Irradiation field angle and intensity optimization method and system based on multi-modal image

Abstract

The invention discloses an irradiation field angle and intensity optimization method and system based on a multi-modal image, and relates to the technical field of medical image processing, and the method comprises the steps: carrying out the fusion processing of a CT image and a functional image of a target region of a patient, and obtaining a fusion image; the information of the fusion image comprises the three-dimensional structure information of the target region and the endangered organ and the gray scale and function information of the internal voxel of each endangered organ; optimizing an irradiation field angle and irradiation field intensity in the radiotherapy plan according to the fused image and a target dose demand; the target dose demand is to preferentially reduce the radiation dose of voxels endangering high internal function coefficients of organs on the premise of ensuring that the target region dose meets the prescription requirement and the organ-endangering dose meets the normal tissue dose constraint condition. According to the invention, the dosage of high-function voxels endangering the interior of organs of a tumor radiotherapy patient can be better reduced, and a radiotherapy physicist can be assisted to formulate a more individualized radiotherapy scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a method and system for optimizing the angle and intensity of a radiation therapy irradiation field based on multimodal medical images. Background Art

[0002] Radiotherapy (abbreviated as radiotherapy) uses ionizing radiation to control or kill tumor cells and is one of the main means of tumor treatment. Compared with traditional three-dimensional conformal radiotherapy (3DCRT), intensity-modulated radiotherapy (IMRT) can significantly reduce the radiation dose received by normal organs. At present, IMRT technology has become the most widely used tumor radiotherapy technology. The radiotherapy equipment corresponding to IMRT includes medical linear accelerators based on X-rays, spiral tomography equipment, CyberKnife, etc., as well as proton and heavy particle radiotherapy equipment based on high-energy beams. Before using the above equipment for IMRT radiotherapy, medical physicists need to use the radiotherapy planning system software (TPS) to determine the angle of each irradiation field, and optimize the intensity of each irradiation field by adjusting the dose constraints of the organ or target area. The irradiation field angle is mainly determined by medical physicists based on previous work experience by observing the anatomical position of organs and tumors. The optimization of irradiation field intensity is achieved by adjusting the constraints of the dose-volume histogram (DVH) and minimizing the physical dose of the organ at risk under the premise that the target area meets the prescription dose. At present, the above process is mainly based on single-modality CT images. CT images represent the density of each voxel point inside the organ with different grayscales, but cannot reflect the functional information of the voxel point. In addition, the above process assumes that the functional distribution of normal organs is uniform. In fact, the functional distribution of normal organs in different cancer patients is highly heterogeneous, and the functional degree of each voxel inside the organ is different. Single-modality imaging cannot incorporate the functional information of voxels into the optimization of the irradiation field, which will cause the voxels with good functions inside the organ to be over-radiated, which will lead to serious radiotherapy complications. Relevant clinical studies on lung cancer radiotherapy patients have confirmed that giving priority to the radiation dose of high-function voxels in the lungs can better protect the lung function after radiotherapy and reduce the incidence of radiation pneumonitis. Therefore, the optimization process of IMRT radiotherapy plan should incorporate the functional information of each voxel of the organ at risk, and provide better protection for the organ by reducing the radiation dose of voxels with higher functional values. Summary of the invention

[0003] To date, a variety of algorithms (simulated annealing, genetic algorithm, exhaustive search algorithm, etc.) have been used to optimize the irradiation field. However, the dose target of the existing irradiation field optimization is only to reduce the physical dose of the organ at risk while ensuring that the target dose meets clinical requirements, without incorporating the functional information of each voxel inside the organ into the optimization process. Irradiation field optimization that does not consider voxel function may set the irradiation angle and intensity in the area where the high-function voxels are located in order to optimize the lowest dose for the entire organ at risk. In addition, irradiation field optimization based on a single CT modality image cannot meet the needs of individualized precision radiotherapy at this stage.

