Radiation therapy device, medical image processing device, radiation therapy method, and storage medium

Abstract

The present application relates to a radiation therapy device, a medical image processing device, a radiation therapy method, and a storage medium. The radiation therapy device of the embodiment includes an acquisition unit, a projection position calculation unit, an element projection image generation unit, and an element projection image synthesis unit. The acquisition unit acquires a condition for X-ray imaging at a therapy stage and a three-dimensional image of a patient imaged before the therapy stage. The projection position calculation unit calculates a projection position of each pixel included in the three-dimensional image when projected onto a two-dimensional X-ray fluoroscopic image generated by the X-ray imaging, based on the condition for the X-ray imaging. The element projection image generation unit generates an element projection image of each pixel when projected onto the X-ray fluoroscopic image. The element projection image synthesis unit synthesizes the element projection image of each pixel based on the projection position, thereby generating a reconstructed image.

Description

Technical Field

[0001] Embodiments of the present invention relate to radiation therapy devices, medical image processing devices, radiation therapy methods, and storage media. Background Technology

[0002] Radiation therapy is a treatment method that destroys a diseased area by irradiating it with radiation. In radiation therapy, to avoid damaging normal tissue, the radiation must be precisely aimed at the diseased area. Therefore, before radiation therapy begins, the location of the diseased area is determined using X-ray fluoroscopy images, and the position and angle of the movable treatment table supporting the patient are appropriately adjusted to ensure the diseased area is precisely within the radiation irradiation range. This alignment is achieved by comparing a digitally reconstructed radiograph (DRR) with the X-ray fluoroscopy image. The DRR is obtained during the treatment planning stage by virtually reconstructing the X-ray fluoroscopy image based on a three-dimensional CT image obtained through prior computed tomography (CT), which is then taken during the treatment phase.

[0003] In the alignment described above, the patient's movement is determined by solving a six-dimensional (3D translation and 3D rotation) search problem, using the similarity between the X-ray fluoroscopy image taken during treatment and the DRR (Different Pathological Rectification). This search problem cannot be solved analytically; therefore, it is usually solved through repeated calculations. Achieving high-precision alignment requires significant processing time. In particular, generating the DRR requires a large amount of computation, consuming a significant portion of the processing time. Therefore, to achieve fast alignment, it is necessary to reduce the number of DRR generation iterations or increase the generation speed.

[0004] To shorten processing time, a method is proposed that reduces the number of DRR generation iterations by evaluating the similarity between the DRR and the X-ray fluoroscopic image only in the direction of significant image variation, thus enabling rapid alignment. While this conventional method reduces the number of DRR generation iterations, the computationally intensive and time-consuming ray tracing method used in DRR generation does not reduce the time required for each DRR generation and therefore cannot achieve truly rapid alignment.

[0005] Existing technical documents

[0006] Patent documents

[0007] Patent Document 1: Japanese Patent Application Publication No. 2013-99431 Summary of the Invention

[0008] The problem that the invention aims to solve

[0009] The problem to be solved by the present invention is to provide a radiotherapy apparatus, a medical image processing apparatus, a radiotherapy method, and a storage medium that enables the rapid generation and processing of DRR (radioradiotherapy response) and the precise localization of patients in a short time.

[0010] Methods for solving problems

[0011] The radiotherapy apparatus of this embodiment includes an acquisition unit, a projection position calculation unit, an element projection image generation unit, and an element projection image synthesis unit. The acquisition unit acquires the conditions for X-ray imaging during the treatment phase, as well as a three-dimensional image of the patient taken before the treatment phase. The projection position calculation unit calculates, based on the X-ray imaging conditions, the projection position of each pixel in the three-dimensional image when projected onto a two-dimensional X-ray fluoroscopic image generated from the X-ray imaging. The element projection image generation unit generates an element projection image of each pixel when the pixels in the three-dimensional image are projected onto the X-ray fluoroscopic image. The element projection image synthesis unit synthesizes the generated element projection images of each pixel based on the calculated projection positions, thereby generating a reconstructed image obtained by virtually reproducing the X-ray fluoroscopic image from the three-dimensional image. Attached Figure Description

[0012] Figure 1 This is a block diagram showing the general configuration of a radiotherapy system including the radiotherapy apparatus of the embodiments.

[0013] Figure 2 This is a diagram illustrating the projection matrix used in the calculation and processing of the projection position in the implementation method.

[0014] Figure 3A This diagram illustrates the generation of DRR using conventional ray tracing methods.

[0015] Figure 3B This diagram illustrates the scenario where a DRR is generated by the DRR generation unit of the embodiment.

[0016] Figure 4 This is a functional block diagram showing the general structure of the DRR generation unit in the implementation method.

[0017] Figure 5 This is a flowchart illustrating an example of the processing flow of a radiotherapy system according to an embodiment.

[0018] Figure 6 This is a flowchart illustrating an example of the DRR generation process of the DRR generation unit in an implementation embodiment.

[0019] Figure 7 This diagram illustrates the situation where the DRR generation unit of the embodiment generates an element projection image.

[0020] Figure 8 This is a schematic diagram of the DRR generated by the DRR generation unit of the implementation method.

[0021] Figure 9 This is a diagram showing the experimental results of the positioning treatment of the radiation therapy device of the embodiment and the device of the comparative example. Detailed Implementation

[0022] Hereinafter, with reference to the accompanying drawings, the radiotherapy apparatus, medical image processing apparatus, radiotherapy method, and procedure of the embodiments will be described.

[0023] Figure 1 This is a block diagram illustrating the general configuration of a radiotherapy system including the radiotherapy apparatus of the embodiments. The radiotherapy system 1, for example, includes a treatment table 10, two radiation sources 20 (radiation source 20-1 and radiation source 20-2), two radiation detectors 30 (radiation detector 30-1 and radiation detector 30-2), a treatment beam irradiation gate 40, and a radiotherapy apparatus 100. The radiotherapy apparatus 100 is an example of a "radiotherapy apparatus" or a "medical image processing apparatus."

