Systems and methods for radiation therapy

Abstract

A method may include obtaining input data related to a target treatment plan for performing radiotherapy on a lesion using a radiation device (1210). The input data may include a first target image of the lesion. The method may also include obtaining a sub-field shape estimation model (1220). The method may also include estimating a plurality of target position combinations of the target treatment plan and a plurality of target sub-field shapes of a collimator of the radiation device based on the sub-field shape estimation model and the input data, each of the plurality of target sub-field shapes corresponding to one of the plurality of target position combinations (1230). One of the plurality of target position combinations may indicate a position of the collimator relative to the lesion.

Description

Technical Field

[0001] The present disclosure relates to radiation therapy and, more particularly, to systems and methods for treatment planning. Background Art

[0002] Radiation therapy is used to treat cancer and other diseases in biological (e.g., human and animal) tissue using radiation devices. Treatment planning is a process that involves determining and / or updating specific radiation therapy parameters to achieve treatment goals. The result of treatment planning is a treatment plan. The sub-field shape of the collimator of the radiation device is an important factor in determining the overall delivery time of the treatment plan. Therefore, it is desirable to provide systems and / or methods to effectively and accurately determine the sub-field shape of the treatment plan. Summary of the invention

[0003] According to one aspect of the present disclosure, a system may include one or more storage devices and one or more processors configured to communicate with the one or more storage devices. The one or more storage devices may include a set of instructions. When the one or more processors execute the instructions, at least one processor may be configured to instruct one or more of the following operations to be performed. The one or more processors may obtain input data related to a target treatment plan for radiotherapy of a lesion using a radiation device, the input data including a first target image of the lesion. The one or more processors may obtain a sub-field shape estimation model. The one or more processors may estimate multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device based on the sub-field shape estimation model and the input data. One of the multiple target position combinations may indicate the position of the collimator relative to the lesion. Each of the multiple target sub-field shapes may correspond to one of the multiple target position combinations.

[0004] In some embodiments, the target treatment plan may include a plurality of control points. Each of the plurality of target position combinations or the plurality of target sub-field shapes may correspond to one of the plurality of control points.

[0005] In some embodiments, the input data may include a second target image of normal tissue surrounding the lesion, a third target image of the lesion, or at least one of target radiation information of the target treatment plan, wherein the target radiation information includes at least one of an output dose, a dose output rate, a dose per pulse, or a dose distribution in the lesion.

[0006] In some embodiments, the target radiation information may be predicted based on the first target image of the lesion, the second target image of normal tissue surrounding the lesion, and the third target image of the lesion.

[0007] In some embodiments, one of the plurality of target position combinations may include a combination of one or more positions at which one or more components of the radiation device operate.

[0008] In some embodiments, one of the plurality of target position combinations may include a rack angle of a rack of the radiation device.

[0009] In some embodiments, one of the plurality of target position combinations may include a combination of a gantry angle of a gantry of the radiation device and a collimator angle of a collimator of the radiation device.

[0010] In some embodiments, one of the plurality of target position combinations may include a combination of a gantry angle of a gantry of the radiation device, a collimator angle of a collimator of the radiation device, and a diagnosis table position of the radiation device.

[0011] In some embodiments, the collimator may include a plurality of pairs of blades. One of the plurality of target sub-field shapes of the collimator includes a blade position of each pair of blades in the plurality of pairs of blades.

[0012] In some embodiments, the blade position of each pair of blades in the plurality of pairs of blades may include a center position of an opening of the pair of blades and an opening width of the pair of blades.

[0013] In some embodiments, the plurality of target location combinations may be within a plurality of discrete candidate location combinations of a generic set of locations.

[0014] In some embodiments, the sub-field shape estimation model can be obtained by executing a training process, including: obtaining the general set of positions containing the multiple candidate position combinations; and iteratively training an initial model based on the general set of positions to determine the sub-field shape estimation model.

[0015] In some embodiments, obtaining the general set of positions including the multiple candidate position combinations may include: obtaining multiple candidate gantry angles, multiple candidate collimator angles, or multiple candidate table positions; and obtaining the general set of positions based on the multiple candidate gantry angles, the multiple candidate collimator angles, or the multiple candidate table positions.

[0016] In some embodiments, the plurality of target sub-field shapes may be within a common set of distances that includes a plurality of discrete candidate leaf positions.

[0017] In some embodiments, the training process may include: obtaining the universal distance set containing the multiple candidate blade positions; and iteratively training the initial model based on the universal distance set to determine the sub-field shape estimation model so that each corresponding candidate sub-field shape in the multiple candidate position combinations output by the sub-field shape estimation model is in the universal distance set.

[0018] In some embodiments, the plurality of candidate blade positions may include a plurality of candidate opening positions and a plurality of candidate opening widths.

[0019] In some embodiments, the training process includes: obtaining training data including a plurality of training sets.

[0020] In some embodiments, obtaining the training data includes: for one of the multiple training sets, obtaining a historical treatment plan previously generated based on a lesion sample; obtaining a first sample image of the lesion sample corresponding to the historical treatment plan; obtaining a sample position combination and a corresponding sample sub-field shape in the historical treatment plan; and obtaining the training set based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape.

[0021] In some embodiments, obtaining the training set based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape may include: obtaining a processed sample position combination located in the position universal set, wherein the processed sample position combination is obtained by processing the sample position combination based on the position universal set; obtaining a processed sample sub-field shape located in the distance universal set, wherein the processed sample sub-field shape is obtained by processing the sample sub-field shape based on the distance universal set; obtaining a sample set including the processed sample sub-field shape and a closed sub-field shape, wherein the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination; and obtaining the training set through the sample set including the first sample image, the processed sample position combination, and the historical treatment plan.

[0022] In some embodiments, the training process may include: initializing the initial model; and updating the initialized initial model by using an iterative process including multiple iterations to obtain the sub-field shape estimation model, at least one iteration of the multiple iterations of the iterative process includes: obtaining one of the multiple training sets; generating an estimated sub-field shape corresponding to the multiple candidate position combinations by inputting the first sample image of the training set into an intermediate model, the intermediate model being the initial model initialized in the first iteration of the multiple iterations of the iterative process or the previously updated model generated in the previous iteration of the iterative process; determining the value of the loss function based on the estimated sub-field shape and the sample set in the training set; determining whether a termination condition is met; in response to not meeting the termination condition, updating the intermediate model based on the value of the loss function to generate an updated model; and starting the next iteration; and designating the intermediate model in the last iteration of the multiple iterations of the iterative process as the sub-field shape estimation model.

[0023] In some embodiments, at least one of the plurality of iterations of the iterative process may include: terminating the iterative process in response to determining that the termination condition is satisfied.

[0024] In some embodiments, the value of the loss function may be determined based on the sparsity of the sample set, where the sparsity of the sample set is related to the closed subfield shape of the sample set.

[0025] In some embodiments, the termination condition may be related to at least one of the value of the loss function or the number of iterations of the iterative process that have been performed.

[0026] In some embodiments, the training set may include a second sample image of normal tissue surrounding the lesion sample, a third sample image of the lesion sample, or sample radiation information of the historical treatment plan, wherein the sample radiation information includes at least one of a sample output dose, a sample dose output rate, a sample dose per pulse, or a sample dose distribution in the lesion sample.

[0027] In some embodiments, the sample radiation information may be predicted based on the first sample image of the lesion sample, the second sample image of normal tissue surrounding the lesion sample, and the third sample image of the lesion sample.

[0028] In some embodiments, at least one of the multiple iterations of the iterative process includes: generating the estimated sub-field shape by inputting at least one of the second sample image, the third sample image, or the sample radiation information of the training set into the intermediate model.

[0029] In some embodiments, at least one of the multiple iterations of the iterative process may include: determining estimated radiation information based on the estimated sub-field shape; comparing the estimated radiation information with the sample radiation information; and updating the intermediate model based on the comparison to generate the updated model.

[0030] According to another aspect of the present disclosure, a method may include one or more of the following operations. One or more processors may obtain input data related to a target treatment plan for radiotherapy of a lesion using a radiation device. The input data may include a first target image of the lesion. One or more processors may obtain a sub-field shape estimation model. One or more processors may estimate multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device based on the sub-field shape estimation model and the input data. One of the multiple target position combinations may indicate the position of the collimator relative to the lesion. Each of the multiple target sub-field shapes may correspond to one of the multiple target position combinations.

[0031] According to another aspect of the present disclosure, a system may include an input acquisition module for acquiring input data related to a target treatment plan for radiotherapy of a lesion using a radiation device, the input data including a first target image of the lesion. The system may also include a model acquisition module for acquiring a sub-field shape estimation model. The system may also include a shape estimation module for estimating multiple target position combinations of the target treatment plan and multiple target sub-field shapes of a collimator of the radiation device based on the sub-field shape estimation model and the input data. One of the multiple target position combinations may indicate the position of the collimator relative to the lesion. Each of the multiple target sub-field shapes may correspond to one of the multiple target position combinations.

