Radiotherapy dose regulation and control method and system based on head and neck positioning error

Abstract

The invention relates to the technical field of medical radiation equipment, in particular to a radiotherapy dose regulation and control method and system based on head and neck positioning errors, and the method comprises the steps: carrying out the structural analysis of a tumor through real-time imaging equipment, obtaining a noise reduction tumor structure diagram, carrying out the activity analysis of the tumor through positron emission tomography equipment, and obtaining a noise reduction tumor structure diagram; obtaining a tumor three-dimensional activity diagram; performing radiotherapy dose prediction operation on the tumor three-dimensional activity diagram to obtain a radiotherapy dose distribution matrix; performing respiration monitoring on the patient to obtain a respiration monitoring result; and when the patient is in a breathing limit position, intercepting the treatment area matrix according to a pre-constructed ray irradiation port position to obtain an offset verification matrix, and performing dose distribution adjustment on the radiotherapy dose distribution matrix according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix for irradiation. According to the invention, the safety and accuracy in the radiotherapy process can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical radiation equipment, and in particular to a method and system for controlling radiotherapy dose based on head and neck positioning errors. Background Art

[0002] Radiotherapy is a medical method that uses high-energy rays to treat cancer. Generally, the location of the cancer tumor is located first, and then the doctor treats the target by controlling the radiotherapy equipment. However, high-energy rays also cause damage to other cells in the body. Usually during radiotherapy, medical staff communicate with patients and expect patients to remain motionless during the radiation process to improve the accuracy of radiation. However, sudden conditions such as discomfort, pain, cramps, and sneezing during the radiation process often cause patients to move suddenly during the radiation process. Medical staff are unable to react in time, and normal cells outside the cancer tumor are radiated, which reduces the treatment effect and causes damage to the patient's body.

[0003] To solve the above problems, adaptive radiation therapy has been developed, which monitors the location of cancer tumors. When changes occur, the radiation position is not aligned with the target, and the radiation dose is immediately reduced, and treatment is stopped if necessary. However, due to the limitations of the accuracy of tumor location recognition and the machine response processing speed of patient movement, adaptive radiation therapy has not achieved good technical results. Summary of the invention

[0004] The present invention provides a radiotherapy dose control method based on head and neck setup error, the main purpose of which is to improve the accuracy of radiation dose control, improve treatment effect and reduce patient damage.

[0005] To achieve the above object, the present invention provides a method for radiotherapy dose control based on head and neck setup error, comprising:

[0006] Using a pre-built real-time imaging device, scanning the tumor area confirmed on the patient to obtain a tumor distribution scan image, and performing a pre-processing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure image;

[0007] According to the denoised tumor structure map, the tumor distribution scan image is configured with matrix values ​​to obtain a treatment area matrix, wherein the value in the denoised tumor structure map corresponding to the treatment area matrix is ​​a preset first value, and the value outside the denoised tumor structure map is a preset second value, and the first value is different from the second value;

[0008] Scanning the tumor region using a pre-built positron emission tomography device to obtain a metabolic activity scan image, and performing a clustering operation based on metabolic activity intensity on the metabolic activity scan image to obtain a metabolic activity heat map;

[0009] Using the de-noised tumor structure map, construct a three-dimensional tumor structure map, and according to the metabolic activity heat map, perform an activity configuration operation on the three-dimensional tumor structure map to obtain a three-dimensional tumor activity map;

[0010] Using a pre-trained radiation dose recognition model, a tumor feature recognition operation is performed on the tumor three-dimensional activity map to obtain a tumor feature sequence set, and a radiotherapy dose prediction operation is performed on the tumor feature sequence set to obtain a radiotherapy dose prediction result, and based on the tumor feature sequence set, a dose distribution operation is performed on the radiotherapy dose prediction result to obtain a radiotherapy dose distribution matrix;

[0011] Using a pre-built breathing control algorithm, monitoring the patient's breathing, obtaining a breathing monitoring result, and determining whether the patient is in a pre-built breathing limit position according to the breathing monitoring result;

[0012] When the patient is in a respiratory limit position, the treatment area matrix is ​​intercepted according to a pre-constructed radiation irradiation port position to obtain an offset verification matrix, and the radiotherapy dose distribution matrix is ​​adjusted according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix;

[0013] A pre-constructed radiation equipment cluster is used to generate high-energy rays according to the updated radiotherapy dose distribution matrix, and the tumor area is irradiated with the high-energy rays.

[0014] Optionally, performing a preprocessing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure image includes:

[0015] Using a pre-built bilateral filtering algorithm, according to a preset parameter configuration, a denoising operation is performed on the tumor distribution scan image to obtain a denoised scan image, wherein the parameter configuration includes a spatial parameter, an intensity parameter and a diameter;

[0016] Performing a convolution operation on the denoised scanned image to obtain a convolution matrix set, and performing a dimension reduction operation based on average pooling and flattening on the convolution matrix set to obtain a feature sequence set;

[0017] Performing a tumor edge classification operation on the feature sequence set to obtain tumor edge features;

[0018] According to the tumor edge features, the tumor distribution scan image is cut to obtain a noise-reduced tumor structure image.

[0019] Optionally, performing a clustering operation based on metabolic activity intensity on the metabolic activity scan image to obtain a metabolic activity heat map includes:

[0020] Performing high-activity area monitoring on the metabolic activity scanning image to obtain a high-activity area;

[0021] According to a pre-constructed K-means clustering algorithm, an adaptive number of activity classification operations are performed on the high-activity region to obtain an activity classification region;

[0022] A heat map visualization operation is performed on the activity classification areas to obtain a metabolic activity heat map.

[0023] Optionally, performing an activity configuration operation on the three-dimensional tumor structure map according to the metabolic activity heat map to obtain a three-dimensional tumor activity map includes:

[0024] Performing structural feature extraction operations on the tumor three-dimensional structure map and metabolic activity heat map respectively to obtain a three-dimensional structure feature set and a metabolic activity feature set;

[0025] Using a pre-built similarity algorithm, a matching score traversal calculation operation based on a key structure is performed on the three-dimensional structural feature set and the metabolic activity feature set to obtain a structural matching result;

[0026] Establish connections between key structures with the highest structural matching scores in the structural matching results to obtain a set of structural connection vectors, and minimize the set of structural connection vectors to obtain an overall overlapping morphology of the metabolic activity heat map and the tumor three-dimensional structure map;

[0027] The metabolic activity heat map is used to perform activity configuration operations on the three-dimensional tumor structure map according to the overall overlapping morphology to obtain a three-dimensional tumor activity map.

