Target region real-time tracking method, electronic equipment, system and storage medium

Abstract

The embodiment of the invention relates to the field of radiotherapy equipment, and discloses a target region real-time tracking method, electronic equipment, a system and a storage medium, and the method comprises the steps: tracking a target region in real time through controlling a main mechanical arm and an auxiliary mechanical arm to move synchronously by using a current prediction model and breathing phase change, the prediction model is used for representing a corresponding relation between the breathing phase and target region motion information, and the target region motion information comprises six-degree-of-freedom information; obtaining a single projection image according to a preset frequency; obtaining a monitoring value of the motion information of the target area by using a single projection image; matching the monitoring value with a predicted value of the same breathing time phase target region motion information in the prediction model; and under the condition of inconsistent matching, the prediction model is re-established. According to the method, the problem of improving the accuracy of real-time target region tracking is solved, and the prediction model is dynamically corrected according to the monitoring information in the treatment stage of real-time target region tracking, so that the target region tracking accuracy is improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of radiotherapy equipment, and in particular to a real-time target area tracking method, electronic equipment, system and storage medium. Background Art

[0002] During radiotherapy of tumors, respiratory movement can cause the tumor to move, so it is necessary to obtain the accurate tumor target location before emitting the treatment beam to achieve precise radiotherapy. Current solutions include: (1) controlling respiratory movement, such as emitting the beam during deep exhalation or deep inspiration, for example, using an abdominal compression plate to reduce the range of movement. However, this method requires human intervention in respiratory movement and cannot guarantee the control effect of respiratory movement; (2) observing the range of movement of the target area through imaging, for example, using 4DCT (Four-dimensional computed tomography) for positioning and 4DCBCT (Four-Dimensional Cone Beam Computed Tomography) for treatment, and then determining the range of movement of the target area and emitting the treatment beam within this range. However, the target area location obtained by this method has errors; (3) respiratory gating technology, in which the patient can breathe freely and wears a respiratory detection device to monitor the respiratory phase, and the treatment beam is only emitted at a specific respiratory phase in the respiratory cycle. However, in this method, the relationship between the respiratory phase and the target area's motion is changing, that is, the target area's motion position may be inconsistent under the same respiratory phase. Therefore, the target area's position is incorrect. (4) Real-time tracking technology: The treatment beam moves with the target area. This method requires predicting the target area's motion position in advance. However, the current real-time tracking technology requires two fixed-position tubes to emit cross-X-rays and determine the target area's position by projecting two orthogonal images at fixed positions. The shooting angle is limited and the calculation is cumbersome. The target area position tracked in real time is not accurate enough. Therefore, the real-time tracking effect is difficult to meet the high-precision requirements of radiotherapy guidance. Summary of the Invention

[0003] The purpose of the present invention is to at least provide a real-time target area tracking method, electronic equipment, system and storage medium, which can at least solve the problem of improving the accuracy of real-time target area tracking, and at least achieve the effect of dynamically correcting the prediction model according to the monitoring information during the treatment stage of real-time target area tracking to improve the accuracy of target area tracking.

[0004] In order to solve the above technical problems, at least one embodiment of the present application provides a real-time target tracking method, which is implemented based on an image-guided radiotherapy system. The image-guided radiotherapy system includes a main robot arm and an auxiliary robot arm. The end of the main robot arm is equipped with a transmitting device for transmitting a cone imaging beam and a treatment device for transmitting a treatment beam, and the end of the auxiliary robot arm is equipped with a receiving device for receiving the cone imaging beam. While the main robot arm drives the treatment device and the transmitting device to move, the auxiliary robot arm drives the receiving device to move synchronously so that the cone imaging beam can pass through the target area to reach the receiving device and the receiving device avoids the irradiation of the treatment beam. The real-time target tracking method includes: utilizing the current prediction model and respiratory phase changes to control the main robot arm and the auxiliary robot arm to move the target area. The target area is tracked in real time by synchronous motion, and the prediction model is used to characterize the correspondence between the respiratory phase and the target area motion information, and the target area motion information includes six-degree-of-freedom information; according to a preset frequency, imaging is performed using a transmitting device and a receiving device to obtain a single projection image; the monitoring value of the target area motion information is obtained using the single projection image; the monitoring value is matched with the predicted value of the target area motion information of the same respiratory phase in the prediction model; when the matching result shows that the monitoring value is inconsistent with the predicted value, the prediction model is re-established, and the re-established prediction model is determined as the current prediction model, so as to continue to use the current prediction model and respiratory phase changes to track the target area in real time by controlling the synchronous movement of the main robotic arm and the auxiliary robotic arm.

[0005] At least one embodiment of the present application also provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the above-mentioned real-time target area tracking method.

[0006] At least one embodiment of the present application further provides an image-guided radiotherapy system, comprising: a main robotic arm, at the end of which is mounted a transmitting device for transmitting a cone-shaped imaging beam and a treatment device for transmitting a treatment beam, the main robotic arm being capable of driving the treatment device and the transmitting device to move synchronously; an auxiliary robotic arm, at the end of which is mounted a receiving device for receiving the cone-shaped imaging beam, the auxiliary robotic arm driving the receiving device to move synchronously while the main robotic arm drives the treatment device and the transmitting device to move, so that the cone-shaped imaging beam can pass through the target area to reach the receiving device and the receiving device avoids the irradiation of the treatment beam; a control device for tracking the target in real time by controlling the synchronous movement of the main robotic arm and the auxiliary robotic arm using the current prediction model and respiratory phase changes area; according to a preset frequency, use the transmitting device and the receiving device to form an image and obtain a single projection image; use the single projection image to obtain a monitoring value of the target area motion information; match the monitoring value with the predicted value of the target area motion information of the same respiratory phase in the prediction model; when the matching result shows that the monitoring value is inconsistent with the predicted value, re-establish the prediction model, and determine the re-established prediction model as the current prediction model, so as to continue to use the current prediction model and the respiratory phase change, and track the target area in real time through the synchronous movement of the main robotic arm and the auxiliary robotic arm; the prediction model is used to characterize the correspondence between the respiratory phase and the target area motion information, and the target area motion information includes six-degree-of-freedom information.

[0007] At least one embodiment of the present application further provides a computer-readable storage medium storing a computer program, which implements the above-mentioned target area real-time tracking method when executed by a processor.

[0008] The embodiments of the present application provide a method, electronic device, system and storage medium for real-time tracking of a target area. The main robotic arm drives the treatment device to project and track the target area in real time. At the same time, it also drives the synchronous rotation of the transmitting device, and drives the receiving device to the corresponding position capable of receiving the imaging beam passing through the target area through the auxiliary robotic arm. In order to improve the accuracy of target area projection, a single projection image is obtained when the set frequency is reached. The monitoring value of the target area motion information can be obtained based on only the projection image, thereby realizing dynamic monitoring of the target area during the treatment stage, and timely correcting the prediction model according to the monitoring results, so that the real-time tracking always adopts the prediction model that can predict the precise target area, thereby improving the accuracy of real-time tracking of the target area.

[0009] In some optional embodiments, the current prediction model and respiratory phase changes are utilized to synchronously control the movement of a primary and secondary robotic arms to track the target area in real time, including: obtaining a real-time monitored respiratory motion curve representing respiratory phase changes; using the current prediction model and respiratory phase changes to predict target area motion information in real time; and controlling the primary robotic arm to move the treatment device based on the predicted target area motion information so that the treatment beam emitted by the treatment device irradiates the target area. By combining respiratory phase changes with the prediction model, dynamic tracking of the target area is achieved, ensuring target accuracy during treatment.

[0010] In some optional embodiments, the target area real-time tracking method further includes: the main robotic arm and the auxiliary robotic arm respectively drive the transmitting device and the receiving device to rotate synchronously, and imaging is performed while rotating to obtain a cone-beam CT projection image sequence; reconstructing a 4D cone-beam CT projection image based on the cone-beam CT projection image sequence; calculating the target area motion information at each respiratory phase based on the 4D cone-beam CT projection image; and establishing a prediction model for characterizing the correspondence between the respiratory phase and the target area motion information based on the target area motion information at each respiratory phase. By reconstructing the 4D cone-beam CT projection image, an accurate prediction model can be constructed for use in dynamic tracking of the target area.

