Device for determining exposure dose, electronic device and computer readable storage medium

By converting the two-dimensional electron field images actually generated during photon radiotherapy into reference images with high similarity, and using neural network models to determine the three-dimensional illuminated dose, the problem of poor accuracy in illuminated dose evaluation in photon radiotherapy is solved. More precise treatment is achieved.

CN118471436BActive Publication Date: 2024-11-01MANTEIA TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410942429.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-15
Publication Date
2024-11-01
Estimated Expiration
2044-07-15

AI Technical Summary

Technical Problem

The prior art cannot accurately evaluate the patient's true radiation dose during photon radiotherapy, resulting in accidental injury to normal tissue during treatment, and lack of effective solutions.

Method used

By converting the actually generated two-dimensional electron field images into target images with a pre-generated reference image that is more similar to the pre-established reference image, and using a neural network model to determine the three-dimensional illuminated dose information of the patient based on the actual flux data.

Benefits of technology

It improves the accuracy of the patient's dose evaluation, helps to achieve accurate treatment, and solves the problem of poor accuracy of the dose evaluation in the prior art.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118471436B_ABST
    Figure CN118471436B_ABST
Patent Text Reader

Abstract

The present application discloses an exposure dose determination device, an electronic device, and a computer-readable storage medium, and relates to the field of medical technology. The exposure dose determination device includes: a first image determination unit, which is used to execute the radiotherapy plan file of the target object according to the medical image of the target object to obtain a first image, wherein the first image is an electronic portal image actually generated during radiotherapy of the target object; a target image determination unit, which is used to convert the first image into a target image having a similarity with at least one reference image greater than a preset similarity; a first determination unit, which is used to determine the actual flux data of the target object during radiotherapy according to the target image; and a second determination unit, which is used to determine the N-dimensional exposure dose information of the target object during treatment according to the actual flux data of the target object during radiotherapy. The present application solves the technical problem of poor accuracy in evaluating the actual exposure dose of patients in the prior art.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of medical technology, and in particular, to a device for determining an exposure dose, an electronic device, and a computer-readable storage medium. Background Art

[0002] The principle of photon therapy is based on the interaction between light energy and biological tissue to affect biological cells. Specifically, when light of a specific wavelength passes through the skin and enters human tissue, it interacts with the pigments and cells in the human tissue. This interaction can transfer energy to cells, thereby changing cell function and metabolic processes.

[0003] Among them, traditional radiotherapy mainly refers to photon radiotherapy, namely X-rays and gamma rays. However, when photon radiotherapy passes through the tumor, it will continue to emit doses, which will inevitably "accidentally injure" the normal tissues around the tumor, leading to various side effects.

[0004] However, the existing technology cannot know the actual dose of the patient's body during the treatment process, and can only estimate the dose distribution in the patient's body before treatment through simulation. However, due to positioning errors and equipment execution errors during the treatment process, the actual dose of the patient is different from the estimated dose before treatment. The existing technology generally verifies the accuracy of the patient's plan execution through EPID (Electronic Portal Imaging Device) images taken during the patient's radiotherapy. However, EPID images are two-dimensional images, and it is difficult to directly estimate the three-dimensional real dose distribution from the two-dimensional EPID images.

[0005] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0006] The present application provides an apparatus for determining an exposure dose, an electronic device, and a computer-readable storage medium, so as to at least solve the technical problem of poor accuracy in assessing the actual exposure dose of a patient in the prior art.

[0007] According to one aspect of the present application, a device for determining an exposure dose is provided, comprising: a first image determination unit, used to execute a radiotherapy plan file of a target object based on a medical image of the target object to obtain a first image, wherein the first image is an electronic portal image actually generated during radiotherapy of the target object, wherein the data in the electronic portal image is two-dimensional data; a target image determination unit, used to convert the first image into a target image having a similarity with at least one reference image greater than a preset similarity, wherein the reference image is a pre-generated electronic portal image; a first determination unit, used to determine actual flux data of the target object during radiotherapy based on the target image; and a second determination unit, used to determine N-dimensional exposure dose information of the target object during treatment based on the actual flux data of the target object during radiotherapy, wherein N is an integer greater than 2.

[0008] Optionally, the target image determination unit in the device for determining the exposure dose includes: a target image determination subunit, used to input the first image into a first prediction model, and generate a target image based on the first image using the first prediction model using prior knowledge learned in model training, wherein the first prediction model is a neural network model trained based on L reference images, and L is an integer greater than 1.

