A radiotherapy robot tumor motion estimation and prediction system and method

Abstract

The present invention discloses a system and method for estimating and predicting tumor motion in a radiotherapy robot, relating to the technical field of radiotherapy robots. The system comprises a state relationship curve analysis unit, a tumor motion tracking and acquisition unit, a model robot integration unit, a radiotherapy parameter modification and adjustment unit, and a real-time prediction and utilization unit. The present invention first uses magnetic resonance imaging technology to obtain the tumor position at different respiratory stages and analyzes its relationship with respiratory state, enabling a precise understanding of the tumor's changing patterns with respiratory motion. Simultaneously, X-ray technology is used to obtain soft tissue subtraction images and tumor motion trajectory data, enhancing understanding of tumor motion characteristics and making later radiotherapy positioning more accurate. This enables the radiotherapy robot to more intelligently track and treat tumors, maximizing the accuracy and efficiency of radiotherapy.

Description

Technical Field

[0001] The present invention relates to the technical field of radiotherapy robots, and in particular to a radiotherapy robot tumor motion estimation and prediction system and method. Background Art

[0002] A tumor is a process in which cells in the body grow and divide uncontrollably. This growth can be benign or malignant. The formation and progression of tumors can be caused by a variety of factors, including genetic factors, environmental exposures, lifestyle choices, and certain viral infections. Treatment options for tumors vary and depend on the type of tumor, its location, its stage of development, and the patient's overall health.

[0003] In recent years, with advances in technology and in-depth medical research, significant progress has been made in the field of cancer treatment, particularly in the development of precision medicine and personalized treatment plans. Common treatment methods include surgical resection, radiotherapy, chemotherapy, targeted therapy, and immunotherapy. For example, radiotherapy robots are often used in affine radiotherapy.

[0004] A radiotherapy robot is a highly advanced medical device that uses robotic technology to precisely locate and deliver radiation to treat cancer and other diseases. This robotic system can automatically perform complex radiotherapy procedures under the control of a doctor, improving the accuracy and efficiency of radiotherapy.

[0005] However, existing radiotherapy robots are unable to estimate the movement trajectory of the tumor based on its current position when in use, making it impossible to predict the development location of the tumor in advance, and thus unable to further perform radiotherapy on possible tumor locations based on the patient's current tumor location, thereby reducing the efficiency of radiotherapy.

[0006] Currently, no effective solutions have been proposed for the problems in related technologies. Summary of the Invention

[0007] In order to solve the above problems, the present invention proposes a radiotherapy robot tumor motion estimation and prediction system and method to maximize the accuracy and efficiency of radiotherapy.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] In a first aspect, the present invention provides a radiotherapy robot tumor motion estimation and prediction system, which includes a state relationship curve analysis unit, a tumor motion tracking and acquisition unit, a model robot integration unit, a radiotherapy parameter modification and adjustment unit, and a real-time prediction and use unit;

[0010] The state relationship curve parsing unit, the tumor motion tracking acquisition unit, the model robot integration unit, the radiotherapy parameter modification and adjustment unit, and the real-time prediction and use unit are sequentially connected;

[0011] A state relationship curve analysis unit is used to obtain the tumor position of the patient under different respiratory stages using magnetic resonance imaging technology, and to analyze the relationship curve between the respiratory state and the displacement of the tumor position;

[0012] Tumor motion tracking acquisition unit, used to use X-ray technology to collect images of the human body's internal structure to obtain soft tissue subtraction images, and use tumor motion tracking algorithms to calculate the tumor's motion trajectory and motion amplitude;

[0013] The model robot integration unit is used to solve the tumor motion characteristics based on the relationship curve, combine the motion trajectory and motion amplitude to build an estimation and prediction model, and integrate the estimation and prediction model with the radiotherapy robot;

[0014] A radiotherapy parameter modification and adjustment unit is used to simulate the output results under different tumor states through an estimated prediction model, and to modify and adjust the control parameters of the radiotherapy robot's radiotherapy measures according to the output results;

[0015] The real-time prediction unit is used to predict the movement position of the tumor based on the patient's real-time magnetic resonance image and X-ray image. The radiotherapy robot receives the movement position and selects the corresponding radiotherapy measures.

[0016] Preferably, the state relationship curve analysis unit includes a magnetic resonance image acquisition module, a respiratory state setting module, a magnetic resonance image processing module and a relationship curve analysis module;

[0017] The magnetic resonance image acquisition module, the respiratory state setting module, the magnetic resonance image processing module and the relationship curve analysis module are connected in sequence;

[0018] A magnetic resonance imaging acquisition module, configured to adjust the parameter settings and position of the magnetic resonance imaging instrument according to the patient's tumor location, and to perform a static scan to obtain the baseline location of the tumor;

[0019] A respiratory state setting module is used to set repeated scanning using a magnetic resonance imaging device under different respiratory states to capture the tumor location under each respiratory state;

[0020] Magnetic resonance image processing module, used to align different scanned magnetic resonance images of the same patient using image registration technology, and to mark the location and size of the tumor;

[0021] The relationship curve analysis module is used to convert the position and size changes of the tumor into dynamic change data, and establish a relationship curve model to identify the trend of tumor movement.

[0022] Preferably, converting the position and size changes of the tumor into dynamic change data and establishing a relationship curve model to identify the trend of tumor movement includes:

[0023] Convert the tumor position and size into a spatial rectangular coordinate system, and solve the mapping relationship between the position and size data based on the spatial rectangular coordinate system;

[0024] The mapping relationship is optimized by the least squares method, and the magnetic resonance image is registered to the spatial rectangular coordinate system according to the torque relationship to obtain dynamic change data;

[0025] The autoregressive model is used to extract dynamic features from dynamically changing data, and the extracted dynamic features are normalized to construct a recognition model.

[0026] Using dynamic features, a support vector data description relationship curve model is established, and the dynamic change data is input into the recognition model and the relationship curve model to obtain two sets of movement indicators;

[0027] The obtained movement index is inputted into the relationship curve model as a feature matrix again to obtain the movement relationship curve of the tumor position and size change, and the movement trend of the tumor is identified based on the movement relationship curve.