[0028] Optionally, before using the pre-trained radiation dose identification model, the method further includes:

[0029] Acquire a radiation dose recognition model including a radiation dose prediction network, a position movement feedback network and a dose output layer, wherein the radiation dose prediction network is a network to be trained, and the position movement feedback network and the dose output layer are fixed parameter networks;

[0030] Acquire a historical tumor assessment sample set, and allocate the historical tumor assessment sample set into a training set and a test set according to a preset allocation ratio;

[0031] According to a pre-constructed cross entropy loss algorithm and a gradient descent algorithm, the radiation dose identification model is trained using the training set to obtain an updated radiation dose identification model;

[0032] Using the test set to perform simulation test on the updated radiation dose identification model to obtain the test accuracy;

[0033] Determine whether the test accuracy is greater than a preset pass threshold;

[0034] When the test accuracy is less than or equal to the qualified threshold, returning to the above step of allocating the historical tumor assessment sample set into a training set and a test set according to a preset allocation ratio, obtaining a reallocated training set and a test set, and using the reallocated training set and the test set to train the radiation dose identification model to obtain an optimized updated radiation dose identification model;

[0035] When the test accuracy is greater than the qualified threshold, the updated radiation dose identification model is determined to be qualified, and a trained radiation dose identification model is obtained.

[0036] Optionally, the radiation dose identification model is expressed as:

[0037] D(x,y,z)=I(x,y,z)·adjust(I M )

[0038] I(x, y, z) = R·A d (x, y, z)

[0039] I M =V mask(P)

[0040] R=β 0 +β 1 T+β 2 S+β 3 A m

[0041] T = argmax(P(T|M))

[0042] Wherein, D(x, y, z) represents the updated radiotherapy dose distribution matrix, I(x, y, z) represents the radiotherapy dose distribution matrix, x, y, z represent the three dimensions of the matrix, and I M represents the offset verification matrix, V represents the treatment area matrix, P represents the position of the irradiation port, mask(·) represents the interception process, adjust(·) represents the tumor movement vector calculation process, R represents the radiotherapy dose prediction result, A d represents the tumor activity distribution, T represents the tumor type, M represents the tumor morphology, P(T|M) represents the probability distribution of tumor type T under tumor morphology M, argmax(·) represents the maximum probability, β 0 , β 1 , β 2 and β 3 represents the model parameter weight, S represents the tumor size, A m represents the average tumor activity, where M, S, and Ad and A m All belong to the tumor characteristic sequence set.

[0043] Optionally, adjusting the dose distribution of the radiotherapy dose distribution matrix according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix includes:

[0044] Performing feature recognition on the offset verification matrix to obtain distribution features of the first value and the second value in the offset verification matrix;

[0045] Performing direction recognition on the distribution characteristics to obtain a tumor movement vector;

[0046] The radiotherapy dose distribution matrix is ​​translated according to the tumor movement vector to obtain an updated radiotherapy dose distribution matrix.

[0047] Optionally, the use of a pre-built respiratory control algorithm to monitor the patient's respiration and obtain a respiratory monitoring result includes:

[0048] Using the pre-built breathing control algorithm, the patient's chest rise and fall is monitored to obtain the chest rise and fall fluctuation curve;

[0049] Derivative calculation is performed on the chest cavity fluctuation curve to obtain a fluctuation variation curve;

[0050] The fluctuation change curve is compared with the preset stable threshold interval to obtain the respiratory monitoring result.

[0051] Optionally, judging whether the patient is in a pre-established respiratory limit position according to the respiratory monitoring result includes:

[0052] When the fluctuation variation in the fluctuation variation curve is greater than the stable threshold interval, marking the respiratory monitoring result as a pre-constructed respiratory stage time position;

[0053] When the fluctuation variation in the fluctuation variation curve is less than or equal to the stable threshold interval, the respiratory monitoring result is marked as a pre-constructed respiratory limit position.

[0054] To achieve the above object, the present invention further provides a radiotherapy dose control system based on head and neck positioning error, comprising:

[0055] An optical image processing module, used to use a pre-built real-time imaging device to scan a tumor area confirmed on a patient to obtain a tumor distribution scan image, and perform a pre-processing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure map, and perform a matrix numerical configuration on the tumor distribution scan image according to the noise-reduced tumor structure map to obtain a treatment area matrix, wherein the numerical value within the noise-reduced tumor structure map corresponding to the treatment area matrix is ​​a preset first numerical value, and the numerical value outside the noise-reduced tumor structure map is a preset second numerical value, and the first numerical value is different from the second numerical value;

[0056] a metabolic activity image processing module, configured to scan the tumor region using a pre-built positron emission tomography device to obtain a metabolic activity scan image, perform a clustering operation on the metabolic activity scan image based on the metabolic activity intensity to obtain a metabolic activity heat map, and construct a three-dimensional tumor structure map using the denoised tumor structure map, and perform an activity configuration operation on the three-dimensional tumor structure map according to the metabolic activity heat map to obtain a three-dimensional tumor activity map;

[0057] a radiation dose prediction module, for performing a tumor feature recognition operation on the tumor three-dimensional activity map using a pre-trained radiation dose recognition model to obtain a tumor feature sequence set, performing a radiotherapy dose prediction operation on the tumor feature sequence set to obtain a radiotherapy dose prediction result, and performing a dose distribution operation on the radiotherapy dose prediction result according to the tumor feature sequence set to obtain a radiotherapy dose distribution matrix;

[0058] The irradiation control module is used to monitor the patient's respiration using a pre-built respiration control algorithm to obtain respiration monitoring results, and determine whether the patient is in a pre-built respiration limit position based on the respiration monitoring results; when the patient is in the respiration limit position, intercept the treatment area matrix according to the pre-built ray irradiation port position to obtain an offset verification matrix, and adjust the dose distribution of the radiotherapy dose distribution matrix according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix; and use a pre-built radiation equipment cluster to generate high-energy rays according to the updated radiotherapy dose distribution matrix, and use the high-energy rays to irradiate the tumor area.

[0059] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:

[0060] a memory storing at least one instruction; and

[0061] The processor executes the instructions stored in the memory to implement the above-mentioned radiotherapy dose control method based on head and neck positioning error.

[0062] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is executed by a processor in an electronic device to implement the above-mentioned radiotherapy dose control method based on head and neck positioning error.