[0032] According to another aspect of the present disclosure, a non-transitory computer-readable medium may include at least one set of instructions. At least one set of instructions may be executed by one or more processors of a computing device. One or more processors may obtain input data related to a target treatment plan for radiotherapy of a lesion using a radiation device. The input data may include a first target image of the lesion. One or more processors may obtain a sub-field shape estimation model. One or more processors may estimate multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device based on the sub-field shape estimation model and the input data. One of the multiple target position combinations may indicate the position of the collimator relative to the lesion. Each of the multiple target sub-field shapes may correspond to one of the multiple target position combinations.

[0033] According to another aspect of the present disclosure, a system may include one or more storage devices and one or more processors configured to communicate with the one or more storage devices. The one or more storage devices may include a set of instructions. When the one or more processors execute the instructions, at least one processor may be configured to instruct the system to perform one or more of the following operations. One or more processors may acquire an initial model. One or more processors may acquire training data. One or more processors may train the initial model based on the training data to obtain a sub-field shape estimation model. The sub-field shape estimation model may be configured to estimate multiple target position combinations in a target treatment plan based on input data, and the target position combination is used to perform radiotherapy on a lesion using a radiation device and multiple target sub-field shapes of a collimator of the radiation device. The input data may include a first target image of the lesion. Each of the multiple target sub-field shapes may correspond to one of the multiple target position combinations. One of the multiple target position combinations may indicate the position of the collimator relative to the lesion.

[0034] According to another aspect of the present disclosure, a method may include one or more of the following operations. One or more processors may acquire an initial model. One or more processors may acquire training data. One or more processors may acquire a sub-field shape estimation model by training the initial model based on the training data. The sub-field shape estimation model may be configured to estimate a plurality of target position combinations in a target treatment plan based on input data, the target position combinations being used to perform radiotherapy on a lesion using a radiation device and a plurality of target sub-field shapes of a collimator of the radiation device. The input data may include a first target image of a lesion. Each of the plurality of target sub-field shapes may correspond to one of the plurality of target position combinations. One of the plurality of target position combinations may indicate a position of the collimator relative to the lesion.

[0035] According to another aspect of the present disclosure, a system may include a model acquisition module for acquiring an initial model. The model acquisition module may also be used to acquire training data. The model acquisition module may also be used to train the initial model based on the training data to acquire a sub-field shape estimation model. The sub-field shape estimation model may be configured to estimate multiple target position combinations in a target treatment plan based on input data, and the target position combinations are used to perform radiotherapy on a lesion using a radiation device and a plurality of target sub-field shapes of a collimator of the radiation device. The input data may include a first target image of a lesion. Each of the multiple target sub-field shapes may correspond to one of the multiple target position combinations. One of the multiple target position combinations may indicate the position of the collimator relative to the lesion.

[0036] According to another aspect of the present disclosure, a non-transitory computer-readable medium may include at least one set of instructions. At least one set of instructions may be executed by one or more processors of a computing device. One or more processors may acquire an initial model. One or more processors may acquire training data. One or more processors may acquire a sub-field shape estimation model by training the initial model based on the training data. The sub-field shape estimation model may be configured to estimate multiple target position combinations in a target treatment plan based on input data, and the target position combination is used to perform radiotherapy on a lesion using a radiation device and a plurality of target sub-field shapes of a collimator of the radiation device. The input data may include a first target image of a lesion. Each of the multiple target sub-field shapes may correspond to one of the multiple target position combinations. One of the multiple target position combinations may indicate the position of the collimator relative to the lesion.

[0037] Some additional features will be set forth in the following description, and some features will become apparent to those skilled in the art upon examination of the following description and drawings, or may be learned by production or operation of the embodiments. The features of the present disclosure may be realized and obtained by practicing or using the various aspects of the methods, means, and combinations set forth in the detailed examples discussed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The present disclosure will be further described in the form of exemplary embodiments, which will be described in detail by way of the accompanying drawings. These embodiments are not limiting, and in these embodiments, the same number represents the same structure, wherein:

[0039] Figure 1 is a schematic diagram of an exemplary medical radiation system according to some embodiments of the present disclosure;

[0040] Figure 2is a schematic diagram of hardware and / or software components of an exemplary computing device according to some embodiments of the present disclosure;

[0041] Figure 3 is a schematic diagram of hardware and / or software components of an exemplary mobile device according to some embodiments of the present disclosure;

[0042] Figure 4 is an angular schematic diagram of an exemplary rack according to some embodiments of the present disclosure;

[0043] Figure 5 is a schematic diagram of angles of exemplary collimators according to some embodiments of the present disclosure;

[0044] Figure 6 is a schematic diagram of exemplary sub-field shapes according to some embodiments of the present disclosure;

[0045] Figure 7 is a flowchart of an exemplary training process for generating a subfield shape estimation model according to some embodiments of the present disclosure;

[0046] Figure 8 is a flowchart of an exemplary process of obtaining training data according to some embodiments of the present disclosure;

[0047] Fig. 9 is a flowchart of an exemplary process of obtaining training data according to some embodiments of the present disclosure;

[0048] Fig.10 is a flowchart of an exemplary process of determining a sub-field shape estimation model according to some embodiments of the present disclosure;

[0049] Fig.11 is a block diagram of a processing device according to some embodiments of the present disclosure;

[0050] Fig.12 FIG. 4 is a flowchart of an exemplary process of estimating a sub-field shape of an MLC based on a sub-field shape estimation model according to some embodiments of the present disclosure. DETAILED DESCRIPTION

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. However, it should be understood by those skilled in the art that the present application can be implemented without these details. In other cases, in order to avoid unnecessarily obscuring the various aspects of the present application, well-known methods, processes, systems, components and / or circuits have been described at a higher level. For those of ordinary skill in the art, it is obvious that various changes can be made to the disclosed embodiments, and without departing from the principles and scope of the present application, the general principles defined in the present application can be applied to other embodiments and application scenarios. Therefore, the present application is not limited to the embodiments shown, but conforms to the broadest scope consistent with the scope of the application.

[0052] The terms used herein are only used for the purpose of describing specific exemplary embodiments and are not intended to be limiting. As used herein, the singular forms "a", "an" and "the" may also be intended to include plural forms, unless the context clearly indicates an exception. It should also be understood that, as in the present application specification, the terms "include" and / or "comprise" only indicate the presence of the features, wholes, steps, operations, components and / or parts, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, components, parts and / or combinations thereof.

[0053] It should be understood that the terms "system", "unit", "module" and / or "block" used herein are methods for distinguishing different components, elements, parts, portions or assemblies at different levels in ascending order. However, if these terms achieve the same purpose, they may be replaced by another term.

[0054] Generally, the terms "module", "unit" or "block" as used herein refer to logic embodied in hardware or firmware, or a collection of software instructions. The modules, units or blocks described herein may be implemented as software and / or hardware, and may be stored in any type of non-transitory computer-readable medium or other storage device. In some embodiments, software modules / units / blocks may be compiled and linked into an executable program. It should be understood that software modules may be called from other modules / units / blocks or from themselves, and / or may be called in response to detected events or interrupts. A module configured to execute a command on a computing device (e.g., Figure 2The software modules / units / blocks executed on the processor 210 shown in the figure can be provided on a computer-readable medium, such as a compact disc, a digital video disc, a flash drive, a disk, or any other tangible medium, or as a digital download (and can be initially stored in a compressed or installable format, which needs to be installed, decompressed or decrypted before execution). The software code here can be stored in part or in whole in the storage device of the computing device performing the operation and applied in the operation of the computing device. The software instructions can be embedded in firmware, such as an erasable programmable read-only memory (EPROM). It should also be understood that the hardware modules / units / blocks can include connected logical components, such as gates and triggers, and / or can include programmable units, such as programmable gate arrays or processors. The modules / units / blocks or computing device functions described here can be implemented as software modules / units / blocks, but can be represented by hardware or firmware. Generally, the modules / units / blocks described here refer to logical modules / units / blocks, which can be combined with other modules / units / blocks or divided into sub-modules / sub-units / sub-blocks, regardless of their physical organization or storage. The present application can be applied to a system, an engine, or a part thereof.

[0055] It is understood that, unless the context clearly indicates otherwise, when a unit, engine, module or block is referred to as being "connected," "connected," or "coupled" to another unit, engine, module or block, it may be directly connected, connected, coupled or communicate with another unit, engine, module or block, or there may be an intermediate unit, engine, module or block, unless the context clearly indicates otherwise. In this application, the term "and / or" includes one or more of the related listed items or combinations.

[0056] These and other features, characteristics and functions and methods of operation of the related structural elements of the present application, as well as the combination of components and manufacturing economy, may become more apparent from the following description of the accompanying drawings, which form a part of the specification of this application. However, it should be understood that the drawings are for illustration and description purposes only and are not intended to limit the scope of the present application. It should be understood that the drawings are not drawn to scale.

[0057] For the purpose of illustration, the following description is provided to help better understand the process of exposure control. It should be understood that it is not intended to limit the scope of the present disclosure. For those of ordinary skill in the art, a certain number of changes, modifications and / or modifications may be made under the guidance of the present disclosure. These changes, modifications and / or modifications do not depart from the scope of the present disclosure.