[0024] Treatment table 10 is an examination table for placing and fixing a subject (patient) P who will receive radiation therapy. Treatment table 10 is equipped with a translation mechanism and a rotation mechanism for changing the direction of the treatment beam irradiated onto the fixed patient P. Treatment table 10 can move along three axes, i.e., along six axes, via the translation mechanism and the rotation mechanism, respectively.

[0025] Radiation source 20-1 irradiates the patient P from a predetermined angle with radiation r-1 for fluoroscopy. Radiation source 20-2 irradiates the patient P from a predetermined angle different from radiation source 20-1 with radiation r-2 for fluoroscopy. Radiation r-1 and radiation r-2 are, for example, X-rays. Figure 1 The image shows X-ray imaging of patient P, fixed on treatment table 10, from two directions. Additionally, in... Figure 1 The diagram of the control unit that controls the irradiation by radiation r from radiation source 20 is omitted.

[0026] Radiation detector 30-1 detects radiation r-1 that has irradiated from radiation source 20-1 and passed through the patient P's body, generating an X-ray fluoroscopic image of the patient P's body corresponding to the energy level of the detected radiation r-1. Radiation detector 30-2 detects radiation r-2 that has irradiated from radiation source 20-2 and passed through the patient P's body, generating an X-ray fluoroscopic image of the patient P's body corresponding to the energy level of the detected radiation r-2.

[0027] The radiation detector 30 includes a plurality of X-ray detectors arranged in a two-dimensional array. The radiation detector 30 generates a digital image representing the energy magnitude of the radiation r that reaches each X-ray detector as a digital value, as an X-ray fluoroscopic image. The radiation detector 30 is, for example, a Flat Panel Detector (FPD). The radiation detectors 30-1 and 30-2 output the generated X-ray fluoroscopic images T1 and T2 to the radiation therapy apparatus 100, respectively. In addition, in Figure 1 , the illustration of the control unit that controls the generation of the X-ray fluoroscopic image using the radiation detector 30 is omitted.

[0028] In the radiation therapy system 1, the positions of the radiation source 20 and the radiation detector 30 are fixed, so the imaging direction of the imaging device constituted by the set of the radiation source 20 and the radiation detector 30 (the direction relative to the fixed coordinate system of the treatment room) is fixed. Therefore, when three-dimensional coordinates are defined in the three-dimensional space where the radiation therapy system 1 is installed, the positions of the radiation source 20 and the radiation detector 30 can be represented by three-axis coordinate values. In the following description, the information of the three-axis coordinate values is referred to as imaging system geometry information of the imaging device constituted by the set of the radiation source 20 and the radiation detector 30. The imaging system geometry information includes information such as the position of the radiation source 20, the position and the inclination of the radiation detector 30. By using the imaging system geometry information, the position of the patient P in the prescribed three-dimensional coordinates can be obtained from the position where the radiation emitted from the radiation source 20 passes through the inside of the patient P and reaches the radiation detector 30.

[0029] The imaging system geometry information can be obtained from the designed installation positions of the radiation source 20 and the radiation detector 30 when the radiation therapy system 1 is installed. Alternatively, the geometry information can also be obtained from the installation positions of the radiation source 20 and the radiation detector 30 measured by a three-dimensional measuring instrument or the like. By obtaining a projection matrix in advance from the imaging system geometry information, the radiation therapy apparatus 100 can calculate at which position (projection position) in the captured two-dimensional fluoroscopic image the patient P in the three-dimensional space is imaged (to which position on the DRR each point in the three-dimensional space is projected).

[0030] Figure 2 is a diagram for explaining the projection matrix used in the projection position calculation processing of the embodiment. The projection matrix P is a matrix representing the correspondence when a point in the three-dimensional space is projected onto a two-dimensional fluoroscopic image. The point X(→)=(X,Y,Z) in the three-dimensional space t and the point u=(u,v) on the two-dimensional fluoroscopic image that is the projection targett The relationship (→) represents a vector) is represented by the following equation (1).

[0031] [Mathematical Expression 1]

[0032]

[0033] The projection matrix P is represented by the following equations (2) and (3). In equations (2) and (3), the position of the radiation source 20 is set as L(→)=(l X ,l Y ,l Z ) t Let the basis vector of the radiation detector 30 (FPD) be u(→)=(u X ,u Y ,u) t v(→)=(v X ,v,v Z ) t w(→)=(w X ,w Y ,w) t Let c(→) = (c u ,c v ) t Let f be the distance from L(→) to c(→), and let su[mm / pixel] and sv[mm / pixel] be the pixel spacing of the radiation detector 30.

[0034] [Mathematical Expression 2]

[0035]

[0036] [Mathematical Expression 3]

[0037] R = [uvw] (3)

[0038] In addition, filming at the same time Figure 1 In the imaging device for two fluoroscopic images of patient P as shown, a projection matrix is ​​pre-calculated for each group of radiation source 20 and radiation detector 30. Therefore, based on the positions of lesions, bones, or other affected areas within patient P captured in the two fluoroscopic images, or the positions of pre-placed markers within patient P, coordinates representing the positions of the affected areas or markers in a predetermined three-dimensional coordinate system can be calculated.

[0039] In addition, Figure 1The diagram illustrates the configuration of a radiotherapy system 1 comprising two sets of radiation sources 20 and radiation detectors 30, i.e., two imaging devices. The radiotherapy system 1 may have three or more imaging devices (three or more sets of radiation sources 20 and radiation detectors 30). Alternatively, the radiotherapy system 1 may also have only one imaging device (one set of radiation sources 20 and radiation detectors 30).