[0004] In view of this, the purpose of the present invention is to provide a method and system for optimizing the angle and intensity of the irradiation field based on multimodal images, so as to introduce the organ function information in the multimodal images into the irradiation field optimization, and avoid arranging the irradiation field in a high-function area in order to reduce the overall organ dose.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] In a first aspect, the present invention provides a method for optimizing the angle and intensity of an irradiation field based on multimodal imaging, comprising:

[0007] Acquire a multimodal image of a target area of ​​a patient; the target area includes a target area and a human organ area located around the target area; the human organ area includes one or more organs at risk; the multimodal image includes at least a CT image and a functional image; the CT image is used to provide three-dimensional structural information of the organs at risk and the target area, and the functional image is used to provide functional distribution information of the organs at risk;

[0008] Performing registration and fusion processing on the CT image and the functional image to obtain a fused image; the fused image includes the target area, three-dimensional structural information of the organ at risk, and grayscale information and functional information of each voxel inside the organ at risk;

[0009] According to the fused image and the target dose requirement, the irradiation field angle and the irradiation field intensity in the radiotherapy plan are optimized; the target dose requirement is to give priority to reducing the radiation dose of voxels with high internal functional coefficients of the organs at risk, under the premise of ensuring that the target area dose meets the prescription requirements and the organ at risk dose meets the normal tissue dose constraint conditions; the functional coefficient is determined according to the functional distribution information.

[0010] Optionally, the acquiring of a multimodal image of a target area of ​​a patient specifically includes:

[0011] Acquire a CT image of the patient's target area;

[0012] Acquire a functional image of a target area of ​​a patient; the functional image includes a medical functional image containing organ function information and / or a virtual functional image generated using deep learning technology;

[0013] The process of determining the virtual function image is as follows:

[0014] Using deep learning technology, CT images and functional images of the sample area are processed to establish an image generation model;

[0015] The image generation model and the CT image of the target area are used to generate a virtual functional image of the target area.

[0016] Optionally, performing registration and fusion processing on the CT image and the functional image to obtain a fused image specifically includes:

[0017] Using a deformable registration technology or a rigid registration technology, the functional images of different modalities are respectively registered with the CT image to obtain registered functional images;

[0018] According to the spatial resolution of the CT image, interpolation or resampling processing is performed on the registered functional image to obtain a processed functional image; wherein the spatial resolution and spatial coordinates of the processed functional image are consistent with those of the CT image;

[0019] The processed functional image is fused with the CT image to obtain a fused image containing functional information and anatomical information.

[0020] Optionally, optimizing the irradiation field angle and irradiation field intensity of the radiotherapy plan according to the fused image and the target dose requirement specifically includes:

[0021] Constructing a candidate irradiation field angle set; the candidate irradiation field angle set includes a plurality of feasible irradiation field angles;

[0022] According to the fused image and the candidate irradiation field set, an iterative optimization algorithm is used to calculate the optimal solution that satisfies the target dose loss function, thereby determining one or more optimal irradiation field angles in the radiotherapy plan, and the intensity distribution corresponding to each optimal irradiation field angle; the target dose loss function is determined according to the target dose requirement; the optimal irradiation field angle is used to determine the gantry angle during treatment; the intensity distribution corresponding to the optimal irradiation field angle is used to determine the motion sequence of the multi-leaf collimator leaves.

[0023] Optionally, the target dose loss function is determined as follows:

[0024] constructing an organ-at-risk dose loss function, wherein the organ-at-risk dose loss function incorporates functional information of each organ-at-risk voxel;

[0025] Construct target dose loss function;

[0026] A total target loss function is determined according to the organ-at-risk dose loss function and the target area dose loss function.

[0027] Optionally, the dose loss function of the organ at risk is:

[0028]

[0029] U(d) represents the objective function of radiation dose to organs at risk;

[0030] N OAR Represents the total number of all organs at risk that need to be considered for radiotherapy planning;

[0031] represents the maximum acceptable dose at voxel k in organ at risk m;

[0032] d mk represents the calculated dose at voxel k in organ at risk m under the beam weight vector d;

[0033] N m represents the total number of voxels in the organ at risk m;

[0034] F mk represents the functional coefficient at voxel k in the organ at risk m;

[0035] a m represents the weight of the functional information of the organ at risk m during optimization; a m is a manually adjustable value, by adjusting a m The size of changes the weight of functional information in the optimization process; if there is no functional information of the endangered organ m, then a m =0;

[0036] + represents the positive operator.

[0037] Optionally, the target area dose loss function is:

[0038]

[0039] T(d) represents the objective function of target area irradiation dose;

[0040] represents the expected irradiation dose at voxel k in the target area;

[0041] N T represents the total number of voxels in the planned target volume;

[0042] d k represents the dose calculated at voxel k;

[0043] n represents the total number of irradiation fields.