[0058] In this disclosure, the terms "radiation therapy," "radiation therapy," "radiotherapy," and "treatment" are used interchangeably to refer to treatment using radiation to treat, for example, cancer and other diseases in biological (e.g., human and animal) tissue. The terms "treatment plan," "plan for treatment," and "radiation therapy plan" are used interchangeably to refer to the plan for performing radiation therapy.

[0059] Figure 1 1 is a schematic diagram of an exemplary medical radiation system according to some embodiments of the present disclosure. In some embodiments, the medical radiation system 100 can be applied to any radiation therapy scenario using a multi-leaf collimator (MLC). As used herein, the terms "therapy", "radiation therapy", "radiotherapy" and "radiotherapy" are used interchangeably. In some embodiments, the medical radiation system 100 can be applied to intensity modulated radiation therapy (IMRT), intensity modulated rotation therapy (IMAT), volumetric intensity modulated rotation therapy (VMAT), image guided radiation therapy (IGRT), single arc radiation therapy, multi-arc radiation therapy, etc.

[0060] like Figure 1 , the medical radiation system 100 may include a radiation device 110, a network 120, one or more terminals 130, a processing device 140, and a storage device 150. In some embodiments, the medical radiation system 100 may further include an imaging device 160. In some embodiments, the radiation device 110 and the imaging device 160 may be integrated into a single device or separate devices. In some embodiments, the imaging device 160 may be omitted in the medical radiation system 100.

[0061] In some embodiments, the components in the medical radiation system 100 can be connected in one or more ways. By way of example only, the radiation device 110 can be connected to the processing device 140 via the network 120. As another example, the radiation device 110 can be directly connected to the processing device 140, as shown by the double-headed arrow in the dashed line connecting the radiation device 110 and the processing device 140. As a further example, the storage device 150 can be directly connected to the processing device 140 or connected to the processing device 140 through the network 120. As yet another example, the terminal 130 can be directly connected to the processing device 140 (as shown by the double-headed arrow in the dashed line connecting the terminal 130 and the processing device 140) or connected to the processing device 140 through the network 120. As yet another example, the imaging device 160 can be connected to the radiation device 110 directly or through the network 120.

[0062] In this disclosure, Figure 1 The X-axis, Y-axis, and Z-axis shown in may form an orthogonal coordinate system. Figure 1The X-axis and Z-axis shown in the figure may be horizontal, and the Y-axis may be vertical. As shown in the figure, the positive X-direction along the X-axis may be from the left side to the right side of the radiation device 110 as seen from the direction facing the front of the radiation device 110; Figure 1 The positive Y direction of the Y axis shown can be from the bottom to the top of the radiation device 110; Figure 1 The positive Z direction of the Z axis shown may refer to a direction in which the subject is moved away from the table 113 of the radiation device 110 .

[0063] In some embodiments, the radiation device 110 can be configured to deliver radiation to a planned target volume (PTV) (also referred to as a target, e.g., a lesion) of a subject (e.g., a patient) to treat the subject. The radiation device 110 may include a gantry 111, a radiation source 112, and a table 113. In some embodiments, the radiation source 112 may be connected to the gantry 111 and may rotate around the subject (e.g., a Z-axis). During radiation therapy of a subject using the radiation device 110, the subject may be placed on the table 113 and moved into a treatment area of ​​the radiation device 110, such as Figure 1 The treatment area 114 is in the treatment area.

[0064] The radiation source 112 may emit radiation rays to an object placed in the examination table 113. The radiation rays may include X-rays, gamma rays, alpha rays, ultraviolet rays, lasers, neutrons, protons, etc. or a combination thereof. A multi-leaf collimator (MLC) 115 (e.g., Figure 4 The MLC 115 may be attached to the radiation source 112. The MLC 115 may include adjustable blades that act as a filter, blocking or allowing the radiation to pass through so as to adjust the shape of the radiation rays to the shape of the target of the subject (e.g., a tumor) while reducing the risk of adjacent organs (OARs) of the exposed target (e.g., normal tissue immediately adjacent to the target) under radiation. In some embodiments, the MLC 115 may rotate with the radiation source 112. In some embodiments, the position of the MLC 115 relative to the radiation source 112 may be constant or substantially constant during rotation.

[0065] In some embodiments, the object can be a living being. By way of example only, the object can include a patient, an artificial object, etc. As another example, the object can include a specific part, organ, and / or tissue of a patient. For example, the object can include the head, brain, neck, body, shoulder, arm, chest, heart, stomach, blood vessel, soft tissue, knee, foot, etc., or any combination thereof. In the present disclosure, "subject" and "object" are used interchangeably.

[0066] In some embodiments, the imaging device 160 may be configured to acquire one or more images of a target of the object. In some embodiments, the imaging device 160 may include a single-modality imaging device and / or a multi-modality imaging device. The single-modality imaging device may include, for example, a magnetic resonance imaging (MRI) device, a computed tomography (CT) device, a positron emission tomography (PET) device, a single-photon emission computed tomography (SPECT) device, an ultrasound device, an X-ray device, etc. For example, the multi-modality imaging device may include a computed tomography-magnetic resonance imaging (MRI-CT) device, a positron emission tomography-magnetic resonance imaging (PET-MRI) device, a single-photon emission computed tomography-magnetic resonance imaging (SPECT-MRI) device, a digital subtraction angiography-magnetic resonance imaging (DSA-MRI) device, a positron emission tomography-computed tomography (PET-CT) device, a single-photon emission computed tomography-computed tomography (SPECT-CT) device, etc.

[0067] The network 120 can facilitate the exchange of information and / or data. In some embodiments, one or more components of the medical radiation system 100 (e.g., the radiation device 110, the terminal 130, the processing device 140, the storage device 150, or the imaging device 160) can send information and / or data to another component(s) in the medical radiation system 100 via the network 120. For example, the processing device 140 can obtain user instructions from the terminal 130 via the network 120. As another example, the processing device 140 can obtain scan data from the imaging device 160 via the network 120. In some embodiments, the network 120 can be any type of wired or wireless network, or a combination thereof. The network 120 may be and / or include a public network (e.g., the Internet), a private network (e.g., a local area network (LAN), a wide area network (WAN), etc.), a wired network (e.g., Ethernet), a wireless network (e.g., an 802.11 network, a Wi-Fi network), a cellular network (e.g., a Long Term Evolution (LTE) network), a frame relay network, a virtual private network (“VPN”), a satellite network, a telephone network, a router, a hub, a switch, a server computer, and / or any combination thereof. By way of example only, the network 120 may include a cable network, a wired network, a fiber optic network, a telecommunications network, an intranet, a wireless local area network (WLAN), a metropolitan area network (MAN), a public switched telephone network (PSTN), a Bluetooth TM Network, ZigBee TMIn some embodiments, the network 120 may include one or more network access points. For example, the network 120 may include a wired or wireless network access point, such as a base station and / or an Internet exchange point, through which one or more components of the medical radiation system 100 may connect to the network 120 to exchange data and / or information.

[0068] The terminal 130 includes a mobile device 131, a tablet computer 132, a laptop computer 133, etc. or any combination thereof. In some embodiments, the mobile device 131 may include a smart home device, a wearable device, a smart mobile device, a virtual reality device, an augmented reality device, etc. or any combination thereof. In some embodiments, the smart home device may include a smart lighting device, a control device for a smart appliance, a smart monitoring device, a smart TV, a smart camera, an intercom, etc. or any combination thereof. In some embodiments, the wearable device may include a bracelet, footwear, glasses, a helmet, a watch, clothing, a backpack, accessories, etc. or any combination thereof. In some embodiments, the smart mobile device may include a smart phone, a personal digital assistant (PDA), a gaming device, a navigation device, a point of sale (POS) device, etc. or any combination thereof. In some embodiments, the virtual reality device and / or the augmented reality device may include a virtual reality helmet, virtual reality glasses, a virtual reality patch, an augmented reality helmet, augmented reality glasses, an augmented reality patch, etc. or any combination thereof. For example, the virtual reality device and / or the augmented reality device may include Google Glass, Oculus Rift, HoloLens, Gear VR, etc. In some embodiments, the terminal 130 may remotely operate the radiation device 110. In some embodiments, the terminal 130 may operate the radiation device 110 through a wireless connection. In some embodiments, the terminal 130 may operate the radiation device 110 through a wireless connection. In some embodiments, the terminal 130 may receive information and / or instructions input by a user, and send the received information and / or instructions to the radiation device 110 or the processing device 140 through the network 120. In some embodiments, the terminal 130 may receive data and / or information from the processing device 140. In some embodiments, the terminal 130 may be a part of the processing device 140. In some embodiments, the terminal 130 may be omitted.

[0069] In some embodiments, the processing device 140 may process data obtained from the radiation device 110, the terminal 130, the storage device 150, or the imaging device 160. For example, the processing device 140 may determine the sub-field shape of the treatment plan based on the sub-field shape estimation model. The processing device 140 may be a central processing unit (CPU), a digital signal processor (DSP), a system on a chip (SoC), a microcontroller unit (MCU), or any combination thereof.

