Lung cancer radiotherapy plan generation method and system based on dose prediction and auxiliary contour
By generating anisotropic auxiliary contours using a deep neural network model based on dose prediction and auxiliary contours, the problem of relying on experience in radiotherapy planning was solved, achieving efficient and scientific optimization of radiotherapy plans and improving the quality and efficiency of lung cancer radiotherapy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CANCER INST & HOSPITAL CHINESE ACADEMY OF MEDICAL SCI
- Filing Date
- 2023-05-29
- Publication Date
- 2026-05-29
AI Technical Summary
Existing radiotherapy planning relies on personal experience and trial-and-error methods, resulting in long optimization times, high resource consumption, and unstable planning quality. Furthermore, the uniformity of commonly used auxiliary contours cannot meet the dose distribution requirements for lung cancer radiotherapy, affecting the design quality and efficiency.
A dose prediction and auxiliary contour-based approach is adopted. An anisotropic auxiliary contour is generated using a trained deep neural network model, and inverse optimization is performed in combination with dose distribution to generate a radiotherapy plan for lung cancer.
It improves the efficiency and quality of radiotherapy planning, reduces planning time, enhances the protection of organs at risk, and is suitable for units with insufficient radiotherapy experience.
Smart Images
Figure CN116688376B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of medical device technology, specifically to a method and system for generating radiotherapy plans for lung cancer based on dose prediction and auxiliary profiles. Background Technology
[0002] Radiotherapy planning is a prerequisite for implementing radiotherapy. Current intensity-modulated radiotherapy (IMRT) planning is based on complex inverse optimization algorithms. First, auxiliary contours and optimization conditions / constraints are manually formulated based on the patient's target location, prescribed dose requirements, etc. The planning system generates a flux map based on these conditions and constraints using complex optimization algorithms. Then, combined with the accelerator's constraints, it generates executable field information (e.g., lead gate position, multi-leaf collimator (MLC) shape, control point MU, etc.).
[0003] Currently, radiotherapy planning systems still have the following shortcomings: First, IMRT (Intra-Invasive Therapy Treatment) planning typically relies on personal experience and trial-and-error methods. Planners need to go through dozens of trials to find a satisfactory design. The quality of the plan largely depends on the planner's subjective experience and the time spent. The complexity of inverse optimization algorithms leads to long optimization times and high hardware resource consumption. Second, there are significant differences in the quality and time taken for IMRT plans designed between different radiotherapy centers, and even among different personnel within the same center. Improving the efficiency and quality of radiotherapy planning is an urgent need in radiotherapy today.
[0004] In inverse intensity-modulated radiotherapy (IMRT) planning optimization, numerous auxiliary contours are needed for dose constraint to minimize radiation dose to organs at risk outside the target area. Currently, commonly used auxiliary contours in clinical practice (such as Ring, Nt, Fan, etc.) are based on the planner's subjective experience and are typically uniformly extended outwards around the target area. In lung cancer planning, due to the need for stronger dose limits in the lung region, the dose drops slowly in the anterior-posterior direction and more rapidly in the lateral direction. The dose exhibits a significant non-uniform drop along all directions. Using isotropic Ring, Nt, and Fan for dose limits in this case can lead to overly strict dose constraints in some areas and overly lenient constraints in others, thus affecting the optimization results. How to scientifically and rationally create auxiliary contours is a crucial factor affecting the quality and efficiency of radiotherapy planning. Summary of the Invention
[0005] The purpose of this invention is to provide a method and system for generating radiotherapy plans for lung cancer based on dose prediction and auxiliary profiles, so as to solve at least one of the technical problems existing in the background art.
[0006] To achieve the above objectives, the present invention adopts the following technical solution:
[0007] On one hand, the present invention provides a method for generating radiotherapy plans for lung cancer based on dose prediction and auxiliary profiles, comprising:
[0008] Obtain case data for lung cancer radiotherapy plans;
[0009] The trained dose prediction model is used to process radiotherapy planning case data to obtain dose distribution results for the target area and organs at risk;
[0010] An anisotropic auxiliary profile is generated based on the predicted dose distribution and the dose constraints required for inverse optimization.
[0011] Dose limits for the target area, organs at risk, and auxiliary contours are obtained based on the auxiliary contours and dose distribution.