[0044] In a second aspect, the present invention provides an irradiation field angle and intensity optimization system based on multimodal imaging, comprising:

[0045] A multimodal image acquisition module, used to acquire multimodal images of a target area of ​​a patient; the target area includes a target area and a human organ area located around the target area; the human organ area includes one or more organs at risk; the multimodal images include at least a CT image and a functional image; the CT image is used to provide three-dimensional structural information of the organs at risk and the target area, and the functional image is used to provide functional distribution information of the organs at risk;

[0046] An image fusion module, used for performing registration and fusion processing on the CT image and the functional image to obtain a fused image; the fused image includes the target area, three-dimensional structural information of the organ at risk, and grayscale information and functional information of each voxel inside the organ at risk;

[0047] The irradiation field angle and intensity optimization module is used to optimize the irradiation field angle and irradiation field intensity in the radiotherapy plan according to the fused image and the target dose requirement; the target dose requirement is to give priority to reducing the radiation dose of voxels with high internal function coefficients of the organs at risk, under the premise of ensuring that the target area dose meets the prescription requirements and the organ at risk dose meets the normal tissue dose constraint conditions; the function coefficient is determined according to the function distribution information.

[0048] Optionally, the multimodal image acquisition module specifically includes:

[0049] A CT image acquisition unit, used to acquire a CT image of a target area of ​​a patient;

[0050] A functional image acquisition unit, used to acquire a functional image of a target area of ​​a patient; the functional image includes a medical functional image containing organ function information and / or a virtual functional image generated using deep learning technology;

[0051] The process of determining the virtual function image is as follows:

[0052] Using deep learning technology, CT images and functional images of the sample area are processed to establish an image generation model;

[0053] The image generation model and the CT image of the target area are used to generate a virtual functional image of the target area.

[0054] Optionally, the image fusion module specifically includes:

[0055] A registration unit, configured to register the functional images of different modalities with the CT image respectively using a deformation registration technology or a rigid registration technology to obtain a registered functional image;

[0056] A processing unit, configured to perform interpolation or resampling processing on the registered functional image according to the spatial resolution of the CT image to obtain a processed functional image; wherein the spatial resolution and spatial coordinates of the processed functional image are consistent with those of the CT image;

[0057] The fusion unit is used to fuse the processed functional image with the CT image to obtain a fused image containing functional information and anatomical information.

[0058] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0059] First, the combined application of multimodal images is realized to optimize the irradiation field angle and intensity.

[0060] Second, the organ function information is incorporated into the irradiation field optimization. During the optimization process, the functional coefficient of each voxel inside the endangered organ is considered, and the voxels with high functional coefficients are protected.

[0061] Third, the individualized irradiation field angle and intensity can be determined based on a patient's medical images to meet the current requirements for precision radiotherapy.

[0062] Fourth, it can help medical physicists determine the irradiation field angle of the organ function protection radiotherapy plan and improve the efficiency and quality of radiotherapy plan formulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0064] Figure 1 A flowchart of a method for optimizing the angle and intensity of an irradiation field based on multimodal imaging provided by an embodiment of the present invention;

[0065] Figure 2 A structural diagram of an irradiation field angle and intensity optimization system based on multimodal imaging provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] With the development of medical imaging technology, more functional images can be obtained before radiotherapy for cancer patients. For example, lung perfusion images that reflect lung function and magnetic resonance perfusion images that reflect blood flow relationships. The development and joint application of these multimodal imaging technologies have laid the foundation for mining the functional distribution information of organs at risk. In addition, deep learning technology has also been widely used in the generation of medical images and has shown broad prospects for clinical application. Virtual functional images generated by generative adversarial networks in deep learning also contain functional information of human organs. In recent years, retrospective studies have shown that prioritizing the reduction of radiation doses of high-function voxels inside organs at risk can better protect organ function, reduce the probability of radiotherapy complications, and bring more significant treatment benefits to patients.

[0068] Incorporating the functional information of each voxel into the angle and intensity optimization of the tumor's radiotherapy irradiation field can better protect the function of the organ. At the same time, in order to achieve personalized precision medicine for tumor patients, the functional information of voxels should also be incorporated into the radiotherapy plan design and optimization process. However, this improvement is bound to increase the difficulty of radiotherapy planning and pose greater challenges to radiation physicists.