[0011] In some optional embodiments, the target area real-time tracking method further includes: the main robotic arm and the auxiliary robotic arm respectively drive the transmitting device and the receiving device to rotate synchronously, and imaging is performed while rotating to obtain a cone-beam CT projection image sequence; the images of the same respiratory phase in the cone-beam CT projection image sequence are divided into a group; the target area motion information is calculated based on each group of images of the same respiratory phase, the calculated target area motion information is averaged, and the average value is determined as the target area motion information under the corresponding respiratory phase; based on the target area motion information under each respiratory phase, a prediction model is established to characterize the correspondence between the respiratory phase and the target area motion information. By calculating the target area motion information of the cone-beam CT projection images of the same respiratory phase and taking the average value, more accurate target area motion information under different respiratory phases is obtained to achieve modeling.

[0012] In some optional embodiments, the monitoring value of the target area motion information is obtained using the single projection image, including: obtaining a reference CT image for generating a treatment plan; performing 2D-3D registration based on the single projection image and the reference CT image to obtain the projection position of the single projection image; performing 2D-3D registration based on the single projection image and the reference CT image to obtain the position change of the target area relative to the reference CT image at the current moment, and determining the target area motion information in the patient coordinate system determined according to the position change of the target area relative to the reference CT image at the current moment as the monitoring value of the target area motion information.

[0013] In some optional embodiments, the position change of the target area relative to the reference CT image at the current moment includes the translation change and rotation change of the target area relative to the reference CT image in the projection coordinate system. The target area motion information in the patient coordinate system is determined by the following geometric relationship between the projection coordinate system and the patient coordinate system:

[0014]

[0015]

[0016]

[0017]

[0018]

[0019] Where, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the projection coordinate system, represents the geometric magnification factor, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the projection coordinate system, Indicates the angle of rotation of the projection plane A around the x-axis.

[0020] In some optional embodiments, obtaining a monitoring value of target area motion information using the single projection image includes: performing 2D registration with each image in the cone-beam CT projection image sequence using the single projection image, determining a second target image from the cone-beam CT projection image sequence that best matches the single projection image; determining target area motion information at the same respiratory phase in the 4D cone-beam CT projection image based on the respiratory phase of the second target image, and determining the target area motion information as the monitoring value of the target area motion information. By determining the image from the cone-beam CT projection image sequence that best matches the currently captured single projection image, the target area motion information at the same respiratory phase can be quickly determined from the 4D cone-beam CT projection image as the monitoring value. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] One or more embodiments are exemplarily described by the figures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments.

[0022] Figure 1 is a flow chart of a target area real-time tracking method provided in an embodiment of the present application; Figure 2 is a schematic diagram of a cross-section of a human body and a target area provided in an embodiment of the present application; Figure 3is a schematic diagram of a projection coordinate system provided in an embodiment of the present application; Figure 4 The patient coordinate system xyz and the projection coordinate system x provided in the embodiment of the present application are A y A z A Schematic diagram of the relationship between Figure 5 is a schematic diagram of the relationship between the three coordinate systems provided in the embodiments of the present application; Figure 6 Schematic diagram of a target area real-time tracking device provided in an embodiment of the present application; Figure 7 is a schematic diagram of an image-guided radiotherapy system provided in an embodiment of the present application; Figure 8 This is a schematic diagram of the relationship between the treatment beam and the imaging beam provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, each embodiment of the present application will be described in detail below with reference to the accompanying drawings. However, it will be understood by those skilled in the art that in each embodiment of the present application, many technical details are proposed to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical solutions claimed in the present application can be implemented. The division of the following embodiments is for convenience of description and should not constitute any limitation on the specific implementation of the present application. The various embodiments can be combined and referenced with each other under the premise of no contradiction.

[0024] In order to facilitate understanding of the embodiments of the present application, relevant content regarding real-time tracking of the target area is first introduced here.

[0025] During radiotherapy of tumors, respiratory movement can cause the movement of the tumor target, so it is necessary to obtain the accurate tumor target location before emitting the treatment beam to achieve precise radiotherapy. Current solutions include: (1) controlling respiratory movement, such as emitting the beam during deep exhalation or deep inspiration, for example, using an abdominal pressure plate to reduce the range of movement. However, this method requires human intervention in respiratory movement and cannot guarantee the control effect of respiratory movement; (2) observing the target area's range of movement through imaging, for example, using 4DCT for positioning and 4DCBCT for treatment, and then determining the target area's range of movement, and emitting the treatment beam within this range of movement, but the target area position obtained by this method has errors; (3) respiratory gating technology, in which the patient can breathe freely and wears a respiratory detection device to monitor the respiratory phase, and the treatment beam is only emitted at a specific respiratory phase in the respiratory cycle. However, in this method, the relationship between the respiratory phase and the target area's movement is variable, that is, the target area's movement position may be inconsistent under the same respiratory phase, so the target area position obtained has errors; (4) real-time tracking technology, in which the treatment beam moves with the target area. This method requires predicting the target area's movement position in advance. However, current real-time tracking technology requires two fixed-position tubes to emit crossed X-rays, and to determine the target location by projecting two orthogonal projection images at fixed positions. The shooting angle is limited and the calculation is cumbersome. The target location of real-time tracking is not accurate enough, so the real-time tracking effect is difficult to meet the high-precision requirements of radiotherapy guidance.

[0026] Currently, real-time tracking technology is a relatively accurate method. Respiratory tracking systems in related technologies predict the target's location by establishing a relationship between the internal target's motion and the external respiratory curve. This predicted target position is then transmitted to a robotic arm, enabling the treatment arm to move synchronously with the internal target. In this approach, the image-guided radiotherapy device consists of a treatment head placed on a robotic arm. Two ceiling-mounted tubes and two corresponding floor-mounted detectors form the imaging device, which is installed so that the X-rays emitted by the two tubes intersect. This device's positioning guidance image quality is poor, and CBCT imaging and the resulting 3D spatial information cannot be achieved. Furthermore, the positioning phase relies solely on dual-orthogonal 2D projection images, requiring multiple patient irradiation exposures. This is time-consuming and significantly increases the patient's radiation exposure. Furthermore, because the patient's position and the dual-orthogonal image positions are fixed, if the tumor target in the captured image is unclear or obscured, it is impossible to obtain an image containing a clear tumor target by changing the shooting angle, thereby affecting the accuracy of target positioning.

[0027] In order to solve the above-mentioned technical problem of improving the accuracy of real-time tracking of the target area, the present invention proposes a real-time target area tracking method. The implementation details of the real-time target area tracking method of this embodiment are specifically described below. The following content is only for the convenience of understanding the implementation details provided by the present invention and is not necessary for implementing this solution.

[0028] Example 1: The real-time target area tracking method of this embodiment is implemented based on an image-guided radiotherapy system, which includes a main robotic arm and an auxiliary robotic arm. The end of the main robotic arm is equipped with a transmitting device for transmitting a cone-shaped imaging beam and a treatment device for transmitting a treatment beam, and the end of the auxiliary robotic arm is equipped with a receiving device for receiving the cone-shaped imaging beam. While the main robotic arm drives the treatment device and the transmitting device to move, the auxiliary robotic arm drives the receiving device to move synchronously so that the cone-shaped imaging beam can pass through the target area and reach the receiving device, and the receiving device avoids the irradiation of the treatment beam.

[0029] In one example, the transmitting device of the image-guided radiotherapy system serves as an X-ray source, and an X-ray tube can be used to transmit a cone-shaped imaging beam. The receiving device serves as an imaging detection device capable of detecting the imaging beam, and a detector can be used. The target area real-time tracking method of this embodiment can be applied to electronic devices with communication, computing and data storage capabilities. The specific process can be as follows: Figure 1 Shown, including: Step 101 , using the current prediction model and respiratory phase changes, the target area is tracked in real time by controlling the synchronous movement of the main robotic arm and the auxiliary robotic arm, wherein the prediction model is used to characterize the correspondence between the respiratory phase and the target area motion information, and the target area motion information includes six degrees of freedom information.

