A gynecological radiotherapy dose self-adaptive regulation system
By combining anatomical feature acquisition and geometric deformation reconstruction modules, and dynamically adjusting the multi-leaf grating and radiation source, the errors and delays caused by large-scale deformation of the pelvic anatomical structure in gynecological radiotherapy are solved, achieving real-time and precise radiotherapy control and reducing the risk of radiation-induced inflammation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JIANGSU CANCER HOSPITAL
- Filing Date
- 2026-03-30
- Publication Date
- 2026-06-26
AI Technical Summary
Existing gynecological radiotherapy techniques are unable to cope with large-scale deformations of pelvic anatomy in real time, leading to unexpected high-dose irradiation of organs and an increased risk of radiation-induced inflammation. Furthermore, existing systems are prone to computational delays and errors under complex operating conditions.
The physical characterization signal of the pelvic tissue interface is obtained by the anatomical feature acquisition module. The real-time displacement deviation of the soft tissue interface is extracted by combining the ultrasound echo signal. The result is converted into a geometric parameter matrix of a virtual control grid by the geometric deformation reconstruction module. The multi-leaf grating and radiation source are dynamically adjusted by the energy field dynamic adjustment module to achieve real-time alignment and dose rate compensation of the radiation energy deposition field.
It enables real-time response to large-scale pelvic deformation, reduces the risk of accidental irradiation of non-target areas, improves the accuracy and stability of radiotherapy, and reduces the probability of radiation-induced inflammation.
Smart Images
Figure CN122273019A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of radiotherapy technology, and particularly relates to a gynecological radiotherapy dose adaptive control system. Background Technology
[0002] Radiotherapy is a core clinical treatment for gynecological malignancies. Its basic goal is to precisely match the radiation dose field to the spatial distribution of the tumor target area by controlling the cross-sectional shape and flux intensity of the high-energy beam. In current clinical practice, a static physical plan is formulated based on the three-dimensional medical images acquired before treatment, and a multi-leaf grating is used to limit the radiation cross-section within a preset contour range. The internal anatomical structure of the gynecological pelvis has a high degree of deformability. Factors such as changes in bladder fullness, intestinal peristalsis, and pelvic gas accumulation cause random displacement of the cervix and endometrial target area.
[0003] To address anatomical deformation, the industry has long employed a static redundancy strategy that expands the planned target area boundary, adding a 5mm to 10mm physical tolerance boundary around the clinical target area. While this approach reduces the risk of off-target exposure, its inherent hidden cost is that organs at risk may be subjected to unexpected high doses of radiation, increasing the probability of radiation proctitis and radiation cystitis. Hardware physical boundary redundancy struggles to balance treatment gain and damage protection, and there are also shortcomings in control software and adaptive algorithms. For example, Chinese invention patent CN114126708B discloses a dose-guided real-time adaptive radiotherapy. The approach, which determines the patient's anatomical structure at the first moment within the treatment segment and compares it with the reference anatomical structure, adjusts subsequent dose delivery accordingly. The scheme is based on single-modal image feedback and linear dose compensation logic. However, it faces inherent limitations in gynecological pelvic conditions: intestinal gas and soft tissue exhibit characteristic overlap in X-ray transmission projection, and single energy attenuation information cannot support accurate identification of dynamic target area boundaries, causing the system to generate erroneous compensation actions; the adjustment logic does not deviate from the voxel-level physical interaction simulation framework, the data acquisition to dose recalculation cycle incurs computational costs, and the instruction generation speed cannot match the large-scale instantaneous deformation frequency of pelvic organs.
[0004] Therefore, the technical problem to be solved by this invention is how to construct a closed-loop control mechanism that can counteract large-scale deformation of the pelvic anatomical interface in real time, eliminate the lag in the inverse physical field calculation, and have the ability to identify gas-liquid boundaries, so as to realize the millisecond-level follow-up adjustment of the multi-leaf grating to the dynamic target area. Summary of the Invention
[0005] In this technical solution, a gynecological radiotherapy dose adaptive control system includes: The anatomical feature acquisition module is used to acquire physical characterization signals of the pelvic tissue interface; The geometric deformation reconstruction module is equipped with a virtual control grid that characterizes the spatial distribution of the initial radiotherapy plan dose field. The geometric deformation reconstruction module is connected to the anatomical feature acquisition module. The energy field dynamic adjustment module is connected to the geometric deformation reconstruction module; the adaptive control logic of the system is as follows: the anatomical feature acquisition module extracts the real-time displacement deviation of the soft tissue interface in the pelvic cavity in three-dimensional phase space based on the energy flow attenuation data generated by high-energy ray transmission projection and real-time ultrasound echo signal, and generates an anatomical deformation vector with vector attributes. The geometric deformation reconstruction module uses the anatomical deformation vector as a boundary constraint and transforms it into the spatial displacement of the topology control nodes in the virtual control mesh through a mapping mechanism. It also uses the Laplacian smoothing algorithm, which maintains the consistency of the topology manifold, to transform the static geometric envelope model corresponding to the initial radiotherapy plan into a geometric parameter matrix that characterizes the dynamic target distortion features. This transforms the inverse solution of the physical dose field into a geometric topology update of the parameterized mesh surface. The dynamic adjustment module of the energy field receives the geometric parameter matrix and analyzes the target area centroid offset trajectory and cross-sectional contour conformal deviation. It collaboratively outputs a joint drive command composed of the transient dose rate coupling of the multi-leaf grating position sequence and the radiation source pulse sequence. During the radiation irradiation process, it dynamically adjusts the radiation energy deposition field. By changing the radiation field geometry and beam flux intensity, it achieves real-time alignment of the three-dimensional geometric cross-section of the radiation energy deposition field with the physical boundary of the wandering target area in both spatial and temporal dimensions.