[0063] The present invention solves the problems described in the background technology. The present invention first analyzes the structure of the tumor through a real-time imaging device to obtain a de-noised tumor structure diagram, and then uses a positron emission tomography device to analyze the activity of the tumor to obtain a three-dimensional tumor activity diagram; then, in the embodiment of the present invention, the three-dimensional tumor activity diagram is analyzed by a pre-trained radiation dose recognition model to obtain a radiotherapy dose prediction result, which greatly improves the accuracy and efficiency of radiotherapy dose assessment; then, the present invention controls the radiation frequency through a breathing control algorithm, and according to the breathing interval, there is enough time to adjust the dose distribution, thereby improving the accuracy of radiotherapy; wherein, the present invention uses an offset verification matrix to identify the offset of the patient's position, and through smaller computing resources, the accuracy of dose adjustment is improved. Therefore, the present invention can improve the accuracy of radiation dose regulation, improve the treatment effect and reduce patient damage. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 A schematic flow chart of a method for controlling radiotherapy dose based on head and neck setup error provided by an embodiment of the present invention;

[0065] Figure 2 A functional module diagram of a radiotherapy dose control system based on head and neck setup error provided by an embodiment of the present invention;

[0066] Figure 3 A schematic diagram of the structure of an electronic device for implementing the radiotherapy dose control method based on head and neck positioning error provided in one embodiment of the present invention.

[0067] Description of figure markings:

[0068] 1. Electronic device; 10. Processor; 11. Memory; 12. Bus.

[0069] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0070] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0071] The embodiment of the present application provides a method for radiotherapy dose control based on head and neck positioning errors. The execution subject of the method for radiotherapy dose control based on head and neck positioning errors includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided in the embodiment of the present application. In other words, the method for radiotherapy dose control based on head and neck positioning errors can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0072] Reference Figure 1 FIG. 1 is a flow chart of a method for controlling the radiotherapy dose based on the head and neck setup error according to an embodiment of the present invention. In this embodiment, the method for controlling the radiotherapy dose based on the head and neck setup error includes:

[0073] S1. Using a pre-built real-time imaging device, scan the tumor area confirmed on the patient to obtain a tumor distribution scan image, and perform a pre-processing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure image.

[0074] In the embodiment of the present invention, the real-time imaging device is a CT scanning device, which can observe the shape, size and structure of the tumor. In addition, the real-time imaging device can also be an MRI, PET scanning device, etc. The present invention is described with a tumor distribution scanning image of the CT image type.

[0075] Specifically, in an embodiment of the present invention, a CT scanning device is used to obtain a tumor distribution scanning image, wherein the tumor distribution scanning image is an image set, and the tumor CT images at each depth position are obtained by scanning each depth of the tumor.

[0076] In detail, in an embodiment of the present invention, the preprocessing operation based on noise reduction and segmentation is performed on the tumor distribution scan image to obtain a noise-reduced tumor structure image, including:

[0077] Using a pre-built bilateral filtering algorithm, according to a preset parameter configuration, a denoising operation is performed on the tumor distribution scan image to obtain a denoised scan image, wherein the parameter configuration includes a spatial parameter, an intensity parameter and a diameter;

[0078] Performing a convolution operation on the denoised scanned image to obtain a convolution matrix set, and performing a dimension reduction operation based on average pooling and flattening on the convolution matrix set to obtain a feature sequence set;

[0079] Performing a tumor edge classification operation on the feature sequence set to obtain tumor edge features;

[0080] According to the tumor edge features, the tumor distribution scan image is cut to obtain a noise-reduced tumor structure image.

[0081] The bilateral filtering algorithm is an algorithm for reducing noise in CT images and improving image quality.

[0082] Specifically, in the embodiment of the present invention, the OpenCV tool is first used to call the bilateral filtering function, and then the tumor distribution scan image is loaded into the memory to ensure that the image format is correct, and then the parameters for setting the bilateral filtering algorithm are obtained, including: spatial parameters: indicating the similarity in the control space domain, usually affecting the size of the filtering area; intensity parameters: indicating the similarity in the control intensity domain, which is related to the gray value difference of the image; diameter: indicating the size of the image filtering window, which can usually be set to 0 and is automatically calculated according to the spatial parameters. In the embodiment of the present invention, according to the above parameters and functions, each image in the tumor distribution scan image is looped and the same bilateral filtering operation is applied to obtain a denoised scan image.

[0083] Furthermore, in the embodiment of the present invention, the convolution layer, pooling layer and flattening layer in the neural network model are used to perform feature extraction operations on the denoised scanned image to obtain a feature sequence set. Among them, the convolution layer is used for preliminary feature extraction, and the pooling layer and flattening layer are used to reduce the dimension of the data until it is one dimension.

[0084] In the embodiment of the present invention, whether each feature is continuous is determined to determine whether it is an edge feature, thereby obtaining the edge feature of the tumor, and finally performing image segmentation according to the edge feature of the tumor to obtain a denoised tumor structure image.

[0085] S2. According to the denoised tumor structure map, the tumor distribution scan image is numerically configured in a matrix to obtain a treatment area matrix, wherein the numerical value within the denoised tumor structure map corresponding to the treatment area matrix is ​​a preset first numerical value, and the numerical value outside the denoised tumor structure map is a preset second numerical value, and the first numerical value is different from the second numerical value.

[0086] Among them, the tumor distribution scanning image is a larger area, and the denoised tumor structure map is a smaller area. In an embodiment of the present invention, the area of ​​the denoised tumor structure map in the tumor distribution scanning image is configured as a preset first value, such as 1, and the area outside the denoised tumor structure map is configured as a preset second value, such as 0.

[0087] It can be foreseen that when the radiation port positions of the radiation equipment cluster are aligned with the noise-reduced tumor structure map, the values ​​of the radiation port positions are all 1, and when the patient moves, the values ​​of the radiation port positions are all interactions between 1 and 0. By observing the direction of 0 in 1, the patient's movement direction can be known, which facilitates the subsequent regulation of the dose of high-energy rays. The dose refers to the value of the radiation energy, in Gray or Gy. 1Gy is equal to 1 joule of radiation energy absorbed per kilogram of tissue.

[0088] S3. Scan the tumor area using a pre-built positron emission tomography device to obtain a metabolic activity scan image, and perform a clustering operation on the metabolic activity scan image based on metabolic activity intensity to obtain a metabolic activity heat map.

[0089] It should be known that tumor cells usually have higher metabolic activity, especially the uptake and utilization of glucose, a phenomenon known as the "Warburg effect". Tumor cells tend to produce energy through anaerobic metabolism, even under aerobic conditions. This metabolic feature can be detected by imaging techniques such as positron emission tomography (PET).

[0090] Specifically, in the embodiment of the present invention, the positron emission tomography device uses a radioactive tracer, such as a radioactive isotope (fluorine-18), which can bind to specific biological molecules and be metabolized in the body. The radioactive isotope releases positrons during the decay process, and the positrons undergo annihilation reactions when they meet the surrounding electrons, generating two gamma rays (γ rays) emitted in opposite directions. The positron emission tomography device can capture these gamma rays and record their source positions, and finally process these data through a computer to generate a three-dimensional image of the tissue in the body, reflecting the distribution of the tracer, thereby obtaining a metabolic activity scanning image.