[0069] Therefore, the method of optimizing the radiation field angle and intensity based on multimodal imaging is of great significance for protecting the function of endangered organs and assisting physicists in making radiation therapy plans. However, in this field, there is no method for optimizing the radiation field angle and intensity of IMRT by combining multimodal imaging.

[0070] In view of this, the present invention provides a method and system for optimizing the angle and intensity of the irradiation field based on multimodal imaging. The overall process of the present invention can be summarized as: image acquisition, image preprocessing, image segmentation, image fusion, and optimization of the angle and intensity of the irradiation field. The purpose of the first four steps is to provide the target area, the three-dimensional structural information of the organs at risk, and the functional information of the internal voxels of each organ at risk for the final optimization.

[0071] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0072] Embodiment 1

[0073] Figure 1A flowchart of a method for optimizing the angle and intensity of an irradiation field based on multimodal imaging is provided in an embodiment of the present invention. Figure 1 As shown, an embodiment of the present invention provides a method for optimizing the angle and intensity of an irradiation field based on multimodal imaging, which includes the following steps.

[0074] Step 101: Acquire a multimodal image of a target area of ​​a patient; the target area includes a target area and a human organ area located around the target area; the human organ area includes one or more organs at risk; the multimodal image includes at least a CT image and a functional image; the CT image is used to provide three-dimensional structural information of the organs at risk and the target area, and the functional image is used to provide functional distribution information of the organs at risk.

[0075] This step specifically includes:

[0076] S1: Acquire a CT image of the target area of ​​the patient.

[0077] An example: Obtaining a CT image of a tumor patient before radiotherapy positioning; the CT image has the structure of the organ at risk and the tumor area outlined by the physician. This organ structure and tumor area are used to construct the three-dimensional target area, the shape and spatial position of the organ.

[0078] S2: Acquire functional images of the target area of ​​the patient; these functional images include but are not limited to medical functional images containing organ function information such as PET, MRI perfusion and SPECT, and virtual functional images generated using deep learning technology.

[0079] S2-1: Medical Functional Images

[0080] The present invention can select corresponding functional medical examination images according to the functional characteristics of different organs, including but not limited to existing three-dimensional medical images such as lung perfusion images, PET images, spectral CT images, and nuclear magnetic perfusion images. Considering the different functional imaging modes of different organs at risk, multiple functional images of different modalities can be obtained in this step.

[0081] S2-2: Virtual functional images generated by deep learning technology

[0082] Since medical functional imaging examinations are expensive and easily restricted by the patient's physical condition, some patients may not be able to undergo high-cost functional imaging scans. Deep learning technology can be used to generate virtual functional images from low-cost CT images based on trained models.

[0083] The example of generating lung perfusion images from low-cost CT images is used for explanation (this is just an example, and the process can be used to generate images of more modalities for use in this application, such as generating multi-spectral CT images from CT images, and generating virtual perfusion MRI images from CT images). The specific implementation of this step is:

[0084] S2-2-1: Scanning and acquiring SPECT and CT images; acquiring the patient's lung perfusion image (SPECT image) and CT image.

[0085] S2-2-2: Image registration: perform rigid or deformation registration on the SPECT image and CT image to determine the relative spatial position of the lung tissue, and ensure that the spatial coordinate information of the voxels between the CT and SPECT images is consistent through image registration.

[0086] S2-2-3: Lung tissue segmentation: segment the left lung tissue and the right lung tissue in the registered image.

[0087] S2-2-4: Image scale correction: Since SPECT images and CT images have different spatial resolutions, the voxel sizes between the images are different. This step can ensure that the spatial resolutions of SPECT images and CT images are the same by image interpolation.

[0088] S2-2-5: training a deep learning network model; using the SPECT images and CT images processed in the above steps as network inputs to train the deep learning network model. The deep learning network models used in this step include but are not limited to generative adversarial networks, convolutional neural networks, residual neural networks, etc.

[0089] S2-2-6: Virtual functional image generation: The CT image is used as input of the trained model to generate and output a SPECT image containing lung function information.

[0090] Step 102: performing registration and fusion processing on the CT image and the functional image to obtain a fused image; the fused image includes the target area, three-dimensional structural information of the organ at risk, and grayscale information and functional information of each voxel inside the organ at risk.

[0091] This step specifically includes:

[0092] Step A: Image registration; Since the spatial coordinate positions of different images are inconsistent, the present invention uses existing deformation registration technology and / or rigid registration technology to register functional images of different modalities with CT images respectively.