[0070] In some embodiments, the processing device 140 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, the processing device 140 may be local or remote. For example, the processing device 140 may access information and / or data stored in the radiation device 110, the terminal 130, the imaging device 160, and / or the storage device 150 through the network 120. As another example, the processing device 140 may be directly connected to the radiation device 110, the terminal 130, the imaging device 160, and / or the storage device 150 to access the stored information and / or data. In some embodiments, the processing device 140 may be implemented on a cloud platform. By way of example only, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, a cross-cloud, a multi-cloud, etc., or any combination thereof. In some embodiments, the processing device 140 may be implemented on a cloud platform in the present disclosure. Figure 2 One or more components illustrated in the diagram are implemented on a computing device 200.

[0071] The storage device 150 may store data and / or instructions. In some embodiments, the storage device 150 may store data obtained from the terminal 130 and / or the processing device 140. For example, the storage device 150 may store one or more images obtained from the imaging device 160. As another example, the storage device 150 may store a sub-field shape estimation model. In some embodiments, the storage device 150 may store data and / or instructions, and the processing device 140 may execute or use these data and / or instructions to perform the exemplary methods described in the present disclosure. For example, the storage device 150 may store instructions, and the processing device 140 may execute or use these instructions to determine the sub-field shape of the treatment plan based on the sub-field shape estimation model. In some embodiments, the storage device 150 may include a mass storage device, a removable storage device, a volatile read-write memory, a read-only memory (ROM), etc., or any combination thereof. Exemplary mass storage devices may include magnetic disks, optical disks, solid-state drives, etc. Exemplary removable storage devices may include flash drives, floppy disks, optical disks, memory cards, compressed disks, tapes, etc. Exemplary volatile read-write memory may include random access memory (RAM). Exemplary RAM may include dynamic random access memory (DRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), static random access memory (SRAM), thyristor random access memory (T-RAM), zero capacitance random access memory (Z-RAM), etc. Exemplary ROM may include mask read-only memory (MROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), compact disk read-only memory (CD-ROM), digital versatile disk read-only memory ROM, etc. In some embodiments, the storage device 150 may be implemented on a cloud platform. By way of example only, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, a cross-cloud, a multi-cloud, etc., or any combination thereof.

[0072] In some embodiments, the storage device 150 can be connected to the network 120 to communicate with one or more components of the medical radiation system 100 (e.g., the radiation device 110, the imaging device 160, the terminal 130, the processing device 140). One or more components of the medical radiation system 100 can access data or instructions stored in the storage device 150 through the network 120. In some embodiments, the storage device 150 can be directly connected to or communicate with one or more components of the medical radiation system 100 (e.g., the terminal 130, the processing device 140). In some embodiments, the storage device 150 can be part of the processing device 140.

[0073] Figure 2is a schematic diagram of hardware and / or software components of an exemplary computing device according to some embodiments of the present disclosure. Figure 2 As shown, computing device 200 may include processor 210, memory 220, input / output (I / O) 230, and communication port 240. In some embodiments, processing device 140 and / or terminal 130 may be implemented on computing device 200.

[0074] The processor 210 can execute computer instructions (program code) and when executing the instructions, cause the processing device 140 to perform the functions of the processing device 140 according to the techniques described herein. For example, the computer instructions may include routines, programs, objects, components, signals, data structures, processes, modules, and functions that perform specific functions described herein. In some embodiments, the processor 210 can process data and / or images obtained from the radiation device 110, the terminal 130, the storage device 150, the imaging device 160, and / or any other component of the medical radiation system 100. For example, the processing device 140 can determine the sub-field shape of the treatment plan based on the sub-field shape estimation model. In some embodiments, the processor 210 may include one or more hardware processors, such as a microcontroller, a microprocessor, a reduced instruction set computer (RISC), an application-specific integrated circuit (ASIC), an application-specific instruction set processor (ASIP), a central processing unit (CPU), a graphics processing unit (GPU), a physical processing unit (PPU), a microcontroller unit, a digital signal processor (DSP), a field programmable gate array (FPGA), an advanced RISC machine (ARM), a programmable logic device (PLD), and any circuits and processors capable of performing one or more functions, or any combination thereof.

[0075] For illustration purposes only, one processor is described in the computing device 200. However, it should be noted that the computing device 200 in the present disclosure may also include multiple processors. Therefore, the operations and / or method steps performed by one processor as described in the present disclosure may also be performed by multiple processors jointly or individually. For example, if in the present disclosure, the processor of the computing device 200 executes process A and process B simultaneously, it should be understood that in the computing device 200, process A and process B may also be executed jointly or separately by two or more different processors (e.g., the first processor executes process A, the second processor executes process B, or the first and second processors jointly execute process A and process B).

[0076] The memory 220 may store data / information obtained from the radiation device 110, the terminal 130, the storage device 150, or any other component of the medical radiation system 100. In some embodiments, the storage device 150 may include a mass storage device, a removable storage device, a volatile read-write memory, a read-only memory (ROM), etc., or any combination thereof. Exemplary mass storage may include a magnetic disk, an optical disk, a solid-state drive, etc. Exemplary removable storage may include a flash drive, a floppy disk, an optical disk, a memory card, a compressed disk, a tape, etc. Exemplary volatile read-write memory may include a random access memory (RAM). Exemplary RAM may include a dynamic random access memory (DRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), a static random access memory (SRAM), a thyristor random access memory (T-RAM), a zero-capacitance random access memory (Z-RAM), etc. Exemplary ROMs may include mask read-only memory (MROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), compact disk read-only memory (CD-ROM), digital versatile disk read-only memory ROM, etc. In some embodiments, the memory 220 may store one or more programs and / or instructions to perform the exemplary methods described in the present disclosure. For example, the storage device 150 may store instructions that the processing device 140 may execute or use to determine the sub-field shape of the treatment plan based on the sub-field shape estimation model.

[0077] Input / output (I / O) 230 may input and / or output signals, data, information, etc. In some embodiments, I / O 230 may enable a user to interact with processing device 140. In some embodiments, I / O 230 may include input devices and output devices. Exemplary input devices may include a keyboard, a mouse, a touch screen, a microphone, etc., or a combination thereof. Exemplary output devices may include a display device, a speaker, a printer, a projector, etc., or a combination thereof. Exemplary display devices may include a liquid crystal display (LCD), a light emitting diode (LED) based display, a flat panel display, a curved screen, a television device, a cathode ray tube (CRT), etc., or a combination thereof.

[0078] The communication port 240 can be connected to a network (e.g., network 120) to facilitate data communication. The communication port 240 can establish a connection between the processing device 140 and the radiation device 110, the terminal 130, or the storage device 150. The connection can be a wired connection, a wireless connection, or a combination of the two that enables data transmission and reception. The wired connection may include an electric cable, an optical cable, a telephone line, etc., or any combination thereof. The wireless connection may include a Bluetooth connection, a Wi-Fi connection, a WiMax connection, a WLAN connection, a ZigBee connection, a mobile network connection (e.g., 3G, 4G, 5G), etc., or any combination thereof. In some embodiments, the communication port 240 may be a standardized communication port, such as RS232, RS485, etc. In some embodiments, the communication port 240 may be a specially designed communication port. For example, the communication port 240 may be designed according to the Digital Imaging and Communications in Medicine (DICOM) protocol.

[0079] Figure 3 1 is a schematic diagram of hardware and / or software components of an exemplary mobile device according to some embodiments of the present disclosure. In some embodiments, the processing device 140 and / or the terminal 130 may be implemented on the computing device 200. Figure 3 As shown, the mobile device 300 may include a display 310, a communication platform 320, a graphics processing unit (GPU) 330, a central processing unit (CPU) 340, an I / O 350, a memory 360, and a storage 390. In some embodiments, any other suitable components, including but not limited to a system bus or a controller (not shown in the figure), may also be included in the mobile device 300. In some embodiments, a mobile operating system 370 (e.g., iOS, Android, Windows Phone, etc.) and one or more applications 380 may be loaded from the storage 390 into the memory 360 for execution by the CPU 340. The application 380 may include a browser or any other suitable mobile application for receiving and rendering information related to image processing or other information from the processing device 140. User interaction with the information stream may be implemented through the I / O 350 and provided to the processing device 140 and / or other components of the medical radiation system 100 through the network 120.

[0080] In order to implement the various modules, units and functions thereof described in the present disclosure, a computer hardware platform can be used as a (multiple) hardware platform for one or more elements described herein. The hardware elements, operating systems and programming languages ​​of such computers are conventional in nature, and it is presumed that those skilled in the art are sufficiently familiar with the technology to adopt these technologies to control radiation in medical radiation as described herein. A computer with user interface elements can be used to implement a personal computer (PC) or other type of workstation or terminal device, but if properly programmed, the computer can also act as a server. Those skilled in the art are familiar with the structure, programming and general operation of such computer equipment, and therefore, the accompanying drawings should be self-explanatory.

