Method and apparatus for optimizing beam angle for proton beam therapy

Abstract

A method for optimizing the beam angle for proton beam therapy includes the steps of: inputting distance information from the patient's body surface to the tumor, information on internal heterogeneity, information on the passage of the proton beam through the body, and information on the patient's internal movement associated with respiration into an analyzer; inputting the distance information from the patient's body surface to the tumor, information on internal heterogeneity, information on the passage of the proton beam through the body, and information on the patient's internal movement associated with respiration into a deep learning model; and calculating the beam angle for proton beam therapy based on the output values ​​of the deep learning model.

Description

Technical Field

[0001] The technology of the present disclosure relates to a method and apparatus for optimizing the proton beam angle during proton beam therapy by utilizing deep learning technology.

Background Art

[0002] Radiation therapy refers to a treatment method that uses radiation such as X-rays, gamma rays, electron beams, and proton beams to destroy tumors. Among radiation therapies, proton beam therapy is a treatment method that uses proton beams to destroy cancer tissues. Proton beam therapy is a treatment method that utilizes the principle of the Bragg Peak, which can minimize damage to normal tissues and destroy tumors using proton beams. The principle of the Bragg Peak refers to the phenomenon in which a proton beam releases a huge amount of radiation energy as soon as it reaches a tumor after passing through normal tissues.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Proton beam therapy may be greatly affected by changes in internal organs due to the Bragg Peak characteristics. Therefore, before performing proton beam therapy, an optimal treatment plan should be established considering the movement of organs so that the radiation for treatment can irradiate the tumor as planned to the maximum extent. Specifically, a treatment plan must be established to minimize the damage caused by radiation to the risk organs around the tumor during radiation therapy and to ensure that the maximum dose of radiation is incident on the tumor site.

[0005] In particular, determining the angle of the proton beam is extremely important when creating a treatment plan in proton therapy. For example, if there are many heterogeneous materials in the path the proton beam travels, the range of the proton beam will vary, which increases the likelihood that the proton beam will not be delivered effectively as planned. The same is true if there are organ regions in the path the proton beam travels that move in accordance with the patient's respiration. In such cases, problems may arise where the proton beam is directed at normal organs instead of the tumor, or where the tumor is not adequately irradiated, leading to a recurrence of cancer in the treated area.

[0006] However, determining the angle of the proton beam requires a very high level of clinical experience and knowledge, and there was a problem in that it took a lot of time to determine the beam angle.

[0007] The technology disclosed herein aims to provide a method for optimizing the proton beam angle during proton therapy by utilizing deep learning technology. [Means for solving the problem]

[0008] A method for optimizing the beam angle for proton beam therapy includes the steps of: inputting distance information from the patient's body surface to the tumor, information on internal heterogeneity, information on the passage of the proton beam through the body, and information on the patient's internal movement associated with respiration into an analyzer; inputting the distance information from the patient's body surface to the tumor, information on internal heterogeneity, information on the passage of the proton beam through the body, and information on the patient's internal movement associated with respiration into a deep learning model; and calculating the beam angle for proton beam therapy based on the output values ​​of the deep learning model. [Effects of the Invention]

[0009] The technology described herein can reduce the time spent on proton therapy planning and improve the quality of proton therapy planning. In this process, it is possible to reduce the waiting time for patients receiving proton therapy or to increase the number of patients receiving proton therapy. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows the overall process by which the analytical instrument (100) optimizes the proton beam angle. [Figure 2] This figure shows one embodiment (200) of a method for optimizing the proton beam angle. [Figure 3] This figure shows one embodiment of distance information from the patient's body surface to the tumor and information on internal body heterogeneity. [Figure 4] This figure shows one embodiment of information about bodily movement. [Figure 5] This figure shows a process for calculating a beam angle, which is a customized method for a patient undergoing proton beam therapy, according to one embodiment of the process. [Figure 6] This figure shows one embodiment of calculating the therapeutic effect of proton beam therapy by calculating the proton beam angle. [Figure 7] This figure shows another embodiment in which an analyzer calculates the proton beam angle for proton therapy. [Figure 8] This figure shows one embodiment of the analytical apparatus (300). [Modes for carrying out the invention]

[0011] The technology disclosed herein can be modified in various ways and may have many different embodiments. Specific embodiments of the technology disclosed herein may be described in the specification and drawings. However, these are merely illustrative of the technology disclosed herein and are not intended to limit the technology disclosed herein to any specific embodiment. Therefore, all modifications, equivalents, or substitutes included in the concept and scope of the technology disclosed herein should be understood as being included in the technology disclosed herein.