[0093] Step B: image preprocessing; since the spatial resolutions of the multi-modal and multi-sequence images registered in step A are different, images of other modalities are interpolated or resampled to ensure that the spatial resolutions between different modalities are consistent with the CT image, that is, according to the spatial resolution of the CT image, the registered functional image is interpolated or resampled to obtain a processed functional image; wherein the spatial resolution of the processed functional image is consistent with the spatial resolution of the CT image; and the spatial coordinates of the processed functional image are consistent with the spatial coordinates of the CT image.

[0094] Step C: Image fusion; fusing the processed functional image with the CT image to obtain a fused image; the information of the fused image includes the three-dimensional structural information of the target area and the organ at risk and the grayscale and functional information of each voxel inside the organ at risk.

[0095] Step 103: Optimize the irradiation field angle and irradiation field intensity in the radiotherapy plan according to the fused image and the target dose requirement; the target dose requirement is to prioritize reducing the radiation dose requirement of voxels with high internal function coefficients of organs at risk, on the premise that the target area dose meets the prescription requirements and the organ at risk dose meets the normal tissue dose constraint; the function coefficient is determined according to the function distribution information.

[0096] According to the above steps, a fused image containing the three-dimensional structural information and functional distribution information of the organs at risk has been obtained, that is, the functional distribution information of the organs at risk will be included in the optimization process of this application. Therefore, in the process of optimizing the angle and intensity of the irradiation field, the radiation dose of voxels with higher functional coefficients is preferentially reduced while ensuring that the radiation dose of the organs at risk is as low as possible. Based on the three-dimensional structural information and functional distribution information of the organs at risk contained in the fused image, the angle and intensity of the irradiation field are optimized and calculated.

[0097] The above steps specifically include: first constructing a set of candidate irradiation field angles; the candidate irradiation field angle set includes multiple feasible irradiation field angles, and the irradiation field angles are used to determine the gantry angle; then, based on the fused image and the candidate irradiation field set, an iterative optimization algorithm is used to calculate the optimal solution that satisfies the target dose loss function, and then one or more optimal irradiation field angles in the radiotherapy plan, and the intensity distribution corresponding to each optimal irradiation field angle are determined; the target dose loss function is determined according to the target dose requirement; the optimal irradiation field angle is used to determine the gantry angle during treatment; the intensity distribution corresponding to the optimal irradiation field angle is used to determine the motion sequence of the multi-leaf collimator leaves.

[0098] The process of constructing the candidate irradiation field set is as follows:

[0099] All possible irradiation field angles are generated at fixed angle intervals (eg, 5°) and included in candidate irradiation field set B, which covers all feasible irradiation field angles but does not include angles that may cause a collision between the treatment couch or the patient and the gantry.

[0100] b: irradiation field b, b∈B.

[0101] D b : Dose flux matrix of irradiation field b.

[0102] Represents the dose contribution of sub-field j of irradiation field b to voxel i under unit intensity.

[0103] X b : The intensity vector of the irradiation field b.

[0104] The intensity of sub-field j of irradiation field b.

[0105] d b : Dose distribution of irradiation field b; d b =D b X b .

[0106] f(d): The target quality function of the radiotherapy plan, which is a scalar function of the dose distribution d.

[0107] To optimize the angle, the embodiment of the present invention introduces a binary vector v, v b ∈{0,1}. When irradiation field b is used, v b =1, 0 when not in use.

[0108] The optimal optimization problem of n irradiation fields can be expressed as

[0109]

[0110] Constraints:

[0111] d=∑ b∈B v b D b X b ;

[0112] ∑ b∈B v b ≤n;

[0113]

[0114] v b ∈{0,1}.

[0115] The determination process of the target dose loss function is:

[0116] Step 1: Calculate the functional coefficient of the organ at risk

[0117] The functional coefficient of each organ-at-risk voxel i is calculated based on the fused image. The functional coefficient is calculated as follows:

[0118]

[0119] in, represents the functional coefficient of voxel i in organ at risk m; represents the functional image value of voxel i in organ at risk m; Min m Represents the minimum value of the functional image in the organ at risk m; Max m Indicates the maximum value of the functional image of the organ at risk m.

[0120] Step 2: Construct target dose loss function.