[0012] Based on the auxiliary profile and dose constraint limits, inverse optimization is performed to generate a lung cancer radiotherapy plan.
[0013] Optionally, the dose prediction model is trained using a training set, which includes multiple lung cancer radiotherapy plan cases. The lung cancer radiotherapy plan cases include at least CT images of the cases, dose distribution, target area and organ at risk contours, and minimum distance distribution from patient voxels to the target area. The model is trained using CT images, target area and organ at risk contours, and minimum distance distribution from patient voxels to the target area as input data and dose distribution as output data.
[0014] Optionally, the closest distance from the i-th point of the patient to the target area is defined as follows:
[0015]
[0016] Among them, D i This represents the shortest distance from the i-th voxel to the target region (x). i y i z i Let (x) be the coordinates of the i-th point, and (x) be the coordinates of the i-th point. k y k z k D represents the coordinates of point k on the target area; when voxel i is located within the target area or outside the patient's body, D... i Defined as 0.
[0017] Optionally, the dose prediction model is trained using a deep neural network framework, including Vgg-Unet, Res-Unet, Trans-Unet, and UNeXt.
[0018] Optionally, based on the dose distribution characteristics of the lung cancer radiotherapy plan, an anisotropic auxiliary profile is generated using the dose predicted by the model.
[0019] Optionally, the auxiliary contour has different widths in various directions around the target area depending on the rate of predicted dose drop, and is the area enclosed by a specific isodose line or patient contour.
[0020] Optionally, the dose constraint limit generation module generates auxiliary contours, dose limit conditions for organs at risk and target areas based on dose distribution, and imports them into the planning system through an interface program for automatic optimization of radiotherapy plans.
[0021] Secondly, the present invention provides a lung cancer radiotherapy planning generation system based on dose prediction and auxiliary contouring, comprising:
[0022] The acquisition module is used to acquire lung cancer radiotherapy plan case data;
[0023] The prediction module is used to process radiotherapy planning case data using a trained dose prediction model to obtain dose distribution results for the target area and organs at risk.
[0024] The first determining module is used to generate anisotropic auxiliary profiles based on the predicted dose distribution and the dose constraints required for inverse optimization.
[0025] The second determining module is used to obtain the dose limits for the target area, organs at risk, and auxiliary contours based on the auxiliary contours and dose distribution.
[0026] The reverse optimization module is used to perform reverse optimization based on the auxiliary profile and dose constraint limits to generate a lung cancer radiotherapy plan.
[0027] Thirdly, the present invention provides a non-transitory computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the lung cancer radiotherapy planning generation method based on dose prediction and auxiliary contour as described above.
[0028] Fourthly, the present invention provides a computer program product, including a computer program that, when run on one or more processors, is used to implement the lung cancer radiotherapy planning generation method based on dose prediction and auxiliary profile as described above.
[0029] Fifthly, the present invention provides an electronic device, comprising: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the lung cancer radiotherapy planning generation method based on dose prediction and auxiliary contour as described above.
[0030] The beneficial effects of this invention are as follows: incorporating the minimum distance distribution from the patient voxel to the target area into the model training improves the accuracy of dose prediction; using the dose distribution to generate auxiliary contours makes the optimization conditions / targets more reasonable, thus greatly shortening the planning time; in terms of normal tissue protection, this automatic planning method is beneficial for the protection of organs at risk, such as the lungs; and automatically completing the radiotherapy planning for lung cancer is beneficial for rapidly improving the planning level of units lacking radiotherapy experience.
[0031] The advantages of additional aspects of the invention will be set forth more clearly in the following description or will be learned by practice of the invention. Attached Figure Description
[0032] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below 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 effort.
[0033] Figure 1 This is a flowchart illustrating the structure of the automatic lung cancer planning method based on dose prediction and auxiliary profile generation as described in an embodiment of the present invention.
[0034] Figure 2 This is a schematic diagram of the dose-based auxiliary profile generation method according to an embodiment of the present invention. Detailed Implementation
[0035] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.
[0036] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0037] It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as here.
[0038] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, and / or combinations thereof.
[0039] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.
[0040] To facilitate understanding of the present invention, the present invention will be further explained and described below with reference to the accompanying drawings and specific embodiments. However, the specific embodiments do not constitute a limitation on the embodiments of the present invention.
[0041] Those skilled in the art should understand that the accompanying drawings are merely schematic diagrams of embodiments, and the components in the drawings are not necessarily essential for implementing the present invention.