[0012] In the terms used below, singular expressions should be understood to include plural expressions unless the context clearly indicates otherwise, and terms such as “includes” should be understood to mean that the described features, quantities, stages, actions, components, parts, or combinations thereof exist, without prejudice to the possibility of the existence or addition of one or more other features, quantities, stages, actions, components, parts, or combinations thereof.

[0013] Before providing a detailed explanation of the drawings, it is important to clarify that the classification of components in this specification is merely based on the primary function that each component is responsible for. That is, two or more components described later may be combined into a single component, or a single component may be divided into two or more components based on more subdivided functions. Furthermore, each component described later may additionally perform some or all of the functions performed by other components, in addition to its own primary function, and some of the functions of the primary function that each component is responsible for may be exclusively performed by other components.

[0014] Furthermore, when performing a method or operation, each step constituting the method may be performed in a different order than specified, unless the context clearly indicates a specific order. That is, each step may be performed in the same order as specified, substantially simultaneously, or in the reverse order.

[0015] The following describes a beam angle calculation method customized for patients undergoing proton beam therapy.

[0016] In the following explanation, it is assumed that the analytical device optimizes the beam angle using a learning model. The analytical device can be implemented using a variety of devices capable of data processing. For example, the analytical device can be implemented using a PC, a server on a network, a smart device, or a chipset with a dedicated program installed.

[0017] Next, the overall process by which the analyzer executes a method for optimizing the proton beam angle will be described.

[0018] FIG. 1 shows the overall process by which the analyzer 100 optimizes the beam angle for proton beam therapy.

[0019] Distance information from the patient's body surface to the tumor, in-vivo heterogeneity information, in-vivo passage information of the proton beam, and in-vivo movement information associated with the patient's respiration can be input into the analyzer. The analyzer can input the distance information from the patient's body surface to the tumor, the in-vivo heterogeneity information, the in-vivo passage information of the proton beam, and the in-vivo movement information associated with the patient's respiration into the deep learning model. The analyzer can calculate the proton beam therapy beam angle based on the output value of the deep learning model. Proton beam therapy can be performed based on the proton beam angle for proton beam therapy calculated in this way.

[0020] Hereinafter, the method for optimizing the proton beam angle will be specifically described.

[0021] FIG. 2 shows an embodiment (200) of the method for optimizing the proton beam angle.

[0022] Distance information from the patient's body surface to the tumor, in-vivo heterogeneity information, in-vivo passage information of the proton beam, and in-vivo movement information associated with the patient's respiration can be input into the analyzer (210).

[0023] The distance information from the patient's body surface to the tumor can mean the distance (Radiological depth) through which the proton beam passes from the patient's body surface to the tumor. That is, the distance information from the patient's body surface to the tumor can mean how far the tumor is from the patient's body surface.

[0024] Internal heterogeneity information can include information about how homogeneous the body is as the proton beam passes through it. Specifically, internal heterogeneity information can be information calculated based on the standard deviation value of Hounsfield Units (HU) obtained from CT images.

[0025] Information about the passage of a proton beam within the body can indicate how well the proton beam passes through the body. Specifically, this information can indicate how much of the proton beam passes through a risk organ within the body. For example, this information may include the penetration weight of organs at risk (OAR), which indicates the degree to which the proton beam passes through a risk organ within the body.

[0026] Information on internal bodily movement associated with a patient's respiration can refer to information about the degree to which internal organs change as the patient breathes. Specifically, this information can include information calculated from deformation vector fields (DVF). This may be obtained from a four-dimensional CT scan in which the patient's respiratory information is stored. In one embodiment, the internal bodily movement information can include DVF values ​​calculated from a four-dimensional CT scan during respiratory phases 0 to 90. Alternatively, the internal bodily movement information can include DVF values ​​calculated from a four-dimensional CT scan during respiratory phases 0 to 50. Or, the internal bodily movement information can include DVF values ​​calculated from a four-dimensional CT scan during respiratory phases 50 to 90.