[0028] Preferably, the dynamic features are used to establish a support vector data description relationship curve model, and the dynamic change data is input into the recognition model and the relationship curve model to obtain two sets of movement indicators including:

[0029] Set the dynamic feature as the center point to establish the overall target area, and describe the dynamic change data and the center point with a support vector;

[0030] The dynamic change data within the overall target area is set as the support vector data, and the dynamic change data outside the overall target area is eliminated;

[0031] Implement error definition operation on support vector data and determine whether the distance from support vector data to the center point is less than the minimum constraint condition for dynamic feature acquisition;

[0032] According to the minimum constraint condition, the relaxation factor and penalty coefficient are introduced to obtain the optimization function of the relationship curve model, and the optimization function is combined with the Lagrangian function to obtain the relationship curve model;

[0033] The recognition model and the relationship curve model are used to set the mobile index threshold, weighted index, number of dynamic change data clusters and initial iteration value, and the membership matrix is obtained by determining the cluster center of the dynamic change data;

[0034] The dynamic change data are input into the recognition model and the relationship curve model respectively to obtain two groups of movement indices: the position index and the movement speed of the tumor at the imaging time point.

[0035] Preferably, the tumor motion tracking acquisition unit includes an X-ray image acquisition module, a subtraction image acquisition module, a position recognition and segmentation module, and a trajectory amplitude calculation module;

[0036] The X-ray image acquisition module, the subtraction image acquisition module, the position recognition and segmentation module, and the trajectory amplitude calculation module are connected in sequence;

[0037] The X-ray image acquisition module is used to adjust the X-ray equipment exposure parameters, radiation dose, and imaging area according to the patient's tumor location to capture images of the human body's internal structures;

[0038] A subtraction image acquisition module is used to process the image of the human body's internal structure according to the soft tissue subtraction and imaging processing technology to increase the soft tissue contrast and obtain a soft tissue subtraction image;

[0039] A position recognition and segmentation module is used to process the soft tissue subtraction image using image recognition and segmentation technology to locate the tumor area and mark the tumor boundary position;

[0040] The trajectory amplitude calculation module is used to analyze the motion trajectory of the tumor and measure the motion amplitude by using the tumor motion tracking algorithm in combination with the tumor target position.

[0041] Preferably, using a tumor motion tracking algorithm in combination with the tumor target position to analyze the motion trajectory of the tumor and measure the motion amplitude includes:

[0042] The shape points, area points and perimeter points of the tumor target location are described using a data point description method as tumor motion features, and the tumor motion features are input into a pre-set coordinate system;

[0043] A residual neural network combined with a tumor motion tracking algorithm is used to collect boundary data of tumor motion features in the coordinate system to extract deep residual features and generate motion boundary lines that match tumor tracking.

[0044] Constructing a target motion boundary model based on the motion boundary line to reflect the trajectory data of the tumor tracking motion boundary, and describing the final point of the motion trajectory of the tumor position by the fast tracking model based on the trajectory data;

[0045] The position change of the tumor in the coordinate system is determined based on the initial and final points of the motion trajectory to obtain the displacement distance, and the speed of the tumor movement is solved based on the displacement distance to obtain the tumor motion amplitude.

[0046] Preferably, constructing a target motion boundary model based on the motion boundary line to reflect the trajectory data of the tumor tracking motion boundary, and describing the final point of the motion trajectory of the fast tracking model to locate the tumor position based on the trajectory data includes:

[0047] According to the tumor target position and the motion boundary line, the coordinates of each tumor point on the boundary line are reflected, and the coordinate position is used to set the coordinate structure sequence set to establish the target motion boundary model to reflect the trajectory data of the tumor tracking motion boundary;

[0048] The similarity between the coordinates of the tumor points is judged to determine the final target threshold of the target motion boundary, and a fast tracking model is constructed to locate the final point of the tumor motion trajectory.

[0049] Preferably, the fast tracking model is expressed as:

[0050] ;

[0051] Where, ε Indicates the final positioning point of the tumor trajectory; θ i The first i Trajectory data points; θ j The first j Trajectory data points; ψ The coordinate points representing the motion boundary line; P The target threshold represents the motion boundary line; L Boundary lines representing tumor target motion.

[0052] Preferably, the model robot integration unit includes an interval statistics processing module, an estimation model construction module, a control integration application module and a protocol integration control module;

[0053] The interval statistics processing module, the estimation model building module, the control integration application module and the protocol integration control module are connected in sequence;

[0054] The interval statistics processing module is used to divide and calculate characteristic parameters based on the relationship curve, motion trajectory and motion amplitude, and multiply the integral interval statistics by the tumor relationship motion change value of the same patient and the corresponding motion trajectory amplitude;

[0055] An estimation model building module is used to implement a clustering operation on the statistical results through a clustering algorithm to obtain an initial prediction level, and to build an estimation prediction model with the interval distribution as input and the initial prediction level as output;

[0056] The control integration application module is used to optimize the parameters of the estimation and prediction model using an optimization algorithm, and integrate and apply the constructed estimation and prediction model with the controller of the radiotherapy robot;

[0057] The protocol integrated control module is used to establish the control protocol of the radiotherapy robot according to the estimated prediction model, calculate the integrated control frequency, and input the intelligent integrated control quantity to complete the automatic control of the radiotherapy robot.

[0058] In a second aspect, the present invention further provides a method for estimating and predicting tumor motion of a radiotherapy robot, the method comprising the following steps:

[0059] S1. Use magnetic resonance imaging to determine the tumor location at different respiratory stages and analyze the relationship between respiratory status and tumor displacement.

[0060] S2. Use X-ray technology to collect images of the human body's internal structures to obtain soft tissue subtraction images, and use tumor motion tracking algorithms to calculate the tumor's motion trajectory and amplitude;

[0061] S3. Solve the tumor motion characteristics based on the relationship curve and combine the motion trajectory and motion amplitude to build an estimation prediction model, and integrate the estimation prediction model with the radiotherapy robot;

[0062] S4. Simulating the output results under different tumor states by estimating the prediction model, and modifying and adjusting the control parameters of the radiotherapy robot's radiotherapy measures according to the output results;

[0063] S5. The tumor movement position is predicted based on the patient's real-time MRI image and X-ray image, and the radiotherapy robot receives the movement position and selects the corresponding radiotherapy measures.