[0042] Example 1
[0043] In this embodiment 1, a lung cancer radiotherapy plan generation system based on dose prediction and auxiliary contours is first provided, including: an acquisition module for acquiring lung cancer radiotherapy plan case data; a prediction module for processing the radiotherapy plan case data using a trained dose prediction model to obtain dose distribution results for the target area and organs at risk; a first determination module for generating anisotropic auxiliary contours based on the predicted dose distribution and the dose constraints required for inverse optimization; a second determination module for obtaining dose limits for the target area, organs at risk, and auxiliary contours based on the auxiliary contours and dose distributions; and an inverse optimization module for performing inverse optimization based on the auxiliary contours and dose constraint limits to generate a lung cancer radiotherapy plan.
[0044] In this embodiment 1, the above-described system is used to implement a method for generating a lung cancer radiotherapy plan based on dose prediction and auxiliary contours. The method includes: using an acquisition module to acquire lung cancer radiotherapy plan case data; using a model training module to train a model using the lung cancer radiotherapy plan case data to obtain a radiotherapy dose prediction model; using a prediction module to process the radiotherapy plan case data using the trained dose prediction model to obtain dose distribution results for the target area and organs at risk; using a first determination module to generate anisotropic auxiliary contours based on the predicted dose distribution and the dose constraints required for inverse optimization; using a second determination module to obtain dose limits for the target area, organs at risk, and auxiliary contours based on the auxiliary contours and dose distributions; and using an inverse optimization module to perform inverse optimization based on the auxiliary contours and dose constraint limits to generate a lung cancer radiotherapy plan.
[0045] The dose prediction model is trained using a training set that includes multiple lung cancer radiotherapy plans. The lung cancer radiotherapy plans include at least CT images of the cases, dose distribution, target area and organ at risk contours, and minimum distance distribution from patient voxels to the target area. The model is trained using CT images, target area and organ at risk contours, and minimum distance distribution from patient voxels to the target area as input data and dose distribution as output data.
[0046] The closest distance between the patient's i-th point and the target area is defined as follows:
[0047]
[0048] Among them, D i This represents the shortest distance from the i-th voxel to the target region (x). i y i z i Let (x) be the coordinates of the i-th point, and (x) be the coordinates of the i-th point. k y k z k D represents the coordinates of point k on the target area; when voxel i is located within the target area or outside the patient's body, D... i Defined as 0.
[0049] The dose prediction model was trained using deep neural network frameworks, including Vgg-Unet, Res-Unet, Trans-Unet, and UNeXt.
[0050] Based on the dose distribution characteristics of lung cancer radiotherapy plans, anisotropic auxiliary contours are generated based on model-predicted doses. The width of these auxiliary contours varies in different directions around the target area depending on the rate of dose drop, and they are regions enclosed by specific isodose lines or patient contours. The dose constraint limit generation module generates auxiliary contours and dose limits for organs at risk and the target area based on the dose distribution, and imports these limits into the planning system via an interface program for automatic optimization of the radiotherapy plan.
[0051] Example 2
[0052] This invention provides an automated radiotherapy planning method based on dose prediction and auxiliary contour generation. It uses deep learning to predict dose distribution and automatically generates auxiliary contours and dose limit constraints for optimization based on the dose distribution. The method utilizes a radiotherapy planning system and its scripts to achieve automated and efficient planning. The automated lung cancer planning method based on dose prediction and auxiliary contour generation includes a dataset creation module, a deep neural network dose prediction module, an anisotropic auxiliary contour generation module, and a dose constraint limit generation module. The dataset module is a lung cancer radiotherapy planning case dataset, which is high-quality inverse intensity-modulated radiotherapy (IMRT) planning data that has been clinically approved and implemented, including CT images, contour structures of the target area and organs at risk, minimum distance distribution from the target area, and dose distribution. The deep neural network dose prediction module uses deep neural network frameworks, including Vgg-Unet, Res-Unet, Trans-Unet, and UNeXt. The anisotropic auxiliary contour generation module generates anisotropic auxiliary contours based on the dose distribution characteristics of the lung cancer radiotherapy plan and the dose predicted by the model.