[0040] The therapeutic beam irradiation gate 40 irradiates the target area (the affected area) within the patient P with radiation as the therapeutic beam B. The therapeutic beam B can be, for example, X-rays, gamma rays, electron beams, proton beams, neutron beams, heavy particle beams, etc. The therapeutic beam B is delivered to the patient P in a straight line from the therapeutic beam irradiation gate 40. Additionally, in... Figure 1 The diagram shows the configuration of a radiation therapy system 1 with a fixed treatment beam irradiation gate 40, but it is not limited to this. The radiation therapy system 1 may also have multiple treatment beam irradiation gates.

[0041] The radiotherapy apparatus 100 controls the operation of various functions of the radiotherapy system 1. The radiotherapy apparatus 100 includes, for example, an input interface 110, a display unit 120, a storage unit 130, and a control unit 140. Furthermore, these functional units can be distributed across multiple devices. For example, the DRR generation function in the control unit 140 can be implemented by a processing unit separate from the radiotherapy apparatus 100. This processing unit is an example of a "medical image processing apparatus".

[0042] Input interface 110 accepts various input operations from practitioners (doctors, technicians, etc.) performing radiation therapy using the radiation therapy system 1, and outputs a signal indicating the accepted input operation to control unit 140. Input interface 110 may be, for example, a keyboard, mouse, touch panel, etc.

[0043] Display unit 120 displays information such as CT images, DRR, X-ray fluoroscopy images, the current position of patient P, and a predetermined appropriate position for radiotherapy (hereinafter referred to as "appropriate position"). Display unit 120 is, for example, a liquid crystal display (LCD). When the input interface 110 is implemented by a touch panel, the functions of display unit 120 can be incorporated into the touch panel.

[0044] Storage unit 130 stores various information required for radiotherapy. For example, storage unit 130 stores three-dimensional images of the patient P obtained during the treatment planning stage, allowing for visualization within the body. These three-dimensional images are, for example, three-dimensional image data obtained by imaging devices such as CT scanners, cone-beam CT scanners, and magnetic resonance imaging (MRI) devices. In the following description, we will use the case where the three-dimensional image is a CT image D1 obtained by imaging the patient P with a CT scanner as an example. In addition, storage unit 130 stores, for example, treatment planning information D2 such as the irradiation position, irradiation direction, irradiation level, and number of irradiations for each patient's radiation beam B, determined during the treatment planning stage, as well as imaging system geometry information D3. Storage unit 130 is implemented, for example, using RAM (Random Access Memory), ROM (Read Only Memory), or HDD (Hard Disk Drive).

[0045] The control unit 140 controls the actions used to realize the various functions of the radiotherapy system 1.

[0046] The control unit 140 includes, for example, a first acquisition unit 151, a second acquisition unit 153, a DRR generation unit 155, a positioning unit 157, an examination bed control unit 159, an irradiation control unit 161, and a display control unit 163.

[0047] The first acquisition unit 151 acquires the CT image D1 of patient P, the treatment plan information D2 of patient P, and the imaging system geometry information D3 from the storage unit 130. Additionally, the first acquisition unit 151 can acquire the CT image D1, etc., based on information input via the input interface 110. Furthermore, the first acquisition unit 151 can acquire the CT image D1, etc., from a database (file server, etc.) connected via a network. Furthermore, the first acquisition unit 151 can acquire the CT image D1 from a storage medium such as a DVD or CD-ROM via a drive device installed in the radiotherapy apparatus 100. In other words, the first acquisition unit 151 acquires the conditions for X-ray imaging during the treatment phase, as well as three-dimensional images of the patient taken before the treatment phase. The first acquisition unit 151 is an example of an "acquisition unit".

[0048] The second acquisition unit 153 acquires X-ray fluoroscopic images T1 and T2 input from radiation detectors 30-1 and 30-2 during the treatment phase.

[0049] The DRR generation unit 155 generates a DRR based on the CT image D1 acquired by the first acquisition unit 151 and the imaging system geometric information D3. Figure 3AThis diagram illustrates the generation of DRR using conventional ray tracing methods. Figure 3B This diagram illustrates the situation where a DRR is generated by the DRR generation unit 155 of the embodiment.

[0050] like Figure 3A As shown, in conventional ray tracing methods, a CT image D1 is virtually positioned between the radiation source 20 and the DRR. The brightness value of each pixel in the DRR can be obtained by integrating the brightness values ​​of each pixel PX in the CT image D1 along the path of the X-ray connecting the radiation source 20 and the pixel. In this case, sampling is performed at short intervals along the X-ray path, and the brightness values ​​of the CT image D1 are summed. That is, referring to the brightness of the pixels in the CT image D1 through which the X-ray passes, integration processing is required for each pixel of the DRR, resulting in a large computational load. Increasing the number of samples by shortening the sampling interval can generate a high-precision DRR, but the processing time increases with the number of samples. Therefore, there is a trade-off between the image quality of the DRR and the processing time. To generate a DRR with sufficient image quality for positioning, a sampling interval smaller than the pixel spacing of the CT image D1 is desired.

[0051] On the other hand, such as Figure 3B As shown, in the DRR generation process using the DRR generation unit 155, instead of utilizing the X-ray path information, the DRR is generated based on the projection position information of the DRR projected onto each pixel of the CT image D1 and the information of the projected pixels (hereinafter referred to as "feature projection images"). Since the DRR is the image after projecting the CT image D1, the DRR generation unit 155 can generate the DRR by superimposing the feature projection images of all pixels of the CT image D1. For example, in... Figure 3B In the example shown, the DRR generation unit 155 generates a feature projection image EP1 corresponding to a representative pixel PX1 (hereinafter referred to as the "reference pixel") in the CT image D1. The generated feature projection image EP1 is then transformed in two dimensions to generate feature projection images (feature projection images EP2, etc.) corresponding to other pixels. Furthermore, when generating DRR using these feature projection images, unlike the case using ray tracing, the computational load depends only on the number of pixels in the CT image D1. Therefore, the processing time for generating DRR can be shortened.