[0009] Optionally, the device for determining the irradiated dose also includes: an acquisition unit, used to acquire medical images and historical radiotherapy plans of a reference object; a first processing unit, used to execute the historical radiotherapy plan file of the reference object based on the medical image of the reference object, and obtain the electronic portal image actually generated by the reference object during radiotherapy; a second processing unit, used to use the electronic portal image actually generated by the reference object during radiotherapy as a training sample; a third processing unit, used to use L reference images as training labels; a first training unit, used to iteratively train the first initial neural network model through training samples and training labels until the iteration end condition is met to obtain a first prediction model.

[0010] Optionally, the device for determining the exposure dose also includes: a setting unit, used to set at least one position information of a flat-panel detector, wherein the flat-panel detector is an instrument that generates an electronic portal image based on the rays emitted by the ray emitting device; and a reference image generating unit, used to generate L reference images through a dose calculation system based on a historical radiotherapy plan file of a reference object and at least one position information of the flat-panel detector.

[0011] Optionally, the first determination unit in the device for determining the irradiated dose includes: a target image processing subunit, used to input the target image into a second prediction model, and determine the actual flux data of the target object during radiotherapy based on the target image using the second prediction model using prior knowledge learned in model training, wherein the second prediction model is a neural network model trained using a reference image as a training sample and the actual flux data of the reference object as a training label.

[0012] Optionally, the device for determining the irradiated dose also includes: a fourth processing unit, used to input the reference image into the second initial neural network model, and obtain estimated flux data through the second initial neural network model based on the image data in the reference image; a fifth processing unit, used to determine the loss function value of the second initial neural network model based on the estimated flux data and the actual flux data of the reference object; a second training unit, used to perform multiple iterative training on the second initial neural network model according to the loss function value, until the second initial neural network model enters a convergence state or reaches a maximum number of training times, and uses the second initial neural network model after the last training as a second prediction model, wherein the convergence state indicates that the loss function values ​​of the second initial neural network model obtained by M consecutive calculations are all less than a preset threshold, and M is an integer greater than or equal to 1.

[0013] Optionally, the first determination unit in the device for determining the irradiated dose includes: a first processing subunit, used to input the target image and the first image into a third prediction model, and determine the actual flux data of the target object during radiotherapy based on the target image and the first image through the third prediction model using prior knowledge learned in model training, wherein the third prediction model is a neural network model trained using the electronic field image and reference image actually generated by the reference object during radiotherapy as training samples and the actual flux data of the reference object as training labels.

[0014] Optionally, the device for determining the exposure dose also includes: a second image determination unit, used to use the electronic portal image actually generated by the reference object during radiotherapy as the second image; a second image adjustment unit, used to perform L adjustment operations on the second image respectively to obtain L reference images, wherein each adjustment operation is used to adjust the image information of the second image to obtain a reference image.

[0015] According to another aspect of the present application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer-readable storage medium, wherein when the computer program runs, the device where the computer-readable storage medium is located controls the operation of the device for determining the exposure dose described in any one of the above items.

[0016] According to another aspect of the present application, an electronic device is also provided, wherein the electronic device includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors control the operation of any one of the above-mentioned devices for determining the exposure dose.

[0017] In the present application, a device for determining an irradiated dose is provided, wherein the device for determining an irradiated dose includes: a first image determination unit, a target image determination unit, a first determination unit, and a second determination unit. The first image determination unit is used to execute the radiotherapy plan file of the target object according to the medical image of the target object to obtain a first image, wherein the first image is an electronic portal image actually generated during radiotherapy of the target object, wherein the data in the electronic portal image is two-dimensional data; the target image determination unit is used to convert the first image into a target image having a similarity greater than a preset similarity with at least one reference image, wherein the reference image is a pre-generated electronic portal image; the first determination unit is used to determine the actual flux data of the target object during radiotherapy according to the target image; the second determination unit is used to determine the N-dimensional irradiated dose information of the target object during treatment according to the actual flux data of the target object during radiotherapy, wherein N is an integer greater than 2.