[0006] Preferably, the anatomical feature acquisition module follows these rules during operation: it utilizes the reflection characteristics of ultrasound at interfaces with different acoustic impedances to identify the soft tissue boundary between the anterior rectal wall and the posterior bladder wall; it uses the soft tissue boundary as a geometric constraint of the anatomical structure to correct the physical boundary of the X-ray transmission projection when identifying the area where intestinal gas and soft tissue overlap, thus eliminating the deformation calculation deviation caused by changes in the filling state of the pelvic organs.
[0007] Preferably, during the reconstruction process, the geometric deformation reconstruction module treats the spatial envelope of the target area and organs at risk in the initial radiotherapy plan as a geometric mesh model composed of several topological nodes. After receiving the anatomical deformation vector, it updates the surface shape of the geometric mesh model by adjusting the coordinates of the mesh nodes, thus transforming the complex physical field dose recalculation process into a geometric node update process of the parameterized mesh, thereby shortening the dose response delay.
[0008] Preferably, during the adjustment process, the energy field dynamic adjustment module performs phase synchronization registration between the motion trajectory of the multi-leaf grating and the dose rate of the radiation source output pulse; based on the preset radiation dose limit logic for the rectum and bladder, combined with the deformation trend fed back by the geometric deformation reconstruction module, it adjusts the subfield radiation flux pointing to the organ at risk, and limits the radiation redundancy boundary within the physical displacement tolerance range corresponding to a 100ms system response delay.
[0009] Preferably, the anatomical feature acquisition module is further used to: acquire residual radiation attenuation data after passing through the target area through X-ray detection data; invert the electron density drift value inside the target area based on the residual radiation attenuation data; and perform online calibration of the three-dimensional spatial components of the anatomical deformation vector by combining the anatomical contour information provided by the ultrasonic echo signal, thereby improving the mapping accuracy between the radiation energy deposition field and the true physical essence of the dynamic target area.
[0010] Preferably, the geometric deformation reconstruction module is further used to: extract the offset characteristic parameters of the target area geometric center in the spatial coordinate system in real time from the geometric parameter matrix; the energy field dynamic adjustment module translates the field center in real time according to the offset characteristic parameters, and automatically blocks the beam output signal when the displacement represented by the offset characteristic parameters exceeds 5mm, so as to prevent mis-irradiation of non-target areas caused by exceeding the mechanical tracking limit.
[0011] Preferably, during operation, the geometric deformation reconstruction module maintains the adjacency relationship between virtual control grid nodes through a homeomorphic mapping mechanism; monitors the relative displacement gradient of adjacent nodes in the virtual control grid, identifies the intensity of nonlinear distortion appearing locally in the target area, and uses this distortion intensity as a weight parameter for smooth correction of the field edge parameters in the joint drive command, so as to ensure the continuity of the physical distribution of radiation dose at the edge of the deformation section.
[0012] Preferably, the energy field dynamic adjustment module is also used to: monitor the dose feedback pulse signal in real time; when the deviation between the dose feedback pulse signal and the target dose in the joint drive command is within the range of 3% to 5%, adjust the leaf stepping speed of the multi-leaf grating to compensate for the influence of dose rate fluctuation on the uniformity of energy deposition in the target area, and ensure that the physical dose distribution of radiation dose in the dynamic target area meets the preset dosimetric target.
[0013] Preferably, the anatomical feature acquisition module continuously stores organ filling records and anatomical displacement data across radiotherapy fractions; the energy field dynamic adjustment module extracts the periodic deviation patterns in the anatomical displacement data and automatically corrects the starting position coordinates of the initial geometric grid between adjacent treatment fractions, thereby achieving predictive dose optimization based on anatomical evolution trends and reducing the transient adaptive load of the system in a single treatment.