[0030] Specifically, the target area motion information includes six-degree-of-freedom information of the current target area relative to the target area in the reference CT image, indicating the offset of the patient's real-time tumor position relative to the tumor position in the original reference CT image.

[0031] In a specific implementation, step 101 is performed during the treatment phase. After entering the treatment phase, the system controls the main robotic arm to drive the treatment device to aim at the target area and emit a treatment beam to perform radiotherapy on the tumor. Before entering the treatment phase, the patient positioning phase must be performed.

[0032] In some examples, during the patient positioning phase, the master robotic arm drives the transmitting device to rotate in a circular path along the cross-section of the human body around the target area, so that the cone imaging beam is directed toward the target area. The slave robotic arm synchronously drives the receiving device to rotate in a circular path along the cross-section of the human body around the target area. The synchronous rotation of the master and slave robotic arms ensures that the center of the cone imaging beam at each rotational position is perpendicular to the center of the receiving device after passing through the human body. During the synchronous rotation of the master and slave robotic arms, the transmitting device is controlled to transmit the imaging beam and the receiving device is controlled to receive the imaging beam passing through the human body. Projected images are acquired at a certain frequency, and the acquired projection images are reconstructed according to the CBCT algorithm to obtain CBCT images. Based on the CBCT image results, the bed or robotic arm is moved to position the radiotherapy system and the bed to a state where the treatment beam can accurately project to the target area.

[0033] Step 102: Using a transmitter and a receiver to perform imaging at a preset frequency to obtain a single projection image.

[0034] During the treatment phase, the current prediction model and respiratory phase changes are used to track the target area in real time by controlling the synchronous movement of the main robotic arm and the auxiliary robotic arm. At the preset frequency, the transmitting device transmits a conical imaging beam at the position where the target area motion information needs to be detected. The auxiliary robotic arm drives the receiving device to receive the imaging beam passing through the human body on the opposite side of the human body to obtain a single projection image. For example, a single projection image is obtained at a frequency of imaging once every 1 minute. The projection position of the single projection image is the position where the target area motion information needs to be detected.

[0035] In specific implementations, the mechanical installation of the radiotherapy system needs to ensure that both the treatment beam and the imaging beam can pass through the target area. Furthermore, in order to obtain good image quality, the receiving device uses a detector with kV energy. However, since the treatment beam is of MV energy, if the MV energy irradiates the detector, it will affect the life of the detector and the image quality. Therefore, it is necessary to avoid the irradiation of the receiving device by the conical imaging beam of the transmitting device while being able to pass through the target area to reach the receiving device.

[0036] Step 103: Obtain a monitoring value of target area motion information using a single projection image.

[0037] Because both the transmitter and treatment device are mounted on the main robotic arm, rotating synchronously and at similar mounting angles, the projection direction of this single projection image is also similar to the direction of the treatment beam. Therefore, the resulting single projection image accurately represents the target motion information within a plane perpendicular to the treatment beam. This information is crucial for accurate and reliable dynamic tracking during the treatment phase. With only a small amount of information (a single projection image), accurate target motion monitoring can be obtained, further accurately determining whether the current prediction model needs to be revised, thereby improving real-time tracking effectiveness.

[0038] Step 104 : Match the monitored value with the predicted value of the target area motion information in the same respiratory phase in the prediction model.

[0039] By matching the monitored values ​​with the predicted values ​​of the target area motion information at the same respiratory phase in the prediction model, the accuracy of the prediction results of the current prediction model can be tested.

[0040] In step 105, if the matching result indicates that the monitored value and the predicted value are inconsistent, the prediction model is re-established and determined as the current prediction model. The current prediction model and respiratory phase changes are then used to continuously track the target volume in real time by controlling the synchronous movement of the main and auxiliary robotic arms. If the matching result indicates that the monitored value and the predicted value are consistent, the current prediction model and respiratory phase changes are then used to continuously track the target volume in real time by controlling the synchronous movement of the main and auxiliary robotic arms. That is, the prediction model remains unchanged.

[0041] In this embodiment, the main robotic arm drives the treatment device to project and track the target area in real time. At the same time, it also drives the synchronous rotation of the transmitting device, and drives the receiving device through the auxiliary robotic arm to reach the corresponding position that can receive the imaging beam passing through the target area. In order to improve the accuracy of target area projection, when the set frequency is reached, the position where the target area motion information needs to be detected is reached, and then a single projection image at the position is obtained. The monitoring value of the target area motion information can be obtained based on only this projection image, thereby realizing dynamic monitoring of the target area during the treatment stage, so that the prediction model can be corrected in time according to the monitoring results, so that the real-time tracking always uses the prediction model that can predict the precise target area, thereby improving the accuracy of real-time tracking of the target area.

[0042] In some embodiments, using the current prediction model and respiratory phase changes, the target area is tracked in real time by controlling the synchronous movement of the main robotic arm and the auxiliary robotic arm, including: Step 1011 , obtaining a respiratory motion curve that represents respiratory phase changes monitored in real time.

[0043] In a specific implementation, the patient's respiratory movement curve can be monitored by wearing a respiratory monitoring device.

[0044] Step 1012: Utilize the current prediction model and respiratory phase changes to predict target area motion information in real time.

[0045] Specifically, the respiratory phase change can be obtained through the respiratory motion curve. Since the prediction model characterizes the correspondence between the respiratory phase and the target area motion information, the target area motion information can be predicted based on the respiratory phase change.

[0046] Step 1013: Based on the predicted target area motion information, the main robotic arm is controlled to drive the treatment device to move so that the treatment beam emitted by the treatment device irradiates the target area.

[0047] In some embodiments, modeling can be achieved by reconstructing 4D cone-beam CT projection images. Specifically, the above-mentioned target area real-time tracking method also includes the following modeling process: In step 1001 , the main robotic arm and the auxiliary robotic arm respectively drive the transmitting device and the receiving device to rotate synchronously, and imaging is performed while rotating to obtain a cone-beam CT projection image sequence.

[0048] Specifically, according to the preset motion trajectory of CBCT, the main and auxiliary robotic arms perform imaging while rotating to obtain CBCT projection sequence images.

[0049] Step 1002 : reconstructing a 4D cone-beam CT projection image based on a cone-beam CT projection image sequence.

[0050] Step 1003 : Calculate the target area motion information at each respiratory phase based on the 4D cone-beam CT projection image.

[0051] Step 1004 : Based on the target area motion information in each respiratory phase, a prediction model is established to characterize the corresponding relationship between the respiratory phase and the target area motion information.

[0052] In other embodiments, the target area motion information can be calculated and averaged for the cone beam CT projection images at the same respiratory phase to obtain more accurate target area motion information at different respiratory phases to achieve modeling. Specifically, the target area real-time tracking method also includes the following modeling process: In step 100a, the main robotic arm and the auxiliary robotic arm respectively drive the transmitting device and the receiving device to rotate synchronously, and imaging is performed while rotating to obtain a cone-beam CT projection image sequence.

[0053] Step 100b: Divide the images of the same respiratory phase in the cone-beam CT projection image sequence into one group.

[0054] Step 100c: Calculate the target area motion information based on each group of images of the same respiratory phase, take an average of the calculated target area motion information, and determine the average as the target area motion information in the corresponding respiratory phase.

[0055] Step 100d: Based on the target area motion information in each respiratory phase, a prediction model is established to characterize the corresponding relationship between the respiratory phase and the target area motion information.

[0056] In a specific implementation, a prediction model can be established using one of the two modeling methods described above, and then the prediction model can be used to track the target area in real time. If the matching results show that the monitored value is inconsistent with the predicted value, the prediction model can also be re-established using one of the two modeling methods described above.