[0081] Radiation therapy can be used to treat, for example, cancer and other diseases in biological (e.g., human and animal) tissues. Radiation therapy can be used with radiation devices (e.g., Figure 1 An exemplary radiation therapy is provided using a radiation device 110 in the apparatus, whereby a target (e.g., a tumor) can be irradiated by high-energy particles (e.g., electrons, photons, ions, etc.). In radiation therapy, multiple radiation rays can be directed at a target from different angles. A target (e.g., a lesion) can be referred to as a planned target volume (PTV). Normal tissue immediately surrounding the target can be referred to as an organ at risk (OAR). A treatment plan can be a process involving determining and / or updating specific radiation therapy parameters to achieve a treatment goal. The result of the treatment plan can be a treatment plan. The treatment planning process can include depicting one or more targets and one or more OARs from one or more medical images of the patient, and specifying radiation therapy parameters based on the depiction results.

[0082] by Figure 1 Taking the target treatment plan corresponding to the target treatment performed by the radiation device 110 on the target (e.g., a lesion) of an object (e.g., a patient) as an example, the radiation therapy parameters of the target treatment plan may include the total time of the target treatment, the total dose within the total time, a plurality of target control points within the total time (also referred to as target control time points), target radiation information corresponding to each of the plurality of target control points or the like, or any combination thereof.

[0083] In some embodiments, the target radiation information corresponding to each of the multiple target control points may include dose output rate, dose output per pulse, output dose, dose distribution in the target, target position combination, sub-field shape, etc., or any combination thereof.

[0084] In some embodiments, the dose distribution in the target corresponding to the target control point may refer to the dose distribution in the target during the time period at which the target control point ends. In some embodiments, the dose output rate corresponding to the target control point may refer to the dose output rate during the time period at which the target control point ends. In some embodiments, the dose output per pulse corresponding to the target control point may refer to the dose output per pulse during the time period at which the target control point ends. In some embodiments, the output dose corresponding to the target control point may refer to the output dose during the time period at which the target control point ends. For example, the time period may be from the beginning of the total time (e.g., the first control point among multiple target control points) to the target control point. As another example, the time period may be between the target control point and the control point before the target control point.

[0085] In some embodiments, a target position combination or a sub-field shape corresponding to a target control point may indicate a position of the radiation device 110 on the target control point.

[0086] In some embodiments, the target position combination corresponding to the target control point may indicate the position of the MLC 115 relative to the target at the target control point.

[0087] In some embodiments, the target position combination corresponding to the target control point may include a combination of one or more positions of one or more components of the radiation device 110 operating at the target control point. For example, the target position combination corresponding to the target control point may include the gantry angle of the radiation source 112 at the target control point. As another example, the target position combination corresponding to the target control point may include a combination of the gantry angle of the radiation source 112 and the collimator angle of the MLC 115 at the target control point (also referred to as a gantry-collimator angle combination). As yet another example, the target position combination corresponding to the target control point may include a combination of the gantry angle of the radiation source 112, the collimator angle of the MLC 115, and the table position of the table 113 at the target control point.

[0088] In some embodiments, during target treatment, the radiation source 112 may be positioned around Figure 1 The Z-axis shown rotates and emits radiation to the target from different angles. The gantry angle (also called beam angle) corresponding to the target control point can indicate the position of the radiation source 112 in the radiation device 110 at the target control point. For example, when the radiation source 112 rotates to the position of the target control point, the corresponding gantry angle can refer to the angle between the center line of the radiation ray emitted by the radiation source 112 at this position and the reference direction (for example, the direction perpendicular to the ground).

[0089] Just as an example, Figure 4 is a schematic diagram of exemplary rack angles according to some embodiments of the present disclosure. Figure 4The radiation device 110 shown in FIG. 1 is viewed from a direction facing the front of the radiation device 110 (e.g., along Figure 1 and 4 A view of the radiation device 110 as seen in the negative Z direction (shown in FIG. 1 ). Figure 4 The X, Y, and Z axes correspond to Figure 1 The X-axis, Y-axis and Z-axis in Figure 4 In the diagram, the positive Z direction is represented as the direction perpendicular to the paper and pointing outward.

[0090] like Figure 4 As shown, the radiation source 112 can rotate around the Z axis along the trajectory 116. The MLC 115 can rotate around the Z axis together with the radiation source 112. The radiation source 112 can emit a radiation line (or radiation beam) 440. The center line 430 of the radiation line 440 is Figure 4 The dashed line 420 perpendicular to the ground (along Figure 1 and Figure 4 The Y direction shown in FIG. 4A can be designated as a reference line. When the radiation source 112 is rotated to the position 410, the angle between the dashed line 420 and the center line 430 of the radiation line 440 is 90°. The gantry angle of the radiation source 112 at the position 410 is 90°.

[0091] In some embodiments, the MLC 115 of the radiation device 110 may be rotated about the Z-axis of the radiation source 112 and about the centerline of the MLC 115 (e.g., coincident with the centerline 430 of the radiation line 440 emitted from the radiation source 112). The collimator angle corresponding to the target control point may indicate the position of the MLC 115 at the target control point relative to the radiation source 112. For example, when the MLC 115 is rotated about its centerline to the position of the target control point, the corresponding collimator angle may refer to the angle between the position and the initial position of the MLC 115.

[0092] Just as an example, Figure 5 is a schematic diagram of exemplary collimator angles according to some embodiments of the present disclosure. Figure 5 The MLC 115 shown in FIG. 1 is perpendicular to the centerline of the MLC 115 (eg, Figure 4 The cross section of the MLC 115 is shown in FIG. 4 (with the center line 430 of the radial line 440 coincident). The MLC 115 may be centered about its center line (e.g., Figure 4 430) of the radial lines 440 shown in FIG. Figure 5 direction R in the rotation).

[0093] The location 520 of the MLC 115 (e.g., at Figure 5) can be designated as the initial position. When the MLC 115 is rotated in the direction R to the position 510 (eg, Figure 5 When the initial position 520 is 90° away from the position 510 (shown by a solid line in FIG. 1 ), the angle between the initial position 520 and the position 510 is 90°. The collimator angle of the MLC 115 at the position 510 is 90°.

[0094] In some embodiments, the examination table 113 can be arranged along Figure 1 In some embodiments, the table 113 can be moved around Figure 1 The position of the table 113 may indicate at least one of a translation of the table 113 along the X direction, a translation of the table 113 along the Y direction, a translation of the table 113 along the Z direction, a rotation of the table 113 around the Y direction, a rotation of the table 113 around the Z direction, and a rotation of the table 113 around the X direction.

[0095] In some embodiments, the MLC 115 may include a plurality of blade pairs, each blade pair including two blades disposed opposite to each other. In some embodiments, at a target control point, the plurality of blade pairs may be individually moved to respective target blade positions, thereby forming an aperture of a specific shape to establish a treatment beam that approximates a target shape. The shape of the aperture of the MLC 115 at the target control point may be referred to as a target subfield shape of the MLC 115 at the target control point.

[0096] In some embodiments, the target sub-field shape of the MLC 115 at the target control point may include a target blade position for each of a plurality of blade pairs. For example, the target blade position of a blade pair may include an opening position (e.g., a center position of an opening) and an opening width formed by the blade pair. As another example, the blade pair may include a first blade and a second blade. The target blade position of the blade pair may include a first target position of the first blade and a second target position of the second blade. In some embodiments, the first target position of the first blade may be represented as the position of the center point of the end of the first blade. The end of the first blade may be directly opposite the second blade. The second target position of the second blade may be represented as the position of the center point of the end of the second blade. The end of the second blade may be directly opposite the first blade. In some embodiments, each of the plurality of blade pairs may be numbered. The target blade position information of the blade pair may also include the serial number of the blade pair.

[0097] For example, Figure 6 is a schematic diagram of an exemplary sub-field shape of the MLC 115 according to some embodiments of the present disclosure. Figure 6 The MLC 115 shown in FIG. 1 is perpendicular to the centerline of the MLC 115 (eg, Figure 4As shown, the MLC 115 includes 7 blade pairs (e.g., Figure 6 ), which can be traced along Figure 6 The 7 pairs of blades are numbered N1-N7. At the target control point, the blades of MLC 115 may form a subfield shape 610 (e.g., Figure 6 The white area surrounded by the shaded block in the figure).

[0098] Please refer to the figure as shown Figure 6 , a coordinate system is established based on MLC 115. The coordinate system includes a coordinate axis 620 along a direction L. The positive direction of the coordinate axis 620 is from left to right. A dashed line 630 is designated as the origin of the coordinate system.

[0099] In some embodiments, the target blade position of the blade pair of MLC 115 at the target control point can be expressed as (N, c, w), where N refers to the number of the blade pair, c refers to the position, such as the center point of the opening formed by the blade pair in a coordinate system (e.g., a portion of the sub-field shape 610), and w refers to the width of the opening along the direction L. In some embodiments, the width w can be formed by positioning the blade pair symmetrically relative to the center c of the blade pair.