[0064] The beneficial effects of the present invention are:

[0065] 1. The present invention first uses magnetic resonance imaging technology to obtain the tumor position at different respiratory stages and analyzes its relationship with the respiratory state, so as to accurately understand the changing pattern of the tumor with respiratory movement. At the same time, X-ray technology is used to obtain soft tissue subtraction images and tumor motion trajectory data, which enhances the understanding of the tumor motion characteristics and makes the subsequent radiotherapy positioning more accurate. Finally, by establishing an estimation and prediction model and integrating it with the radiotherapy robot, an individualized treatment plan based on the patient's specific situation is realized, and the radiotherapy robot can track and treat the tumor more intelligently, thereby maximizing the accuracy and efficiency of radiotherapy.

[0066] 2. The present invention uses magnetic resonance imaging technology to obtain tumor positions under different respiratory stages and analyzes the relationship curve between respiratory status and tumor position displacement. By implementing static scanning and repeated scanning under different respiratory states, the specific position changes of the tumor with respiratory movement can be captured in detail. At the same time, the changes in tumor position and size are converted into dynamic data, and a relationship curve model is established. This can not only identify the movement trend of the tumor, but also provide a scientific basis for formulating personalized treatment plans and adjusting treatment strategies.

[0067] 3. By adjusting the exposure parameters, radiation dose, and imaging area of the X-ray equipment, the present invention can optimize the image acquisition process according to the specific location and characteristics of the patient's tumor, ensuring the acquisition of high-quality internal structure images. At the same time, soft tissue subtraction and imaging processing technologies are used to enhance the contrast of soft tissue, making the identification of tumors and surrounding tissues clearer. Image recognition and segmentation technologies are applied to process soft tissue subtraction images to accurately locate the tumor area and mark the tumor boundary. The tumor motion tracking algorithm is combined with the precise tumor target location to analyze the tumor's motion trajectory in detail and accurately measure its motion amplitude.

[0068] 4. The present invention uses relationship curves and motion trajectory data to quantify the motion characteristics of tumors, which helps to accurately predict the position of tumors during radiotherapy, thereby adjusting the irradiation direction and position of the radiation beam to ensure that the radiation beam is accurately aimed at the tumor and reduce damage to surrounding normal tissues. At the same time, by classifying tumor motion data through clustering algorithms, tumor groups with similar motion characteristics can be identified, and more personalized and optimized treatment plans can be designed for different types of tumors. By using optimization algorithms to optimize the parameters of the prediction model, the prediction accuracy of the model can be further improved, and the adjustment time required during radiotherapy can be reduced. The estimated prediction model can be integrated with the control system of the radiotherapy robot to realize automatic control of the radiotherapy robot, reduce errors and uncertainties in human operation, and make the radiotherapy process more stable and reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0070] Figure 1 This is a principle block diagram of a radiotherapy robot tumor motion estimation and prediction system according to an embodiment of the present invention;

[0071] Figure 2 This is a principle block diagram of a state relationship curve parsing unit in a tumor motion estimation and prediction system of a radiotherapy robot according to an embodiment of the present invention;

[0072] Figure 3 This is a principle block diagram of a tumor motion tracking and acquisition unit in a tumor motion estimation and prediction system of a radiotherapy robot according to an embodiment of the present invention;

[0073] Figure 4 This is a principle block diagram of a model robot integration unit in a radiotherapy robot tumor motion estimation and prediction system according to an embodiment of the present invention;

[0074] Figure 5 The present invention is a flowchart of a method for estimating and predicting tumor motion of a radiotherapy robot according to an embodiment of the present invention.

[0075] In the picture:

[0076] 1. State relationship curve analysis unit; 101. Magnetic resonance image acquisition module; 102. Respiratory state setting module; 103. Magnetic resonance image processing module; 104. Relationship curve analysis module; 2. Tumor motion tracking acquisition unit; 201. X-ray image acquisition module; 202. Subtraction image acquisition module; 203. Position recognition and segmentation module; 204. Trajectory amplitude calculation module; 3. Model robot integration unit; 301. Interval statistics processing module; 302. Estimation model construction module; 303. Control integration application module; 304. Protocol integration control module; 4. Radiotherapy parameter modification and adjustment unit; 5. Real-time prediction and use unit. DETAILED DESCRIPTION

[0077] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0078] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0079] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0080] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0081] See also Figure 1 The present invention provides a radiotherapy robot tumor motion estimation and prediction system, which includes a state relationship curve analysis unit 1, a tumor motion tracking and acquisition unit 2, a model robot integration unit 3, a radiotherapy parameter modification and adjustment unit 4 and a real-time prediction and use unit 5.

[0082] Among them, the state relationship curve analysis unit 1, the tumor motion tracking acquisition unit 2, the model robot integration unit 3, the radiotherapy parameter modification and adjustment unit 4 and the real-time prediction and use unit 5 are connected in sequence;

[0083] See also Figure 2 The state relationship curve analysis unit 1 is used to obtain the tumor position of the patient under different breathing stages using magnetic resonance imaging technology, and to analyze the relationship curve between the breathing state and the tumor position displacement.

[0084] In this embodiment, the state relationship curve analysis unit 1 includes a magnetic resonance image acquisition module 101 , a respiratory state setting module 102 , a magnetic resonance image processing module 103 and a relationship curve analysis module 104 .

[0085] The magnetic resonance image acquisition module 101 , the respiratory state setting module 102 , the magnetic resonance image processing module 103 and the relationship curve analysis module 104 are connected in sequence.

[0086] The magnetic resonance image acquisition module 101 is used to adjust the parameter settings and position of the magnetic resonance imaging instrument according to the patient's tumor position, and perform static scanning to obtain the tumor baseline position.

[0087] Specifically, adjusting the parameter settings and position of the magnetic resonance imaging instrument according to the patient's tumor location and performing a static scan to obtain the tumor baseline location includes the following steps:

[0088] Assess the patient's historical tumor condition, including the type, location, size of the tumor and the patient's health status. Adjust the scanning parameters of the magnetic resonance imaging device, including magnetic field strength, echo time and repetition time, according to the location and characteristics of the patient's tumor to obtain the best imaging effect.