[0121] To select a suitable irradiation angle, both the dose of the target area and the dose limit of the organ at risk should be considered. The embodiment of the present invention uses the maximum entropy method based on information theory to construct the target area dose loss function of the radiotherapy plan.

[0122] Target area irradiation requires a minimum dose lower limit (prescription dose) and a maximum dose upper limit. The prescription dose is known when designing the radiotherapy plan, and the target area dose loss function is constructed based on the prescription dose. The target area irradiation dose target dose function T(d) can be defined as:

[0123]

[0124] T(d): objective function of target area irradiation dose

[0125] The desired irradiation dose (prescription dose) for voxel k in the target volume.

[0126] N T : The total number in the planning target volume (PTV).

[0127] d k : The dose calculated at voxel k is calculated by formula (4).

[0128]

[0129] n: represents the total number of irradiation fields.

[0130] Step 3: Construct the dose loss function for organs at risk

[0131] Based on the functional information matrix, the dose loss function U(d) of the organ at risk including the functional loss is calculated:

[0132]

[0133] U(d): objective function of radiation dose to organs at risk;

[0134] N OAR : The total number of organs at risk that need to be considered in radiotherapy planning.

[0135] Maximum acceptable dose to voxel k in organ at risk m.

[0136] d mk : The calculated dose to voxel k in organ at risk m under the beam weight vector d.

[0137] N m : the total number of voxels in the organ at risk m;

[0138] F mk : Functional coefficient of voxel k in organ at risk m.

[0139] a m : The weight of the functional information of the organ at risk m during optimization; a m is a manually adjustable value, by adjusting a m The size of changes the weight of functional information in the optimization process. If no effective functional image information is collected for the organ at risk m, then a m =0.

[0140] +: Positive operator; ensures that only threshold dose violations contribute to the OAR dose penalty function.

[0141] Step 4: Construct target dose loss function

[0142] Combining formulas (3) and (5), the target dose loss function is constructed:

[0143] f(d)=T(d)+λU(d) (6)

[0144] In formula (6), the weight between the target area and the organ at risk is adjusted by λ. If λ is small, the target area has a large weight, and if λ is large, the organ at risk has a large weight. In this case, the result is more inclined to protect the organ at risk. f(d) represents the deviation between the calculated dose of the target area and the expected dose, and the reduction between the calculated dose of the organ at risk and the maximum allowable dose.

[0145] The optimal solution of the above formula is calculated by an iterative method or a gradient descent method. The optimal irradiation field angle set and the intensity distribution d of each irradiation field in the optimal irradiation field angle set are obtained by solving f(d). Iterative optimization algorithms include but are not limited to simulated annealing, genetic algorithm, artificial neural network algorithm, and Newton gradient algorithm.

[0146] To achieve the above objectives, the present invention also provides an irradiation field angle and intensity optimization system based on multimodal imaging, such as Figure 2 As shown, including:

[0147] The multimodal image acquisition module 201 is used to acquire multimodal images of a target area of ​​a patient; the target area includes a target area and a human organ area located around the target area; the human organ area includes one or more organs at risk; the multimodal images include at least a CT image and a functional image; the CT image is used to provide three-dimensional structural information of the organs at risk and the target area, and the functional image is used to provide functional distribution information of the organs at risk.

[0148] The image fusion module 202 is used to perform registration and fusion processing on the CT image and the functional image to obtain a fused image; the fused image includes the target area, three-dimensional structural information of the organ at risk, and the grayscale information and functional information of each voxel inside the organ at risk.

[0149] The irradiation field angle and intensity optimization module 203 is used to optimize the irradiation field angle and irradiation field intensity in the radiotherapy plan according to the fused image and the target dose requirement; the target dose requirement is to give priority to reducing the radiation dose of voxels with high internal function coefficients of the organs at risk, on the premise of ensuring that the target area dose meets the prescription requirements and the organ at risk dose meets the normal tissue dose constraint conditions; the function coefficient is determined according to the function distribution information.

[0150] The multimodal image acquisition module 201 specifically includes:

[0151] The CT image acquisition unit is used to acquire a CT image of a target area of ​​a patient.

[0152] A functional image acquisition unit is used to acquire a functional image of a target area of ​​a patient; the functional image includes a medical functional image containing organ function information and / or a virtual functional image generated using deep learning technology.

[0153] The process of determining the virtual function image is as follows:

[0154] Using deep learning technology, the CT images and functional images of the sample area are processed to establish an image generation model.