[0057] In some embodiments, a specific implementation method for obtaining a monitoring value of target area motion information using a single projection image is provided, including: Step 1031 : Acquire a reference CT image for generating a treatment plan.

[0058] Specifically, the reference CT image may refer to an original CT image obtained before formulating a treatment plan, and the target area may be located based on the original image to generate a treatment plan.

[0059] Step 1032: Perform 2D-3D registration based on the single projection image and the reference CT image to obtain the position change of the target area relative to the reference CT image at the current moment, and determine the target area motion information in the patient coordinate system determined based on the position change of the target area relative to the reference CT image at the current moment as the monitoring value of the target area motion information.

[0060] In some cases, a single 2D projection image (X-ray image) is the current image, containing the real-time position and pose of the target. 2D-3D registration of a single 2D projection image (X-ray image) to a 3D reference CT image requires the generation of digitally reconstructed radiographs (DRRs) from the 3D CT volume. Intensity-based 2D-3D image registration modifies the positional translation and rotation of the CT volume, generating a set of DRR images for each position. Similarity measurements are then calculated between the X-ray image and this set of DRR images, identifying the DRR image with the greatest similarity to the X-ray image. The translation and rotation of the CT volume used to generate this DRR image represent the rigid transformation (displacement and rotation) of the X-ray image relative to the reference CT image. The region of interest (ROI) for similarity measurement may vary depending on the scenario. For example, during setup or for organs unaffected by respiratory motion, the ROI may be bone tissue. Alternatively, for organs affected by respiratory motion, the ROI may be tissue or the target volume. Figure 2 The cross-section of the human body and the target area is shown, and the 3D patient (body) coordinate system xyz is defined, where the x-axis points into the paper, the z-axis is perpendicular to the patient's body, and the y-axis is parallel to the patient's body. The position of the patient is determined by three translations and three rotations ( ) These six degrees of freedom are represented. Assume that the projection plane A represents a single projection image, S A is the emission source of a single projected image, Figure 3 is the projected coordinate system x of the projection plane A A y A z A , Figure 4 is the patient coordinate system xyz and the projection coordinate system x A y A z A The relationship between xA The direction is the same as the x direction, and the projection plane A is equivalent to rotating θ around the x axis A Angle, so that y and y A Parallel, z and z A Parallel. According to 2D-3D registration, traverse the reference CT image ( ) transform to generate a set of DRR images, and the DRR image with the greatest similarity is used by ( ) represents the position change of the target area at the current moment relative to the target area of ​​the reference CT image.

[0061] Traversing the six degrees of freedom to generate a set of DRR images and then calculating the similarity separately is a relatively time-consuming process. Therefore, in other examples, in order to establish the relationship between the 3D patient coordinate system and the projection coordinate system, to achieve accelerated calculations and reduce the amount of traversal to generate DRR images, another 3D coordinate system x' y'z' is introduced, z' and z A The same direction, such as Figure 5 As shown. In the 2D projection coordinate system where the projection plane A is located, the 3D rigid transformation is decomposed into the in-plane transformation ( , ) and two out-of-plane rotations ( , ),z A Direction position Geometric magnification factor representation. For projection plane A, only θx' and θy' of the CT volume need to be changed to generate a set of reference images, these references The image corresponds to two out-of-plane rotations of the projection plane A ( , ) different combinations, two-dimensional plane transformation ( ) is estimated by comparing 2D-2D images (X-ray image and DRR image with the greatest similarity), two out-of-plane rotations ( , ) is calculated by the best match between the X-ray image and the DRR image. After identifying the reference DRR image whose out-of-plane rotation is close to the actual value, the in-plane transformation estimate can be accurately obtained. In this way, the position change of the target area in the projection coordinate system relative to the target area of ​​the reference CT image can be obtained, which realizes accelerated calculation, reduces the amount of traversal for generating the DRR image, and directly obtains a more accurate target area motion in the plane perpendicular to the treatment beam, which can track the lesion more accurately. Furthermore, the position change of the target area relative to the reference CT image at the current moment includes the translation change and rotation change of the target area relative to the reference CT image in the projection coordinate system. The geometric relationship between the projection coordinate system of the projection plane A and the patient coordinate system can be expressed by the following geometric relationship formula. The target area motion information (translation change and rotation change) in the patient coordinate system can be determined by the following geometric relationship formula:

[0062]

[0063]

[0064]

[0065]

[0066]

[0067] Where, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the projection coordinate system, represents the geometric magnification factor, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the patient coordinate system, θ A Indicates the angle of rotation of the projection plane A around the x-axis.

[0068] In one example, the target area relative motion information in the patient coordinate system obtained by the above transformation is transformed into the treatment beam direction, and the robotic arm is controlled to follow the target area in the treatment beam direction, so that the treatment beam tracks and irradiates the target area. In another example, the geometric transformation relationship is the same as the above geometric relationship, except that It represents the angle in the y direction between the treatment beam and the imaging beam, thereby directly obtaining the relative motion information of the target area in the direction of the treatment beam, and controlling the robotic arm to follow the target area in the direction of the treatment beam, that is, realizing the tracking and irradiation of the target area by the treatment beam.

[0069] In some embodiments, another specific implementation method for obtaining a monitoring value of target area motion information using a single projection image is provided, including: Step 103a: Perform 2D registration between the single projection image and each image in the cone-beam CT projection image sequence, and determine a second target image that best matches the single projection image from the cone-beam CT projection image sequence.

[0070] Step 103b: determining target area motion information in the 4D cone-beam CT projection image at the same respiratory phase according to the respiratory phase of the second target image, and determining the target area motion information as a monitoring value of the target area motion information.

[0071] By performing 2D-2D registration of the monitored single projection image with the cone-beam CT projection image sequence obtained in the aforementioned modeling process, the most matching cone-beam CT projection image is found. Then, based on the cone-beam CT projection image, the target area motion value with the same respiratory phase as the cone-beam CT projection image is found in the reconstructed 4D-CBCT projection image sequence, thus determining the monitoring value of the target area motion information.

[0072] In addition, in the process of modeling by calculating the target area motion information of cone-beam CT projection images of the same respiratory phase and then taking the average value, the target area motion information is calculated based on each group of images of the same respiratory phase, and one of the two specific implementation methods of obtaining the monitoring value of the target area motion information using a single projection image can also be used.

[0073] Compared to existing radiotherapy systems using ring or C-arm gantry systems, this embodiment utilizes the primary and secondary robotic arms to implement the aforementioned method for real-time target tracking, achieving a wider treatment angle and significantly minimizing normal tissue damage while treating the tumor target. CBCT images obtained using the primary and secondary robotic arms in this manner offer improved image quality, faster patient placement, and real-time tracking of the target location, enabling more precise treatment. Compared to existing technologies, this method achieves closer angles between the treatment beam and the imaging beam, providing more accurate information on the tumor target's motion in the direction of the treatment beam compared to orthogonal imaging.

[0074] Example 2: Another embodiment of the present application relates to a real-time target tracking device. The implementation details of the real-time target tracking device of this embodiment are described in detail below. The following content is only the implementation details provided for easy understanding and is not necessary for the implementation of this solution. The real-time target tracking device of this embodiment is implemented based on an image-guided radiotherapy system. The image-guided radiotherapy system includes a main robotic arm and an auxiliary robotic arm. The end of the main robotic arm is equipped with a transmitting device for transmitting a cone-shaped imaging beam and a treatment device for transmitting a treatment beam. The end of the auxiliary robotic arm is equipped with a receiving device for receiving the cone-shaped imaging beam. While the main robotic arm drives the treatment device and the transmitting device to move, the auxiliary robotic arm drives the receiving device to move synchronously so that the cone-shaped imaging beam can pass through the target area to reach the receiving device and the receiving device avoids the irradiation of the treatment beam. The schematic diagram of the real-time target tracking device can be as follows Figure 6 As shown, it includes: a real-time tracking module 201, an image acquisition module 202, a monitoring calculation module 203, an information matching module 204 and a model building module 205.