[0091] In detail, in an embodiment of the present invention, the clustering operation based on the metabolic activity intensity is performed on the metabolic activity scan image to obtain a metabolic activity heat map, including:

[0092] Performing high-activity area monitoring on the metabolic activity scanning image to obtain a high-activity area;

[0093] According to a pre-constructed K-means clustering algorithm, an adaptive number of activity classification operations are performed on the high-activity region to obtain an activity classification region;

[0094] A heat map visualization operation is performed on the activity classification areas to obtain a metabolic activity heat map.

[0095] Specifically, in an embodiment of the present invention, the metabolic activity scan image includes normal tissue and tumor tissue, but the activity level of tumor tissue is significantly higher than that of other normal tissues, so the high activity area is first extracted. Then, considering the starting position of the tumor, the surrounding environment and other factors, the center point of the tumor's initial development is not the center point of the physical structure, so it is necessary to further divide the activity intensity. In an embodiment of the present invention, the K-means clustering algorithm is used for activity classification. Since the specific K value is not set in the present invention, the clustering algorithm adaptively allocates the K value, obtains the most appropriate division by continuous traversal calculation, and finally outputs the activity classification area. Finally, the visualization tool Tableau is used to perform heat distribution to obtain a metabolic activity heat map. Among them, the K-means clustering algorithm is a commonly used unsupervised learning algorithm for dividing a data set into K different clusters (clusters), and then the goal is to divide the data points into each cluster so that the similarity between the data points in the same cluster is as high as possible, and the similarity between the data points in different clusters is as low as possible.

[0096] S4. Using the denoised tumor structure map, construct a three-dimensional tumor structure map, and according to the metabolic activity heat map, perform activity configuration operation on the three-dimensional tumor structure map to obtain a three-dimensional tumor activity map.

[0097] Specifically, in an embodiment of the present invention, the denoised tumor structure image is a hierarchical image set. The present invention calculates and predicts the mean of the layer gaps in the hierarchical image set to fill in the image at the vacant position, and then performs smoothing to obtain a three-dimensional tumor structure image.

[0098] In detail, in an embodiment of the present invention, performing an activity configuration operation on the three-dimensional tumor structure diagram according to the metabolic activity heat map to obtain a three-dimensional tumor activity map includes:

[0099] Performing structural feature extraction operations on the tumor three-dimensional structure map and metabolic activity heat map respectively to obtain a three-dimensional structure feature set and a metabolic activity feature set;

[0100] Using a pre-built similarity algorithm, a matching score traversal calculation operation based on a key structure is performed on the three-dimensional structural feature set and the metabolic activity feature set to obtain a structural matching result;

[0101] Establish connections between key structures with the highest structural matching scores in the structural matching results to obtain a set of structural connection vectors, and minimize the set of structural connection vectors to obtain an overall overlapping morphology of the metabolic activity heat map and the tumor three-dimensional structure map;

[0102] The metabolic activity heat map is used to perform activity configuration operations on the three-dimensional tumor structure map according to the overall overlapping morphology to obtain a three-dimensional tumor activity map.

[0103] In the embodiment of the present invention, although the image types are different, the tumor contour can be seen. The present invention calculates a three-dimensional structural feature set and a metabolic activity feature set, both of which include a tumor protrusion, a linear part, and a smooth part. The three-dimensional structural feature set and the metabolic activity feature set can be matched and connected through a cosine similarity algorithm to obtain a structural connection vector set. When the value of each vector in the structural connection vector set is a minimum value, such as 0 (when the layer field of view is different, the minimum value may be non-0), indicating that the metabolic activity heat map and the tumor three-dimensional structure map have overlapped, the three-dimensional structure map of the tumor can be subjected to an activity configuration operation according to the metabolic activity heat map to obtain a tumor three-dimensional activity map.

[0104] S5. Using the pre-trained radiation dose recognition model, perform a tumor feature recognition operation on the tumor three-dimensional activity map to obtain a tumor feature sequence set, perform a radiotherapy dose prediction operation on the tumor feature sequence set to obtain a radiotherapy dose prediction result, and perform a dose distribution operation on the radiotherapy dose prediction result based on the tumor feature sequence set to obtain a radiotherapy dose distribution matrix.

[0105] In an embodiment of the present invention, the radiation dose identification model is an image processing neural network model, which is used to predict the tumor type according to the tumor characteristics, and then predict the radiotherapy dose according to the tumor characteristics and tumor type. The output layer of the model also includes a position movement feedback network, which is used to redistribute the radiotherapy dose according to the movement of the tumor position.

[0106] In detail, in an embodiment of the present invention, before using the pre-trained radiation dose identification model, the method further includes:

[0107] Acquire a radiation dose recognition model including a radiation dose prediction network, a position movement feedback network and a dose output layer, wherein the radiation dose prediction network is a network to be trained, and the position movement feedback network and the dose output layer are fixed parameter networks;

[0108] Acquire a historical tumor assessment sample set, and allocate the historical tumor assessment sample set into a training set and a test set according to a preset allocation ratio;

[0109] According to a pre-constructed cross entropy loss algorithm and a gradient descent algorithm, the radiation dose identification model is trained using the training set to obtain an updated radiation dose identification model;

[0110] Using the test set to perform simulation test on the updated radiation dose identification model to obtain the test accuracy;

[0111] Determine whether the test accuracy is greater than a preset pass threshold;

[0112] When the test accuracy is less than or equal to the qualified threshold, returning to the above step of allocating the historical tumor assessment sample set into a training set and a test set according to a preset allocation ratio, obtaining a reallocated training set and a test set, and using the reallocated training set and the test set to train the radiation dose identification model to obtain an optimized updated radiation dose identification model;

[0113] When the test accuracy is greater than the qualified threshold, the updated radiation dose identification model is determined to be qualified, and a trained radiation dose identification model is obtained.

[0114] Specifically, in the embodiment of the present invention, a historical tumor assessment sample set is first obtained, wherein the historical tumor assessment sample set includes historical tumor imaging and expert diagnosis opinions, and then grouped according to a preset allocation ratio, such as 7:3, to obtain a training set and a test set.

[0115] In the specific training process, the radiation dose prediction network of the present invention is a regression prediction network, and therefore can be trained by a cross entropy loss algorithm and a gradient descent algorithm to update the radiation dose recognition model.

[0116] The cross entropy loss algorithm is a commonly used loss function that can measure the difference between two probability distributions and is usually used to compare the distribution of true labels and the distribution predicted by the model. The gradient descent algorithm is an optimization algorithm used to find the local minimum of a function and is widely used in machine learning and deep learning, especially to minimize the loss function when training models.