[0100] For example, the target leaf position of leaf pair N1 at the target control point can be expressed as (N1, 0, 0) because leaf pair N1 is closed at the origin 630. As another example, the target leaf position of leaf pair N2 at the target control point can be expressed as (N2, c1, 0) because leaf pair N2 is closed at position c1 in the coordinate system. The value of c1 is positive. For another example, the target leaf position of leaf pair N3 at the target control point can be expressed as (N3, c4, d2), where c4 represents the position of the center point of the opening (at a portion of the sub-field shape 610) formed by leaf pair N3 in the coordinate system and is positive, and d2 refers to the opening width along the direction L. As yet another example, the target leaf position of leaf pair N4 at the target control point can be expressed as (N4, c3, d1), where c3 represents the position of the center point of the opening (at a portion of the sub-field shape 610) formed by leaf pair N4 in the coordinate system and is negative, and d1 refers to the opening width along the direction L. As another example, the target blade position of blade pair N7 at the target control point may be expressed as (N7, c2, 0) because blade pair N7 is closed at position c2 in the coordinate system. The value of c2 is negative.

[0101] In some embodiments, one of the seven blade pairs may include a first blade and a second blade. The target blade position of the blade pair at the target control point may be expressed as (N, p1, p2), where p1 refers to the position of the first blade in the coordinate system, and p2 refers to the position of the second blade in the coordinate system. The position of the first blade may be expressed as the center point position of the end of the first blade. The end of the first blade may face the second blade. The position of the second blade may be expressed as the center point position of the end of the second blade. The end of the second blade may face the first blade.

[0102] For example, since blade pair N1 is closed at the target control point, the blade positions of blade pair N1 at the target control point may be the same. As another example, the difference between the coordinates of the blade positions of blade pair N3 along axis 620 at the target control point may be equal to d2.

[0103] In some embodiments, the present disclosure provides a training process (e.g., in conjunction with Figure 7-10 ) to determine a sub-field shape estimation model, the sub-field shape estimation model is used to determine a target sub-field shape corresponding to each target position combination at each point in a plurality of target control points of a target treatment plan. In some embodiments, the present disclosure also provides a process (for example, in combination with Fig.12 ), which is used to use the sub-field shape estimation model to determine the target sub-field shape corresponding to each target position combination of multiple target control points in the target treatment plan.

[0104] Figure 7 FIG. 1 is a flowchart of an exemplary training process for generating a subfield shape estimation model according to some embodiments of the present disclosure. Figure 7 One or more operations of the process 700 shown may be performed at Figure 1 The medical radiation system 100 shown in the figure or a third party (e.g., an external device) is implemented. For example, Figure 7 The process 700 shown in FIG. 7 can be stored in the form of instructions in a storage device (eg, storage device 150 and / or memory 220) of the medical radiation system 100 and processed by a processing device 140 (eg, Figure 2 The processor 210 of the computing device 200 shown in FIG. Figure 3 The CPU 340 of the mobile device 300 shown, or Figure 3 One or more modules of the processing device 140 shown (e.g., Fig.11 The operations of the illustrated process presented below are intended to be illustrative. In some embodiments, process 700 may be accomplished by one or more additional operations not described and / or one or more operations not discussed. In addition, Figure 7The order of operations of the process 700 shown in FIG. 7 is the same. Figure 7 The order in which the operations of process 700 are illustrated and described below is not limiting.

[0105] Step 710, an initial model may be constructed. In some embodiments, a framework of the initial model may be constructed based on the position universal set and the distance universal set, so that the obtained sub-field shape estimation model may be used to output multiple target position combinations within multiple discrete candidate position combinations in the position universal set, and multiple target sub-field shapes corresponding to the multiple target position combinations and within the distance universal set.

[0106] In some embodiments, the position general set may include multiple discrete candidate position combinations. In some embodiments, multiple candidate gantry angles, multiple candidate table positions, or multiple candidate collimator angles may be obtained. The position general set may be obtained based on multiple candidate gantry angles, multiple candidate collimator angles, or multiple table positions. For example, each of the multiple candidate position combinations in the position general set may include one of the multiple candidate gantry angles. As another example, each of the multiple candidate position combinations in the position general set may include one of the multiple candidate gantry angles and the multiple candidate collimator angles. As another example, each of the multiple candidate position combinations in the position general set may include a combination from one of the multiple candidate gantry angles, the multiple candidate collimator angles, and the multiple candidate table positions.

[0107] In some embodiments, the element type in the position general set may be determined based on the type of treatment plan to which the sub-field shape estimation model is expected to be applied. For example, if a sub-field shape estimation model needs to be generated to determine a sub-field shape corresponding to a gantry-collimator angle combination, the position general set may include a combination of multiple candidate gantry angles and multiple candidate collimator angles.

[0108] Taking the gantry-collimator angle combination as an example, a candidate gantry angle range and a candidate collimator angle range can be obtained. Multiple discrete candidate gantry angles and multiple discrete candidate collimator angles can be obtained from the candidate gantry angle range and the candidate collimator angle range, respectively. For example, multiple discrete candidate gantry angles can be selected from the range of candidate gantry angles based on a first angular resolution. The difference between any two adjacent candidate gantry angles in the multiple candidate gantry angles can be equal to the first angular resolution. As another example, multiple discrete candidate collimator angles can be selected from the range of candidate collimator angles based on a second angular resolution. The difference between any two adjacent candidate collimator angles in the multiple candidate collimator angles can be equal to the second angular resolution. The first angular resolution and the second angular resolution can be the same or different.

[0109] As an example only, the range of candidate gantry angles and the range of candidate collimator angles may be set to 360°. A plurality of discrete candidate gantry angles may be selected from the range of candidate gantry angles, each angle being spaced 0.5° (a first angular resolution). For example, a plurality of candidate gantry angles may be represented as {0°, 0.5°, 1.0°, ..., 359.5°}. A plurality of discrete candidate collimator angles may be selected from the range of candidate collimator angles, each angle being spaced 0.5° (a second angular resolution). For example, a plurality of candidate collimator angles may be represented as {0°, 0.5°, 1.0°, ..., 359.5°}. The position universal set may include 720*720 gantry-collimator angle combinations of 720 candidate gantry angles and 720 candidate collimator angles.

[0110] In some embodiments, a larger candidate gantry angle range and a candidate collimator angle range, and a smaller first and second angular resolution value (such as a higher first and second angular resolution) can increase the output of the obtained sub-field shape estimation model, thereby increasing the application and accuracy of the obtained sub-field shape estimation model. For example, if the range of candidate gantry angles in the position universal set is less than 360°, the generated sub-field shape estimation model may not be applicable to single arc treatment plans and multi-arc treatment plans. For another example, if the first angular resolution of the position universal set is 5°, and the gantry angle in the target treatment plan is accurate to 0.5°, the target sub-field shape determined for the target treatment plan using the acquired sub-field shape estimation model may be inaccurate. For another example, if the first angular resolution of the position universal set is 0.5°, the obtained sub-field shape estimation model can accurately determine the target segment shape of the target treatment plan, where the gantry angle can be accurate to 0.5°, 1°, 2°, 5°, etc.

[0111] In some embodiments, the distance universal set may include multiple discrete candidate blade positions. Assuming that each of the multiple candidate blade positions includes a candidate opening position and a candidate opening width, in order to obtain the distance universal set, multiple candidate opening positions and multiple candidate opening widths can be obtained. Each of the multiple candidate blade positions may include one of the multiple candidate opening positions and one of the multiple candidate opening widths. In some embodiments, a series of candidate opening positions and a series of candidate opening widths can be obtained. Multiple candidate opening positions and multiple candidate opening widths can be obtained from the range of candidate opening positions and the range of candidate opening widths, respectively. For example, multiple candidate opening positions can be selected from the range of candidate opening positions based on a first distance resolution. The difference between any two adjacent candidate opening positions in the multiple candidate opening positions can be equal to the first distance resolution. As another example, multiple candidate opening widths can be selected from the range of candidate opening widths based on a second distance resolution. The difference between any two adjacent candidate opening widths in the multiple candidate opening widths can be equal to the second distance resolution.

[0112] For example, the range of candidate opening positions can be set to -200mm to +200mm, and the range of candidate opening widths can be set to 0-150mm. Multiple discrete candidate opening positions can be selected from the range of candidate opening positions, each with an interval of 0.5mm (first distance resolution). For example, multiple candidate opening positions can be expressed as {-200, -199.5, ..0, 0.5, ..., 200}. Multiple discrete candidate opening widths can be selected from the range of candidate opening widths, each with an interval of 0.5mm (second distance resolution). For example, multiple candidate opening widths can be expressed as {0, 0.5, 1.0, ...150}.

[0113] In some embodiments, a larger range of candidate opening positions and a range of candidate opening widths, as well as smaller values ​​of the first and second distance resolutions (e.g., higher first and second distance resolutions) can increase the application and accuracy of the resulting sub-field shape estimation model. For example, if the range of candidate opening positions and candidate opening widths in the distance universal set is narrow, the resulting sub-field shape estimation model may not be applicable to treatment plans for larger lesions. For another example, if the first and second distance resolutions in the distance universal set are relatively high, the target sub-field shape determined for the target treatment plan using the resulting sub-field shape estimation model can more accurately approximate the shape of the target.