[0075] A real-time tracking module 201 is configured to track the target area in real time by controlling the synchronous movement of the main and auxiliary manipulators using the current prediction model and respiratory phase changes. The prediction model is configured to characterize the correspondence between the respiratory phase and target area motion information, which includes six degrees of freedom information. The image acquisition module 202 is used to acquire a single projection image by using the transmitting device and the receiving device to perform imaging according to a preset frequency; A monitoring calculation module 203 is used to obtain a monitoring value of target area motion information using a single projection image; An information matching module 204 is used to match the monitored value with the predicted value of the target area motion information of the same respiratory phase in the prediction model; The model building module 205 is used to re-establish the prediction model when the matching result shows that the monitoring value is inconsistent with the predicted value, and determine the re-established prediction model as the current prediction model, so as to continue to use the current prediction model and respiratory phase changes to track the target area in real time by controlling the synchronous movement of the main robotic arm and the auxiliary robotic arm.

[0076] Specifically, the target area motion information includes six-degree-of-freedom information of the current target area relative to the target area in the reference CT image, indicating the offset of the patient's real-time tumor position relative to the tumor position in the original reference CT image.

[0077] In practice, after entering the treatment phase, the system controls the main robotic arm to drive the treatment device to align with the target area and emit a treatment beam to treat the tumor. Before entering the treatment phase, the patient needs to be positioned and then tracked in real time.

[0078] In some examples, during the patient positioning phase, the master robotic arm drives the transmitter to rotate in a circular path along the cross-section of the human body around the target area, directing the conical imaging beam toward the target area. The slave robotic arm synchronously drives the receiver to rotate in a circular path along the cross-section of the human body around the target area. The synchronous rotation of the master and slave robotic arms ensures that the center of the conical imaging beam at each rotational position is perpendicular to the center of the receiver after passing through the human body. During the synchronous rotation of the master and slave robotic arms, the transmitter is controlled to transmit the imaging beam and the receiver is controlled to receive the imaging beam after passing through the human body. Projected images are acquired at a certain frequency, and the acquired projection images are converted into CBCT images according to the CBCT reconstruction algorithm. Based on the CBCT image results, the bed or robotic arm is moved to position the radiotherapy system and bed so that the treatment beam can accurately illuminate the target area. This positioning method can quickly and accurately locate the target area without the need to move the bed and system multiple times. Compared with existing technologies, the target area position determined by positioning is accurate, and the positioning time is significantly reduced, resulting in high positioning efficiency.

[0079] In this embodiment, the main robotic arm drives the treatment device to project and track the target area in real time. At the same time, it also drives the synchronous rotation of the transmitting device, and drives the receiving device to the corresponding position that can receive the imaging beam passing through the target area through the auxiliary robotic arm. In order to improve the accuracy of the target area projection, when the set frequency is reached, the position where the target area motion information needs to be detected is reached, and then a single projection image at the position is obtained. The monitoring value of the target area motion information can be obtained only based on the projection image, thereby realizing dynamic monitoring of the target area during the treatment stage, and timely correcting the prediction model according to the monitoring results, so that the real-time tracking always uses the prediction model that can predict the precise target area, thereby improving the accuracy of the real-time tracking of the target area.

[0080] In some embodiments, the target area is tracked in real time by controlling the synchronous movement of the main robotic arm and the auxiliary robotic arm using the current prediction model and respiratory phase changes, including: obtaining a real-time monitored respiratory motion curve that characterizes respiratory phase changes; using the current prediction model and respiratory phase changes to predict the target area motion information in real time; and controlling the main robotic arm to drive the treatment device to move based on the predicted target area motion information so that the treatment beam emitted by the treatment device irradiates the target area.

[0081] In some embodiments, modeling can be achieved by reconstructing 4D cone-beam CT projection images, and the model building module is also used to implement the following modeling process: the main robotic arm and the auxiliary robotic arm respectively drive the transmitting device and the receiving device to rotate synchronously, and imaging is performed while rotating to obtain a cone-beam CT projection image sequence; 4D cone-beam CT projection images are reconstructed according to the cone-beam CT projection image sequence; the target area motion information in each respiratory phase is calculated according to the 4D cone-beam CT projection images; and a prediction model for characterizing the correspondence between the respiratory phase and the target area motion information is established according to the target area motion information in each respiratory phase.

[0082] In other embodiments, the target area motion information can be calculated and averaged for the cone-beam CT projection images of the same respiratory phase to obtain more accurate target area motion information under different respiratory phases to achieve modeling. The model building module is also used to implement the following modeling process: the main robotic arm and the auxiliary robotic arm respectively drive the transmitting device and the receiving device to rotate synchronously, and imaging is performed while rotating to obtain a cone-beam CT projection image sequence; the images of the same respiratory phase in the cone-beam CT projection image sequence are divided into a group; the target area motion information is calculated based on each group of images of the same respiratory phase, the calculated target area motion information is averaged, and the average value is determined as the target area motion information under the corresponding respiratory phase; based on the target area motion information under each respiratory phase, a prediction model is established to characterize the corresponding relationship between the respiratory phase and the target area motion information.

[0083] In a specific implementation, a prediction model can be established using one of the two modeling methods described above, and then the prediction model can be used to track the target area in real time. If the matching results show that the monitored value is inconsistent with the predicted value, the prediction model can also be re-established using one of the two modeling methods described above.

[0084] In some embodiments, a specific implementation method for obtaining a monitoring value of target area motion information using a single projection image is provided, including: obtaining a reference CT image used to generate a treatment plan; performing 2D-3D registration based on the single projection image and the reference CT image to obtain a position change of the target area relative to the reference CT image at the current moment, and determining the target area motion information in the patient coordinate system determined according to the position change of the target area relative to the reference CT image at the current moment as the monitoring value of the target area motion information.

[0085] In some cases, a single 2D projection image (X-ray image) is the current image, containing the real-time position and pose of the target. 2D-3D registration of a single 2D projection image (X-ray image) to a 3D reference CT image requires the generation of digitally reconstructed radiographs (DRRs) from the 3D CT volume. Intensity-based 2D-3D image registration modifies the positional translation and rotation of the CT volume, generating a set of DRR images for each position. Similarity measurements are then calculated between the X-ray image and this set of DRR images, identifying the DRR image with the greatest similarity to the X-ray image. The translation and rotation of the CT volume used to generate this DRR image represent the rigid transformation (displacement and rotation) of the X-ray image relative to the reference CT image. The region of interest (ROI) for similarity measurement may vary depending on the scenario. For example, during setup or for organs unaffected by respiratory motion, the ROI may be bone tissue. Alternatively, for organs affected by respiratory motion, the ROI may be tissue or the target volume. Figure 2 The cross-section of the human body and the target area is shown, and the 3D patient (body) coordinate system xyz is defined, where the x-axis points into the paper, the z-axis is perpendicular to the patient's body, and the y-axis is parallel to the patient's body. The position of the patient is determined by three translations and three rotations ( ) These six degrees of freedom are represented. Assume that the projection plane A represents a single projection image, S A is the emission source of a single projected image, Figure 3 is the projected coordinate system x of the projection plane A A y A z A , Figure 4 is the patient coordinate system xyz and the projection coordinate system x A y A z A The relationship between x AThe direction is the same as the x direction, and the projection plane A is equivalent to rotating θ around the x axis A Angle, so that y and y A Parallel, z and z A Parallel. According to 2D-3D registration, traverse the reference CT image ( ), The transformation generates a set of DRR images, and the DRR image with the greatest similarity is used ( ) represents the position change of the target area at the current moment relative to the target area of ​​the reference CT image.