[0117] Then, the updated radiation dose identification model is simulated and tested using the test set to obtain the test accuracy. When the test accuracy is greater than a preset qualified threshold, such as 95%, it indicates that the training is qualified, thereby preventing the training from failing.

[0118] It is understandable that, considering the data scarcity of the historical tumor assessment sample set, when the training fails, the model is trained by disrupting the sample composition and sequence until a trained radiation dose recognition model is obtained.

[0119] In detail, in the embodiment of the present invention, the radiation dose identification model is expressed as:

[0120] D(x,y,z)=I(x,y,z)·sdjust(I M )

[0121] I(x, y, z) = R·A d (x, y, z)

[0122] I M =V mask(P)

[0123] R=β 0 +β 1 T+β 2 S+β 3 A m

[0124] T = argmax(P(T|M))

[0125] Wherein, D(x, y, z) represents the updated radiotherapy dose distribution matrix, I(x, y, z) represents the radiotherapy dose distribution matrix, x, y, z represent the three dimensions of the matrix, and I M represents the offset verification matrix, V represents the treatment area matrix, P represents the position of the irradiation port, mask(·) represents the interception process, adjust(·) represents the tumor movement vector calculation process, R represents the radiotherapy dose prediction result, A d represents the tumor activity distribution, T represents the tumor type, M represents the tumor morphology, P(T|M) represents the probability distribution of tumor type T under tumor morphology M, argmax(·) represents the maximum probability, β 0 , β 1 , β 2 and β 3 represents the model parameter weight, S represents the tumor size, A m represents the average tumor activity, where M, S, and A d and A m All belong to the tumor characteristic sequence set.

[0126] Specifically, in the embodiment of the present invention, a tumor feature sequence set is obtained by performing a tumor feature recognition operation on the tumor three-dimensional activity map, and the tumor feature sequence set is [tumor morphology M, tumor size S, tumor activity distribution A d and mean tumor activity A m Among them, tumor type T can be predicted based on tumor morphology M, and tumor type T and average tumor activity A m and tumor size S, the radiotherapy dose prediction results can be predicted, and the radiotherapy dose prediction results are combined with the tumor activity distribution A d , we can get the radiotherapy dose distribution matrix I(x, y, z).

[0127] In detail, in an embodiment of the present invention, the step of adjusting the dose distribution of the radiotherapy dose distribution matrix according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix includes:

[0128] Performing feature recognition on the offset verification matrix to obtain distribution features of the first value and the second value in the offset verification matrix;

[0129] Performing direction recognition on the distribution characteristics to obtain a tumor movement vector;

[0130] The radiotherapy dose distribution matrix is ​​translated according to the tumor movement vector to obtain an updated radiotherapy dose distribution matrix.

[0131] In the embodiment of the present invention, according to the offset verification matrix I provided by the mobile feedback network M , we can find out the offset direction by analyzing the position of data 0 in the matrix and data 1. M ), according to the offset direction adjust(I M ) translates the radiotherapy dose distribution matrix I(x, y, z) to obtain an updated radiotherapy dose distribution matrix D(x, y, z).

[0132] S6. Use a pre-built breathing control algorithm to monitor the patient's breathing, obtain a breathing monitoring result, and determine whether the patient is in a pre-built breathing limit position based on the breathing monitoring result.

[0133] In detail, in the embodiment of the present invention, the use of a pre-built breathing control algorithm to monitor the patient's breathing and obtain a breathing monitoring result includes:

[0134] Using the pre-built breathing control algorithm, the patient's chest rise and fall is monitored to obtain the chest rise and fall fluctuation curve;

[0135] Derivative calculation is performed on the chest cavity fluctuation curve to obtain a fluctuation variation curve;

[0136] The fluctuation change curve is compared with the preset stable threshold interval to obtain the respiratory monitoring result.

[0137] Among them, the breathing control algorithm plays an important role in the medical field, especially in radiotherapy and imaging. Its main purpose is to dynamically adjust the treatment or imaging process during the patient's breathing cycle to ensure that the target area (such as a tumor) remains in the optimal position during the radiation or imaging process.

[0138] In the embodiment of the present invention, monitoring is first performed to obtain a chest fluctuation curve, and then the curve is differentiated to obtain a fluctuation change curve, and then a preset stable threshold interval, such as (-0.1, 0.1), is compared to obtain a respiratory monitoring result.

[0139] Further, in an embodiment of the present invention, judging whether the patient is in a pre-established respiratory limit position according to the respiratory monitoring result includes:

[0140] When the fluctuation variation in the fluctuation variation curve is greater than the stable threshold interval, marking the respiratory monitoring result as a pre-constructed respiratory stage time position;

[0141] When the fluctuation variation in the fluctuation variation curve is less than or equal to the stable threshold interval, the respiratory monitoring result is marked as a pre-constructed respiratory limit position.

[0142] In the embodiment of the present invention, considering that the patient's tumor may be located in the head and neck, which is greatly affected by breathing, when the patient is in the time position of the breathing stage, the tumor position changes all the time, so high-energy ray excitation is not performed.

[0143] S7. When the patient is in the respiratory limit position, the treatment area matrix is ​​intercepted according to the pre-constructed radiation irradiation port position to obtain an offset verification matrix, and the radiotherapy dose distribution matrix is ​​adjusted according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix.

[0144] It is understandable that the position of the radiation irradiation port is affected by the radiation equipment and is usually fixed or moves within a small range. In the embodiment of the present invention, it is assumed that the position of the radiation irradiation port is fixed and is aligned with the tumor position of the patient.

[0145] Furthermore, when the patient moves, the tumor position shifts from the position of the radiation irradiation port. The present invention uses a method of viewing the treatment area matrix to view the distance and direction of the patient's movement.

[0146] Exemplarily, when the patient does not move, the intercepted offset verification matrix is:

[0147]

[0148] Among them, 2 is assumed to be the tumor active center.

[0149] Exemplarily, when the patient moves, the offset verification matrix that may be intercepted is:

[0150]

[0151] It is understandable that the value 1 is evenly distributed in the lower left corner of the value 0. Since the value 1 represents the noise-reduced tumor structure image, it can be known that the patient position moves a distance to the lower left corner.

[0152] Therefore, according to the distance moved, the radiotherapy dose distribution matrix can be adjusted so that the high-energy ray with the largest dose is still aimed at position 2, while the high-energy ray at position 0 is cancelled, and an updated radiotherapy dose distribution matrix is ​​obtained. This can ensure the treatment effect at the center position and avoid damage to ordinary tissues.

