a dynamic adaptive multi-objective dose optimization algorithm model, specifically including:

[0107] 1. CTV dose coverage deficiency penalty function :

[0108]

[0109] Among them, represents the CTV prescription dose; represents the actual dose at the th dose point within the CTV; represents the total number of dose points within the CTV. It is used to enhance the sensitivity of the low-dose region.

[0110] 2. CTV internal high-dose region volume penalty function :

[0111]

[0112] Among them, represents the V150 threshold, such as 135 Gy. It is used to strengthen the penalty for overdose hot spots and avoid excessive local doses.

[0113] 3. OAR dose overlimit penalty function :

[0114]

[0115] Among them, represents all dose reference points of the organ at risk; represents the limiting dose of the organ at risk, Dmax of the urethra = 150 Gy, D2cc of the rectum = 100 Gy; represents the dynamic weight, , which linearly increases with the number of iterations t (α = 0.02). Through the segmented dynamic weight, the protection of the OAR is gradually strengthened.

[0116] Applying the optimized dynamic adaptive multi-objective dose optimization algorithm model of this embodiment, compared with the multi-objective genetic algorithm of Embodiment 1, the comparison results are shown in Tables 2 and 3.

[0117] Table 2. Comparison of dosimetric indices

[0118]

[0119] Table 3. Comparison of algorithm performances

[0120]

[0121] The optimized dynamic adaptive multi-objective dose optimization algorithm model in this embodiment enhances the sensitivity to extreme dose deviations through exponential and square terms, optimizes the linear average of the multi-objective genetic algorithm, prevents the population from falling into local optima, and has a more uniform Pareto solution distribution, showing significant improvements in convergence speed, solution quantity, operation duration, and accuracy.

[0122] Embodiment 4:

[0123] A computer-readable storage medium is also provided. The computer-readable storage medium stores a computer program executable by a processor. When the computer program is executed by the processor, it realizes the functions of each module in the aforementioned multi-objective position optimization system for low-dose-rate radioactive seed implantation for prostate cancer and can achieve the same technical effects. To avoid repetition, this embodiment will not be elaborated further.

[0124] Embodiment 5:

[0125] A computer device includes a memory, a processor, a communication interface, and a communication bus. Among them, the memory, the processor, and the communication interface communicate with each other through the communication bus. The memory is used to store a computer program. The processor is used to execute the computer program stored on the memory. When the processor executes the computer program, it realizes the functions of each module in the aforementioned multi-objective position optimization system for low-dose-rate radioactive seed implantation for prostate cancer and can achieve the same technical effects. To avoid repetition, this embodiment will not be elaborated further.

[0126] Computer-readable media include both permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassette tapes, magnetic disk storage, or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. As defined herein, computer-readable media do not include transitory media such as modulated data signals and carrier waves.

[0127] The conventional technologies in the above embodiments are well-known prior arts to those skilled in the art, so they will not be elaborated in detail here.

[0128] The specific embodiments described herein are merely illustrative of the spirit of the present invention. Those skilled in the art to which the present invention pertains may make various modifications or supplements to the described specific embodiments or use similar means for substitution, but will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

[0129] Although a detailed description of the present invention has been given and some specific embodiments have been cited, it is obvious that various changes or modifications can be made by those skilled in the art without departing from the spirit and scope of the present invention.

[0130] Although the above specific embodiments have shown, described and pointed out the novel features applicable to various embodiments, it should be understood that various omissions, substitutions and changes in the form and details of the described apparatus or method can be made without departing from the spirit of the present disclosure. Additionally, the above various features and methods can be used independently of each other or can be combined in various ways. All possible combinations and sub - combinations are intended to fall within the scope of the present disclosure. Many of the above embodiments include similar components and, therefore, these similar components can be interchanged in different embodiments. Although the present invention has been disclosed in the context of certain embodiments and examples, those skilled in the art should understand that the present invention can extend beyond the specifically disclosed embodiments to other alternative embodiments and / or applications and their obvious modifications and equivalents. Therefore, the present invention is not intended to be limited by the specific disclosure of the preferred embodiments herein.

[0131] Matters not covered by the present invention are all well - known techniques.