[0052] Figure 4 This is a functional block diagram showing the general configuration of the DRR generation unit 155 in the implementation method.

[0053] The DRR generation unit 155 includes, for example, a projection position calculation unit 201, an element projection image generation unit 203, and an element projection image synthesis unit 205.

[0054] The projection position calculation unit 201 calculates the projection position of each pixel in the CT image D1 when it is projected onto the DRR, based on the imaging system geometric information D3. Based on the treatment plan, information such as three-dimensional position and rotation angle is set in the CT image D1. The projection position calculation unit 201 calculates the projection position of the three-dimensional image coordinate system x(→)=(x,y,z) set in the CT image D1 using the following equation (4). t Convert to room coordinate system X(→)=(X,Y,Z) t The system can calculate the position on the DRR where a point in the room coordinate system is projected based on the camera's geometric information D3. In equation (4), A is a predetermined transformation matrix set based on the camera's geometric information D3. Furthermore, the projection position calculation unit 201 calculates the DRR coordinate system u(→)=(u,v) based on the room coordinate system X(→) using the following equation (5). t In equation (5), P is the projection matrix.

[0055] [Mathematical Expression 4]

[0056]

[0057] [Mathematical Expression 5]

[0058]

[0059] That is, the projection position calculation unit 201 calculates the projection position of each pixel contained in the three-dimensional image when it is projected onto the two-dimensional X-ray perspective image generated by the X-ray imaging, based on the conditions of X-ray imaging.

[0060] The feature projection image generation unit 203 generates feature projection images when each pixel of the CT image D1 is projected onto the DRR. However, generating accurate feature projection images for all pixels contained in the CT image D1 requires processing time. Therefore, the feature projection image generation unit 203 first generates a feature projection image for a reference pixel, performs a two-dimensional transformation on the generated feature projection image, thereby approximating and generating feature projection images for other pixels. That is, the feature projection image generation unit 203 generates a feature projection image for each pixel when each pixel of the three-dimensional image is projected onto the X-ray fluoroscopic image. The feature projection image generation unit 203 generates a feature projection image for a reference pixel contained in the three-dimensional image, performs a two-dimensional transformation on the generated feature projection image of the reference pixel, and generates feature projection images for other pixels in the three-dimensional image besides the reference image.

[0061] The feature projection image compositing unit 205 generates a DRR by pasting the feature projection images generated by the feature projection image generation unit 203 onto the projection position and compositing them. Basically, the size of the feature projection image is one pixel or more, and multiple feature projection images overlap with each pixel of the DRR. Therefore, the feature projection image compositing unit 205 adds the brightness values ​​during compositing. That is, the feature projection image compositing unit 205 composites the feature projection images generated for each pixel based on the calculated projection position, thereby generating a reconstructed image obtained by virtually reproducing the X-ray fluoroscopic image from the three-dimensional image. The details of the processing by the projection position calculation unit 201, the feature projection image generation unit 203, and the feature projection image compositing unit 205 will then be described.

[0062] Back Figure 1 The positioning unit 157 compares the DRR generated by the DRR generation unit 155 with the X-ray fluoroscopic images T1 and T2 acquired by the second acquisition unit 153 to determine the appropriate position of the patient P for radiotherapy. Then, the positioning unit 157 calculates the amount of movement required to move the current position of the patient P, fixed on the treatment table 10, to the appropriate position for radiotherapy. In other words, the positioning unit 157 calculates the amount of movement of the treatment table 10 required to irradiate the treatment beam B from the current position of the patient P towards the treatment site from the irradiation direction predetermined relative to the CT image D1 during the planning stage. The positioning unit 157 outputs the calculated movement amount to the examination table control unit 159. That is, the positioning unit 157 positions the patient based on the generated reconstructed image.

[0063] Based on the movement information output by the positioning unit 157, the examination table control unit 159 controls the translation and rotation mechanisms installed on the treatment table 10 to change the position and orientation of the patient P fixed on the treatment table 10. The examination table control unit 159 outputs a signal S1 indicating the movement amount to the treatment table 10. For example, the examination table control unit 159 controls the translation and rotation mechanisms of the treatment table 10 along three axes, i.e., six axes, respectively.

[0064] The irradiation control unit 161 controls the irradiation of the treatment beam B using the treatment beam irradiation gate 40. Based on the treatment plan information D2 acquired by the first acquisition unit 151 and the X-ray fluoroscopic images T1 and T2 acquired in real time by the second acquisition unit 153 during the treatment phase, the irradiation control unit 161 outputs a signal S2 indicating the irradiation timing of the treatment beam B to the treatment beam irradiation gate 40.

[0065] The display control unit 163 controls the display unit 120 to display CT images, DRR, X-ray fluoroscopy images, the current position of patient P, appropriate position and other information.

[0066] Some or all of the functions of the control unit 140 of the aforementioned radiotherapy device 100 may include, for example, a hardware processor such as a CPU (Central Processing Unit) and a storage device (a storage device with a non-volatile storage medium) storing programs (software). Various functions can be implemented by the processor executing the programs. Furthermore, some or all of the functions of the control unit 140 of the aforementioned radiotherapy device 100 may be implemented using hardware such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), GPU (Graphics Processing Unit), etc. (including the circuitry section), or various functions may be implemented through the cooperation of software and hardware. Additionally, some or all of the functions of the control unit 140 of the aforementioned radiotherapy device 100 may also be implemented using a dedicated LSI. The program (software) can be stored in the storage unit 130, or it can be stored on a removable storage medium (non-volatile storage medium) such as a DVD or CD-ROM. It is installed in the storage unit 130 by installing the storage medium into the drive device of the radiation therapy system 1. Alternatively, the program (software) can be pre-downloaded from other computer devices via a network and installed in the storage unit 130.