[0114] In some embodiments, an initial model can be further constructed based on the number of blade pairs, so that the acquired sub-field shape estimation model can be used to output blade positions corresponding to multiple candidate position combinations for each blade pair, and applied to MLCs with the same number (or count) of blade pairs. For example, an initial model can be constructed based on a number equal to 60 blade pairs and a position universal set including 720*720 candidate position combinations, and the acquired sub-field shape estimation model is applied to an MLC with 60 blade pairs, and is configured to output a total of 720*720*60, a total of 31,104,000 blade positions corresponding to 720*720 candidate position combinations.

[0115] In some embodiments, the framework of the initial model can be constructed based on a discrete finite position universal set and a discrete finite distance universal set, so that the resulting sub-field shape estimation model can be configured to estimate multiple target position combinations and corresponding target sub-field shapes from the position universal set and the distance universal set. The gantry angle, collimator angle, diagnostic table, and blade position are "discretized" (such as 0.5 degrees or 1 mm) in the space of the position universal set and the distance universal set, so that the overall set of the position universal set and the distance universal set is larger, but still a finite set, which makes the prediction of the target position combination and the target sub-field shape using the sub-field shape estimation model have a controllable operation amount, efficient operation speed, and accurate operation results. For example, the position universal set can include 720*720 gantry-collimator angles, which are a combination of 720 discrete candidate gantry angles {0°, 0.5°, 1.0°, ..., 359.5°} and 720 discrete candidate collimator angles {0°, 0.5°, 1.0°, ..., 359.5°}. The distance universal set may include 801*301 blade positions, which are combinations of 801 discrete candidate opening positions {-200, -199.5, ..0, 0.5, ..., 200} and 301 discrete candidate opening widths {0, 0.5, 1.0, ...150}. The overall set of position universal sets and distance universal sets is large, but still a finite set.

[0116] At step 720, training data may be obtained based on the initial model. In some embodiments, the training data may include multiple training sets. In some embodiments, one of the multiple training data may be obtained based on information about historical treatment plans generated for a sample (e.g., a patient's lesion, also referred to as a sample) based on previous data. Details regarding obtaining training data may be found elsewhere in this disclosure (e.g., with Figure 8 relevant content).

[0117] In step 730 , a subfield shape estimation model may be determined by training an initial model based on training data.

[0118] In some embodiments, the training process can be an offline process in which a database of training data with ground truth measurements is assembled. For example, a database of treatment plans generated from previous treatments of different patients can be constructed. In this database, the training data can be represented by the number (or count) of features. The training process can then learn or train the mapping between features and ground truth by achieving the best fit between predictions and ground truth in a portion or the entire training database.

[0119] During the training process, the training data (including the training input and the corresponding known output) can be input into the constructed initial model, so that the constructed initial model learns how to generalize the information learned in the training process from the training data and provide output for new input data.

[0120] In some embodiments, the initial model may include multiple parameters. The parameter values ​​in the initial model may be updated by performing an iterative back-propagation training procedure (e.g., a stochastic gradient descent back-propagation training technique) to determine the subfield shape estimation model. That is, in order to adjust the parameters of the initial model, the initial model may back-propagate the error determined by the output of the initial model.

[0121] In some embodiments, the sub-field shape estimation model may be a machine learning model. In some embodiments, the sub-field shape estimation model may be a deep learning model. In some embodiments, the sub-field shape estimation model may be a deep convolution-deconvolution (e.g., encoder-decoder) network, such as a U-shaped convolutional neural network (U-Net), a V-shaped convolutional neural network (V-Net), a residual network (Res-Net), or a residual dense network (Red-Net).

[0122] For details related to the training process for determining the subfield shape estimation model, reference may be made to other contents of the present disclosure (e.g., in conjunction with Fig.10 content).

[0123] In some embodiments, as more data is collected, the size of the training database containing the training data may increase. The updated database may be used to retrain the subfield shape estimation model, for example, regularly or irregularly. New samples in the training database may come from unseen treatment plans (e.g., treatment plans that were not used for training or prediction in the past), or from treatment plans that were used for prediction in the past but are now updated.

[0124] In some embodiments, the subfield shape estimation model may be provided by the medical radiation system 100 (eg, the processing device 140, Fig.11The sub-field shape estimation model may be determined and / or updated by the model acquisition module 1120 of the processing device 140 in the medical radiation system 100 or a third party (e.g., an external device). In some embodiments, the medical radiation system 100 may determine and / or update the sub-field shape estimation model offline and store the sub-field shape estimation model in a storage device (e.g., the storage device 150, the memory 220, or the memory 390). In some embodiments, the sub-field shape estimation model may be determined and / or updated (or maintained) by, for example, a manufacturer or supplier of the radiation device 110. For example, the manufacturer or supplier may load the sub-field shape estimation model into the medical radiation system 100 or a part thereof (e.g., the processing device 140) before or during the installation of the radiation device 110 and / or the processing device 140, and maintain or update the sub-field shape estimation model from time to time (regularly or irregularly). The maintenance or update may be implemented by installing a program stored on a storage device (e.g., a CD, a USB drive, etc.), or by a program retrieved from an external source (e.g., a server maintained by a manufacturer or supplier) via the network 120. The procedure may include a new sub-field shape estimation model, or a portion of the model that replaces or supplements a corresponding portion of the sub-field shape estimation model.

[0125] It should be noted that the above description is provided for illustrative purposes only and is not intended to limit the scope of the present disclosure. For those of ordinary skill in the art, various changes and modifications can be made under the guidance of the present disclosure. However, these changes and modifications do not depart from the scope of the present disclosure.

[0126] Figure 8 is a flowchart of an exemplary process of obtaining training data according to some embodiments of the present disclosure. In some embodiments, Figure 8 One or more steps of the process 800 shown in FIG. Figure 1 or a third party (e.g., an external device). Figure 8 The process 800 shown in FIG. 8 may be stored in a storage device (eg, storage device 150 and / or memory 220) of the medical radiation system 100 in the form of instructions and executed by a processing device 140 (eg, Figure 2 The processor 210 of the computing device 200 shown in Figure 3 ) or one or more modules (e.g., a CPU 340 of the mobile device 300 shown in FIG. Fig.11 The steps of the illustrated process presented below are intended to be illustrative. In some embodiments, process 800 may be completed with one or more additional steps not described and / or one or more steps not discussed. In addition, Figure 8The order of the steps of process 800 illustrated in FIG. 8 and described below is not intended to be limiting. In some embodiments, Figure 7 Step 720 of process 700 may be performed based on process 800 .

[0127] In some embodiments, the training data may include multiple training sets. One of the multiple training sets may be obtained based on process 900. In some embodiments, the rest of the multiple training sets may also be obtained based on process 800.

[0128] Step 810, a historical treatment plan previously generated based on a lesion sample (e.g., a sample) can be obtained. In some embodiments, the historical treatment plan can be obtained from a storage device (e.g., storage device 150, memory 220, memory 390, etc.). In some embodiments, the type of the historical treatment plan can correspond to the initial model. For example, the number (or count) of leaf pairs of the MLC corresponding to the historical treatment plan can be equal to the number (or count) of leaf pairs corresponding to the initial model. For another example, if the general set of positions used to construct the initial model includes multiple gantry-collimator angle combinations, the position combination (also referred to as a sample position combination) in the historical treatment plan can be a gantry-collimator angle combination. For another example, the range of sample position combinations in the historical treatment plan can be within the range of candidate position combinations in the general set of positions used to construct the initial model.

[0129] In step 820, one or more first sample images of the lesion sample corresponding to the historical treatment plan can be obtained. In some embodiments, the historical treatment plan is previously generated based on one or more first sample images of the sample. In some embodiments, the one or more first sample images may include an identification of the sample. For example, the outline of the sample can be annotated in the one or more first sample images. In some embodiments, the identification of the sample can be manually performed or modified by a user (e.g., a doctor). In some embodiments, the identification of the sample can be automatically performed using an image segmentation algorithm or an image segmentation model.

[0130] In step 830, a sample position combination and a sample sub-field shape corresponding to each of a plurality of sample control points in the historical treatment plan may be obtained.

[0131] Step 840, a training set may be obtained based on the first sample image, sample position combination, and sample subfield shape of the historical treatment plan. In some embodiments, each corresponding sample position combination and sample subfield shape of the plurality of sample control points may be processed based on the position general set and the distance general set to obtain a training set. For details on obtaining training data, reference may be made to other contents of the present disclosure (e.g., in conjunction with Fig. 9 content).

[0132] In some embodiments, one or more second sample images of normal tissue around the sample, one or more third sample images, and sample radiation information of the historical treatment plan may be obtained. The training set may also be obtained by including one or more second sample images, one or more third sample images, and / or sample radiation information. In some embodiments, one or more first sample images of the sample may be acquired by scanning the sample. For example, the first sample image may include a single-modality image and / or a multi-modality image. For example, a single-modality image may include a magnetic resonance (MR) image, a CT image, a PET image, a SPECT image, an ultrasound image, an X-ray image, and the like. For example, a multi-modality image may include a computed tomography-magnetic resonance imaging (MRI-CT) system, a PET-MR image, a SPECT-MR image, a DSA-MR image, a PET-CT image, a SPECT-CT image, and the like. In some embodiments, the first sample image may be a reconstructed image, for example, a three-dimensional (3D) rendered image, a multi-planar reconstruction (MPR) image, a curved surface reconstruction (CPR) image, and the like.