[0018] As can be seen from the above content, the present application realizes the conversion of the patient's real EPID (i.e., the first image) into a virtual EPID (i.e., the target image) by converting the first image into a target image whose similarity with at least one reference image is greater than a preset similarity, and then determines the patient's actual flux data based on the virtual EPID, thereby finally determining the patient's three-dimensional exposure dose information (i.e., N-dimensional exposure dose information) based on the actual flux data. It can be seen that the technical solution of the present application realizes the purpose of converting EPID image data into three-dimensional real patient exposure dose data, thereby improving the accuracy of the assessment of the patient's real exposure dose, which is conducive to the subsequent precise treatment of the patient, thereby solving the technical problem of poor accuracy in the assessment of the patient's real exposure dose in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0020] Figure 1 is a schematic diagram of an optional device for determining an irradiation dose according to an embodiment of the present application;

[0021] Figure 2 is a flowchart of an optional training method of a first prediction model according to an embodiment of the present application;

[0022] Figure 3 is a schematic diagram of the position of a flat panel detector according to an embodiment of the present application;

[0023] Figure 4 is an optional model connection architecture diagram according to an embodiment of the present application;

[0024] Figure 5 It is a flow chart of an optional training method of a second prediction model according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0026] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] It should also be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) collected by this application are information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with the relevant laws, regulations and standards of the relevant regions, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation portals are provided for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or institutions. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or institution through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or institution.

[0028] In addition, this application collects and analyzes customer information and provides users with corresponding operation entrances for them to choose to agree or reject automated decision-making results.

[0029] According to an embodiment of the present application, an optional embodiment of a device for determining an irradiated dose is provided, wherein: Figure 1 is a schematic diagram of an optional device for determining an exposure dose according to an embodiment of the present application, such as Figure 1 As shown, the device for determining the exposure dose includes: a first image determining unit 101 , a target image determining unit 102 , a first determining unit 103 , and a second determining unit 104 .

[0030] Among them, the first image determination unit 101 is used to execute the radiotherapy plan file of the target object according to the medical image of the target object to obtain a first image, wherein the first image is an electronic portal image actually generated during radiotherapy of the target object, wherein the data in the electronic portal image is two-dimensional data.

[0031] Optionally, the target subject is a patient who is preparing to receive radiotherapy or a patient who has received radiotherapy for a period of time but needs to continue radiotherapy later.

[0032] Optionally, the medical image of the target object may be a single-modality image or a multi-modality image, wherein the medical image includes but is not limited to various types of medical images such as MR images, CT images, and PET images.

[0033] Optionally, the radiotherapy plan file of the target object is a radiotherapy plan file that has been developed, and the medical staff will perform radiotherapy on the target object according to the radiotherapy plan file.

[0034] Optionally, the first image is an electronic portal image actually generated during radiotherapy of the target object, wherein the electronic portal image is an EPID image. For example, the target object is a patient whose actual exposure dose distribution in the body is to be calculated. By inputting the radiotherapy plan file and medical image of the target object into the exposure dose determination device, the exposure dose determination device can execute the radiotherapy plan file of the target object based on the medical image of the target object, thereby obtaining the actual EPID image (i.e., the first image) of the target object treated on that day.

[0035] Optionally, the device for determining the exposure dose may be a software system or an embedded system combining software and hardware. In addition, the device for determining the exposure dose may also be run on an electronic device, wherein the electronic device includes but is not limited to a smart tablet, a smart phone, a laptop computer, a desktop computer, and the like.

[0036] In addition, it should be noted that the data in the EPID image is two-dimensional data, but the radiation dose distribution information in the patient's body is usually three-dimensional data. Therefore, there is a large potential error in directly estimating the radiation dose distribution information in the patient's body based on the EPID image.

[0037] In an optional embodiment, the device for determining the exposure dose in the present application also includes a target image determination unit 102, wherein the target image determination unit 102 is used to convert the first image into a target image having a similarity with at least one reference image greater than a preset similarity, wherein the reference image is a pre-generated electronic portal image.

[0038] Optionally, there can be multiple reference images, wherein the reference image is a pre-generated EPID image, and the reference image does not necessarily have to be an EPID image generated during the patient's actual radiotherapy process. The reference image can also be an EPID image constructed based on the actual EPID image taken. For example, the reference image can be an EPID image obtained by adjusting the actual EPID image taken, or it can be a reference image generated by using neural network model technology, with the actual EPID image taken as model input data, and the neural network model predicts the actual EPID image.