[0067] Next, the processing of the radiation therapy system 1 will be explained. Figure 5 This is a flowchart illustrating an example of the processing flow of the radiotherapy system according to an embodiment. Furthermore, in the following description, during the treatment planning stage, a CT image D1 of patient P taken by a CT device and treatment plan information D2 are pre-stored in the storage unit 130.

[0068] First, the first acquisition unit 151 acquires the CT image D1 of the patient P, the target of treatment, from the storage unit 130 (step S101). The first acquisition unit 151 outputs the acquired CT data D1 to the DRR generation unit 155.

[0069] Next, the second acquisition unit 153 acquires the current X-ray fluoroscopic image of the patient P output by the radiation detector 30 (step S103). The second acquisition unit 153 outputs the acquired X-ray fluoroscopic image to the positioning unit 157.

[0070] Next, the DRR generation unit 155 and the positioning unit 157 begin sparse search processing of the position (hereinafter referred to as "CT position") of the CT image D1 virtually configured in the three-dimensional space of the treatment room. In the sparse search processing of the CT position, the DRR generation unit 155 generates a DRR based on the CT image D1 output by the first acquisition unit 151 (step S105). The DRR generation unit 155 outputs the generated DRR to the positioning unit 157. The details of the DRR generation process using the DRR generation unit 155 will then be described.

[0071] Next, the positioning unit 157 searches for the CT location with the highest similarity to the current DRR and X-ray fluoroscopy image based on the DRR output by the DRR generation unit 155 and the X-ray fluoroscopy image output by the second acquisition unit 153 (step S107).

[0072] Next, the positioning unit 157 determines whether the positional offset of the patient P at the searched CT position is within a specified range (step S109). The positional offset represents the offset between the CT position of the CT image D1 (the position of the patient P in the CT image D1) and the current position of the patient P fixed on the treatment table 10.

[0073] If the positioning unit 157 determines that the positional offset of patient P at the searched CT location is not within a specified range, it outputs the information of the searched CT location to the DRR generation unit 155, causing the process to return to step S105. Therefore, in step S105, the DRR generation unit 155 generates a new DRR based on the CT location information output by the positioning unit 157. In step S107, the positioning unit 157 searches for the CT location with the highest similarity to the new DRR and X-ray fluoroscopy image generated by the DRR generation unit 155. Thus, the DRR generation unit 155 and the positioning unit 157 cooperate to repeat the sparse search process for CT locations until the positional offset of patient P at the searched CT location falls within a specified range, i.e., the similarity between the DRR and the X-ray fluoroscopy image becomes higher than a specified similarity threshold.

[0074] On the other hand, in step S109, if it is determined that the positional offset of patient P at the searched CT location is within a specified range, the DRR generation unit 155 and the positioning unit 157 begin a dense search process to search for the CT location with the least positional offset of patient P in greater detail. During the dense search of CT locations, the DRR generation unit 155 generates a DRR based on the CT location searched in the sparse search process (step S111). The DRR generation unit 155 outputs the generated DRR to the positioning unit 157.

[0075] Next, the positioning unit 157, using the CT position found in the sparse search process as a reference, searches for the final CT position based on the DRR output by the DRR generation unit 155 and the X-ray fluoroscopic image output by the second acquisition unit 153 (step S113). For example, the positioning unit 157 searches for the CT position with the least positional offset of the patient P by moving the CT position along rotational and translational directions based on the three-dimensional coordinates of the treatment room, based on the DRR and X-ray fluoroscopic image with the CT position found in the sparse search process. In other words, the positioning unit 157 moves the CT position according to six parameters representing the amount of rotation and translation based on the three-dimensional coordinates of the treatment room, and searches for the CT position with the highest similarity between the DRR and the X-ray fluoroscopic image.

[0076] Next, based on the final CT position found, the positioning unit 157 calculates the amount of movement (six control parameters) for rotating and translating the treatment table 10 with reference to the three-dimensional coordinates in the treatment room (step S115). The positioning unit 157 outputs the calculated amount of movement to the examination table control unit 159.

[0077] Next, the examination table control unit 159 moves the treatment table 10 according to the movement amount output by the positioning unit 157 (step S117). Afterwards, the irradiation control unit 161 controls the treatment beam irradiation gate 40, causing the treatment beam B to irradiate the affected area of ​​the patient P. This completes the processing of this flowchart.

[0078] Next, the details of the DRR generation process in steps 105 and 111 above will be explained. Figure 6 This is a flowchart illustrating an example of the DRR generation process of the DRR generation unit 155 in an embodiment.

[0079] First, the projection position calculation unit 201 calculates the projection position of each pixel of the CT image D1 when it is projected onto the DRR based on the imaging system geometry information D3 (step S201). The projection position calculation unit 201 converts the image coordinate system set in the CT image D1 into the room coordinate system, multiplies the room coordinate system by the projection matrix based on the imaging system geometry information D3, and thereby calculates the DRR coordinate system as the position on the DRR.

[0080] Next, the feature projection image generation unit 203 generates feature projection images of each pixel of the CT image D1 when projected onto the DRR (step S203). Even if the pixels contained in the CT image D1 are all of the same shape, the feature projection image will change when their positions change in three-dimensional space. If a rigorous calculation is performed, an accurate feature projection image needs to be generated for all pixels contained in the CT image D1, but the computational cost is high and the DRR cannot be generated quickly. Therefore, the feature projection image generation unit 203 first generates the feature projection image corresponding to the reference pixel, transforms the generated feature projection image in two dimensions, and then performs feature projection images on pixels other than the reference pixel.

[0081] Next, the feature projection image synthesis unit 205 generates DRR by pasting multiple feature projection images generated by the feature projection image generation unit 203 onto the projection position and synthesizing them (step S205).