[0089] The positioning function of the magnetic resonance imaging instrument is used to ensure that the patient and the specific tumor area are in the optimal imaging position. The patient remains still for a static scan to obtain a baseline position image of the tumor.

[0090] The respiratory state setting module 102 is used to set the repeated scanning by the magnetic resonance imaging device under different respiratory states to capture the tumor position under each respiratory state.

[0091] Specifically, performing repeated scans using a magnetic resonance imaging device under different respiratory states to capture the tumor location under each respiratory state includes the following steps:

[0092] Patients are trained on how to control their breathing to ensure they can remain still in a specific breathing state as instructed during the scan. MRI scan parameters, including scan sequence, image resolution, and scan time, are adjusted based on the patient's tumor location and expected respiratory state.

[0093] Using respiratory synchronization technology, the timing of scanning is adjusted according to the patient's breathing pattern to ensure that images are acquired during the same breathing phase. During the scanning process, sound or visual signals are used to guide the patient to remain still in a specific breathing state (including inhalation, exhalation, and breath holding).

[0094] Repeated MRI scans were performed under various breathing conditions to capture the changing position of the tumor during different breathing phases.

[0095] The magnetic resonance image processing module 103 is used to align different scanned magnetic resonance images of the same patient using image registration technology, and to mark the location and size of the tumor.

[0096] Specifically, using image registration technology to align different MRI scans of the same patient and annotate the location and size of the tumor includes the following steps:

[0097] The magnetic resonance image is cropped to remove irrelevant background areas and focus on the tumor and its surrounding tissues. De-noising algorithms, including Gaussian filtering and median filtering, are applied to reduce random noise in the image. One image is selected as a reference image and the other images are initially aligned with it to align the anatomical structures in the image.

[0098] The least squares method is used to refine the alignment between images, ensuring that the same anatomical structures in different images are completely spatially aligned. The tumor is accurately located on the registered image with the assistance of an automatic detection algorithm. The size of the tumor, including parameters such as length, width, and volume, is measured and annotated to facilitate subsequent analysis and treatment planning.

[0099] The relationship curve analysis module 104 is used to convert the position and size changes of the tumor into dynamic change data, and establish a relationship curve model to identify the trend of tumor movement.

[0100] Specifically, the changes in tumor location and size are converted into dynamic change data, and a relationship curve model is established to identify the trend of tumor movement, including:

[0101] Convert the tumor position and size into a spatial rectangular coordinate system, and solve the mapping relationship between the position and size data based on the spatial rectangular coordinate system;

[0102] The mapping relationship is optimized by the least squares method, and the magnetic resonance image is registered to the spatial rectangular coordinate system according to the torque relationship to obtain dynamic change data.

[0103] The autoregressive model is used to extract dynamic features from dynamically changing data, and the extracted dynamic features are normalized to construct a recognition model.

[0104] Dynamic features are used to establish a support vector data description relationship curve model, and the dynamically changing data are input into the recognition model and the relationship curve model to obtain two sets of movement indicators.

[0105] Among them, the dynamic features are used to establish a support vector data description relationship curve model, and the dynamic change data is input into the recognition model and the relationship curve model to obtain two sets of movement indicators including:

[0106] Set the dynamic feature as the center point to establish the overall target area, and describe the dynamic change data and the center point with a support vector;

[0107] The dynamic change data within the overall target area is set as the support vector data, and the dynamic change data outside the overall target area is eliminated;

[0108] Implement error definition operation on support vector data and determine whether the distance from support vector data to the center point is less than the minimum constraint condition for dynamic feature acquisition;

[0109] According to the minimum constraint condition, the relaxation factor and penalty coefficient are introduced to obtain the optimization function of the relationship curve model, and the optimization function is combined with the Lagrangian function to obtain the relationship curve model;

[0110] The recognition model and the relationship curve model are used to set the mobile index threshold, weighted index, number of dynamic change data clusters and initial iteration value, and the membership matrix is obtained by determining the cluster center of the dynamic change data;

[0111] The dynamic change data are input into the recognition model and the relationship curve model respectively to obtain two groups of movement indices: the position index and the movement speed of the tumor at the imaging time point.

[0112] The obtained movement index is inputted into the relationship curve model as a feature matrix again to obtain the movement relationship curve of the tumor position and size change, and the movement trend of the tumor is identified based on the movement relationship curve.

[0113] Therefore, by using magnetic resonance imaging technology to obtain the tumor position under different respiratory stages and analyzing the relationship curve between respiratory state and tumor position displacement, static scanning and repeated scanning under different respiratory states can be performed to capture in detail the specific position changes of the tumor with respiratory movement.

[0114] See also Figure 3 The tumor motion tracking acquisition unit 2 is used to use X-ray technology to collect images of the internal structure of the human body to obtain soft tissue subtraction images, and use a tumor motion tracking algorithm to calculate the tumor motion trajectory and motion amplitude.

[0115] In this embodiment, the tumor motion tracking and acquisition unit 2 includes an X-ray image acquisition module 201 , a subtraction image acquisition module 202 , a position recognition and segmentation module 203 , and a trajectory amplitude calculation module 204 .

[0116] The X-ray image acquisition module 201 , the subtraction image acquisition module 202 , the position recognition and segmentation module 203 and the trajectory amplitude calculation module 204 are connected in sequence.

[0117] The X-ray image acquisition module 201 is used to adjust the exposure parameters, radiation dose and imaging area of the X-ray device according to the patient's tumor location to acquire images of the internal structure of the human body.

[0118] The steps of adjusting the exposure parameters, radiation dose, and imaging area of the X-ray device according to the location of the patient's tumor to acquire images of the internal structure of the human body include the following:

[0119] Collect information related to the patient's medical history, tumor location, and size, and adjust the exposure parameters of the X-ray equipment according to the patient's body shape, tumor location, and type. Based on the patient's specific condition and examination area, calculate and set the appropriate radiation dose to ensure image quality while protecting the patient from unnecessary radiation.