[0153] S8. Using a pre-built radiation equipment cluster, generate high-energy rays according to the updated radiotherapy dose distribution matrix, and irradiate the tumor area with the high-energy rays.

[0154] Finally, the embodiment of the present invention inputs the calculated updated radiotherapy dose distribution matrix into the control matrix of the radiation equipment cluster, and uses the radiation equipment cluster to perform high-energy ray excitation, thereby achieving accurate and safe treatment.

[0155] The present invention solves the problems described in the background technology. The present invention first analyzes the structure of the tumor through a real-time imaging device to obtain a de-noised tumor structure diagram, and then uses a positron emission tomography device to analyze the activity of the tumor to obtain a three-dimensional tumor activity diagram; then, in the embodiment of the present invention, the three-dimensional tumor activity diagram is analyzed by a pre-trained radiation dose recognition model to obtain a radiotherapy dose prediction result, which greatly improves the accuracy and efficiency of radiotherapy dose assessment; then, the present invention controls the radiation frequency through a breathing control algorithm, and according to the breathing interval, there is enough time to adjust the dose distribution, thereby improving the accuracy of radiotherapy; wherein, the present invention uses an offset verification matrix to identify the offset of the patient's position, and through smaller computing resources, the accuracy of dose adjustment is improved. Therefore, the present invention can improve the accuracy of radiation dose regulation, improve the treatment effect and reduce patient damage.

[0156] like Figure 2 , is a functional module diagram of a radiotherapy dose control system based on head and neck positioning error provided by an embodiment of the present invention.

[0157] The radiotherapy dose control system 100 based on head and neck setup error of the present invention can be installed in an electronic device. According to the functions to be implemented, the radiotherapy dose control system 100 based on head and neck setup error can include an optical image processing module 101, a metabolic activity image processing module 102, a radiation dose prediction module 103 and an irradiation control module 104. The module of the present invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0158] The optical image processing module 101 is used to use a pre-built real-time imaging device to scan the tumor area confirmed on the patient to obtain a tumor distribution scan image, and perform a pre-processing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure map, and perform a matrix numerical configuration on the tumor distribution scan image according to the noise-reduced tumor structure map to obtain a treatment area matrix, wherein the numerical value in the noise-reduced tumor structure map corresponding to the treatment area matrix is ​​a preset first numerical value, and the numerical value outside the noise-reduced tumor structure map is a preset second numerical value, and the first numerical value is different from the second numerical value;

[0159] The metabolic activity image processing module 102 is used to scan the tumor area using a pre-built positron emission tomography device to obtain a metabolic activity scan image, perform a clustering operation based on the metabolic activity intensity on the metabolic activity scan image to obtain a metabolic activity heat map, and use the denoised tumor structure map to construct a three-dimensional tumor structure map, and perform an activity configuration operation on the three-dimensional tumor structure map according to the metabolic activity heat map to obtain a three-dimensional tumor activity map;

[0160] The radiation dose prediction module 103 is used to use the pre-trained radiation dose recognition model to perform a tumor feature recognition operation on the tumor three-dimensional activity map to obtain a tumor feature sequence set, and to perform a radiotherapy dose prediction operation on the tumor feature sequence set to obtain a radiotherapy dose prediction result, and to perform a dose distribution operation on the radiotherapy dose prediction result according to the tumor feature sequence set to obtain a radiotherapy dose distribution matrix;

[0161] The irradiation control module 104 is used to use a pre-constructed breathing control algorithm to monitor the patient's breathing, obtain breathing monitoring results, and determine whether the patient is in a pre-constructed breathing limit position based on the breathing monitoring results; when the patient is in the breathing limit position, the treatment area matrix is ​​intercepted according to the pre-constructed irradiation port position to obtain an offset verification matrix, and the dose distribution of the radiotherapy dose distribution matrix is ​​adjusted according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix; and a pre-constructed radiation equipment cluster is used to generate high-energy rays according to the updated radiotherapy dose distribution matrix, and the tumor area is irradiated with the high-energy rays.

[0162] In detail, the modules in the radiotherapy dose control system 100 based on head and neck positioning error in the embodiment of the present invention are used in the same manner as described above. Figure 1 The same technical means are used as the radiotherapy dose control method based on head and neck positioning error described in, and can produce the same technical effects, so they will not be repeated here.

[0163] like Figure 3 , is a schematic diagram of the structure of an electronic device for implementing a radiotherapy dose control method based on head and neck positioning errors provided by an embodiment of the present invention.

[0164] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a radiotherapy dose control method program based on head and neck positioning errors.

[0165] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 can also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 1. Further, the memory 11 also includes an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can not only be used to store application software and various types of data installed in the electronic device 1, such as the code of the radiotherapy dose control method program based on the head and neck positioning error, but also can be used to temporarily store data that has been output or is to be output.

[0166] In some embodiments, the processor 10 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc. The processor 10 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect the various components of the entire electronic device, and executes or executes the programs or modules stored in the memory 11 (such as a radiotherapy dose control method program based on head and neck positioning errors, etc.), and calls the data stored in the memory 11 to execute various functions of the electronic device 1 and process data.

[0167] The bus 12 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 may be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.

[0168] Figure 3 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 3The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0169] For example, although not shown, the electronic device 1 may also include a power source (such as a battery) for supplying power to each component. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include any components such as one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, etc. The electronic device 1 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0170] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device 1 and other electronic devices.

[0171] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device 1 and to display a visual user interface.

[0172] The radiotherapy dose control method program based on head and neck setup error stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can achieve:

[0173] Using a pre-built real-time imaging device, scanning the tumor area confirmed on the patient to obtain a tumor distribution scan image, and performing a pre-processing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure image;

[0174] According to the denoised tumor structure map, the tumor distribution scan image is configured with matrix values ​​to obtain a treatment area matrix, wherein the value in the denoised tumor structure map corresponding to the treatment area matrix is ​​a preset first value, and the value outside the denoised tumor structure map is a preset second value, and the first value is different from the second value;

[0175] Scanning the tumor region using a pre-built positron emission tomography device to obtain a metabolic activity scan image, and performing a clustering operation based on metabolic activity intensity on the metabolic activity scan image to obtain a metabolic activity heat map;

[0176] Using the de-noised tumor structure map, construct a three-dimensional tumor structure map, and according to the metabolic activity heat map, perform an activity configuration operation on the three-dimensional tumor structure map to obtain a three-dimensional tumor activity map;

[0177] Using a pre-trained radiation dose recognition model, a tumor feature recognition operation is performed on the tumor three-dimensional activity map to obtain a tumor feature sequence set, and a radiotherapy dose prediction operation is performed on the tumor feature sequence set to obtain a radiotherapy dose prediction result, and based on the tumor feature sequence set, a dose distribution operation is performed on the radiotherapy dose prediction result to obtain a radiotherapy dose distribution matrix;

[0178] Using a pre-built breathing control algorithm, monitoring the patient's breathing, obtaining a breathing monitoring result, and determining whether the patient is in a pre-built breathing limit position according to the breathing monitoring result;

[0179] When the patient is in a respiratory limit position, the treatment area matrix is ​​intercepted according to a pre-constructed radiation irradiation port position to obtain an offset verification matrix, and the radiotherapy dose distribution matrix is ​​adjusted according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix;

[0180] A pre-constructed radiation equipment cluster is used to generate high-energy rays according to the updated radiotherapy dose distribution matrix, and the tumor area is irradiated with the high-energy rays.