[0039] It should be noted that the reason why a large number of reference images need to be generated in advance is that the number of EPID images actually taken is limited. In addition, due to the patient's privacy and security, it is usually difficult to obtain the EPID images actually taken. Multiple factors have resulted in a small number of available EPID images actually taken.

[0040] However, since the reference image is an image constructed based on the actually shot EPID image, the number of reference images can be unlimited. According to actual needs, the number of reference images can be infinite in theory (that is, reference images can be generated without limit). It can be seen that by constructing a large number of reference images, the problem of the small number of actually shot EPID images available in the prior art can be overcome, and the technical effect of expanding the number of EPID images can be achieved.

[0041] It should also be noted that the first image is converted into a target image having a similarity with at least one reference image greater than a preset similarity. The conversion method that can be used may be conversion through a neural network model, or image adjustment of the first image, thereby adjusting the first image to the target image.

[0042] In an optional embodiment, the device for determining the exposure dose in the present application further includes a first determination unit 103, wherein the first determination unit 103 is used to determine actual flux data of the target object during radiotherapy according to the target image.

[0043] Optionally, the first determination unit 103 can determine the actual flux data of the target object during radiotherapy according to the target image through a pre-trained neural network model. The flux data refers to the number of rays or particle beams passing through a unit area per unit time. In radiotherapy, the flux data is usually related to the dose rate, that is, the dose delivered to the biological tissue per unit time. The balance between the flux data and the dose rate is crucial for the control of the radiotherapy effect and side effects.

[0044] In an optional embodiment, the device for determining the exposure dose in the present application also includes a second determination unit 104, wherein the second determination unit 104 is used to determine the N-dimensional exposure dose information of the target object during treatment based on the actual flux data of the target object during radiotherapy, wherein N is an integer greater than 2.

[0045] Optionally, the second determination unit 104 may input the actual flux data into a dose calculation engine, and the dose calculation engine calculates actual exposure dose information in the target object based on the actual flux data, wherein the actual exposure dose information may be three-dimensional or higher data, for example, four-dimensional data including a time dimension.

[0046] Optionally, the dose calculation engine includes but is not limited to a calculation engine using a Monte Carlo algorithm and a calculation engine using a PB (Pencil Beam) algorithm.

[0047] It can be seen from the above content that the present application converts the patient's real EPID image (i.e., the first image) into a virtual EPID image (i.e., the target image) by converting the first image into a target image whose similarity with at least one reference image is greater than a preset similarity, and then determines the patient's actual flux data based on the virtual EPID image, thereby finally determining the patient's three-dimensional exposure dose information (i.e., N-dimensional exposure dose information) based on the actual flux data. It can be seen that the technical solution of the present application achieves the purpose of converting EPID image data into three-dimensional real patient exposure dose data, thereby improving the accuracy of the assessment of the patient's real exposure dose, which is facilitating the subsequent precise treatment of the patient, and thus solving the technical problem of poor accuracy in the assessment of the patient's real exposure dose in the prior art.

[0048] In an optional embodiment, the target image determination unit 102 further includes: a target image determination subunit. The target image determination subunit is used to input the first image into the first prediction model, and generate the target image based on the first image by using the prior knowledge learned in the model training through the first prediction model, wherein the first prediction model is a neural network model trained based on L reference images, and L is an integer greater than 1.

[0049] Optionally, the first prediction model can be various types of neural network models, such as a convolutional neural network, a graph neural network model, a LSTM neural network model, or an attention mechanism neural network model.

[0050] Optionally, since the first prediction model is a neural network model obtained by training based on L reference images, and the target image is a target image obtained by converting the first image and having a similarity with at least one reference image greater than a preset similarity, the process of converting the first image to the target image can actually be understood as converting the first image to an image with a high similarity (or infinitely close to the reference image) to the reference image used in the training phase of the first prediction model. However, the target image also needs to maintain relevance to the first image. In other words, in the process of generating the target image, there are actually two constraints: one is to generate an image that is as similar or even identical to the reference image used in the training phase of the first prediction model as possible; the other is that the first image needs to be used as reference information in the generation process and cannot be completely separated from the influence of the first image.

[0051] For example, assuming that the first image represents an image of a patient suffering from lung cancer, the target image generated based on the first image should also be an image of lung cancer, and not an image of other cancers. On this basis, the reference image whose similarity with the target image is greater than the preset similarity should also be an image of lung cancer.