[0120] The positioning system of the X-ray equipment is used to accurately determine the location of the tumor, and based on this, the optimal imaging area is selected. The position of the X-ray emitter and imaging plate (or detector) is adjusted according to the required imaging area.

[0121] The patient is asked to maintain a specific posture, the X-ray equipment is started, and one or more images are taken, which may include different angles or positions as needed to fully capture the tumor and its surrounding structures.

[0122] The subtraction image acquisition module 202 is used to process the image of the human body internal structure according to the soft tissue subtraction and imaging processing technology to increase the soft tissue contrast and obtain a soft tissue subtraction image.

[0123] The steps of processing the image of the internal structure of the human body to increase the soft tissue contrast and obtain the soft tissue subtraction image according to the soft tissue subtraction and imaging processing technology include the following steps:

[0124] Image denoising technology is used to reduce noise in the image, improve image clarity, adjust image contrast, and enhance the visualization of soft tissue in the image. Soft tissue subtraction algorithms (including digital subtraction angiography and dual-energy subtraction imaging) are selected based on the imaging purpose and imaging technology used.

[0125] The selected subtraction technique is applied to the pre-processed images to enhance the contrast of target soft tissues (such as blood vessels, tumors, etc.) by weakening or removing the images of non-target soft tissues (such as bones, etc.).

[0126] The position recognition and segmentation module 203 is used to process the soft tissue subtraction image using image recognition and segmentation technology to locate the tumor area and mark the tumor boundary position.

[0127] The use of image recognition and segmentation technology to process the soft tissue subtraction image to locate the tumor area and mark the tumor boundary position includes the following steps:

[0128] The image is preprocessed, including adjusting the brightness and contrast to improve image quality. Based on the image characteristics and the target task (tumor localization and boundary marking in this case), the level set method image segmentation technique is selected to identify and segment the tumor area in the soft tissue subtraction image.

[0129] After successfully identifying the tumor area, the boundary of the tumor is clearly marked, including drawing a contour line on the image or highlighting the tumor area with a specific color.

[0130] The trajectory amplitude calculation module 204 is used to analyze the motion trajectory of the tumor and measure the motion amplitude by using a tumor motion tracking algorithm in combination with the tumor target position.

[0131] Among them, using tumor motion tracking algorithms combined with tumor target position to analyze the motion trajectory of tumor motion and measure the motion amplitude includes:

[0132] The shape points, area points and perimeter points of the tumor target location are described using a data point description method as tumor motion features, and the tumor motion features are input into a pre-set coordinate system;

[0133] A residual neural network combined with a tumor motion tracking algorithm is used to collect boundary data of tumor motion features in the coordinate system to extract deep residual features and generate motion boundary lines that match tumor tracking.

[0134] Constructing a target motion boundary model based on the motion boundary line to reflect the trajectory data of the tumor tracking motion boundary, and describing the final point of the motion trajectory of the tumor position by the fast tracking model based on the trajectory data;

[0135] The position change of the tumor in the coordinate system is determined based on the initial and final points of the motion trajectory to obtain the displacement distance, and the speed of the tumor movement is solved based on the displacement distance to obtain the tumor motion amplitude.

[0136] Specifically, the target motion boundary model is constructed based on the motion boundary line to reflect the trajectory data of the tumor tracking motion boundary, and the final point of the motion trajectory of the tumor position is described by the fast tracking model based on the trajectory data, including:

[0137] According to the tumor target position and the motion boundary line, the coordinates of each tumor point on the boundary line are reflected, and the coordinate position is used to set the coordinate structure sequence set to establish the target motion boundary model to reflect the trajectory data of the tumor tracking motion boundary;

[0138] The similarity between the coordinates of the tumor points is judged to determine the final target threshold of the target motion boundary, and a fast tracking model is constructed to locate the final point of the tumor motion trajectory.

[0139] The expression of the fast tracking model is:

[0140] ;

[0141] Where, ε Indicates the final positioning point of the tumor trajectory; θ i The first i Trajectory data points; θ j The first j Trajectory data points; ψ The coordinate points representing the motion boundary line; P The target threshold represents the motion boundary line; L Boundary lines representing tumor target motion.

[0142] Therefore, by adjusting the exposure parameters, radiation dose and imaging area of the X-ray equipment, high-quality internal structure images can be obtained according to the specific location and characteristics of the patient's tumor. At the same time, the tumor motion tracking algorithm is used to combine the precise tumor target position to analyze the tumor's motion trajectory in detail and accurately measure its motion amplitude.

[0143] See also Figure 4 The model robot integration unit 3 is used to solve the tumor motion characteristics based on the relationship curve and combine the motion trajectory and motion amplitude to build an estimation and prediction model, and integrate the estimation and prediction model with the radiotherapy robot.

[0144] In this embodiment, the model robot integration unit 3 includes an interval statistics processing module 301 , an estimation model construction module 302 , a control integration application module 303 and a protocol integration control module 304 .

[0145] Among them, the interval statistics processing module 301, the estimation model construction module 302, the control integration application module 303 and the protocol integration control module 304 are connected in sequence.

[0146] The interval statistics processing module 301 is used to divide and calculate characteristic parameters according to the relationship curve, motion trajectory and motion amplitude, and multiply the tumor relationship motion change value of the same patient by the corresponding motion trajectory amplitude to obtain the integral interval statistics.

[0147] Specifically, the characteristic parameters are divided and calculated according to the relationship curve, motion trajectory and motion amplitude, and the integral interval statistics of the tumor relationship motion change value of the same patient and the corresponding motion trajectory amplitude are multiplied. The following steps are included:

[0148] Based on the tumor's motion relationship curve, the changing pattern of the tumor with respiration or other physiological movements is determined, and image processing technology is used to extract the tumor's motion trajectory, including its path and direction, from the serial images. Based on the tumor's starting and final positions, its motion amplitude at different time points is calculated.

[0149] According to the relationship curve and the motion trajectory, the relationship motion change value of the tumor in a specific time period is calculated, and the tumor motion change value at each time point is multiplied by the corresponding motion trajectory amplitude to obtain the product value.