[0181] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to Figures 1 to 3 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0182] Furthermore, if the module / unit integrated in the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0183] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:

[0184] Using a pre-built real-time imaging device, scanning the tumor area confirmed on the patient to obtain a tumor distribution scan image, and performing a pre-processing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure image;

[0185] According to the denoised tumor structure map, the tumor distribution scan image is configured with matrix values ​​to obtain a treatment area matrix, wherein the value in the denoised tumor structure map corresponding to the treatment area matrix is ​​a preset first value, and the value outside the denoised tumor structure map is a preset second value, and the first value is different from the second value;

[0186] Scanning the tumor region using a pre-built positron emission tomography device to obtain a metabolic activity scan image, and performing a clustering operation based on metabolic activity intensity on the metabolic activity scan image to obtain a metabolic activity heat map;

[0187] Using the de-noised tumor structure map, construct a three-dimensional tumor structure map, and according to the metabolic activity heat map, perform an activity configuration operation on the three-dimensional tumor structure map to obtain a three-dimensional tumor activity map;

[0188] Using a pre-trained radiation dose recognition model, a tumor feature recognition operation is performed on the tumor three-dimensional activity map to obtain a tumor feature sequence set, and a radiotherapy dose prediction operation is performed on the tumor feature sequence set to obtain a radiotherapy dose prediction result, and based on the tumor feature sequence set, a dose distribution operation is performed on the radiotherapy dose prediction result to obtain a radiotherapy dose distribution matrix;

[0189] Using a pre-built breathing control algorithm, monitoring the patient's breathing, obtaining a breathing monitoring result, and determining whether the patient is in a pre-built breathing limit position according to the breathing monitoring result;

[0190] When the patient is in a respiratory limit position, the treatment area matrix is ​​intercepted according to a pre-constructed radiation irradiation port position to obtain an offset verification matrix, and the radiotherapy dose distribution matrix is ​​adjusted according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix;

[0191] A pre-constructed radiation equipment cluster is used to generate high-energy rays according to the updated radiotherapy dose distribution matrix, and the tumor area is irradiated with the high-energy rays.

[0192] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0193] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0194] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for radiotherapy dose control based on head and neck positioning error, characterized in that: The method comprises: Using a pre-built real-time imaging device, scanning the tumor area confirmed on the patient to obtain a tumor distribution scan image, and performing a pre-processing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure image; According to the denoised tumor structure map, the tumor distribution scan image is configured with matrix values ​​to obtain a treatment area matrix, wherein the value in the denoised tumor structure map corresponding to the treatment area matrix is ​​a preset first value, and the value outside the denoised tumor structure map is a preset second value, and the first value is different from the second value; Scanning the tumor region using a pre-built positron emission tomography device to obtain a metabolic activity scan image, and performing a clustering operation based on metabolic activity intensity on the metabolic activity scan image to obtain a metabolic activity heat map; Using the de-noised tumor structure map, construct a three-dimensional tumor structure map, and according to the metabolic activity heat map, perform an activity configuration operation on the three-dimensional tumor structure map to obtain a three-dimensional tumor activity map; Using a pre-trained radiation dose recognition model, a tumor feature recognition operation is performed on the tumor three-dimensional activity map to obtain a tumor feature sequence set, and a radiotherapy dose prediction operation is performed on the tumor feature sequence set to obtain a radiotherapy dose prediction result, and based on the tumor feature sequence set, a dose distribution operation is performed on the radiotherapy dose prediction result to obtain a radiotherapy dose distribution matrix; Using a pre-built breathing control algorithm, monitoring the patient's breathing, obtaining a breathing monitoring result, and determining whether the patient is in a pre-built breathing limit position according to the breathing monitoring result; When the patient is in a respiratory limit position, the treatment area matrix is ​​intercepted according to a pre-constructed radiation irradiation port position to obtain an offset verification matrix, and the radiotherapy dose distribution matrix is ​​adjusted according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix; A pre-constructed radiation equipment cluster is used to generate high-energy rays according to the updated radiotherapy dose distribution matrix, and the tumor area is irradiated with the high-energy rays.

2. The method for radiotherapy dose control based on head and neck positioning error according to claim 1, characterized in that: The step of performing a preprocessing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure image includes: Using a pre-built bilateral filtering algorithm, according to a preset parameter configuration, a denoising operation is performed on the tumor distribution scan image to obtain a denoised scan image, wherein the parameter configuration includes a spatial parameter, an intensity parameter and a diameter; Performing a convolution operation on the denoised scanned image to obtain a convolution matrix set, and performing a dimension reduction operation based on average pooling and flattening on the convolution matrix set to obtain a feature sequence set; Performing a tumor edge classification operation on the feature sequence set to obtain tumor edge features; According to the tumor edge features, the tumor distribution scan image is cut to obtain a noise-reduced tumor structure image.

3. The method for radiotherapy dose control based on head and neck positioning error according to claim 2, characterized in that: The performing of a clustering operation based on metabolic activity intensity on the metabolic activity scan image to obtain a metabolic activity heat map includes: Performing high-activity area monitoring on the metabolic activity scanning image to obtain a high-activity area; According to a pre-constructed K-means clustering algorithm, an adaptive number of activity classification operations are performed on the high-activity region to obtain an activity classification region; A heat map visualization operation is performed on the activity classification areas to obtain a metabolic activity heat map.

4. The method for controlling the radiotherapy dose based on the head and neck positioning error according to claim 3, characterized in that: The step of performing an activity configuration operation on the three-dimensional tumor structure diagram according to the metabolic activity heat map to obtain a three-dimensional tumor activity map includes: Performing structural feature extraction operations on the tumor three-dimensional structure map and metabolic activity heat map respectively to obtain a three-dimensional structure feature set and a metabolic activity feature set; Using a pre-built similarity algorithm, a matching score traversal calculation operation based on a key structure is performed on the three-dimensional structural feature set and the metabolic activity feature set to obtain a structural matching result; Establish connections between key structures with the highest structural matching scores in the structural matching results to obtain a set of structural connection vectors, and minimize the set of structural connection vectors to obtain an overall overlapping morphology of the metabolic activity heat map and the tumor three-dimensional structure map; The metabolic activity heat map is used to perform activity configuration operations on the three-dimensional tumor structure map according to the overall overlapping morphology to obtain a three-dimensional tumor activity map.