[0082] Figure 7 This diagram illustrates the scenario where the DRR generation unit 155 of the embodiment generates a feature projection image. The DRR generation unit 155 generates a feature projection image corresponding to a reference pixel, performs a two-dimensional transformation on the generated feature projection image, thereby generating feature projection images of other pixels, and pastes the generated feature projection images to the projection position and combines them to generate a DRR.

[0083] Specifically, this will be achieved by positioning the object at a point in three-dimensional space X(→) = (X,Y,Z). t The image generated by projecting one pixel of the CT image D1 onto the DRR plane (radiation detector 30) is denoted as feature projection image e(u,v). The center position of e(u,v) on the DRR is taken as the coordinate ec(→) = (eu,ev) projected onto the DRR from X(→). t In X(→) and ec(→), the following relationship (6) holds.

[0084] [Mathematical Expression 6]

[0085]

[0086] In equation (6) above, P is the projection matrix calculated from the shooting system geometry D3.

[0087] Consider I(u,v) generated by coinciding with e(u,v). The pixels of the CT image D1 are arranged stereoscopically in three-dimensional space, and the size of most feature projection images is larger than one pixel. Therefore, in general, there are multiple feature projection images that overlap with the coordinates (u,v) on the DRR. Therefore, I(u,v) is calculated by the following equation (7).

[0088] [Mathematical Expression 7]

[0089]

[0090] In the above equation (7), E uv It is the set of projected images of features that overlap with coordinates (u,v), w i h i It is E uv The image size of the i-th element projection image within the image, s eu [mm / pixel], s ev [mm / pixel] is the pixel spacing of the feature's projected image.

[0091] Next, the method for generating e(u,v) will be explained. Strictly speaking, it is necessary to generate feature projection images for all pixels of the CT image D1, but in this case, the computational load is the same as that required for generating DRR using conventional ray tracing methods. Therefore, the DRR generation unit 155 performs a two-dimensional transformation of the feature projection image corresponding to the reference pixel and approximates the feature projection images of other pixels, thereby simplifying the process. The reference pixel is, for example, the pixel that is the isocenter (Japanese: アイソセンタ) of the area where radiation is concentrated. Hereinafter, the case where the reference pixel is the isocenter will be explained as an example.

[0092] The brightness value of the element projection image is determined (proportional) to the brightness value V (X, Y, Z) of the base CT image D1. Therefore, by multiplying the brightness value of the element projection image of the isocentric pixel by a constant multiple, the brightness values ​​of the element projection images of other pixels can be calculated. That is, the element projection image generation unit 203 calculates the ratio of the brightness value of each other pixel to the brightness value of the isocentric pixel in the CT image D1. Then, the element projection image generation unit 203 multiplies the calculated ratio by the brightness value of the element projection image of the isocentric pixel, thereby calculating the brightness values ​​of the element projection images of other pixels.

[0093] Furthermore, the closer the pixels in the CT image D1 are to the radiation source 20, the larger the feature projection image; the closer they are to the radiation detector 30 (and the farther away from the radiation source 20), the smaller the feature projection image. In other words, compared to the position of isocentric pixels, the closer the positions of other pixels are to the radiation source 20, the larger the feature projection image; compared to the position of isocentric pixels, the closer the positions of other pixels are to the radiation detector 30 (DRR), the smaller the feature projection image. The size of the feature projection image, taking into account this tendency, can be geometrically calculated. Therefore, the feature projection image generation unit 203 can perform conversion considering the different positions of each pixel in the CT image D1 by enlarging or reducing the size of the feature projection image.

[0094] The ray tracing method will be used to locate the isocentric position (X). iso ,Y iso Ziso The pixel-generated feature projection image is set as the reference feature projection image e. iso (u,v). This will be applied to positions (X) other than the isocenter. i ,Y i Z i The pixel element projection image is obtained through e iso The approximation obtained by the two-dimensional transformation of (u,v) is denoted as e. i When (u,v), the conversion formula is represented by the following formulas (8), (9), and (10).

[0095] [Mathematical Expression 8]

[0096]

[0097] [Mathematical Expression 9]

[0098]

[0099] [Mathematical Expression 10]

[0100]

[0101] In equations (8) to (10) above, w and h are the dimensions of the projected images of each element. α is the ratio of pixel values ​​in the CT image D1, which forms the basis for the projected images of each element. The pixel values ​​of the CT image D1 are independent of pixel position. When only the dimensions of the reference element projected image are adjusted, the brightness values ​​of the element projected images of other pixels of the processing target (hereinafter also referred to as "pixels of interest") cannot be approximated. Therefore, correction is performed by the ratio of pixel values. λ is calculated using equation (6) above. i and λ iso β represents the ratio of the depth from the radiation source 20 to the pixel of interest to the depth from the radiation source 20 to the isocenter. The depth from the radiation source 20 to the pixel of interest is, for example, the distance W1 between the point obtained by drawing a perpendicular line from the position of the pixel of interest (e.g., pixel 1) to the line L1 connecting the radiation source 20 and the isocenter (see reference). Figure 7 The depth from the radiation source 20 to the isocenter is, for example, the straight-line distance W0 from the radiation source 20 to the isocenter (refer to...). Figure 7 ).

[0102] exist Figure 7In the example shown, firstly, the feature projection image generation unit 203 projects the isocenter pixels from the pixels contained in the CT image D1 to generate a reference feature projection image EP10. Next, the feature projection image generation unit 203 generates a feature projection image EP11 of pixel 1, which is another pixel in the pixels contained in the CT image D1. Here, pixel 1 is located closer to the radiation source 20 than the isocenter. Therefore, based on the aforementioned depth ratio, the feature projection image generation unit 203 generates a feature projection image EP11 (a magnified feature projection image) that is larger than the size of the reference feature projection image EP10. Furthermore, the feature projection image generation unit 203 calculates the brightness value of the feature projection image EP11 by multiplying the reference feature projection image EP10 by the ratio of the brightness value of pixel 1 to the pixel at the isocenter position.