[0086] Traversing the six degrees of freedom to generate a set of DRR images and then calculating the similarity separately is a relatively time-consuming process. Therefore, in other examples, in order to establish the relationship between the 3D patient coordinate system and the projection coordinate system, to achieve accelerated calculations and reduce the amount of traversal to generate DRR images, another 3D coordinate system x'y'z' is introduced, z' and z A The same direction, such as Figure 5 As shown. In the 2D projection coordinate system where the projection plane A is located, the 3D rigid transformation is decomposed into the in-plane transformation ( ) and two out-of-plane rotations ( , ),z A Direction position The geometric magnification factor is expressed as follows. For projection plane A, only θx' and θy' of the CT volume need to be changed to generate a set of reference DRR images. These reference DRR images correspond to two out-of-plane rotations of projection plane A ( , ) different combinations, two-dimensional plane transformation ( ) is estimated by comparing 2D-2D images (X-ray image and DRR image with the greatest similarity), two out-of-plane rotations ( , ) is calculated by the best match between the X-ray image and the DRR image. After identifying the reference DRR image whose out-of-plane rotation is close to the actual value, the in-plane transformation estimate can be accurately obtained. In this way, the position change of the target area in the projection coordinate system relative to the target area of ​​the reference CT image can be obtained, which realizes accelerated calculation, reduces the amount of traversal for generating the DRR image, and directly obtains a more accurate target area motion in the plane perpendicular to the treatment beam, which can track the lesion more accurately. Furthermore, the position change of the target area relative to the reference CT image at the current moment includes the translation change and rotation change of the target area relative to the reference CT image in the projection coordinate system. The geometric relationship between the projection coordinate system of the projection plane A and the patient coordinate system can be expressed by the following geometric relationship formula. The target area motion information (translation change and rotation change) in the patient coordinate system can be determined by the following geometric relationship formula:

[0087]

[0088]

[0089]

[0090]

[0091]

[0092] Where, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the projection coordinate system, represents the geometric magnification factor, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the patient coordinate system, Indicates the angle of rotation of the projection plane A around the x-axis.

[0093] Through the above geometric relationship, the target area motion information in the projection coordinate system can be converted into the patient coordinate system to obtain the monitoring value of the target area motion information in the patient coordinate system.

[0094] In one example, the target area relative motion information in the patient coordinate system obtained by the above transformation is transformed into the treatment beam direction, and the robotic arm is controlled to follow the target area in the treatment beam direction, so that the treatment beam tracks and irradiates the target area. In another example, the geometric transformation relationship is the same as the above geometric relationship, except that It represents the angle in the y direction between the treatment beam and the imaging beam, thereby directly obtaining the relative motion information of the target area in the direction of the treatment beam, and controlling the robotic arm to follow the target area in the direction of the treatment beam, that is, realizing the tracking and irradiation of the target area by the treatment beam.

[0095] In some embodiments, another specific implementation method for obtaining a monitoring value of target area motion information using a single projection image is provided, including: using the single projection image to perform 2D alignment with each image in a cone-beam CT projection image sequence, and determining a second target image that best matches the single projection image from the cone-beam CT projection image sequence; determining the target area motion information under the same respiratory phase in the 4D cone-beam CT projection image based on the respiratory phase of the second target image, and determining the target area motion information as the monitoring value of the target area motion information.

[0096] By performing 2D-2D registration of the monitored single projection image with the cone-beam CT projection image sequence obtained in the aforementioned modeling process, the most matching cone-beam CT projection image is found. Then, based on the cone-beam CT projection image, the target area motion value with the same respiratory phase as the cone-beam CT projection image is found in the reconstructed 4D-CBCT projection image sequence, thus determining the monitoring value of the target area motion information.

[0097] In addition, in the process of modeling by calculating the target area motion information of cone-beam CT projection images of the same respiratory phase and then taking the average value, the target area motion information is calculated based on each group of images of the same respiratory phase, and one of the two specific implementation methods of obtaining the monitoring value of the target area motion information using a single projection image can also be used.

[0098] Compared to existing radiotherapy systems using ring or C-arm gantry systems, this embodiment utilizes the primary and secondary robotic arms to implement the aforementioned method for real-time target tracking, achieving a wider treatment angle and significantly minimizing normal tissue damage while treating the tumor target. CBCT images obtained using the primary and secondary robotic arms in this manner offer improved image quality, faster patient placement, and real-time tracking of the target location, enabling more precise treatment. Compared to existing technologies, this method achieves closer angles between the treatment beam and the imaging beam, providing more accurate information on the tumor target's motion in the direction of the treatment beam compared to orthogonal imaging.

[0099] It is worth mentioning that all modules involved in this embodiment are logical modules. In actual applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, to highlight the innovation of this application, this embodiment does not include units that are not closely related to solving the technical problem proposed by this application. However, this does not mean that other units do not exist in this embodiment.

[0100] Example 3: Another embodiment of the present application relates to an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the real-time target area tracking method of the above-mentioned embodiments.

[0101] The memory and processor are connected using a bus, which can include any number of interconnected buses and bridges. The bus connects various circuits of one or more processors and memories. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits. These are all well known in the art and are therefore not described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single component or multiple components, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over a wireless medium via an antenna. Furthermore, the antenna receives data and transmits it to the processor.

[0102] The processor is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory can be used to store data used by the processor when performing operations.

[0103] Example 4: Another embodiment of the present application relates to an image-guided radiotherapy system, such as Figure 7 Shown, including: A main robotic arm 301, at the end of which are mounted an emitting device 304 for emitting a cone-shaped imaging beam and a treatment device 303 for emitting a treatment beam. The main robotic arm 301 is capable of driving the treatment device 303 and the emitting device 304 to move synchronously. The auxiliary manipulator 302 has a receiving device 305 installed at its end for receiving the cone imaging beam. When the main manipulator 301 drives the treatment device 303 and the transmitting device 304 to move, the auxiliary manipulator 302 drives the receiving device 305 to move synchronously, so that the cone imaging beam can pass through the target area to reach the receiving device 305 and the receiving device 305 can avoid the irradiation of the treatment beam. Figure 8 As shown; The control device is used to use the current prediction model and respiratory phase changes to track the target area in real time by controlling the synchronous movement of the main robotic arm and the auxiliary robotic arm; to use the transmitting device and the receiving device to perform imaging at a preset frequency to obtain a single projection image; to use the single projection image to obtain a monitoring value of the target area motion information; to match the monitoring value with the predicted value of the target area motion information at the same respiratory phase in the prediction model; if the matching result indicates that the monitoring value and the predicted value are inconsistent, to re-establish the prediction model and determine the re-established prediction model as the current prediction model, so as to continue to use the current prediction model and respiratory phase changes to track the target area in real time by synchronous movement of the main robotic arm and the auxiliary robotic arm; the prediction model is used to characterize the correspondence between the respiratory phase and the target area motion information, and the target area motion information includes six-degree-of-freedom information. The specific implementation of the control device can refer to the aforementioned method and device embodiments.

[0104] The following describes the working mode of the image-guided radiotherapy system of this embodiment by taking the transmitting device as an X-ray tube and the receiving device as a detector as an example.

[0105] Working method 1: (1) During the patient positioning phase, the main robotic arm 301 drives the X-ray tube to rotate along the cross-section of the human body in an arc trajectory around the target area of ​​the tumor, so that the X-rays are directed toward the target area. Correspondingly, the auxiliary robotic arm drives the detector to rotate along the cross-section of the human body in an arc trajectory around the target area. The main and auxiliary robotic arms rotate synchronously so that the center of the X-ray at each rotation position is perpendicular to the center of the detector after passing through the human body. The system controls the emission of the X-ray tube and the reception of the detector, and acquires images at a certain frequency. The acquired images are reconstructed according to the CBCT algorithm to obtain CBCT images. Based on the CBCT image results, the bed 306 or the main and auxiliary robotic arms are moved so that the system can know the accurate location of the tumor target area.

[0106] (2) During the treatment phase of real-time tracking of the target area, the system controls the main robotic arm 301 to carry the treatment device 303 to the target area and emit a treatment beam to perform radiotherapy on the tumor. Figure 8 As shown, the main robotic arm 301 carries the treatment device and emits a treatment beam 307 toward the target area. Simultaneously, the X-ray tube emits an imaging beam 308 at the target area where motion detection is required. The auxiliary robotic arm 302 drives the detector to receive the emitted X-rays on the opposite side of the body. Mechanical installation ensures that both treatment beam 307 and imaging beam 308 pass through the target area. To achieve better image quality, the detector uses kV energy. However, the treatment beam has MV energy. MV energy can damage the detector life and image quality, so the detector must be kept clear of the treatment beam.