[0014] Compared with existing technologies, the gynecological radiotherapy dose adaptive control system of the present invention has the following advantages: 1. In adaptive dose control for gynecological radiotherapy, the system instantiates the beam configuration parameters into an initial three-dimensional geometric topology mesh in a computer-aided design environment, and converts the soft tissue interface deformation vector extracted from the ultrasound radio frequency time-series echo matrix into the traction force of the mesh control nodes. This enables a shift from the complex full reconstruction of the physical dose field to the control logic of pure geometric mesh homeomorphic deformation. This cross-modal information geometric parameterization processing mode, while preserving the key topological features of the anatomical structure, eliminates the nonlinear iterative deadlock and computational delay caused by conventional voxel-level integral operations. It reduces the dynamic follow-up response time for high-frequency, large-scale deformable target areas to the physical limit of system communication, providing a reliable real-time control basis for precise radiotherapy with ultra-narrow boundaries.
[0015] 2. The integrated coordinate system manifold relaxation mechanism can automatically suspend absolute coordinate system registration and switch to relative topology deduction based on the high-frequency gradient of the characteristic pixels of transmitted radiation flux when the signal-to-noise ratio of ultrasound signals deteriorates due to accelerator rotation vibration or radio frequency noise. This control redundancy design based on manifold relaxation overcomes the technical defects of single sensor source being prone to lock-up under complex electromagnetic and mechanical conditions, ensures the physical continuity of dose compensation during treatment, avoids unplanned beam interruption caused by transient signal interference, and improves the system's operational stability in real complex diagnostic and treatment environments.
[0016] 3. By deeply coupling the transmission radiation flux matrix with the ultrasound radio frequency time-series echo matrix, the system utilizes the characteristic of ultrasound being sensitive to differences in acoustic impedance to compensate for the physical identification blind spot when high-energy X-ray transmission projection identifies the overlapping area of intestinal gas and soft tissue. This boundary constraint mechanism based on heterogeneous information complementarity introduces the real soft tissue interface deformation vector into the reverse calculation process of three-dimensional energy deposition, thereby eliminating the calculation deviation caused by the instantaneous changes in gas and fluid distribution inside the pelvic cavity, and ensuring the accuracy of the multi-leaf grating execution array in restoring the real geometric cross section of the dynamic target area. Attached Figure Description
[0017] Figure 1 This is a flowchart of the multi-module collaborative operation of the radiotherapy dose adaptive control system of the present invention; Figure 2 This is the logic diagram for offline calibration calculation of mechanical interference stripping and anatomical deformation vector of the probe of the present invention. Detailed Implementation
[0018] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.
[0019] It should be noted that all directional and positional terms used in this invention, such as: up, down, left, right, front, back, vertical, horizontal, inner, outer, top, bottom, transverse, longitudinal, center, etc., are only used to explain the relative positional relationships and connection situations between components in a specific state (as shown in the accompanying drawings), and are only for the convenience of describing this invention, not to require that this invention be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention. Furthermore, the descriptions of "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated.
[0020] In the description of this invention, unless otherwise explicitly specified and limited, the terms installation, connection, and linking should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.
[0021] In the description of this specification, references to the terms "an embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0022] A gynecological radiotherapy dose adaptive control system includes: The anatomical feature acquisition module is used to acquire physical characterization signals of the pelvic tissue interface; The geometric deformation reconstruction module is equipped with a virtual control grid that characterizes the spatial distribution of the initial radiotherapy plan dose field. The geometric deformation reconstruction module is connected to the anatomical feature acquisition module. The energy field dynamic adjustment module is connected to the geometric deformation reconstruction module; the adaptive control logic of the system is as follows: the anatomical feature acquisition module extracts the real-time displacement deviation of the soft tissue interface in the pelvic cavity in three-dimensional phase space based on the energy flow attenuation data generated by high-energy ray transmission projection and real-time ultrasound echo signal, and generates an anatomical deformation vector with vector attributes. The geometric deformation reconstruction module uses the anatomical deformation vector as a boundary constraint and transforms it into the spatial displacement of the topology control nodes in the virtual control mesh through a mapping mechanism. It also uses the Laplacian smoothing algorithm, which maintains the consistency of the topology manifold, to transform the static geometric envelope model corresponding to the initial radiotherapy plan into a geometric parameter matrix that characterizes the dynamic target distortion features. This transforms the inverse solution of the physical dose field into a geometric topology update of the parameterized mesh surface. The dynamic adjustment module of the energy field receives the geometric parameter matrix and analyzes the target area centroid offset trajectory and cross-sectional contour conformal deviation. It collaboratively outputs a joint drive command composed of the transient dose rate coupling of the multi-leaf grating position sequence and the radiation source pulse sequence. During the radiation irradiation process, it dynamically adjusts the radiation energy deposition field. By changing the radiation field geometry and beam flux intensity, it achieves real-time alignment of the three-dimensional geometric cross-section of the radiation energy deposition field with the physical boundary of the wandering target area in both spatial and temporal dimensions.