[0103] Similarly, the element projection image generation unit 203 generates an element projection image EP12 of pixel 2, which is one of the other pixels in the CT image D1. Here, pixel 2 is located closer to the radiation detector 30 than the isocenter. Therefore, based on the aforementioned depth ratio, the element projection image generation unit 203 generates an element projection image EP12 that is a scaled-down version of the reference element projection image EP10. Furthermore, the element projection image generation unit 203 calculates the brightness value of the element projection image EP12 by multiplying the reference element projection image EP10 by the ratio of the brightness value of pixel 2 to the pixel at the isocenter position.

[0104] The feature projection image generation unit 203 also generates feature projection images for the remaining pixels contained in the CT image D1. The feature projection image compositing unit 205 can generate a composite image by pasting multiple feature projection images generated by the feature projection image generation unit 203 to the projection position and compositing them. Figure 8 The DRR as shown.

[0105] Specifically, the feature projection image generation unit 203 virtually positions the 3D image between the radiation source and the radiation detector used for X-ray imaging. When other pixels are closer to the radiation source than the reference pixel, it performs a conversion process that enlarges the feature projection image of the reference pixel to generate feature projection images of other pixels. When other pixels are closer to the radiation detector than the reference pixel, it performs a conversion process that reduces the feature projection image of the reference pixel to generate feature projection images of other pixels. Furthermore, the feature projection image generation unit 203 calculates the brightness value of the feature projection images of other pixels based on the ratio of the brightness value of the reference pixel to the brightness value of other pixels in the 3D image. The feature projection image generation unit 203 calculates the brightness value of the feature projection images of other pixels by multiplying the brightness value of the feature projection image of the reference pixel by the ratio of the brightness value of other pixels to the brightness value of the reference pixel.

[0106] As described above, since a strict transformation is impossible in two-dimensional conversion, the further a pixel is from the isocenter towards the radiation detector 30 in the horizontal direction, the greater the approximation error of the feature projection image. In patient localization, the affected area is located at the isocenter, which is the position where the highest accuracy of localization is desired. That is, if the feature projection image generated from the pixels at the isocenter is used as a reference to generate the DRR, as in this embodiment, the error near the isocenter can be minimized.

[0107] Figure 9 This diagram shows the experimental results of the positioning processing of the radiotherapy apparatus 100 of the embodiment and the apparatus of the comparative example. In this experiment, positioning processing (sparse search, dense search) was performed from an appropriate initial position, and the movement was calculated, respectively, using a computer with defined processing capabilities, generating a DRR using the element projection image of the embodiment, and generating a DRR using the conventional ray casting method of the comparative example. Figure 9 In this context, tx, ty, and tz represent the three-axis translational movements in a translational mechanism, while rx, ry, and rz represent the three-axis rotational movements in a rotary mechanism. For example... Figure 9 As shown, compared with the processing time of the conventional ray projection method using the comparative example, it was confirmed that the processing time can be significantly reduced when using the element projection image processing of the embodiment.

[0108] According to the implementation method described above, by generating element projection images based on three-dimensional images and synthesizing them, the generation and processing of DRR can be accelerated, and the patient can be located in a short time and with high accuracy.

[0109] Alternatively, when the CT image D1 is a cubic 3D image, the DRR generation unit 155 performs feature projection image generation processing after performing isotropic (converting the CT image D1 into a cube) processing to make the CT image D1 isotropic. Compared to a cuboid, a cube is more likely to produce similar feature projection images regardless of which angle it is projected from, thus suppressing deviations in feature projection images.

[0110] Furthermore, the DRR generation unit 155 generates a DRR by setting a pixel with a brightness of 1 at a reference position (e.g., the isocenter position) as a projection image of a single pixel. This reduces the image error near the isocenter required for high-precision DRR. In this case, the brightness value can be represented by a constant multiple, and the movement in the depth direction can be represented by scale changes. Furthermore, the pixel size of the CT image D1 can be reduced. Therefore, high-quality DRR can be generated.

[0111] According to at least one embodiment described above, the system includes: an acquisition unit (151) that acquires the conditions for X-ray imaging during the treatment phase and a three-dimensional image of the patient taken before the treatment phase; a projection position calculation unit (201) that calculates the projection position of each pixel contained in the three-dimensional image when it is projected onto a two-dimensional X-ray fluoroscopic image generated by the X-ray imaging, based on the conditions for X-ray imaging; an element projection image generation unit (203) that generates an element projection image of each pixel when it is projected onto the X-ray fluoroscopic image; and an element projection image synthesis unit (205) that synthesizes the generated element projection images of each pixel based on the calculated projection position, thereby generating a reconstructed image (DRR) obtained by virtually reproducing the X-ray fluoroscopic image based on the three-dimensional image. This enables the DRR generation process to be accelerated and the patient to be located quickly and with high accuracy.

[0112] Several embodiments of the present invention have been described, but these embodiments are merely illustrative and are not intended to limit the scope of the invention. These embodiments can be implemented in various other ways, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are equally included within the scope of the invention and its equivalents as defined in the claims.

[0113] Explanation of reference numerals in the attached figures

[0114] 1: Radiotherapy system; 10: Treatment table; 20, 20-1, 20-2: Radiation sources; 30, 30-1, 30-2: Radiation detectors; 40: Treatment beam irradiation gate; 100: Radiotherapy device; 110: Input interface; 120: Display unit; 130: Storage unit; 140: Control unit; 151: First acquisition unit; 153: Second acquisition unit; 155: DRR generation unit; 157: Positioning unit; 159: Examination bed control unit; 161: Irradiation control unit; 163: Display control unit; 201: Projection position calculation unit; 203: Element projection image generation unit; 205: Element projection image synthesis unit.