[0150] The product value range is divided into different statistical intervals, and the product value in each interval is statistically analyzed.

[0151] The estimation model construction module 302 is used to perform clustering operations on the statistical results through a clustering algorithm to obtain an initial prediction level, and to construct an estimation prediction model with the interval distribution as input and the initial prediction level as output.

[0152] Specifically, the statistical results are clustered by a clustering algorithm to obtain an initial prediction level, and an estimation prediction model is constructed with the interval distribution as input and the initial prediction level as output, including the following steps:

[0153] The statistical results are preprocessed to eliminate different dimensions, and the appropriate K-means clustering algorithm is selected based on the data characteristics. The algorithm parameters are set according to the characteristics of the data and the expected clustering effect.

[0154] Use the selected algorithm to perform clustering operations on the data, divide the statistical results into several groups according to the similarity of their characteristic parameters, analyze the clustering results, and evaluate the characteristics of each cluster group, including the center point of each group, the distance within the group, etc., to obtain the initial prediction level.

[0155] According to the nature of the problem, random forest is selected as the prediction model, and interval distribution data is used as input and initial prediction level as output, and the prediction model is trained using the training data set.

[0156] The control integration application module 303 is used to perform parameter optimization operations on the estimation prediction model using an optimization algorithm, and integrate and apply the constructed estimation prediction model with the controller of the radiotherapy robot.

[0157] Specifically, the following steps are involved: using an optimization algorithm to perform parameter optimization on the estimation prediction model, and integrating and applying the constructed estimation prediction model with the controller of the radiotherapy robot:

[0158] A comprehensive evaluation of the performance of the estimated prediction model is conducted, including accuracy, recall, and precision. Based on the results of the performance evaluation, the optimization goal is clarified, and the gradient descent optimization algorithm is selected to optimize the model parameters.

[0159] The gradient descent optimization algorithm is used to optimize the parameters of the estimation prediction model to achieve the pre-set optimization goal. After the optimization is completed, the interface between the estimation prediction model and the radiotherapy robot controller is defined to ensure that the two can communicate effectively and achieve integrated connection.

[0160] The protocol integrated control module 304 is used to establish a control protocol for the radiotherapy robot according to the estimation prediction model, calculate the integrated control frequency, and input the intelligent integrated control quantity to complete the automatic control of the radiotherapy robot.

[0161] Specifically, the control protocol of the radiotherapy robot is established based on the estimated prediction model to calculate the integrated control frequency, and the intelligent integrated control quantity is input to complete the automatic control of the radiotherapy robot, which includes the following steps:

[0162] Clarify the goals of radiotherapy robot control, including accurately tracking tumor location, adjusting radiation angle and intensity, etc. to maximize treatment efficacy and minimize damage to surrounding healthy tissues, and design control strategies based on the output of the estimated predictive model.

[0163] Determine the control frequency required by the radiotherapy robot to ensure that it can respond to the updated information provided by the estimated prediction model in a timely manner, and design the control loop, including the cycle frequency of steps such as data acquisition, prediction model calculation, control quantity generation, and execution of control instructions.

[0164] Based on the output of the estimated prediction model and the control strategy, the specific control quantities are calculated, including the movement speed, direction, radiation intensity, etc. of the radiotherapy robot. The calculated control quantities are input into the control system of the radiotherapy robot for real-time adjustment of the robot's behavior. The radiotherapy robot performs corresponding operations based on the received control quantities, including adjusting the position, angle, radiation dose, etc.

[0165] Therefore, quantifying the motion characteristics of tumors using relationship curves and motion trajectory data helps to accurately predict the position of tumors during radiotherapy, thereby adjusting the irradiation direction and position of the radiation beam to ensure that the radiation beam is precisely aimed at the tumor and reduce damage to surrounding normal tissues. At the same time, by classifying tumor motion data through clustering algorithms, tumor groups with similar motion characteristics can be identified, and more personalized and optimized treatment plans can be designed for different types of tumors.

[0166] The radiotherapy parameter modification and adjustment unit 4 is used to simulate the output results under different tumor states through the estimation prediction model, and modify and adjust the control parameters of the radiotherapy measures of the radiotherapy robot according to the output results.

[0167] In this embodiment, simulating the output results of different tumor states by estimating the prediction model and modifying and adjusting the control parameters of the radiotherapy robot's radiotherapy measures according to the output results include the following steps:

[0168] Clarify the goals of the simulation, including simulating the behavior of tumors when they change in size, shape, or location, and prepare the input data required for the simulation based on the goals, including various hypothetical states of the tumor and related physiological parameters.

[0169] Use the estimated prediction model to simulate the prepared input data, generate predicted output results under different tumor states, collect the output results of the model simulation, analyze the simulation results, and identify the key features and change trends under different tumor states.

[0170] Based on the simulation results, the control parameters that may need to be adjusted in the radiotherapy measures are identified, such as radiation dose, angle, duration, etc., and the control parameter adjustment plan of the radiotherapy robot is designed according to the simulation results and analysis.

[0171] The real-time prediction and utilization unit 5 is used to predict the movement position of the tumor based on the patient's real-time magnetic resonance image and X-ray image. The radiotherapy robot receives the movement position and selects the corresponding radiotherapy measure.

[0172] In this embodiment, the tumor motion position is predicted based on the patient's real-time MRI and X-ray images, and the radiotherapy robot receives the motion position and selects the corresponding radiotherapy measures, including the following key steps:

[0173] Use magnetic resonance imaging equipment and X-ray imaging equipment to collect the patient's internal image data in real time to ensure that the latest status of the tumor and surrounding tissues is obtained, and synchronously process magnetic resonance and X-ray image data to identify the location of the tumor.

[0174] Feature information about the tumor is extracted from the preprocessed image and input into the estimation prediction model to predict the motion trajectory and position of the tumor. The control protocol of the radiotherapy robot is updated according to the predicted position of the tumor, including parameters such as radiation dose, angle and duration.

[0175] The radiotherapy robot automatically adjusts its operating parameters according to the updated control protocol, including moving to the appropriate position and adjusting the radiation angle.