[0023] Preferably, the anatomical feature acquisition module follows these rules during operation: it utilizes the reflection characteristics of ultrasound at interfaces with different acoustic impedances to identify the soft tissue boundary between the anterior rectal wall and the posterior bladder wall; it uses the soft tissue boundary as a geometric constraint of the anatomical structure to correct the physical boundary of the X-ray transmission projection when identifying the area where intestinal gas and soft tissue overlap, thus eliminating the deformation calculation deviation caused by changes in the filling state of the pelvic organs.
[0024] Preferably, during the reconstruction process, the geometric deformation reconstruction module treats the spatial envelope of the target area and organs at risk in the initial radiotherapy plan as a geometric mesh model composed of several topological nodes. After receiving the anatomical deformation vector, it updates the surface shape of the geometric mesh model by adjusting the coordinates of the mesh nodes, thus transforming the complex physical field dose recalculation process into a geometric node update process of the parameterized mesh, thereby shortening the dose response delay.
[0025] Preferably, during the adjustment process, the energy field dynamic adjustment module performs phase synchronization registration between the motion trajectory of the multi-leaf grating and the dose rate of the radiation source output pulse; based on the preset radiation dose limit logic for the rectum and bladder, combined with the deformation trend fed back by the geometric deformation reconstruction module, it adjusts the subfield radiation flux pointing to the organ at risk, and limits the radiation redundancy boundary within the physical displacement tolerance range corresponding to a 100ms system response delay.
[0026] Preferably, the anatomical feature acquisition module is further used to: acquire residual radiation attenuation data after passing through the target area through X-ray detection data; invert the electron density drift value inside the target area based on the residual radiation attenuation data; and perform online calibration of the three-dimensional spatial components of the anatomical deformation vector by combining the anatomical contour information provided by the ultrasonic echo signal, thereby improving the mapping accuracy between the radiation energy deposition field and the true physical essence of the dynamic target area.
[0027] Preferably, the geometric deformation reconstruction module is further used to: extract the offset characteristic parameters of the target area geometric center in the spatial coordinate system in real time from the geometric parameter matrix; the energy field dynamic adjustment module translates the field center in real time according to the offset characteristic parameters, and automatically blocks the beam output signal when the displacement represented by the offset characteristic parameters exceeds 5mm, so as to prevent mis-irradiation of non-target areas caused by exceeding the mechanical tracking limit.
[0028] Preferably, during operation, the geometric deformation reconstruction module maintains the adjacency relationship between virtual control grid nodes through a homeomorphic mapping mechanism; monitors the relative displacement gradient of adjacent nodes in the virtual control grid, identifies the intensity of nonlinear distortion appearing locally in the target area, and uses this distortion intensity as a weight parameter for smooth correction of the field edge parameters in the joint drive command, so as to ensure the continuity of the physical distribution of radiation dose at the edge of the deformation section.
[0029] Preferably, the geometric deformation reconstruction module determines the displacement constraint force of each node in the virtual control mesh according to the following logic. : ,in, For displacement constraint force, The preset mesh topology stiffness coefficients, The soft tissue's current spatial coordinates are fed back by the anatomical feature acquisition module. This refers to the baseline coordinate data for the corresponding node in the initial radiotherapy plan.
[0030] Preferably, the energy field dynamic adjustment module is also used to: monitor the dose feedback pulse signal in real time; when the deviation between the dose feedback pulse signal and the target dose in the joint drive command is within the range of 3% to 5%, adjust the leaf stepping speed of the multi-leaf grating to compensate for the influence of dose rate fluctuation on the uniformity of energy deposition in the target area, and ensure that the physical dose distribution of radiation dose in the dynamic target area meets the preset dosimetric target.
[0031] Preferably, the anatomical feature acquisition module continuously stores organ filling records and anatomical displacement data across radiotherapy fractions; the energy field dynamic adjustment module extracts the periodic deviation patterns in the anatomical displacement data and automatically corrects the starting position coordinates of the initial geometric grid between adjacent treatment fractions, thereby achieving predictive dose optimization based on anatomical evolution trends and reducing the transient adaptive load of the system in a single treatment.
[0032] Example 1: When the system faces the condition of irregular high-frequency filling of the rectum accompanied by changes in gas-fluid distribution during cervical tumor radiotherapy, the physical identification aliasing state caused by the transmission projection of a single high-energy ray is induced. The anatomical feature acquisition module integrates the ray transmission data and the ultrasound echo signal. Based on the reflection parameters of ultrasound at the acoustic impedance interface between the anterior rectal wall and the posterior bladder wall, it extracts the displacement bias in three-dimensional phase space from the soft tissue interface and generates an anatomical deformation vector containing spatial direction and displacement magnitude. The anatomical deformation vector provides the geometric deformation reconstruction module with input data with definite spatial coordinate constraints, constraining the spatial tolerance limit of the output coordinate data of the single ray detector.