Claims

1. A radiation therapy device, wherein, have: The acquisition unit acquires the conditions for X-ray imaging during the treatment phase, as well as three-dimensional images of the patient taken prior to the treatment phase. The projection position calculation unit calculates the projection position of each pixel in the three-dimensional image when it is projected onto the two-dimensional X-ray fluoroscopic image generated by the X-ray imaging, based on the conditions of the X-ray imaging. The element projection image generation unit generates an element projection image of each pixel when the individual pixels contained in the three-dimensional image are projected onto the X-ray fluoroscopic image; as well as The feature projection image synthesis unit synthesizes the feature projection images of each generated pixel based on the calculated projection positions, thereby generating a reconstructed image obtained by virtually reproducing the X-ray fluoroscopic image based on the three-dimensional image. The element projection image generation unit Generate feature projection images of the reference pixels contained in the three-dimensional image. By performing a two-dimensional transformation on the generated feature projection image of the reference pixel, the feature projection image of the reference pixel is converted into feature projection images of other pixels besides the reference pixel contained in the three-dimensional image, thereby generating feature projection images of the other pixels from the feature projection image of the reference pixel.

2. The radiation therapy device according to claim 1, wherein, It also has: The positioning unit locates the patient based on the generated reconstructed image.

3. The radiation therapy device according to claim 1 or 2, wherein, The reference pixel is the pixel at the isocenter position in radiotherapy.

4. The radiation therapy device according to claim 1 or 2, wherein, The element projection image generation unit The three-dimensional image is virtually positioned between the radiation source and the radiation detector that perform the X-ray imaging. If the other pixels are closer to the radiation source than the reference pixel, a transformation process is performed to magnify the feature projection image of the reference pixel to generate feature projection images of the other pixels. If the other pixels are closer to the radiation detector than the reference pixel, a transformation process is performed to reduce the feature projection image of the reference pixel to generate feature projection images of the other pixels.

5. The radiation therapy device according to claim 1 or 2, wherein, The element projection image generation unit calculates the brightness value of the element projection image of the other pixels based on the ratio of the brightness value of the reference pixel in the three-dimensional image to the brightness value of the other pixels.

6. The radiation therapy device according to claim 5, wherein, The element projection image generation unit calculates the brightness value of the element projection image of the other pixels by multiplying the brightness value of the element projection image of the reference pixel by the ratio of the brightness value of the other pixels to the brightness value of the reference pixel.

7. A medical image processing device, wherein, have: The acquisition unit acquires the conditions for X-ray imaging during the treatment phase, as well as three-dimensional images of the patient taken prior to the treatment phase. The projection position calculation unit calculates the projection position of each pixel in the three-dimensional image when it is projected onto the two-dimensional X-ray fluoroscopic image generated by the X-ray imaging, based on the conditions of the X-ray imaging. The element projection image generation unit generates an element projection image of each pixel when the individual pixels contained in the three-dimensional image are projected onto the X-ray fluoroscopic image; as well as The feature projection image synthesis unit synthesizes the feature projection images of each generated pixel based on the calculated projection positions, thereby generating a reconstructed image obtained by virtually reproducing the X-ray fluoroscopic image based on the three-dimensional image. The element projection image generation unit Generate feature projection images of the reference pixels contained in the three-dimensional image. By performing a two-dimensional transformation on the generated feature projection image of the reference pixel, the feature projection image of the reference pixel is converted into feature projection images of other pixels besides the reference pixel contained in the three-dimensional image, thereby generating feature projection images of the other pixels from the feature projection image of the reference pixel.

8. A radiation therapy method, wherein, The computer performs the following processing: The conditions for obtaining X-ray images during the treatment phase, as well as three-dimensional images of the patient taken prior to the treatment phase, Based on the conditions of the X-ray imaging, the projection positions of each pixel in the three-dimensional image when projected onto the two-dimensional X-ray fluoroscopic image generated by the X-ray imaging are calculated. The generation of the elemental projection image of each pixel when the individual pixels contained in the three-dimensional image are projected onto the X-ray fluoroscopic image. Based on the calculated projection positions, the element projection images of each generated pixel are synthesized to generate a reconstructed image obtained by virtually reproducing the X-ray fluoroscopic image from the three-dimensional image. Generating the feature projection image includes: Generate feature projection images of the reference pixels contained in the three-dimensional image; and By performing a two-dimensional transformation on the generated feature projection image of the reference pixel, the feature projection image of the reference pixel is converted into feature projection images of other pixels besides the reference pixel contained in the three-dimensional image, thereby generating feature projection images of the other pixels from the feature projection image of the reference pixel.

9. A storage medium storing a program, wherein, The program causes the computer to perform the following processing: The conditions for obtaining X-ray images during the treatment phase, as well as three-dimensional images of the patient taken prior to the treatment phase, Based on the conditions of the X-ray imaging, the projection positions of each pixel in the three-dimensional image when projected onto the two-dimensional X-ray fluoroscopic image generated by the X-ray imaging are calculated. The generation of the elemental projection image of each pixel when the individual pixels contained in the three-dimensional image are projected onto the X-ray fluoroscopic image. Based on the calculated projection positions, the element projection images of each generated pixel are synthesized to generate a reconstructed image obtained by virtually reproducing the X-ray fluoroscopic image from the three-dimensional image. Generating the feature projection image includes: Generate feature projection images of the reference pixels contained in the three-dimensional image; and By performing a two-dimensional transformation on the generated feature projection image of the reference pixel, the feature projection image of the reference pixel is converted into feature projection images of other pixels besides the reference pixel contained in the three-dimensional image, thereby generating feature projection images of the other pixels from the feature projection image of the reference pixel.

Images