[0027] Distance information from the patient's body surface to the tumor, information on internal body heterogeneity, information on the passage of the proton beam through the body, and information on internal body movement associated with the patient's respiration can be obtained from CT (Computed Tomography) images. Specifically, this information may be obtained from 4D CT images that reflect the patient's movement associated with respiration.

[0028] The distance information from the patient's body surface to the tumor, the information on internal heterogeneity, the information on the passage of the proton beam through the body, and the information on internal movement associated with the patient's respiration may be information corresponding to the proton beam angle. For example, the distance information from the patient's body surface to the tumor, the information on internal heterogeneity, the information on the passage of the proton beam through the body, and the information on internal movement associated with the patient's respiration may be in the form of graphs, as shown in Figures 3 and 4. Figure 3(a) shows the distance information from the body surface to the tumor. Figure 3(b) shows the information on internal heterogeneity. Figure 3(c) shows the information on the passage of the proton beam through the body. Figure 4 shows the information on internal movement associated with respiration.

[0029] The analyzer can input information on the distance from the patient's body surface to the tumor, information on internal heterogeneity, information on the passage of the proton beam through the body, and information on internal movement associated with the patient's respiration into a deep learning model (220).

[0030] The deep learning model may be a model that has been trained to output the proton beam angle for proton therapy based on the training data.

[0031] In one embodiment, the deep learning model may include an artificial neural network-based model. For example, the deep learning model may include CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), RBM (Restricted Boltzmann Machine), DBN (Deep Belief Network), GAN (Generative Adversarial Network), RL (Relation Networks), and Transformer Neural Network.

[0032] The analyzer can calculate the proton beam angle for proton therapy based on the output values ​​of a deep learning model (230).

[0033] In one embodiment, the analyzer can calculate the proton beam angle that is deemed optimal for proton therapy within a 360-degree angle. For example, the analyzer can derive the portion of the 360-degree angle that is evaluated as having the highest therapeutic effect as a peak shape or a specific region of the predicted graph.

[0034] Next, we will describe one embodiment of how to construct a deep learning model.

[0035] First, data can be collected from patients who have received proton beam therapy. This patient data may include 4D CT images containing the patient's respiratory information.

[0036] Data from patients who have received proton beam therapy may be manipulated.

[0037] In one embodiment, heterogeneous region information of patient tissue can be extracted from a CT image. Heterogeneous region information may include air, bone, metallic materials (gold fiducials, stents, etc.), and tissue. Alternatively, specifically, regions containing organs or regions containing tumors can be separated in a CT image. A segmentation model can be used for this purpose. The segmentation model may be a pre-trained model that has been trained to segment heterogeneous regions in a CT image or to distinguish regions containing tumors. For example, the segmentation model may be a U-net-based or transformer-based model.

[0038] Alternatively, in one embodiment, the weight of the proton beam path through tumors and normal tissues can be adjusted based on CT images and organ segmentation information.

[0039] Alternatively, in one embodiment, after setting the center of mass position in the tumor information, the distance from the surface to the center of mass position can be calculated, and then the distance information from the body surface to the tumor can be calculated.

[0040] Deep learning models can be trained using processed patient data.

[0041] The learning process can proceed after the processed patient data has been converted to one-dimensional data.

[0042] After learning is complete, the model can be evaluated by comparing the derived beam angle with the results of actual treatments.

[0043] Figure 5 illustrates one embodiment of a beam angle calculation method customized for a patient undergoing proton therapy. As shown in Figure 5, the beam angle calculation method customized for a patient undergoing proton therapy may include contouring medical images, preprocessing the contouring results, and then training or testing a deep learning model such as a DCNN (Deep Convolutional Neural Network). Subsequently, a beam angle customized for the patient undergoing proton therapy can be determined based on the deep learning model such as the DCNN.