[0033] Faced with the computational overload caused by the recalculation of the full three-dimensional voxel physical dose integration in online adaptive control, the geometric deformation reconstruction module receives the anatomical deformation vector and converts it into the spatial displacement of the topology control nodes in the mesh. It then calls the Laplace smoothing algorithm to adjust the three-dimensional coordinates of each mesh node according to the formula... Calculate the spatial displacement constraint force at each node in the virtual control mesh, where F is the spatial displacement constraint force on the mesh node, and k is the mesh topological stiffness coefficient. The feedback is the current spatial coordinate data of the soft tissue. The reference coordinate data of the corresponding control node in the initial radiotherapy plan is used. The algorithm engine maps the spatial displacement constraint force into the geometric topology update amount of the parameterized mesh surface to generate a geometric parameter matrix that characterizes the dynamic target distortion features. This mapping logic reconstructs the nonlinear dose inverse deduction into the translation deformation of the coordinates of the three-dimensional virtual control mesh node, thereby compressing the computational processing latency of the underlying hardware architecture.
[0034] The physical mapping steps for the energy field dynamic adjustment module to receive the geometric parameter matrix and output joint drive commands are as follows: The controller extracts the spatial rotation angle parameters of the current accelerator gantry and establishes a two-dimensional projection plane of the beam angle perpendicular to the beam center axis. Then, the three-dimensional topology control nodes in the geometric parameter matrix are orthogonally projected along the ray divergence trajectory onto the two-dimensional projection plane of the beam angle to generate a two-dimensional distortion contour line of the target area. The processor retrieves the motion axis equations of each physical blade of the multi-leaf grating pre-stored in the underlying memory, calculates the geometric intersection coordinates of the two-dimensional distortion contour line of the target area and each motion axis equation, and sets these geometric intersection coordinates as the target aperture physical coordinates of the corresponding blade. The drive speed of a single blade is based on the formula v=(P target -P current ) / D eltat Calculate, where v is the driving speed of a single blade, P target For the target aperture physical coordinates, P current D is the current physical coordinate of the blade, fed back in real time by the servo encoder. eltat The inherent system response delay is set for this embodiment.
[0035] The closed-loop calculation logic of the transient dose rate parameter in the joint drive instruction is as follows: The processor calculates the two-dimensional geometric transmission area of the current radiation field based on the physical coordinate integration of the target opening of each blade, extracts the baseline radiation field area and baseline dose rate at the corresponding gantry angle in the initial radiotherapy plan, and the controller calculates the dose rate according to the formula R. act =R ref ×(S act / S ref ), calculate and output the transient dose rate signal to the radiation source pulse generator, where R act R is the output transient dose rate. ref S is the baseline dose rate. act S represents the two-dimensional geometric transmission area of the current firing field. ref Using the baseline field area, the above quantitative formula establishes a linear mathematical mapping between the change in the physical transmission window caused by the deformation of the target area cross section and the accelerator beam output intensity, so that the multi-leaf grating mechanical kinematics scheduling and the underlying beam pulse output form a deterministic action combination at the physical execution level.
[0036] The energy field dynamic adjustment module receives the geometric parameter matrix and resolves the offset trajectory of the target area's centroid in the spatial coordinate system and the conformal deviation of the cross-sectional profile. Based on the deviation parameters, it outputs a joint drive command composed of the transient dose rate coupling of the multi-leaf grating position sequence and the radiation source pulse sequence. During the continuous operation cycle of the system, this module monitors the dose feedback pulse signal output by the accelerator. When the quantization deviation ratio between the calculated actual pulse and the target dose value in the joint drive command reaches 4%, the controller adjusts the leaf stepping speed of the corresponding sector multi-leaf grating in reverse according to the quantization deviation ratio. At the same time, when the physical displacement represented by the offset characteristic parameter of the target area's geometric center reaches 5mm, the underlying hardware interface cuts off the beam output signal. The synergistic effect of heterogeneous physical signal fusion and manifold dimensionality reduction operation transforms the pelvic organ filling disturbance into the underlying geometric grid translation parameter, maintaining the alignment state of the radiation energy deposition field with the physical boundary of the dynamic target area within the 100ms system response delay tolerance range.