[0107] The first real-time target tracking solution: a) A respiratory monitoring device is installed outside the patient’s body to display the respiratory movement curve.

[0108] b) Following the motion trajectory of the CBCT, the dual robotic arms perform X-ray imaging while rotating synchronously to obtain a CBCT projection image sequence.

[0109] c) Respiratory modeling: Based on the CBCT projection image sequence obtained in b), a 4D-CBCT image is reconstructed. The 4D-CBCT image is related to the respiratory phase. The target area motion value at each respiratory phase is calculated based on the 4D-CBCT image to obtain a prediction model that characterizes the relationship between the respiratory phase and the target area motion.

[0110] d) During the treatment process of real-time target tracking, the position of the target at the next moment is predicted based on the relationship between the respiratory motion curve and the known prediction model. The main robotic arm is controlled to drive the treatment device and the target area to move together, achieving precise irradiation of the target area.

[0111] e) To ensure the accuracy of the target area, single X-ray images are acquired at a regular frequency, and dynamic monitoring of the treatment process is performed with real-time target tracking. The target area position is determined based on the single X-ray image and then matched with the model. If the degree of match exceeds the threshold (inconsistent), the model needs to be adjusted and rebuilt according to steps b) through c). If the match is consistent, treatment and monitoring continue.

[0112] One method of obtaining the target area location based on a single X-ray image: Calculate the tumor center location based on a single X-ray image and match it with the target area motion information corresponding to the corresponding respiratory curve. The specific method is as follows: The position change of the target area relative to the reference CT image at the current moment is obtained, and the target area motion information in the patient coordinate system determined according to the position change of the target area relative to the reference CT image at the current moment is determined as the monitoring value of the target area motion information.

[0113] In some cases, a single 2D projection image (X-ray image) is the current image, containing the real-time position and pose of the target. 2D-3D registration of a single 2D projection image (X-ray image) to a 3D reference CT image requires the generation of digitally reconstructed radiographs (DRRs) from the 3D CT volume. Intensity-based 2D-3D image registration modifies the positional translation and rotation of the CT volume, generating a set of DRR images for each position. Similarity measurements are then calculated between the X-ray image and this set of DRR images, identifying the DRR image with the greatest similarity to the X-ray image. The translation and rotation of the CT volume used to generate this DRR image represent the rigid transformation (displacement and rotation) of the X-ray image relative to the reference CT image. The region of interest (ROI) for similarity measurement may vary depending on the scenario. For example, during setup or for organs unaffected by respiratory motion, the ROI may be bone tissue. Alternatively, for organs affected by respiratory motion, the ROI may be tissue or the target volume. Figure 2 The cross-section of the human body and the target area is shown, and the 3D patient (body) coordinate system xyz is defined, where the x-axis points into the paper, the z-axis is perpendicular to the patient's body, and the y-axis is parallel to the patient's body. The position of the patient is determined by three translations and three rotations ( ) These six degrees of freedom are represented. Assume that the projection plane A represents a single projection image, S A is the emission source of a single projected image, Figure 3 is the projected coordinate system x of the projection plane A A y A z A , Figure 4 is the patient coordinate system xyz and the projection coordinate system x A y A zA The relationship between x A The direction is the same as the x direction, and the projection plane A is equivalent to rotating around the x axis Angle, so that y and y A Parallel, z and z A Parallel. According to 2D-3D registration, traverse the reference CT image ( ) transform to generate a set of DRR images, and the DRR image with the greatest similarity is used by ( ) represents the position change of the target area at the current moment relative to the target area of ​​the reference CT image.

[0114] Traversing the six degrees of freedom to generate a set of DRR images and then calculating the similarity separately is a relatively time-consuming process. Therefore, in other examples, in order to establish the relationship between the 3D patient coordinate system and the projection coordinate system, to achieve accelerated calculations and reduce the amount of traversal to generate DRR images, another 3D coordinate system x' y'z' is introduced, z' and z A The same direction, such as Figure 5 As shown. In the 2D projection coordinate system where the projection plane A is located, the 3D rigid transformation is decomposed into the in-plane transformation ( ) and two out-of-plane rotations ( , ),z A Direction position The geometric magnification factor is expressed as follows. For projection plane A, only θx' and θy' of the CT volume need to be changed to generate a set of reference DRR images. These reference DRR images correspond to two out-of-plane rotations of projection plane A ( , ) different combinations, two-dimensional plane transformation ( ) is estimated by comparing 2D-2D images (X-ray image and DRR image with the greatest similarity), two out-of-plane rotations ( , ) is calculated by the best match between the X-ray image and the DRR image. After identifying the reference DRR image whose out-of-plane rotation is close to the actual value, the in-plane transformation estimate can be accurately obtained. In this way, the position change of the target area in the projection coordinate system relative to the target area of ​​the reference CT image can be obtained, which realizes accelerated calculation, reduces the amount of traversal for generating the DRR image, and directly obtains a more accurate target area motion in the plane perpendicular to the treatment beam, which can track the lesion more accurately. Furthermore, the position change of the target area relative to the reference CT image at the current moment includes the translation change and rotation change of the target area relative to the reference CT image in the projection coordinate system. The geometric relationship between the projection coordinate system of the projection plane A and the patient coordinate system can be expressed by the following geometric relationship formula. The target area motion information (translation change and rotation change) in the patient coordinate system can be determined by the following geometric relationship formula:

[0115]

[0116]

[0117]

[0118]

[0119]

[0120] Where, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the projection coordinate system, represents the geometric magnification factor, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the patient coordinate system, Indicates the angle of rotation of the projection plane A around the x-axis.

[0121] In one example, the target area relative motion information in the patient coordinate system obtained by the above transformation is transformed into the treatment beam direction, and the robotic arm is controlled to follow the target area in the treatment beam direction, so that the treatment beam tracks and irradiates the target area. In another example, the geometric transformation relationship is the same as the above geometric relationship, except that It represents the angle in the y direction between the treatment beam and the imaging beam, thereby directly obtaining the relative motion information of the target area in the direction of the treatment beam, and controlling the robotic arm to follow the target area in the direction of the treatment beam, that is, realizing the tracking and irradiation of the target area by the treatment beam.

[0122] The second real-time target tracking solution: a) A respiratory monitoring device is installed outside the patient’s body to display the respiratory movement curve.

[0123] b) Following the motion trajectory of the CBCT, the dual robotic arms perform X-ray imaging while rotating synchronously to obtain a CBCT projection image sequence.

[0124] c) Respiratory modeling: By calculating the target motion information of cone-beam CT projection images of the same respiratory phase and taking the average value, more accurate target motion information under different respiratory phases is obtained to achieve modeling. Specifically, the main robotic arm and the auxiliary robotic arm respectively drive the transmitting device and the receiving device to rotate synchronously, and imaging is performed while rotating to obtain a cone-beam CT projection image sequence. The images of the same respiratory phase in the cone-beam CT projection image sequence are divided into a group, and the target motion information is calculated based on each group of images with the same respiratory phase. The calculated target motion information is averaged and the average value is determined as the target motion information under the corresponding respiratory phase. Based on the target motion information under each respiratory phase, a prediction model is established to characterize the correspondence between the respiratory phase and the target motion information.

[0125] d) During the treatment process of real-time target tracking, the position of the target at the next moment is predicted based on the relationship between the respiratory motion curve and the known prediction model. The main robotic arm is controlled to drive the treatment device and the target area to move together, achieving precise irradiation of the target area.

[0126] e) To ensure the accuracy of the target area, single X-ray images are acquired at a regular frequency, and dynamic monitoring of the treatment process is performed with real-time target tracking. The target area position is determined based on the single X-ray image and then matched with the model. If the degree of match exceeds the threshold (inconsistent), the model needs to be adjusted and rebuilt according to steps b) through c). If the match is consistent, treatment and monitoring continue.