5. The method for radiotherapy dose control based on head and neck positioning error according to claim 4, characterized in that: Prior to using the pre-trained radiation dose identification model, the method further includes: Acquire a radiation dose recognition model including a radiation dose prediction network, a position movement feedback network and a dose output layer, wherein the radiation dose prediction network is a network to be trained, and the position movement feedback network and the dose output layer are fixed parameter networks; Acquire a historical tumor assessment sample set, and allocate the historical tumor assessment sample set into a training set and a test set according to a preset allocation ratio; According to a pre-constructed cross entropy loss algorithm and a gradient descent algorithm, the radiation dose identification model is trained using the training set to obtain an updated radiation dose identification model; Using the test set to perform simulation test on the updated radiation dose identification model to obtain the test accuracy; Determine whether the test accuracy is greater than a preset pass threshold; When the test accuracy is less than or equal to the qualified threshold, returning to the above step of allocating the historical tumor assessment sample set into a training set and a test set according to a preset allocation ratio, obtaining a reallocated training set and a test set, and using the reallocated training set and the test set to train the radiation dose identification model to obtain an optimized updated radiation dose identification model; When the test accuracy is greater than the qualified threshold, the updated radiation dose identification model is determined to be qualified, and a trained radiation dose identification model is obtained.

6. The method for radiotherapy dose control based on head and neck positioning error according to claim 5, characterized in that: The radiation dose identification model is expressed as: D(x,y,z)=I(x,y,z)·adjust(I M ) I(x,y,z)=R·A d (x,y,z) I M =V·mask(P) R=β0+β1T+β2S+β3A m T = argmax(P(T|M)) Wherein, D(x, y, z) represents the updated radiotherapy dose distribution matrix, I(x, y, z) represents the radiotherapy dose distribution matrix, x, y, z represent the three dimensions of the matrix, and I M represents the offset verification matrix, V represents the treatment area matrix, P represents the position of the irradiation port, mask(·) represents the interception process, adjust(·) represents the tumor movement vector calculation process, R represents the radiotherapy dose prediction result, A d represents the tumor activity distribution, T represents the tumor type, M represents the tumor morphology, P(T|M) represents the probability distribution of tumor type T under tumor morphology M, argmax(·) represents the maximum probability, β0, β1, β2 and β3 represent the model parameter weights, S represents the tumor size, A m represents the average tumor activity, where M, S, and A d and A m All belong to the tumor characteristic sequence set.

7. The method for radiotherapy dose control based on head and neck positioning error according to claim 6, characterized in that: The step of adjusting the dose distribution of the radiotherapy dose distribution matrix according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix includes: Performing feature recognition on the offset verification matrix to obtain distribution features of the first value and the second value in the offset verification matrix; Performing direction recognition on the distribution characteristics to obtain a tumor movement vector; The radiotherapy dose distribution matrix is ​​translated according to the tumor movement vector to obtain an updated radiotherapy dose distribution matrix.

8. The method for radiotherapy dose control based on head and neck positioning error according to claim 7, characterized in that: The method of using a pre-built respiratory control algorithm to monitor the patient's respiration and obtain a respiratory monitoring result includes: Using the pre-built breathing control algorithm, the patient's chest rise and fall is monitored to obtain the chest rise and fall fluctuation curve; Derivative calculation is performed on the chest cavity fluctuation curve to obtain a fluctuation variation curve; The fluctuation change curve is compared with the preset stable threshold interval to obtain the respiratory monitoring result.

9. The method for radiotherapy dose control based on head and neck positioning error according to claim 8, characterized in that: The determining whether the patient is in a pre-established respiratory limit position according to the respiratory monitoring result includes: When the fluctuation variation in the fluctuation variation curve is greater than the stable threshold interval, marking the respiratory monitoring result as a pre-constructed respiratory stage time position; When the fluctuation variation in the fluctuation variation curve is less than or equal to the stable threshold interval, the respiratory monitoring result is marked as a pre-constructed respiratory limit position.

10. A radiotherapy dose control system based on head and neck positioning error, characterized in that: The system comprises: An optical image processing module, used to use a pre-built real-time imaging device to scan a tumor area confirmed on a patient to obtain a tumor distribution scan image, and perform a pre-processing operation based on noise reduction and segmentation on the tumor distribution scan image to obtain a noise-reduced tumor structure map, and perform a matrix numerical configuration on the tumor distribution scan image according to the noise-reduced tumor structure map to obtain a treatment area matrix, wherein the numerical value within the noise-reduced tumor structure map corresponding to the treatment area matrix is ​​a preset first numerical value, and the numerical value outside the noise-reduced tumor structure map is a preset second numerical value, and the first numerical value is different from the second numerical value; a metabolic activity image processing module, configured to scan the tumor region using a pre-built positron emission tomography device to obtain a metabolic activity scan image, perform a clustering operation on the metabolic activity scan image based on the metabolic activity intensity to obtain a metabolic activity heat map, and construct a three-dimensional tumor structure map using the denoised tumor structure map, and perform an activity configuration operation on the three-dimensional tumor structure map according to the metabolic activity heat map to obtain a three-dimensional tumor activity map; a radiation dose prediction module, for performing a tumor feature recognition operation on the tumor three-dimensional activity map using a pre-trained radiation dose recognition model to obtain a tumor feature sequence set, performing a radiotherapy dose prediction operation on the tumor feature sequence set to obtain a radiotherapy dose prediction result, and performing a dose distribution operation on the radiotherapy dose prediction result according to the tumor feature sequence set to obtain a radiotherapy dose distribution matrix; The irradiation control module is used to monitor the patient's respiration using a pre-built respiration control algorithm to obtain respiration monitoring results, and determine whether the patient is in a pre-built respiration limit position based on the respiration monitoring results; when the patient is in the respiration limit position, intercept the treatment area matrix according to the pre-built ray irradiation port position to obtain an offset verification matrix, and adjust the dose distribution of the radiotherapy dose distribution matrix according to the offset verification matrix to obtain an updated radiotherapy dose distribution matrix; and use a pre-built radiation equipment cluster to generate high-energy rays according to the updated radiotherapy dose distribution matrix, and use the high-energy rays to irradiate the tumor area.

Images