[0037] Example 2: In the industrial testing environment of pelvic radiotherapy, a pelvic physical phantom with dynamic deformation properties is configured. An adjustable-speed fluid injection pump is installed inside the phantom to simulate the fullness of the rectum and bladder. A Gaussian white noise source with a signal-to-noise ratio of 20dB is superimposed in the ultrasound probe echo receiving link to introduce electromagnetic interference. The fluid injection pump injects liquid into the physical phantom at a flow rate of 15mL / min, inducing a gradient physical displacement deviation in the target area from 1.2mm to 5.8mm. Under this condition of background noise and dynamic physical deformation, the anatomical feature acquisition module extracts the noisy ultrasound echo signal and radiation attenuation data of the X-ray detection array, outputs the soft tissue interface displacement, and generates an anatomical deformation vector. A spatial coordinate matrix is established to provide input data for the energy field modulation process. The geometric deformation reconstruction module receives the anatomical deformation vector containing noise disturbance and gradient displacement parameters. A homeomorphic mapping mechanism is used to convert the anatomical deformation vector into the spatial displacement of the topological control nodes in the three-dimensional virtual control mesh. The Laplace smoothing algorithm is called to adjust the three-dimensional coordinates of each mesh node according to the formula... Calculate the spatial displacement constraint force at each node, where F is the spatial displacement constraint force on the mesh node, and k is the mesh topological stiffness coefficient. The feedback is the current spatial coordinate data of the soft tissue. The reference coordinate data for the corresponding control nodes in the initial radiotherapy plan is used; the algorithm engine converts the spatial displacement constraint force into the geometric topology update quantity of the parameterized mesh surface, outputs the geometric parameter matrix representing the physical boundary of the target area, the energy field dynamic adjustment module extracts the conformal deviation data in the geometric parameter matrix, calculates the quantization deviation ratio between the actual pulse and the target dose value by combining the dose feedback pulse signal monitored by the internal bus, and outputs a joint drive command containing the multi-leaf grating position sequence to adjust the beam energy field.
[0038] The conformity index is defined as the ratio of the volume of the prescription dose package to the actual physical volume of the target area. The testing procedure verifies the system performance using objective operating parameters of the sample group under different displacement gradients. When the physical displacement deviation of the target area is at a normal operating condition of 2.1 mm, the spatial calculation delay of the system output is 41.5 ms and the conformity index reaches 98.2%, with the multi-leaf grating maintaining a low-frequency stable adjustment state of 0.8 Hz. As the filling displacement increases to a high-pressure condition of 4.3 mm, the system uses ultrasonic boundary constraints to identify aliasing of the transmitted projection of the ray, and the output spatial calculation delay fluctuates slightly to 42.8 ms, at which point the conformity index remains at 96. 5%, the mechanical oscillation frequency of the multi-leaf grating is within the controlled range of 1.1Hz; as the filling displacement induced by the fluid injection pump continues to increase and reaches the limit condition, the system monitors and obtains the physical displacement of the geometric center of the target area as 5.6mm, which exceeds the preset displacement limit of 5mm. The underlying hardware interface of the sample group of this invention cuts off the beam output signal, and the radiation flux of the target area and the surrounding area falls back to the background level within 15.3ms. The above-mentioned kinematic scheduling of the multi-leaf grating and the synchronous registration logic of the transient dose rate of the beam pulse are combined with the quantization deviation ratio adjustment threshold set in the range of 3% to 5%, and output the alignment state of the three-dimensional geometric section of the target area and the physical boundary of the energy field.
[0039] Example 3: When the anatomical feature acquisition module fuses heterogeneous physical signals to extract anatomical deformation vectors, the system faces the technical condition of initial mechanical distortion of the pelvic tissue interface caused by ultrasound probe contact stress. The system controller is configured with an offline compensation procedure for probe contact stress. Before the radiotherapy plan is implemented, the operator applies the ultrasound probe to the surface of the standard pelvic physical model with a gradient contact stress of 2N to 10N. The force sensor simultaneously acquires the normal stress data applied by the probe, and the ultrasound echo receiving link extracts the depth offset data of the reference target point within the standard pelvic physical model. The processor receives the normal stress data and the depth offset data, and calculates the mechanical compressibility coefficient α under a specific normal stress. The specific calculation formula is as follows: Where α is the mechanical compressibility coefficient, and Δd is the depth offset data of the reference target point. This is the normal stress data.
[0040] The processor establishes a nonlinear mapping function matrix between normal stress data and mechanical compressibility coefficient α and writes it into the underlying memory. During the extraction of clinical anatomical deformation vectors, the anatomical feature acquisition module retrieves the current normal stress output by the force sensor in real time, matches the corresponding mechanical compressibility coefficient α by looking up a table, and calculates the stress-induced deviation by multiplying the current normal stress with the mechanical compressibility coefficient α through a multiplier. Then, the stress-induced deviation is subtracted from the three-dimensional spatial coordinates extracted by the ultrasound echo signal to output the anatomical deformation vector stripped of the probe's mechanical interference.