[0052] From the above content, it can be seen that since the first prediction model is a neural network model, multiple constraints can be deployed by setting loss functions, etc. Therefore, during the model training process, the first prediction model can be controlled to have prior knowledge of converting the first image into a target image whose similarity with at least one reference image is greater than a preset similarity.

[0053] In addition, it should be noted that, as mentioned above, the actual number of EPID images is small and difficult to obtain. Therefore, using actual EPID images as model training data to train the neural network model is actually a small sample training method, and the model prediction accuracy obtained by training in this way is poor. However, this application mainly obtains the first prediction model based on L reference images. Since the reference image is an image constructed based on the actual EPID image, the number of reference images can be unlimited. According to actual needs, the number of reference images can be infinite in theory, thereby overcoming the technical problem of poor prediction accuracy of the model trained in the small sample training method.

[0054] In addition, it should be noted that in the process of generating the target image, by constraining the first image to be used as reference information and not being completely separated from the influence of the first image, it can be ensured that the generated target image will not be a completely fabricated image, and there is still a large correlation between it and the first image. In this way, the prediction accuracy of the subsequent prediction of the flux data of the target object based on the target image can be guaranteed.

[0055] In an optional embodiment, the device for determining the exposure dose further includes: an acquisition unit, a first processing unit, a second processing unit, a third processing unit and a first training unit.

[0056] Among them, the acquisition unit is used to acquire the medical image and historical radiotherapy plan of the reference object; the first processing unit is used to execute the historical radiotherapy plan file of the reference object according to the medical image of the reference object, and obtain the electronic field image actually generated by the reference object during radiotherapy; the second processing unit is used to use the electronic field image actually generated by the reference object during radiotherapy as a training sample; the third processing unit is used to use L reference images as training labels; the first training unit is used to iteratively train the first initial neural network model through training samples and training labels until the iteration end condition is met to obtain a first prediction model.

[0057] Optionally, Figure 2 is a flowchart of an optional training method of a first prediction model according to an embodiment of the present application, such as Figure 2 As shown, the following steps are included:

[0058] Step S201, obtaining medical images and historical radiotherapy plans of N historical patients.

[0059] In step S201, N is an integer greater than 1, and the historical patient represents the above-mentioned reference object, and the medical images of the historical patient include but are not limited to various types or modal images such as MR images, CT images, and PET images. The historical radiotherapy plan of the historical patient is the radiotherapy plan executed by the historical patient during radiotherapy.

[0060] Step S202 , executing the historical radiotherapy plan file of each historical patient according to the medical image of the historical patient, so as to obtain the actual EPID image of the historical patient during radiotherapy.

[0061] Step S203: generating L reference images according to the historical radiotherapy plan of the historical patient.

[0062] Step S204: establishing a first prediction model based on mapping the actual EPID image to the virtual EPID image.

[0063] Optionally, in step S204, the actual EPID images of historical patients during radiotherapy can be used as training samples, and L reference images can be used as training labels, so as to train the first initial neural network model to learn how to generate a virtual EPID image based on the actually taken EPID image, whose similarity with at least one reference image is greater than a preset similarity, wherein the virtual EPID image mentioned in the embodiment of the present application refers to an EPID image that is not an EPID image actually taken, but an EPID image constructed based on a neural network model or other methods.

[0064] Optionally, the iteration end condition may be that the number of training times of the first initial neural network model exceeds the maximum number of training times, or that the loss function values ​​calculated by the first initial neural network model H times in a row are all less than a preset value, where H is an integer greater than or equal to 1.

[0065] In an optional embodiment, the device for determining the irradiated dose further includes: a setting unit and a reference image generating unit. The setting unit is used to set at least one position information of a flat panel detector, wherein the flat panel detector is an instrument for generating an electronic portal image according to the radiation emitted by the radiation emitting device; and the reference image generating unit is used to generate L reference images according to the historical radiotherapy plan file of the reference object and at least one position information of the flat panel detector through a dose calculation system.

[0066] Optionally, Figure 3 is a schematic diagram of the position of a flat panel detector according to an embodiment of the present application, such as Figure 3 As shown, there can be multiple positions of the flat panel detector, for example, position 1, position 2, position 3, etc. below the patient.

[0067] It should be noted that the generation of EPID images is not only affected by the historical radiotherapy plan file, but also by the position of the flat panel detector. For example, when executing the same historical radiotherapy file, the EPID image generated at position 1 is different from the EPID image generated at position 2. Therefore, by setting different positions of the flat panel detector, different reference images can also be constructed.