[0053] Anisotropic auxiliary contours are generated, including areas enclosed by specific isodose lines or patient contours around the target region, where the width of the auxiliary contour varies in different directions depending on the rate of predicted dose drop. The dose constraint limit generation module generates auxiliary contours and dose limits for organs at risk and the target region based on the dose distribution, and imports these limits into the radiotherapy planning system via an interface program for automatic optimization of the radiotherapy plan.
[0054] Specifically:
[0055] A case dataset is established, which is intensity-modulated radiotherapy (IMRT) planning data. The IRT planning data includes at least the case CT images, dose distribution, target area and organ at risk contours, and minimum distance distribution from the patient voxel to the target area.
[0056] The closest distance between the patient's i-th point and the target area is defined as follows:
[0057]
[0058] Among them, D i This represents the shortest distance from the i-th voxel to the target region (x). i y i z i Let (x) be the coordinates of the i-th point, and (x) be the coordinates of the i-th point. k y k z k D represents the coordinates of point k on the target area. When voxel i is located within the target area or outside the patient's body, D... i Defined as 0.
[0059] A three-dimensional dose prediction model is constructed by deep learning training on CT images, target area and organ at risk contours, voxel-to-target minimum distance distribution information and dose distribution in the dataset.
[0060] The case data was input into a trained deep learning-based prediction model to obtain the predicted dose for the target area and organs at risk.
[0061] The auxiliary contour generation is performed by generating anisotropic auxiliary contours based on the predicted dose distribution and the dose constraints required for inverse optimization. During auxiliary contour generation, the generated auxiliary contour Ring1 is the region of interest enclosed by the 95% and 90% prescription dose lines. Similarly, the generated auxiliary contour Ring2 is the region of interest enclosed by the 90% and 85% prescription dose lines.
[0062] During the generation of the auxiliary contour, the generated auxiliary contour Nt is the region of interest enclosed by the 70% prescription dose line and the body surface contour. During the generation of the auxiliary contour, the generated auxiliary contour Fan is the region of interest enclosed by the 50% prescription dose line and the body surface contour.
[0063] Based on the dose distribution, the dose constraints for auxiliary contours, organs at risk, and target areas are calculated using a program algorithm. The dose constraints are then transformed into system optimization conditions through a script program in the planning system, completing subsequent optimization.
[0064] In summary, the automatic planning method based on dose prediction and auxiliary contour generation described in this embodiment incorporates the minimum distance distribution from the patient voxel to the target area into the model training, thereby improving the accuracy of dose prediction; using the dose distribution to generate auxiliary contours makes the optimization conditions / targets more reasonable, thus greatly shortening the planning design time; in terms of normal tissue protection, this automatic planning method is beneficial for the protection of organs at risk, such as the lungs; and the automatic completion of lung cancer radiotherapy planning is beneficial for rapidly improving the planning design level of units lacking radiotherapy experience.
[0065] Example 3
[0066] like Figure 1In this embodiment 3, a flowchart of an automatic lung cancer planning method based on dose prediction and auxiliary profile generation is provided, including functional modules such as establishing a dataset, model training, generating auxiliary profiles, and generating dose constraint limits.
[0067] The dataset is established using intensity-modulated radiotherapy (IMRT) plans for lung cancer patients. CT images, contour structures (including target areas and organs at risk), and dose distributions are extracted from the radiotherapy plans, converted into three-dimensional matrix forms, and aligned with coordinates. These are then stored in files of a specified format (such as ".h5" or ".npz") for model training.
[0068] The CT images are obtained through simulated positioning CT scans; the contour structure can be obtained by manual drawing on the CT images by the physician or by automatic drawing software. In this embodiment, the contour structure of the target area and organs at risk is obtained by manual drawing by the physician; the dose distribution is the dose information calculated by the planning system; to increase the accuracy of model prediction, the minimum distance distribution is introduced as input data for model training.
[0069] The closest distance from the i-th point to the target area is defined as follows:
[0070]
[0071] Among them, D i This represents the shortest distance from the i-th voxel to the target region, (x i y i z i Let (x) be the coordinates of the i-th point, and (x) be the coordinates of the i-th point. k y k z k D represents the coordinates of point k on the target area. When voxel i is located within the target area or outside the patient's body, D... i Defined as 0.