[0041] Example 4: When the system faces the heterogeneous physical signal fusion and solution situation during pelvic radiotherapy, the controller establishes a heterogeneous data underlying mapping mechanism. The system controller defines the same three-dimensional Cartesian reference coordinate system and synchronously receives ultrasound echo point cloud data with spatial location labels and X-ray transmission projection matrix containing radiation attenuation intensity through the communication bus. The algorithm engine calculates the spatial gradient matrix of the ultrasound echo point cloud data to extract the soft tissue interface contour with acoustic impedance differences. At the same time, based on the X-ray transmission projection matrix, the back projection reconstruction algorithm is applied to obtain the three-dimensional voxel distribution sequence reflecting the electron density inside the pelvis. The processor registers the soft tissue interface contour and the three-dimensional voxel distribution sequence in the three-dimensional Cartesian reference coordinate system, calculates the inverse variance of the two sets of data at each coordinate node as the data confidence weight, and uses the data confidence weight to weight and sum the two sets of source data to output the anatomical deformation vector pointing to the physical boundary of the dynamic target area.
[0042] The energy field dynamic adjustment module is configured with a physical displacement limit calibration loop. Before the radiotherapy plan is executed, the system sends a step test command to the multi-leaf grating controller. The average time difference between the multi-leaf grating blades receiving the command and reaching the designated physical blocking position is recorded by the servo encoder as the system's inherent response delay. The maximum migration velocity vector of the tumor tissue in the target area within a specific imaging cycle is extracted. The processor determines the dynamic distortion compensation baseline value by multiplying the system's inherent response delay by the maximum migration velocity vector. The target area external geometric tolerance set in the prescription is extracted. The difference between the target area external geometric tolerance and the dynamic distortion compensation baseline value is established as the beam cutoff safety limit value. The system's inherent response delay extracted in the physical parameter calibration verification is 100ms and the maximum migration velocity vector is 10mm / s. Based on this, the dynamic distortion compensation baseline value is calculated to be 1mm. Combined with the 6mm target area external geometric tolerance set in the prescription, the difference between the two determines the physical displacement limit value to be 5mm.
[0043] Example 5: When the system faces the situation of lacking a spatial reference system for heterogeneous sensor hardware during the initial assembly stage, a spatial alignment calibration reference for heterogeneous physical signals is established. By configuring a dual-modal physical phantom with embedded acoustic reflective microbeads and high-density tungsten alloy target points on the treatment bed, the X-ray detection array and the ultrasound probe synchronously acquire static image data of the dual-modal physical phantom. The processor extracts the three-dimensional coordinate set of the marked points in the two sets of image data and uses the singular value decomposition algorithm to calculate the rigid transformation matrix from the ultrasound coordinate system to the X-ray coordinate system. The geometric deformation reconstruction module writes the rigid transformation matrix into the underlying register, providing a three-dimensional Cartesian reference coordinate system for the registration operation of the anatomical deformation vector and the three-dimensional voxel distribution sequence.
[0044] When individual differences in the mechanical properties of the pelvic soft tissue cause deviations in the topological deformation calculation from the physical displacement trajectory, the anatomical feature acquisition module calibrates the mesh topological stiffness parameters. During the radiotherapy planning phase, it controls the ultrasound probe to emit low-frequency shear waves towards the pelvic target area and acquires the propagation velocity signal of the shear waves within the soft tissue. The processor calculates the Young's modulus value of the adjacent region within the pelvic target area based on the propagation velocity signal. The algorithm engine converts the Young's modulus value into a mesh topological stiffness coefficient k for the Laplace smoothing algorithm based on the initial mesh side length and cross-sectional area parameters. The geometric deformation reconstruction module extracts the mesh topological stiffness coefficient k and injects it into the aforementioned formula. The computational logic is used to establish a three-dimensional virtual control mesh that matches the local mechanical boundary conditions of the target area.
[0045] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.
Claims
1. A gynecological radiotherapy dose adaptive control system, characterized in that, include: The anatomical feature acquisition module is used to acquire physical characterization signals of the pelvic tissue interface; The geometric deformation reconstruction module is equipped with a virtual control grid that characterizes the spatial distribution of the initial radiotherapy plan dose field. The geometric deformation reconstruction module is connected to the anatomical feature acquisition module. The energy field dynamic adjustment module is connected to the geometric deformation reconstruction module; the adaptive control logic of the system is as follows: the anatomical feature acquisition module extracts the real-time displacement deviation of the soft tissue interface in the pelvic cavity in three-dimensional phase space based on the energy flow attenuation data generated by high-energy ray transmission projection and real-time ultrasound echo signal, and generates an anatomical deformation vector with vector attributes. The geometric deformation reconstruction module uses the anatomical deformation vector as a boundary constraint and transforms it into the spatial displacement of the topology control nodes in the virtual control mesh through a mapping mechanism. It also uses the Laplacian smoothing algorithm, which maintains the consistency of the topology manifold, to transform the static geometric envelope model corresponding to the initial radiotherapy plan into a geometric parameter matrix that characterizes the dynamic target distortion features. This transforms the inverse solution of the physical dose field into a geometric topology update of the parameterized mesh surface. The dynamic adjustment module of the energy field receives the geometric parameter matrix and analyzes the target area centroid offset trajectory and cross-sectional contour conformal deviation. It collaboratively outputs a joint drive command composed of the transient dose rate coupling of the multi-leaf grating position sequence and the radiation source pulse sequence. During the radiation irradiation process, it dynamically adjusts the radiation energy deposition field. By changing the radiation field geometry and beam flux intensity, it achieves real-time alignment of the three-dimensional geometric cross-section of the radiation energy deposition field with the physical boundary of the wandering target area in both spatial and temporal dimensions.