[0044] Figure 6 shows one embodiment in which the analytical instrument calculates the proton beam angle for proton therapy. After calculating the effect corresponding to the angle, the analytical instrument displayed this result in a graph.

[0045] The results predicted by the analytical instrument show a significant agreement between the results with high peaks and the actual treatment planning angles (reference data). This indicates that the beam angles for proton therapy predicted by the analytical instrument are reliable.

[0046] Figure 7 shows another embodiment in which the analyzer calculates the proton beam angle for proton therapy. After calculating the effect corresponding to the beam angle, the analyzer displays this in a graph. From the effects corresponding to the beam angle predicted by the analyzer, a beam angle that has an effect equal to or greater than a preset standard can be selected. Thus, as shown in Figure 7, three optimal beam angles (first angle, second angle, and third angle) can be selected. Proton therapy can then be performed based on the beam angles selected in this way.

[0047] Next, I will explain the analytical equipment.

[0048] Figure 8 shows the configuration of one embodiment of the analytical apparatus 300.

[0049] The analytical apparatus 300 may correspond to the analytical apparatus 100 described in Figure 1. The analytical apparatus 300 may also be an apparatus for performing the proton beam optimization method described above.

[0050] The analytical device 300 can be physically implemented in various forms. For example, the analytical device 300 can take the form of a PC, notebook computer, smart device, server, or dedicated data processing chipset.

[0051] The analysis device 300 may include an input device 310, a storage device 320, a calculation device 330, an output device 340, an interface device 350, and a communication device 360.

[0052] The input device 310 may include an interface device (such as a keyboard, mouse, or touchscreen) into which certain commands or data are input. The input device 310 may also include a configuration in which information is input via a separate storage device (such as a USB drive, CD drive, or hard disk drive). Data may be input to the input device 310 via a separate measuring device or via a separate database. Data may be input to the input device 310 via wired or wireless communication.

[0053] The input device 310 may receive information and models necessary to perform the proton beam optimization method described above. The input device 310 may receive distance information from the patient's body surface to the tumor, information on internal body heterogeneity, information on the proton beam's passage through the body, and information on internal body movement associated with the patient's respiration. The input device 310 may also receive a deep learning model.

[0054] The storage device 320 can store information input via the input device 310. The storage device 320 can store information generated during the calculation process performed by the computing device 330. In other words, the storage device 320 can include memory. The storage device 320 can store the results calculated by the computing device 330.

[0055] The memory device 320 can store the information and models necessary to perform the aforementioned proton beam optimization method. The memory device 320 can store distance information from the patient's body surface to the tumor, information on internal heterogeneity, information on the passage of the proton beam through the body, and information on internal movement associated with the patient's respiration.

[0056] The arithmetic unit 330 may be a device such as a chip incorporating a processor, AP, and program that processes data and performs certain calculations. The arithmetic unit 330 can generate control signals to control the analysis device 300.

[0057] The computing device 330 can perform the calculations necessary to execute the proton beam optimization method described above. The computing device 330 can input distance information from the patient's body surface to the tumor, information on internal body heterogeneity, information on the passage of the proton beam through the body, and information on internal body movement associated with the patient's respiration into the deep learning model. Based on the output values ​​of the deep learning model, the computing device 330 can calculate the beam angle for proton therapy.

[0058] The output device 340 may be a device that outputs certain information. The output device 340 can also output interfaces necessary for the data processing process, input data, analysis results, etc. The output device 340 can be realized in a variety of physical forms, such as a display or a device that outputs documents.

[0059] The interface device 350 may be a device that receives certain commands or data from an external source. Information and models necessary for executing the proton beam optimization method may be input to the interface device 350 from a physically connected input device or an external storage device. Control signals for controlling the analyzer 300 may be input to the interface device 350. The interface device 350 can output the results analyzed by the analyzer 300.

[0060] The communication device 360 ​​can mean a configuration for sending and receiving certain information via a wired or wireless network. The communication device 360 ​​can receive control signals necessary for controlling the analysis device 300. The communication device 360 ​​can transmit the results of the analysis performed by the analysis device 300.

[0061] The aforementioned method for optimizing the beam angle for proton therapy can be implemented in a program (or application) that includes a computer-executable algorithm.