[0072] The model training involves establishing a deep learning network for dose prediction. The model network is a 3D network structure, and can be selected from networks such as Vgg-Unet, Res-Unet, Trans-Unet, and UNeXt. This embodiment selects Vgg-Unet. CT images, contour structures, and minimum distance distributions are used as the model's input data, and the dose distribution is used as the model's output data. The dataset is divided into two parts: a training set and a validation set. The training set is used to train and optimize the deep neural network structure parameters, and the validation set is used to evaluate the model's prediction performance and prevent overfitting.
[0073] The deep neural network dose prediction module extracts abstract features from the input data through multiple hidden layers, and the output layer makes predictions based on these extracted features. A neural network consists of neurons and connections between them. It is divided into input layers, hidden layers, and output layers. Compared to shallow neural networks, the "depth" of deep neural networks is reflected in more hidden layers, more flexible and complex connections, and more powerful non-linear representations. It can extract more essential features from the input image, thus achieving more accurate predictions.
[0074] The generated auxiliary contour is obtained by using the dose distribution from the prediction model and then using that dose distribution to obtain the anisotropic auxiliary contour required for inverse optimization, such as... Figure 2 As shown.
[0075] In lung cancer treatment planning, due to the need for stronger dose limits in the lung region, the rate of dose drop differs in the anterior-posterior and lateral directions. Using isotropic auxiliary contours such as Ring, Nt, and Fan for dose limits in this situation can lead to a deterioration in treatment planning quality. In this embodiment, an anisotropic auxiliary contour generated based on dose distribution is as follows:
[0076] Ring 1 is the region of interest enclosed by the 95% and 90% prescription dose lines.
[0077] Ring 2 is the region of interest enclosed by the 90% and 85% prescription dose lines.
[0078] Nt is the region of interest enclosed by the 70% prescription dose line and the body surface contour.
[0079] The generated dose constraint limits are calculated using a program algorithm based on the dose distribution to determine the dose limits for auxiliary contours, organs at risk, and target areas. The dose constraints are then transformed into system optimization conditions through a script program in the planning system, completing subsequent optimization.
[0080] Example 4
[0081] This embodiment 4 provides a non-transitory computer-readable storage medium for storing computer instructions. When executed by a processor, the computer instructions implement the lung cancer radiotherapy plan generation method based on dose prediction and auxiliary contour as described above. The method includes:
[0082] Obtain case data for lung cancer radiotherapy plans;
[0083] The trained dose prediction model is used to process radiotherapy planning case data to obtain dose distribution results for the target area and organs at risk;
[0084] An anisotropic auxiliary profile is generated based on the predicted dose distribution and the dose constraints required for inverse optimization;
[0085] Dose limits for the target area, organs at risk, and auxiliary contours are obtained based on the auxiliary contours and dose distribution.
[0086] Based on the auxiliary profile and dose constraint limits, inverse optimization is performed to generate a lung cancer radiotherapy plan.
[0087] Example 5
[0088] This embodiment 5 provides a computer program product, including a computer program that, when run on one or more processors, is used to implement the lung cancer radiotherapy planning generation method based on dose prediction and auxiliary profile as described above. The method includes:
[0089] Obtain case data for lung cancer radiotherapy plans;
[0090] The trained dose prediction model is used to process radiotherapy planning case data to obtain dose distribution results for the target area and organs at risk;
[0091] An anisotropic auxiliary profile is generated based on the predicted dose distribution and the dose constraints required for inverse optimization;
[0092] Dose limits for the target area, organs at risk, and auxiliary contours are obtained based on the auxiliary contours and dose distribution.
[0093] Based on the auxiliary profile and dose constraint limits, inverse optimization is performed to generate a lung cancer radiotherapy plan.
[0094] Example 6
[0095] This embodiment 6 provides an electronic device, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the lung cancer radiotherapy planning generation method based on dose prediction and auxiliary contour as described above, the method including:
[0096] Obtain case data for lung cancer radiotherapy plans;
[0097] The trained dose prediction model is used to process radiotherapy planning case data to obtain dose distribution results for the target area and organs at risk;
[0098] An anisotropic auxiliary profile is generated based on the predicted dose distribution and the dose constraints required for inverse optimization;
[0099] Dose limits for the target area, organs at risk, and auxiliary contours are obtained based on the auxiliary contours and dose distribution.
[0100] Based on the auxiliary profile and dose constraint limits, inverse optimization is performed to generate a lung cancer radiotherapy plan.