[0176] See also Figure 5 The present invention also provides a method for estimating and predicting tumor motion of a radiotherapy robot, the method comprising the following steps:

[0177] S1. Use magnetic resonance imaging to determine the tumor location at different respiratory stages and analyze the relationship between respiratory status and tumor displacement.

[0178] S2. Use X-ray technology to collect images of the human body's internal structures to obtain soft tissue subtraction images, and use tumor motion tracking algorithms to calculate the tumor's motion trajectory and amplitude;

[0179] S3. Solve the tumor motion characteristics based on the relationship curve and combine the motion trajectory and motion amplitude to build an estimation prediction model, and integrate the estimation prediction model with the radiotherapy robot;

[0180] S4. Simulating the output results under different tumor states by estimating the prediction model, and modifying and adjusting the control parameters of the radiotherapy robot's radiotherapy measures according to the output results;

[0181] S5. The tumor movement position is predicted based on the patient's real-time MRI image and X-ray image, and the radiotherapy robot receives the movement position and selects the corresponding radiotherapy measures.

[0182] In summary, with the help of the above technical solutions of the present invention, the present invention first uses magnetic resonance imaging technology to obtain the tumor position at different respiratory stages and analyzes its relationship with the respiratory state, so as to accurately understand the changing pattern of tumor movement with respiratory motion. At the same time, X-ray technology is used to obtain soft tissue subtraction images and tumor motion trajectory data to enhance the understanding of tumor motion characteristics and make the later radiotherapy positioning more accurate. Finally, by establishing an estimation prediction model and integrating it with the radiotherapy robot, an individualized treatment plan based on the patient's specific situation is achieved, and the radiotherapy robot can track and treat the tumor more intelligently, thereby maximizing the accuracy and efficiency of radiotherapy. The present invention uses magnetic resonance imaging technology to obtain the tumor position at different respiratory stages and analyzes the relationship curve between respiratory state and tumor position displacement. Static scanning and repeated scanning under different respiratory states can capture the specific position changes of the tumor with respiratory motion in detail. At the same time, the changes in the position and size of the tumor are converted into dynamic data and a relationship curve model is established. This can not only identify the movement trend of the tumor, but also provide a scientific basis for formulating personalized treatment plans and adjusting treatment strategies.

[0183] The present invention optimizes the image acquisition process according to the specific location and characteristics of the patient's tumor by adjusting the exposure parameters, radiation dose and imaging area of the X-ray equipment, ensuring the acquisition of high-quality internal structure images. At the same time, soft tissue subtraction and imaging processing technology are used to enhance the contrast of soft tissue, making the identification of tumor and surrounding tissue clearer. Image recognition and segmentation technology are used to process soft tissue subtraction images to accurately locate the tumor area and mark the tumor boundary. The tumor motion tracking algorithm is used, combined with the precise tumor target location, to analyze the tumor's motion trajectory in detail and accurately measure its motion amplitude. The present invention uses relationship curves and motion trajectory data to quantify the motion characteristics of tumors, which helps to accurately predict the position of tumors during radiotherapy, thereby adjusting the irradiation direction and position of the radiation beam, ensuring that the radiation beam is accurately aimed at the tumor, and reducing damage to surrounding normal tissues. At the same time, by classifying tumor motion data through clustering algorithms, tumor groups with similar motion characteristics can be identified, and more personalized and optimized treatment plans can be designed for different types of tumors. Using optimization algorithms to optimize the parameters of the prediction model can further improve the prediction accuracy of the model and reduce the adjustment time required during radiotherapy. The estimated prediction model is integrated with the control system of the radiotherapy robot to realize automated control of the radiotherapy robot, reduce errors and uncertainties in human operation, and make the radiotherapy process more stable and reliable.

[0184] Those skilled in the art will appreciate that the units, i.e., algorithm steps, of the various examples described in conjunction with this embodiment can be implemented using electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0185] Although the above describes the specific embodiments of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without any creative work are still within the scope of protection of the present invention.

Claims

1. A radiotherapy robot tumor motion estimation and prediction system, characterized in that: include: A state relationship curve analysis unit is used to obtain the tumor position of the patient under different respiratory stages using magnetic resonance imaging technology, and to analyze the relationship curve between the respiratory state and the displacement of the tumor position; Tumor motion tracking acquisition unit, used to use X-ray technology to collect images of the human body's internal structure to obtain soft tissue subtraction images, and use tumor motion tracking algorithms to calculate the tumor's motion trajectory and motion amplitude; The model robot integration unit is used to solve the tumor motion characteristics based on the relationship curve, combine the motion trajectory and motion amplitude to build an estimation and prediction model, and integrate the estimation and prediction model with the radiotherapy robot; A radiotherapy parameter modification and adjustment unit is used to simulate the output results under different tumor states through an estimated prediction model, and to modify and adjust the control parameters of the radiotherapy robot's radiotherapy measures according to the output results; A real-time prediction unit is used to predict the movement position of the tumor based on the patient's real-time magnetic resonance image and X-ray image. The radiotherapy robot receives the movement position and selects the corresponding radiotherapy measures. Wherein, the state relationship curve analysis unit includes: The relationship curve analysis module is used to convert the tumor position and size into a spatial rectangular coordinate system, and solve the mapping relationship between the position and size data based on the spatial rectangular coordinate system; implement torque relationship optimization operations on the mapping relationship according to the principle of least squares method, and align the magnetic resonance image to the spatial rectangular coordinate system according to the torque relationship to obtain dynamic change data; use the autoregressive model to extract dynamic features of the dynamic change data, and perform normalization processing on the extracted dynamic features to construct a recognition model; use dynamic features to establish a support vector data description relationship curve model, and input the dynamic change data into the recognition model and the relationship curve model to obtain two sets of movement indicators.

2. A radiotherapy robot tumor motion estimation and prediction system according to claim 1, characterized in that: The state relationship curve analysis unit includes: A magnetic resonance image acquisition module is used to adjust the parameter position of the magnetic resonance imaging instrument according to the tumor location and perform static scanning to obtain the baseline position of the tumor; A respiratory state setting module is used to set repeated scanning using a magnetic resonance imaging device under different respiratory states to capture the tumor location under each respiratory state; Magnetic resonance image processing module, used to align different scanned magnetic resonance images of the same patient using image registration technology, and to mark the location and size of the tumor; The magnetic resonance image acquisition module, the respiratory state setting module, the magnetic resonance image processing module and the relationship curve analysis module are connected in sequence.