[0062] The program may be provided stored on a temporary or non-transitory computer-readable medium.

[0063] Non-temporary computer-readable storage media refers to storage media that store data semi-permanently and are readable by devices, rather than media that store data for short periods, such as registers, caches, and memory. Specifically, the various applications or programs mentioned above may be provided stored on non-temporary computer-readable storage media such as CDs, DVDs, hard disks, Blu-ray discs, USB memory sticks, memory cards, ROMs (read-only memory), PROMs (programmable read-only memory), EPROMs (Erasable PROMs, EPROMs), or EEPROMs (Electrically EPROMs), or flash memory.

[0064] Temporarily computer-readable storage media refers to various types of RAM, including static RAM (SRAM), dynamic RAM (DRAM), SDRAM (Synchronous DRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synclink DRAM (SLDRAM), and direct Rambus RAM (DRRAM).

[0065] Furthermore, the embodiments and accompanying drawings of the present invention only clearly illustrate a part of the technical concept of the present invention described above, and it is obvious that all modifications and specific embodiments that can be easily inferred by a person skilled in the art within the scope of the technical concept contained in the aforementioned specification and drawings fall within the scope of the rights of the aforementioned technology.

Claims

1. The analysis device receives information on the distance from the patient's body surface to the tumor, information on internal body heterogeneity, information on the passage of the proton beam through the body, and information on internal body movement associated with the patient's respiration. The analysis device inputs distance information from the patient's body surface to the tumor, information on internal body heterogeneity, information on the passage of the proton beam through the body, and information on internal body movement associated with the patient's respiration into a deep learning model. A method for optimizing a beam angle for proton beam therapy, characterized in that the analytical device includes the step of calculating a beam angle for proton beam therapy based on the output value of the deep learning model.

2. The method for optimizing the beam angle for proton beam therapy according to claim 1, wherein the in-body heterogeneity information includes information calculated based on the standard deviation value of the Hounsfield Unit (HU) obtained from CT (Computed Tomography) images.

3. The method for optimizing the beam angle for proton beam therapy according to claim 1, wherein the information on internal bodily movements associated with the patient's respiration includes information calculated from DVF (Deformation Vector Fields).

4. The method for optimizing the beam angle for proton beam therapy according to claim 1, wherein the distance information from the patient's body surface to the tumor, information on internal body heterogeneity, information on the passage of the proton beam through the body, and information on internal body movement associated with the patient's respiration include information corresponding to the proton beam angle.

5. The method for optimizing the beam angle for proton beam therapy according to claim 1, wherein calculating the beam angle for the aforementioned proton beam therapy involves deriving the effect corresponding to a 360-degree angle in peak form.

6. An input device receives information on the distance from the patient's body surface to the tumor, information on internal body heterogeneity, information on the passage of the proton beam through the body, and information on internal body movement associated with the patient's respiration. A computing device that inputs distance information from the patient's body surface to the tumor, information on internal body heterogeneity, information on the passage of the proton beam through the body, and information on internal body movement associated with the patient's respiration into a deep learning model, and calculates the beam angle for proton beam therapy based on the output values ​​of the deep learning model, A beam angle optimization apparatus for proton beam therapy, comprising a memory device for storing the deep learning model.

7. The beam angle optimization apparatus for proton beam therapy according to claim 6, wherein the internal heterogeneity information includes information calculated based on the standard deviation value of HU (Hounsfield Unit) obtained from CT images.

8. The beam angle optimization apparatus for proton beam therapy according to claim 6, wherein the information on internal movement associated with the patient's respiration includes information calculated based on the value of DVF (deformation vector fields).

9. The beam angle optimization apparatus for proton beam therapy according to claim 6, wherein the distance information from the patient's body surface to the tumor, information on internal body heterogeneity, information on the passage of the proton beam through the body, and information on internal body movement associated with the patient's respiration include information corresponding to the proton beam angle.

10. The beam angle optimization apparatus for proton beam therapy according to claim 6, wherein calculating the beam angle for the aforementioned proton beam therapy involves deriving the effect corresponding to a 360-degree angle in the form of a peak.

Images