[0101] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0102] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0103] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0104] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment, whereby a series of operational steps are performed to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0105] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that, based on the technical solutions disclosed in the present invention, various modifications or variations that can be made by those skilled in the art without creative effort should be included within the scope of protection of the present invention.
Claims
1. A method for generating radiotherapy plans for lung cancer based on dose prediction and auxiliary profiles, characterized in that, include: Obtain case data for lung cancer radiotherapy plans; A trained dose prediction model is used to process radiotherapy planning case data to obtain dose distribution results for the target area and organs at risk. The dose prediction model is trained from a training set, which includes multiple lung cancer radiotherapy planning cases. The lung cancer radiotherapy planning cases include at least CT images of the cases, dose distribution, contours of the target area and organs at risk, and the minimum distance distribution from the patient voxels to the target area. The model is trained using CT images, contours of the target area and organs at risk, and the minimum distance distribution from the patient voxels to the target area as input data and dose distribution as output data. Anisotropic auxiliary contours are generated based on the predicted dose distribution and the dose constraints required for inverse optimization. Specifically, based on the dose distribution characteristics of the lung cancer radiotherapy plan, anisotropic auxiliary contours are generated based on the dose predicted by the model. The width of the auxiliary contours around the target area varies in different directions depending on the rate of drop of the predicted dose, and is the area enclosed by a specific isodose line or patient contour. Based on the auxiliary profile and dose distribution, dose constraint limits for the target area, organs at risk, and auxiliary profile are obtained; Based on the auxiliary profile and dose constraint limits, reverse optimization is performed to generate a radiotherapy plan for lung cancer. The dose constraint limit generation module generates auxiliary profiles and dose limits for organs at risk and target areas based on the dose distribution, and imports them into the planning system through an interface program for automatic optimization of the radiotherapy plan.
2. The method for generating a lung cancer radiotherapy plan based on dose prediction and auxiliary contours according to claim 1, characterized in that, The closest distance between the patient's i-th point and the target area is defined as follows: ； Among them, D i This represents the shortest distance from the i-th voxel to the target region (xi). i y i z i Let (x) be the coordinates of the i-th point, and (x) be the coordinates of the i-th point. k y k z k D represents the coordinates of point k on the target area; when voxel i is located within the target area or outside the patient's body, D... i Defined as 0.
3. The method for generating a lung cancer radiotherapy plan based on dose prediction and auxiliary contours according to claim 1, characterized in that, The dose prediction model was trained using deep neural network frameworks, including Vgg-Unet, Res-Unet, Trans-Unet, and UNeXt.
4. A lung cancer radiotherapy planning system based on dose prediction and auxiliary profile, characterized in that, include: The acquisition module is used to acquire lung cancer radiotherapy plan case data; The prediction module processes radiotherapy planning case data using a trained dose prediction model to obtain dose distribution results for the target area and organs at risk. The dose prediction model is trained from a training set, which includes multiple lung cancer radiotherapy planning cases. These cases include at least CT images, dose distributions, target area and organ at risk contours, and minimum distance distributions from patient voxels to the target area. The model is trained using CT images, target area and organ at risk contours, and minimum distance distributions from patient voxels to the target area as input data and dose distributions as output data. The first determining module is used to generate an anisotropic auxiliary contour based on the predicted dose distribution and the dose constraints required for inverse optimization. The anisotropic auxiliary contour is generated based on the dose predicted by the model according to the dose distribution characteristics of the lung cancer radiotherapy plan. The width of the auxiliary contour is different in each direction around the target area as the predicted dose falls at different rates. It is an area enclosed by a specific isodose line or patient contour. The second determining module is used to obtain dose constraint limits for the target area, organs at risk, and auxiliary contours based on the auxiliary contours and dose distribution. The reverse optimization module is used to perform reverse optimization based on auxiliary profiles and dose constraint limits to generate a radiotherapy plan for lung cancer. The dose constraint limit generation module generates auxiliary profiles and dose limit conditions for organs at risk and target areas based on dose distribution, and imports them into the planning system through an interface program for automatic optimization of the radiotherapy plan.
5. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the lung cancer radiotherapy planning generation method based on dose prediction and auxiliary contour as described in any one of claims 1-3.
6. An electronic device, characterized in that, include: The electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the lung cancer radiotherapy planning generation method based on dose prediction and auxiliary profile as described in any one of claims 1-3.