[0127] The second method for obtaining the target area position based on a single X-ray image is as follows: the monitored single projection image is 2D-2D aligned with the cone-beam CT projection image sequence obtained in the aforementioned modeling process to find the most matching cone-beam CT projection image. Then, based on the cone-beam CT projection image, the target area motion value with the same respiratory phase as the cone-beam CT projection image is found in the reconstructed 4D-CBCT projection image sequence, thus determining the monitoring value of the target area motion information.

[0128] Compared to existing radiotherapy systems using ring or C-arm gantry systems, this embodiment utilizes the primary and secondary robotic arms to implement the aforementioned method for real-time target tracking, achieving a wider treatment angle and significantly minimizing normal tissue damage while treating the tumor target. CBCT images obtained using the primary and secondary robotic arms in this manner offer improved image quality, faster patient placement, and real-time tracking of the target location, enabling more precise treatment. Compared to existing technologies, this method achieves closer angles between the treatment beam and the imaging beam, providing more accurate information on the tumor target's motion in the direction of the treatment beam compared to orthogonal imaging.

[0129] Embodiment 5: Another embodiment of the present application relates to a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned embodiment of the method for real-time target area tracking.

[0130] That is, those skilled in the art will understand that all or part of the steps in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a program. The program is stored in a storage medium and includes a number of instructions for causing a device (which may be a single-chip microcomputer, chip, etc.) or a processor to execute all or part of the steps in the methods described in the various embodiments of this application. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0131] Those skilled in the art will appreciate that the above embodiments are specific embodiments for implementing the present application, and that in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present application.

Claims

1. A method for real-time target tracking, implemented based on an image-guided radiotherapy system, wherein the image-guided radiotherapy system comprises a main robotic arm and an auxiliary robotic arm, wherein the main robotic arm is provided with a transmitter for transmitting a cone-shaped imaging beam and a treatment device for transmitting a treatment beam at its distal end, and the auxiliary robotic arm is provided with a receiver for receiving the cone-shaped imaging beam at its distal end; While the main robotic arm drives the treatment device and the transmitting device to move, the auxiliary robotic arm drives the receiving device to move synchronously so that the cone imaging beam can pass through the target area to reach the receiving device and the receiving device avoids the irradiation of the treatment beam; The target area real-time tracking method comprises: Using the current prediction model and respiratory phase changes, the target area is tracked in real time by controlling the synchronous movement of the main and auxiliary robotic arms. The prediction model is used to characterize the correspondence between the respiratory phase and the target area motion information, and the target area motion information includes six degrees of freedom information. According to the preset frequency, the transmitting device and the receiving device are used to form an image to obtain a single projection image; Obtaining a monitoring value of target area motion information using the single projection image; Matching the monitored value with the predicted value of the target area motion information in the same respiratory phase in the prediction model; When the matching result indicates that the monitored value is inconsistent with the predicted value, the prediction model is re-established and the re-established prediction model is determined as the current prediction model to continue to use the current prediction model and respiratory phase changes to track the target area in real time by controlling the synchronous movement of the main robotic arm and the auxiliary robotic arm.

2. The target area real-time tracking method according to claim 1, characterized in that: Utilizing the current prediction model and respiratory phase changes, the target area is tracked in real time by controlling the synchronous movement of the main and auxiliary robotic arms, including: Acquire a real-time monitored respiratory motion curve representing respiratory phase changes; Utilize the current prediction model and respiratory phase changes to predict target area motion information in real time; According to the predicted target area motion information, the main robotic arm is controlled to drive the treatment device to move, so that the treatment beam emitted by the treatment device irradiates the target area.

3. The target area real-time tracking method according to claim 1, characterized in that: Also includes: The main robotic arm and the auxiliary robotic arm drive the transmitting device and the receiving device to rotate synchronously respectively, and imaging is performed while rotating to obtain a cone-beam CT projection image sequence; reconstructing a 4D cone-beam CT projection image according to the cone-beam CT projection image sequence; Calculating target area motion information at each respiratory phase according to the 4D cone-beam CT projection image; According to the target area motion information in each respiratory phase, a prediction model is established to characterize the corresponding relationship between the respiratory phase and the target area motion information.

4. The target area real-time tracking method according to claim 1, characterized in that: Also includes: The main robotic arm and the auxiliary robotic arm drive the transmitting device and the receiving device to rotate synchronously respectively, and imaging is performed while rotating to obtain a cone-beam CT projection image sequence; dividing images of the same respiratory phase in the cone-beam CT projection image sequence into a group; Calculating target area motion information according to each group of images of the same respiratory phase, taking an average of the calculated target area motion information, and determining the average as the target area motion information at the corresponding respiratory phase; According to the target area motion information in each respiratory phase, a prediction model is established to characterize the corresponding relationship between the respiratory phase and the target area motion information.

5. The target area real-time tracking method according to claim 1, characterized in that: Obtaining a monitoring value of target area motion information using the single projection image includes: Acquiring reference CT images for generating treatment plans; Based on the single projection image and the reference CT image, 2D-3D alignment is performed to obtain the position change of the target area relative to the reference CT image at the current moment, and the target area motion information in the patient coordinate system determined according to the position change of the target area relative to the reference CT image at the current moment is determined as the monitoring value of the target area motion information.

6. The target area real-time tracking method according to claim 5, characterized in that: The current position change of the target area relative to the reference CT image includes the translation change and rotation change of the target area relative to the reference CT image in the projection coordinate system. The target area motion information in the patient coordinate system is determined by the following geometric relationship between the projection coordinate system and the patient coordinate system: Where, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the projection coordinate system, represents the geometric magnification factor, ( ) represents the translation and rotation changes of the target area relative to the reference CT image in the patient coordinate system, Indicates the angle of rotation of the projection plane A around the x-axis.

7. The target area real-time tracking method according to claim 3, characterized in that: Obtaining a monitoring value of target area motion information using the single projection image includes: performing 2D registration using the single projection image and each image in the cone-beam CT projection image sequence, and determining a second target image from the cone-beam CT projection image sequence that best matches the single projection image; The target area motion information at the same respiratory phase in the 4D cone-beam CT projection image is determined according to the respiratory phase of the second target image, and the target area motion information is determined as a monitoring value of the target area motion information.

8. An electronic device, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the real-time target area tracking method according to any one of claims 1 to 7.

9. An image-guided radiotherapy system, characterized in that: include: A main robotic arm, at the end of which is mounted a launch device for launching a cone-shaped imaging beam and a treatment device for launching a treatment beam, and the main robotic arm is capable of driving the treatment device and the launch device to move synchronously; An auxiliary robotic arm, at the end of which is mounted a receiving device for receiving the cone-shaped imaging beam. The auxiliary robotic arm drives the receiving device to move synchronously with the main robotic arm while driving the treatment device and the transmitting device, so that the cone-shaped imaging beam can pass through the target area to reach the receiving device while the receiving device avoids the irradiation of the treatment beam. A control device, configured to track a target area in real time by controlling the synchronous movement of a main robotic arm and an auxiliary robotic arm using a current prediction model and respiratory phase changes; According to a preset frequency, the transmitting device and the receiving device are used to perform imaging to obtain a single projection image; and the monitoring value of the target area motion information is obtained using the single projection image; The monitoring value is matched with the predicted value of the target area motion information of the same respiratory phase in the prediction model; when the matching result shows that the monitoring value is inconsistent with the predicted value, the prediction model is re-established, and the re-established prediction model is determined as the current prediction model, so as to continue to use the current prediction model and respiratory phase changes to track the target area in real time through the synchronous movement of the main robotic arm and the auxiliary robotic arm; the prediction model is used to characterize the correspondence between the respiratory phase and the target area motion information, and the target area motion information includes six-degree-of-freedom information.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the target area real-time tracking method according to any one of claims 1 to 7 is implemented.

Images