Claims

1. A multi-target position optimization system for low-dose rate radioactive seed implantation in prostate cancer, characterized in that Comprising: A processing module configured to perform the following steps: receiving the contour data of the CTV and the dose constraint conditions of the OAR; generating an initial population of candidate particle positions based on the contour data, wherein the candidate particle positions are distributed at a preset interval within a range of 0.5 cm to 1 cm outside the CTV; iteratively optimizing the initial population using a multi-objective genetic algorithm, the multi-objective genetic algorithm including the NSGA-II framework, and the optimization objectives being a fitness function for minimizing the fitness that the dose outside the CTV is lower than the prescribed dose, a fitness function for the dose inside the CTV exceeding the high dose limit, and a fitness function for the dose of the OAR exceeding the limit; An output module configured to perform the following steps: outputting a Pareto optimal solution set that satisfies all dose constraint conditions, the solution set including multiple particle spatial coordinate distribution schemes.

2. The system according to claim 1, wherein The step of generating the initial population includes: generating candidate position points at an interval of 0.5 cm in a three-dimensional space; randomly selecting a subset of the candidate position points as initial individuals, and each individual is encoded as a binary or real number sequence of particle coordinates.

3. The system according to claim 1, characterized in that The multi-objective genetic algorithm further includes: A crossover recombination module configured to generate new solutions by randomly single-point crossover recombination of individual chromosomes; A crowding degree calculation module configured to screen Pareto front solutions based on the distribution density of solutions in the objective space.

4. The system according to any one of claims 1-3, characterized in that: The fitness function includes: For the control points outside the CTV, calculating a penalty score when the dose is lower than the prescribed dose: Among them, represents the control point fitness function outside the CTV, which calculates the degree of deviation of all control points from the target; represents the number of control points outside the CTV; represents the point dose at a distance of from the radiation source, represents the CTV dose control point; represents the CTV prescription dose; represents the implant dose; and / or For the control points inside the CTV, calculating a penalty score when the dose is higher than the preset high dose threshold: Among them, represents the control point fitness function inside the CTV, which calculates the degree of deviation of all control points from the target; represents the number of control points inside the CTV; represents the distance from the radiation source as the point dose at that distance, represents the CTV dose control point; represents the high dose that needs to be controlled inside the CTV; represents the implant dose; and / or For the control points of the OAR, calculating a penalty score when the dose exceeds the limit: Among them, represents the control point fitness function of the organ at risk; represents all dose reference points of the organ at risk; represents the limiting dose of the organ at risk; represents the implant dose; the dose at the th dose point is calculated according to the above formula , and the objective function is minimized when each dose reference point of the OAR is less than the limiting dose .

5. The system according to claim 4, wherein The high dose threshold is set to 150% of the prescribed dose, and the OAR dose limit includes that the maximum dose of the urethra is less than 150 Gy and the D2cc dose of the rectum is less than 100 Gy.

6. The system according to claim 4, wherein The system further includes a dose calculation module for calculating the dose contribution of a spatial point according to the following formula: Among them, represents the dose rate at a distance of ; are all constants; is the average dose rate constant; is the radial dose function; is the anisotropy function.

7. The system according to claim 5, wherein The processing module is further configured to: normalize the output Pareto optimal solution set to visualize the solution distribution of multiple objectives; screen the feasible solutions that simultaneously satisfy CTV D90 > 90 Gy, V150 < 50%, urethra Dmax < 150 Gy, and rectal D2cc < 100 Gy.

8. The system according to claim 1, wherein The system further includes a user interface module configured to receive the weight hyperparameter input by the user and dynamically adjust the convergence direction of the optimization process based on the parameter.

9. A computer-readable storage medium, characterized in that, Stored with computer program instructions, when the computer program instructions are executed by a processor, the functions of each module in the system according to any one of claims 1-8 are implemented.

10. A computer device, the computer device comprising a memory, a processor, a communication interface, and a communication bus; wherein, The memory, the processor, and the communication interface communicate with each other through the communication bus; the memory is used for storing a computer program; the processor is used for executing the computer program stored on the memory, and is characterized in that: when the processor executes the computer program, the functions of each module in the system according to any one of claims 1-8 are implemented.

Images