[0068] Optionally, in an embodiment of the present application, a dose calculation system (also referred to as a dose calculation engine) can be used to generate L reference images based on a historical radiotherapy plan of a reference object and at least one position information of a flat-panel detector, wherein the dose calculation engine includes but is not limited to a calculation engine using a Monte Carlo algorithm and a calculation engine using a PB algorithm.

[0069] It can be seen from the above content that by preparing multiple historical radiotherapy plans and setting at least one position information of the flat panel detector, the technical effect of constructing multiple reference images can be achieved.

[0070] In an optional embodiment, the first determination unit 103 further includes: a target image processing subunit. The target image processing subunit is used to input the target image into the second prediction model, and determine the actual flux data of the target object during radiotherapy based on the target image by using the prior knowledge learned in the model training through the second prediction model, wherein the second prediction model is a neural network model trained with the reference image as a training sample and the actual flux data of the reference object as a training label.

[0071] Optionally, Figure 4 is an optional model connection architecture diagram according to an embodiment of the present application, such as Figure 4 As shown, in the device for determining the irradiated dose, at least a first prediction model and a second prediction model are deployed, wherein there is a connection relationship between the first prediction model and the second prediction model. In actual application, the model input data of the first prediction model is the actual EPID image (corresponding to the first image mentioned above) of the target object (the patient whose actual irradiated dose information is to be evaluated), and the model output data of the first prediction model is a target image generated based on the first image and having a similarity greater than a preset similarity with at least one reference image used in training the first prediction model (it can also be understood as a predicted EPID image generated by the first prediction model). In addition, the predicted EPID image output by the first prediction model is used as the model input data of the second prediction model, and the second prediction model predicts the actual flux data of the target object during radiotherapy based on the predicted EPID image output by the first prediction model.

[0072] In an optional embodiment, the device for determining the exposure dose further includes: a fourth processing unit, a fifth processing unit, and a second training unit.

[0073] Among them, the fourth processing unit is used to input the reference image into the second initial neural network model, and the estimated flux data is calculated by the second initial neural network model according to the image data in the reference image; the fifth processing unit is used to determine the loss function value of the second initial neural network model according to the estimated flux data and the actual flux data of the reference object; the second training unit is used to perform multiple iterative training on the second initial neural network model according to the loss function value until the second initial neural network model enters a convergence state or reaches a maximum number of training times, and the second initial neural network model after the last training is used as the second prediction model, wherein the convergence state characterizes that the loss function values ​​of the second initial neural network model obtained by M consecutive calculations are all less than a preset threshold, and M is an integer greater than or equal to 1.

[0074] Optionally, Figure 5 is a flow chart of an optional training method of a second prediction model according to an embodiment of the present application, such as Figure 5 As shown, the following steps are included:

[0075] Step S501, inputting the reference image into the second initial neural network model.

[0076] Optionally, in step S501, the second initial neural network model may be various types of neural network models, such as a convolutional neural network, a graph neural network model, a LSTM neural network model, or an attention mechanism neural network model.

[0077] Step S502, calculating estimated flux data based on image data in the reference image using a second initial neural network model.

[0078] Optionally, estimated flux data can be calculated based on the reference image through the second initial neural network model. It should be noted that the estimated flux data may be different from the actual flux data used as training labels. Therefore, the training goal of the second initial neural network model is to control the second initial neural network model to learn as much as possible how to generate estimated flux data that is as similar as possible to the training labels based on the reference image.

[0079] Step S503: establishing a second prediction model that maps the reference EPID image to the flux data.

[0080] Optionally, the difference between the estimated flux data and the training label (i.e., the actual flux data of the reference object) can be represented by a loss function value. Therefore, the second training unit can perform multiple iterative training on the second initial neural network model according to the loss function value until the second initial neural network model enters a convergence state or reaches a maximum number of training times, and the second initial neural network model after the last training is used as the second prediction model, wherein the convergence state indicates that the loss function values ​​of the second initial neural network model obtained by M consecutive calculations are all less than a preset threshold, and M is an integer greater than or equal to 1.