3. The radiotherapy robot tumor motion estimation and prediction system according to claim 2, characterized in that: The method further includes: using dynamic features to establish a support vector data description relationship curve model, and inputting the dynamic change data into the recognition model and the relationship curve model to obtain two sets of movement indicators; The obtained movement index is inputted into the relationship curve model as a feature matrix again to obtain the movement relationship curve of the tumor position and size change, and the movement trend of the tumor is identified based on the movement relationship curve.

4. A radiotherapy robot tumor motion estimation and prediction system according to claim 3, characterized in that: The dynamic features are used to establish a support vector data description relationship curve model, and the dynamic change data is input into the recognition model and the relationship curve model to obtain two sets of movement indicators including: Set the dynamic feature as the center point to establish the overall target area, and describe the dynamic change data and the center point with a support vector; The dynamic change data within the overall target area is set as the support vector data, and the dynamic change data outside the overall target area is eliminated; Implement error definition operation on support vector data and determine whether the distance from support vector data to the center point is less than the minimum constraint condition for dynamic feature acquisition; According to the minimum constraint condition, the relaxation factor and penalty coefficient are introduced to obtain the optimization function of the relationship curve model, and the optimization function is combined with the Lagrangian function to obtain the relationship curve model; The recognition model and the relationship curve model are used to set the mobile index threshold, weighted index, number of dynamic change data clusters and initial iteration value, and the membership matrix is obtained by determining the cluster center of the dynamic change data; The dynamic change data are input into the recognition model and the relationship curve model respectively to obtain two groups of movement indices: the position index and the movement speed of the tumor at the imaging time point.

5. The radiotherapy robot tumor motion estimation and prediction system according to claim 1, characterized in that: The tumor motion tracking acquisition unit includes an X-ray image acquisition module, a subtraction image acquisition module, a position recognition and segmentation module, and a trajectory amplitude calculation module; The X-ray image acquisition module, the subtraction image acquisition module, the position recognition and segmentation module, and the trajectory amplitude calculation module are connected in sequence; The X-ray image acquisition module is used to adjust the exposure parameters, radiation dose and imaging area of the X-ray equipment according to the location of the patient's tumor to acquire images of the internal structure of the human body; The subtraction image acquisition module is used to process the human body internal structure image according to the soft tissue subtraction and imaging processing technology to increase the soft tissue contrast and obtain a soft tissue subtraction image; The position recognition and segmentation module is used to process the soft tissue subtraction image using image recognition and segmentation technology to locate the tumor area and mark the tumor boundary position; The trajectory amplitude calculation module is used to analyze the motion trajectory of the tumor and measure the motion amplitude by using a tumor motion tracking algorithm in combination with the tumor target position.

6. The radiotherapy robot tumor motion estimation and prediction system according to claim 5, characterized in that: The method of analyzing the motion trajectory of the tumor and measuring the motion amplitude by using a tumor motion tracking algorithm in combination with the tumor target position includes: The shape points, area points and perimeter points of the tumor target location are described using a data point description method as tumor motion features, and the tumor motion features are input into a pre-set coordinate system; A residual neural network combined with a tumor motion tracking algorithm is used to collect boundary data of tumor motion features in the coordinate system to extract deep residual features and generate motion boundary lines that match tumor tracking. Constructing a target motion boundary model based on the motion boundary line to reflect the trajectory data of the tumor tracking motion boundary, and describing the final point of the motion trajectory of the tumor position by the fast tracking model based on the trajectory data; The position change of the tumor in the coordinate system is determined based on the initial and final points of the motion trajectory to obtain the displacement distance, and the speed of the tumor movement is solved based on the displacement distance to obtain the tumor motion amplitude.

7. The radiotherapy robot tumor motion estimation and prediction system according to claim 6, characterized in that: The target motion boundary model is constructed based on the motion boundary line to reflect the trajectory data of the tumor tracking motion boundary, and the final point of the motion trajectory of the tumor position is described based on the trajectory data by the fast tracking model, including: According to the tumor target position and the motion boundary line, the coordinates of each tumor point on the boundary line are reflected, and the coordinate position is used to set the coordinate structure sequence set to establish the target motion boundary model to reflect the trajectory data of the tumor tracking motion boundary; The similarity between the coordinates of the tumor points is judged to determine the final target threshold of the target motion boundary, and a fast tracking model is constructed to locate the final point of the tumor motion trajectory.

8. The radiotherapy robot tumor motion estimation and prediction system according to claim 7, characterized in that: The fast tracking model is expressed as: ； Where, Indicates the final positioning point of the tumor trajectory; The first Trajectory data points; The first Trajectory data points; The coordinate points representing the motion boundary line; The target threshold represents the motion boundary line; Boundary lines representing tumor target motion.

9. The radiotherapy robot tumor motion estimation and prediction system according to claim 1, characterized in that: The model robot integration unit includes an interval statistics processing module, an estimation model construction module, a control integration application module and a protocol integration control module; The interval statistics processing module, the estimation model building module, the control integration application module and the protocol integration control module are connected in sequence; The interval statistics processing module is used to divide and calculate characteristic parameters based on the relationship curve, motion trajectory and motion amplitude, and multiply the integral interval statistics by the tumor relationship motion change value of the same patient and the corresponding motion trajectory amplitude; The estimation model construction module is used to perform a clustering operation on the statistical results through a clustering algorithm to obtain an initial prediction level, and to construct an estimation prediction model with the interval distribution as input and the initial prediction level as output; The control integration application module is used to perform parameter optimization operations on the estimation and prediction model using an optimization algorithm, and integrate and apply the constructed estimation and prediction model with the controller of the radiotherapy robot; The protocol integrated control module is used to establish a control protocol for the radiotherapy robot according to the estimation prediction model, calculate the integrated control frequency, and input the intelligent integrated control quantity to complete the automatic control of the radiotherapy robot.

Images