[0155] The image generation model and the CT image of the target area are used to generate a virtual functional image of the target area.

[0156] The image fusion module 202 specifically includes:

[0157] The registration unit is used to register the functional images of different modalities with the CT image respectively by using a deformation registration technology or a rigid registration technology to obtain a registered functional image.

[0158] A processing unit is used to interpolate or resample the registered functional image according to the spatial resolution of the CT image to obtain a processed functional image; wherein the spatial resolution and spatial coordinates of the processed functional image are consistent with those of the CT image.

[0159] The fusion unit is used to fuse the processed functional image with the CT image to obtain a fused image containing functional information and anatomical information.

[0160] Compared with the traditional irradiation field intensity optimization scheme, the present invention has the following innovations:

[0161] First, the present invention proposes a concept and method for optimizing the angle and intensity of the irradiation field based on the joint application of multimodal images.

[0162] Second, the present invention incorporates organ function information into the optimization of irradiation field angle and intensity, and determines the gantry angle and multi-leaf collimator leaf motion sequence accordingly, which can better reduce the radiation dose of high-function coefficient voxels endangering the organs.

[0163] Compared with the traditional angle optimization scheme, the present invention has the following advantages:

[0164] First, the combination of multimodal images is used to determine the gantry angle of the radiotherapy accelerator and the multi-leaf collimator leaf motion sequence.

[0165] Second, functional information is introduced into the gantry angle optimization. During the optimization process, the functional information of each voxel inside the organ at risk is considered, and the radiation dose of high-functional voxels is preferentially reduced.

[0166] Third, according to the differences in organ functions among different patients, the weight of the functional information can be adjusted during the optimization process to achieve better dose distribution.

[0167] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0168] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for optimizing the angle and intensity of an irradiation field based on multimodal imaging, characterized in that: include: Acquire a multimodal image of a target area of ​​a patient; the target area includes a target area and a human organ area located around the target area; the human organ area includes one or more organs at risk; the multimodal image includes at least a CT image and a functional image; the CT image is used to provide three-dimensional structural information of the organs at risk and the target area, and the functional image is used to provide functional distribution information of the organs at risk; Performing registration and fusion processing on the CT image and the functional image to obtain a fused image; the fused image includes the target area, three-dimensional structural information of the organ at risk, and grayscale information and functional information of each voxel inside the organ at risk; According to the fused image and the target dose requirement, the irradiation field angle and the irradiation field intensity in the radiotherapy plan are optimized; the target dose requirement is to give priority to reducing the radiation dose of voxels with high internal functional coefficients of the organs at risk, under the premise of ensuring that the target area dose meets the prescription requirements and the organ at risk dose meets the normal tissue dose constraint conditions; the functional coefficient is determined according to the functional distribution information.

2. The method for optimizing the angle and intensity of the irradiation field based on multimodal imaging according to claim 1, characterized in that: The step of acquiring a multimodal image of a target area of ​​a patient specifically includes: Acquire a CT image of the patient's target area; Acquire a functional image of a target area of ​​a patient; the functional image includes a medical functional image containing organ function information and / or a virtual functional image generated using deep learning technology; The process of determining the virtual function image is as follows: Using deep learning technology, CT images and functional images of the sample area are processed to establish an image generation model; The image generation model and the CT image of the target area are used to generate a virtual functional image of the target area.

3. The method for optimizing the angle and intensity of the irradiation field based on multimodal imaging according to claim 1, characterized in that: The registering and fusing the CT image and the functional image to obtain a fused image specifically includes: Using a deformable registration technology or a rigid registration technology, the functional images of different modalities are respectively registered with the CT image to obtain registered functional images; According to the spatial resolution of the CT image, interpolation or resampling processing is performed on the registered functional image to obtain a processed functional image; wherein the spatial resolution and spatial coordinates of the processed functional image are consistent with those of the CT image; The processed functional image is fused with the CT image to obtain a fused image containing functional information and anatomical information.