[0081] In an optional embodiment, the first determination unit 103 further includes: a first processing subunit. The first processing subunit is used to input the target image and the first image into a third prediction model, and determine the actual flux data of the target object during radiotherapy based on the target image and the first image by using the third prediction model using the prior knowledge learned in model training, wherein the third prediction model is a neural network model trained using the electronic portal image and the reference image actually generated by the reference object during radiotherapy as training samples and the actual flux data of the reference object as training labels.

[0082] Optionally, in an embodiment of the present application, a third prediction model may also be deployed in the device for determining the irradiated dose, wherein the training samples of the third prediction model are EPID images actually generated of the reference object during radiotherapy and virtual EPID images (i.e., reference images) constructed based on the actually generated EPID images, and the training labels of the third prediction model are the actual flux data of the reference object.

[0083] In an optional embodiment, the device for determining the irradiated dose also includes: a sixth processing unit, used to input the reference image and the EPID image actually generated by the reference object during radiotherapy into the third initial neural network model, and calculate the intermediate flux data through the third initial neural network model based on the reference image and the EPID image actually generated by the reference object during radiotherapy; a seventh processing unit, used to determine the loss function value of the third initial neural network model based on the intermediate flux data and the actual flux data of the reference object; a third training unit, used to perform multiple iterative training on the third initial neural network model according to the loss function value of the third initial neural network model until the third initial neural network model enters a convergence state or reaches the maximum number of training times, and the third initial neural network model after the last training is used as the third prediction model.

[0084] In an optional embodiment, the device for determining the irradiated dose further includes: a second image determination unit and a second image adjustment unit. The second image determination unit is used to use the electronic portal image actually generated by the reference object during radiotherapy as the second image; the second image adjustment unit is used to perform L adjustment operations on the second image respectively to obtain L reference images, wherein each adjustment operation is used to adjust the image information of the second image to obtain a reference image.

[0085] Optionally, in addition to generating reference images according to the position information of the flat panel detector and the historical radiotherapy plan through the dose calculation system, L reference images can be obtained by adjusting the actual EPID image of the reference object during radiotherapy. The adjustment operation includes but is not limited to image registration, reduction, enlargement, modification and other operations on the actual EPID image of the reference object. The purpose of the adjustment operation is also to generate more virtual EPID images based on the actual EPID of the reference object, so as to achieve the technical effect of expanding the EPID image.

[0086] According to another aspect of an embodiment of the present application, a computer-readable storage medium is further provided, wherein a computer program is stored in the computer-readable storage medium, wherein when the computer program is executed, the device where the computer-readable storage medium is located controls the operation of the device for determining the exposure dose described in any one of the above items.

[0087] According to another aspect of an embodiment of the present application, an electronic device is also provided, wherein the electronic device includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors control the operation of any one of the above-mentioned devices for determining the exposure dose.

[0088] As can be seen from the above content, in the present application, a device for determining an irradiated dose is provided, wherein the device for determining an irradiated dose includes: a first image determination unit, a target image determination unit, a first determination unit, and a second determination unit. The first image determination unit is used to execute the radiotherapy plan file of the target object according to the medical image of the target object to obtain a first image, wherein the first image is an electronic portal image actually generated during radiotherapy of the target object, wherein the data in the electronic portal image is two-dimensional data; the target image determination unit is used to convert the first image into a target image having a similarity greater than a preset similarity with at least one reference image, wherein the reference image is a pre-generated electronic portal image; the first determination unit is used to determine the actual flux data of the target object during radiotherapy according to the target image; the second determination unit is used to determine the N-dimensional irradiated dose information of the target object during treatment according to the actual flux data of the target object during radiotherapy, wherein N is an integer greater than 2.

[0089] It can be seen that the present application converts the patient's real EPID (i.e., the first image) into a virtual EPID (i.e., the target image) by converting the first image into a target image whose similarity with at least one reference image is greater than a preset similarity, and then determines the patient's actual flux data based on the virtual EPID, thereby finally determining the patient's three-dimensional exposure dose information (i.e., N-dimensional exposure dose information) based on the actual flux data. It can be seen that the technical solution of the present application achieves the purpose of converting the EPID into a three-dimensional real patient exposure dose, improves the accuracy of the assessment of the patient's real exposure dose, is facilitating subsequent precise treatment of the patient, and thus solves the technical problem of poor accuracy in the assessment of the patient's real exposure dose in the prior art.

[0090] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0091] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0092] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0093] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0094] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0095] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc., which can store program code.