2. The gynecological radiotherapy dose adaptive control system according to claim 1, characterized in that, The anatomical feature acquisition module follows these rules during operation: it utilizes the reflection characteristics of ultrasound at interfaces with different acoustic impedances to identify the soft tissue boundary between the anterior rectal wall and the posterior bladder wall; and it uses the soft tissue boundary as a geometric constraint of the anatomical structure to correct the physical boundary of X-ray transmission projection when identifying the area where intestinal gas and soft tissue overlap.
3. The gynecological radiotherapy dose adaptive control system according to claim 1, characterized in that, During the reconstruction process, the geometric deformation reconstruction module treats the spatial envelope of the target area and organs at risk in the initial radiotherapy plan as a geometric mesh model composed of several topological nodes. After receiving the anatomical deformation vector, it updates the surface shape of the geometric mesh model by adjusting the coordinates of the mesh nodes, thus transforming the complex physical field dose recalculation process into a geometric node update process of the parameterized mesh.
4. The gynecological radiotherapy dose adaptive control system according to claim 1, characterized in that, During the adjustment process, the energy field dynamic adjustment module performs phase synchronization registration between the motion trajectory of the multi-leaf grating and the dose rate of the radiation source output pulse; Based on the preset radiation dose limit logic for the rectum and bladder, and combined with the deformation trend fed back by the geometric deformation reconstruction module, the radiation flux of the subfield pointing to the organ at risk is adjusted, and the radiation redundancy boundary is limited to the physical displacement tolerance range corresponding to a system response delay of 100ms.
5. The gynecological radiotherapy dose adaptive control system according to claim 1, characterized in that, The anatomical feature acquisition module is further used to: acquire residual radiation attenuation data after passing through the target area through X-ray detection data; invert the electron density drift value inside the target area based on the residual radiation attenuation data; and perform online calibration of the three-dimensional spatial components of the anatomical deformation vector by combining the anatomical contour information provided by the ultrasonic echo signal, thereby improving the mapping accuracy between the radiation energy deposition field and the true physical nature of the dynamic target area.
6. The gynecological radiotherapy dose adaptive control system according to claim 1, characterized in that, The geometric deformation reconstruction module is further used to: extract the offset characteristic parameters of the target area geometric center in the spatial coordinate system in real time from the geometric parameter matrix; the energy field dynamic adjustment module translates the field center in real time according to the offset characteristic parameters, and automatically blocks the beam output signal when the displacement represented by the offset characteristic parameters exceeds 5mm.
7. The gynecological radiotherapy dose adaptive control system according to claim 1, characterized in that, During operation, the geometric deformation reconstruction module maintains the adjacency relationship between virtual control grid nodes through a homeomorphic mapping mechanism; monitors the relative displacement gradient of adjacent nodes in the virtual control grid, identifies the intensity of nonlinear distortion appearing locally in the target area, and uses this distortion intensity as a weight parameter for smooth correction of the field edge parameters in the joint drive command, so as to ensure the continuity of the physical distribution of radiation dose at the edge of the deformation section.
8. The gynecological radiotherapy dose adaptive control system according to claim 1, characterized in that, The energy field dynamic adjustment module is also used to: monitor the dose feedback pulse signal in real time; when the deviation between the dose feedback pulse signal and the target dose in the joint drive command is within the range of 3% to 5%, adjust the leaf stepping speed of the multi-leaf grating to compensate for the influence of dose rate fluctuation on the uniformity of energy deposition in the target area, and ensure that the physical dose distribution of radiation dose in the dynamic target area meets the preset dosimetric target.
9. The gynecological radiotherapy dose adaptive control system according to claim 1, characterized in that, The anatomical feature acquisition module continuously stores organ filling records and anatomical displacement data across radiotherapy fractions; The energy field dynamic adjustment module extracts the periodic deviation patterns in the anatomical displacement data and automatically corrects the starting position coordinates of the initial geometric grid between adjacent treatment sessions. This enables predictive dose optimization based on anatomical evolution trends, reducing the transient adaptive load of the system in a single treatment session.