4. The method for optimizing the angle and intensity of the irradiation field based on multimodal imaging according to claim 1, characterized in that: Optimizing the irradiation field angle and irradiation field intensity of the radiotherapy plan according to the fused image and the target dose requirement specifically includes: Constructing a candidate irradiation field angle set; the candidate irradiation field angle set includes a plurality of feasible irradiation field angles; According to the fused image and the candidate irradiation field set, an iterative optimization algorithm is used to calculate the optimal solution that satisfies the target dose loss function, thereby determining one or more optimal irradiation field angles in the radiotherapy plan, and the intensity distribution corresponding to each optimal irradiation field angle; the target dose loss function is determined according to the target dose requirement; the optimal irradiation field angle is used to determine the gantry angle during treatment; the intensity distribution corresponding to the optimal irradiation field angle is used to determine the motion sequence of the multi-leaf collimator leaves.

5. The method for optimizing the angle and intensity of the irradiation field based on multimodal imaging according to claim 4, characterized in that: The determination process of the target dose loss function is: constructing an organ-at-risk dose loss function, wherein the organ-at-risk dose loss function incorporates functional information of each organ-at-risk voxel; Construct target dose loss function; A total target loss function is determined according to the organ-at-risk dose loss function and the target area dose loss function.

6. The method for optimizing the angle and intensity of the irradiation field based on multimodal imaging according to claim 5, characterized in that: The dose loss function of the organ at risk is: U(d) represents the objective function of radiation dose to organs at risk; N OAR Represents the total number of organs at risk that need to be considered for radiotherapy planning; represents the maximum acceptable dose at voxel k in organ at risk m; d mk represents the calculated dose at voxel k in organ at risk m under the beam weight vector d; N m represents the total number of voxels in the organ at risk m; F mk represents the functional coefficient at voxel k in the organ at risk m; a m represents the weight of the functional information of the organ at risk m during optimization; a m is a manually adjustable value, by adjusting a m The size of changes the weight of functional information in the optimization process; If there is no functional information on the organ at risk, then a m =0; + represents the positive operator.

7. The method for optimizing the angle and intensity of the irradiation field based on multimodal imaging according to claim 5, characterized in that: The target area dose loss function is: T(d) represents the objective function of target area irradiation dose; represents the expected irradiation dose at voxel k in the target area; N T represents the total number of voxels in the planned target volume; d k represents the dose calculated at voxel k; n represents the total number of irradiation fields.

8. A system for optimizing the angle and intensity of an irradiation field based on multimodal imaging, characterized in that: include: A multimodal image acquisition module, used to acquire multimodal images of a target area of ​​a patient; the target area includes a target area and a human organ area located around the target area; the human organ area includes one or more organs at risk; the multimodal images include at least a CT image and a functional image; the CT image is used to provide three-dimensional structural information of the organs at risk and the target area, and the functional image is used to provide functional distribution information of the organs at risk; An image fusion module, used for performing registration and fusion processing on the CT image and the functional image to obtain a fused image; the fused image includes the target area, three-dimensional structural information of the organ at risk, and grayscale information and functional information of each voxel inside the organ at risk; The irradiation field angle and intensity optimization module is used to optimize the irradiation field angle and irradiation field intensity in the radiotherapy plan according to the fused image and the target dose requirement; the target dose requirement is to give priority to reducing the radiation dose of voxels with high internal function coefficients of the organs at risk, under the premise of ensuring that the target area dose meets the prescription requirements and the organ at risk dose meets the normal tissue dose constraint conditions; the function coefficient is determined according to the function distribution information.

9. The irradiation field angle and intensity optimization system based on multimodal imaging according to claim 8, characterized in that: The multimodal image acquisition module specifically includes: A CT image acquisition unit, used to acquire a CT image of a target area of ​​a patient; A functional image acquisition unit, used to acquire a functional image of a target area of ​​a patient; the functional image includes a medical functional image containing organ function information and / or a virtual functional image generated using deep learning technology; The process of determining the virtual function image is as follows: Using deep learning technology, CT images and functional images of the sample area are processed to establish an image generation model; The image generation model and the CT image of the target area are used to generate a virtual functional image of the target area.

10. The irradiation field angle and intensity optimization system based on multimodal imaging according to claim 8, characterized in that: The image fusion module specifically includes: A registration unit, configured to register the functional images of different modalities with the CT image respectively using a deformation registration technology or a rigid registration technology to obtain a registered functional image; A processing unit, configured to perform interpolation or resampling processing on the registered functional image according to the spatial resolution of the CT image to obtain a processed functional image; wherein the spatial resolution and spatial coordinates of the processed functional image are consistent with those of the CT image; A fusion unit is used to fuse the processed functional image with the CT image to obtain a fused image containing functional information and anatomical information.

Images