[0096] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A device for determining an exposure dose, It is characterized in that include: A first image determination unit is used to execute the radiotherapy plan file of the target object according to the medical image of the target object to obtain a first image, wherein the first image is an electronic portal image actually generated during the radiotherapy of the target object, wherein the data in the electronic portal image is two-dimensional data; a target image determination unit, configured to convert the first image into a target image having a similarity with at least one reference image greater than a preset similarity, wherein the reference image is a pre-generated electronic portal image; a first determining unit, configured to determine actual flux data of the target object during radiotherapy according to the target image, or to determine actual flux data of the target object during radiotherapy according to the target image and the first image; A second determining unit, configured to determine N-dimensional irradiation dose information of the target object during treatment according to actual flux data of the target object during radiotherapy, wherein N is an integer greater than 2; Among them, the target image determination unit includes: a target image determination subunit, which is used to input the first image into a first prediction model, and generate the target image based on the first image by using the first prediction model to utilize the prior knowledge learned in model training, wherein the first prediction model is a neural network model trained based on L reference images, and L is an integer greater than 1.

2. The device for determining the exposure dose according to claim 1, It is characterized in that The device for determining the irradiated dose further comprises: an acquisition unit for acquiring medical images and historical radiotherapy plans of a reference subject; A first processing unit, configured to execute a historical radiotherapy plan file of the reference subject according to the medical image of the reference subject, and obtain an electronic portal image actually generated during radiotherapy of the reference subject; A second processing unit is used to use the electronic portal image actually generated by the reference object during radiotherapy as a training sample; A third processing unit, configured to use the L reference images as training labels; The first training unit is used to iteratively train the first initial neural network model through the training samples and the training labels until an iteration end condition is met to obtain the first prediction model.

3. The device for determining the exposure dose according to claim 2, It is characterized in that The device for determining the irradiated dose further comprises: A setting unit, used to set at least one position information of a flat panel detector, wherein the flat panel detector is an instrument for generating an electronic portal image according to the rays emitted by the ray emitting device; The reference image generating unit is used to generate L reference images according to the historical radiotherapy plan file of the reference object and at least one position information of the flat panel detector through a dose calculation system.

4. The device for determining the exposure dose according to claim 3, It is characterized in that The first determining unit includes: The target image processing subunit is used to input the target image into a second prediction model, and determine the actual flux data of the target object during radiotherapy based on the target image by using the prior knowledge learned in model training through the second prediction model, wherein the second prediction model is a neural network model trained using the reference image as a training sample and the actual flux data of the reference object as a training label.

5. The device for determining the exposure dose according to claim 4, It is characterized in that The device for determining the irradiated dose further comprises: a fourth processing unit, configured to input the reference image into a second initial neural network model, and obtain estimated flux data by calculating the image data in the reference image through the second initial neural network model; a fifth processing unit, configured to determine a loss function value of the second initial neural network model according to the estimated flux data and the actual flux data of the reference object; A second training unit is used to perform multiple iterative training on the second initial neural network model according to the loss function value until the second initial neural network model enters a convergence state or reaches a maximum number of training times, and uses the second initial neural network model after the last training as the second prediction model, wherein the convergence state represents that the loss function values ​​of the second initial neural network model obtained by M consecutive calculations are all less than a preset threshold, and M is an integer greater than or equal to 1.

6. The device for determining the exposure dose according to claim 3, It is characterized in that The first determining unit includes: The first processing subunit is used to input the target image and the first image into a third prediction model, and determine the actual flux data of the target object during radiotherapy based on the target image and the first image by using the third prediction model and the prior knowledge learned in the model training, wherein the third prediction model is a neural network model trained using the electronic portal image actually generated by the reference object during radiotherapy and the reference image as training samples and the actual flux data of the reference object as training labels.

7. The device for determining the exposure dose according to claim 2, It is characterized in that The device for determining the irradiated dose further comprises: A second image determination unit, configured to use the electronic portal image actually generated during radiotherapy of the reference object as the second image; The second image adjustment unit is used to perform L adjustment operations on the second image respectively to obtain L reference images, wherein each adjustment operation is used to adjust the image information of the second image to obtain one reference image.

8. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located controls the operation of the device for determining the exposure dose according to any one of claims 1 to 7.

9. An electronic device, It is characterized in that It comprises one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors control the operation of the device for determining the exposure dose as described in any one of claims 1 to 7.