[0133] In some embodiments, the second sample image of normal tissue surrounding the sample may be a mask image of the surrounding normal tissue separated from at least one of the one or more first sample images. In some embodiments, the third sample image of the sample may be a mask image of the sample separated from at least one of the one or more first sample images. In some embodiments, the sample radiation information may include at least one of a sample output dose, a sample dose output rate, a sample dose distribution in the sample, a sample dose output per pulse, an acceptable dose of normal tissue surrounding the sample, a minimum dose of the sample, etc., or any combination thereof. In some embodiments, the sample radiation information may be determined based on one or more first sample images, one or more second sample images, and / or one or more third sample images. In some embodiments, the sample radiation information may be determined using a dose prediction model.

[0134] It should be noted that the above description is provided for illustrative purposes only and is not intended to limit the scope of the present disclosure. For those of ordinary skill in the art, various changes and modifications can be made under the guidance of the present disclosure. However, these changes and modifications do not depart from the scope of the present disclosure.

[0135] Fig. 9 is a flowchart of an exemplary process of obtaining training data according to some embodiments of the present disclosure. In some embodiments, Fig. 9 One or more steps of the process 900 shown in FIG. Figure 1 or a third party (e.g., an external device). Fig. 9The process 900 shown in FIG. 1 may be stored in a storage device (eg, storage device 150 and / or memory 220) of the medical radiation system 100 in the form of instructions and executed by a processing device 140 (eg, Figure 2 The processor 210 of the computing device 200 shown in Figure 3 ) or one or more modules (e.g., a CPU 340 of the mobile device 300 shown in FIG. Fig.11 The steps of the illustrated process presented below are intended to be illustrative. In some embodiments, process 900 may be completed with one or more additional steps not described and / or one or more steps not discussed. In addition, Fig. 9 The order of the steps of process 900 illustrated in and described below is not intended to be limiting. In some embodiments, Figure 8 Step 840 of process 800 may be performed based on process 900 .

[0136] Step 910, a processed sample position combination located in the position universal set may be obtained. In some embodiments, the sample position combination of the historical treatment plan obtained in step 830 may be processed based on the position universal set to obtain the processed sample position combination.

[0137] As an example only, an initial model may be constructed based on a generic set of positions including 720*720 gantry-collimator angle combinations. 0 , 0.5 0 , 1.0 0 , …, 359.5 0} and 720 candidate collimator angles {0 0 , 0.5 0 , 1.0 0 , …, 359.5 0} Determine 720*720 rack-collimator angle combinations. The sample rack angle and sample collimator angle of the sample rack-collimator angle combination can be modeled to their respective nearest 0.5° points, including the sample rack angle and the sample collimator angle. For example, the sample rack angle of 40.1° can be processed and set to 40°.

[0138] Step 920, the processed sample sub-field shape located within the distance universal set can be obtained. In some embodiments, the sample sub-field shape of the historical treatment plan obtained in step 830 can be processed based on the distance universal set to obtain the processed sample sub-field shape. As an example only, the initial model can be established on a distance universal set containing candidate opening positions {-200, -199.5, ..0, 0.5, ..., 200} and candidate opening widths {0, 0.5, 1.0, ...150}. The sample opening position and sample opening width of the sample sub-field shape can be simulated to the respective nearest 0.5mm points, which include the sample opening position and the sample opening width. For example, a sample opening width of 10.6mm can be processed and set to 10.5mm.

[0139] Step 930, obtaining a sample set including processed sample sub-field shapes and closed sub-field shapes, and the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination. In some embodiments, the closed sub-field shape of the MLC may indicate that all leaf pairs of the MLC are closed. In the sample set, each of a plurality of candidate position combinations may have a corresponding sub-field shape. In some embodiments, there may be 180-240 sample control points in the historical treatment plan. If the position universal set includes 720*720 candidate position combinations, most of the sub-field shapes in the sample set may be closed sub-field shapes. Since the sub-field shapes are closed, the sample set may be considered to be sparse.

[0140] In step 940 , a training set may be obtained through a sample set including sample position combinations and historical treatment plans.

[0141] In some embodiments, at least one of steps 910-930 may be performed outside the initial model. In some embodiments, at least one of steps 910-930 may be performed by the initial model. For example, the sample position combination or the sample subfield shape may be input into the initial model. The processed sample subfield shape or the processed sample position combination may be obtained by the initial model processing the sample subfield shape or the sample position combination. Alternatively, the initial model may obtain a sample set based on the processed sample subfield shape and the processed sample position combination.

[0142] It should be noted that the above description is provided for illustrative purposes only and is not intended to limit the scope of the present disclosure. For those of ordinary skill in the art, various changes and modifications can be made under the guidance of the present disclosure. However, these changes and modifications do not depart from the scope of the present disclosure.

[0143] Fig.10 is a flowchart of an exemplary process for determining a subfield shape estimation model according to some embodiments of the present disclosure. Fig.10 One or more steps of the process 1000 shown in FIG. Figure 1 or a third party (e.g., an external device). Fig.10 The process 1000 shown in FIG. 1 may be stored in a storage device (eg, storage device 150 and / or memory 220) of the medical radiation system 100 in the form of instructions and executed by a processing device 140 (eg, Figure 2 The processor 210 of the computing device 200 shown in Figure 3 ) or one or more modules (e.g., a CPU 340 of the mobile device 300 shown in FIG. Fig.11 The steps of the illustrated process presented below are intended to be illustrative. In some embodiments, process 1000 may be completed with one or more additional steps not described and / or one or more steps not discussed. In addition, Fig.10 The order of the steps of process 1000 illustrated in FIG. 1 and described below is not intended to be limiting. In some embodiments, Figure 7 Step 730 of process 700 may be performed based on process 1000 .

[0144] Step 1010, the initial model can be initialized. In some embodiments, the constructed initial model can include multiple parameters. Each of the multiple parameters can be assigned an initial value. The initialized initial model can be trained using training data to determine the training values ​​of the multiple parameters. That is, the initialized initial model is trained, and the multiple parameters of the initial values ​​are updated to the training values, thereby generating a subfield shape estimation model.

[0145] In some embodiments, the sub-field shape estimation model may be determined by updating the initialized initial model using an iterative process including multiple iterations. At least one of the multiple iterations of the iterative process may include steps 1020-1060.

[0146] In step 1020, one of the plurality of training sets may be obtained. In some embodiments, a set of features may be extracted from the training set. In some embodiments, the extracted features may be binary, numerical, categorical, ordered, binomial, interval, text-based, or any combination thereof, depending on the source and type of data in the training set.

[0147] In addition to the above features, one or more derived features may be determined from the extracted features. For example, the derived feature(s) may be represented as a linear or nonlinear combination of the extracted features.

[0148] In some embodiments, extracting features from the training set can be fully automatic, semi-automatic, manual, or a combination thereof. For example, in a fully automatic feature extraction method, one or more image processing algorithms can first detect areas corresponding to the sample in one or more first sample images, and then extract features from the detected areas. Under a semi-automatic method, some features can be automatically extracted, while other features and / or automatically extracted features can be annotated, edited, or corrected by a user. Under a manual method, features can be annotated or measured by a user.

[0149] In some embodiments, the initial model can learn features as part of the training. For example, the initial model can be a deep learning model. Deep learning (e.g., deep structured learning, hierarchical learning, or deep machine learning) can model high-level abstractions in the data by using multiple processing layers with a structure consisting of multiple nonlinear transformations, where the input data features are not explicitly designed. A deep neural network can process inputs through multiple layers of feature extraction to produce features for output. Machine training can be unsupervised when learning the features to use and how to classify a given training sample (e.g., a feature vector). With deep learning, the input features required for machine learning are not explicitly designed by the user.

[0150] Step 1030, by inputting the first sample of the training set into the intermediate model, an estimated sub-field shape corresponding to a plurality of candidate position combinations is generated. In some embodiments, one or more second sample images, one or more third sample images and / or sample radiation information in the training set may also be input into the intermediate model. In some embodiments, the intermediate model may be an initial model initialized in the first iteration of multiple iterations of the iterative process, or a previously updated model generated in the previous iteration of the iterative process. In some embodiments, the data in the training set input into the initial model may be in the form of extracted features shown in step 1020.

[0151] In some embodiments, the intermediate model can select some of the multiple candidate position combinations based on the input, and estimate the corresponding sub-field shapes with true shapes from the distance universal set. The sub-field shapes of the remaining multiple candidate position combinations can be determined as closed shapes. The estimated sub-field shapes output by the intermediate model may include sub-field shapes with true shapes and corresponding to the selected candidate position combinations, and closed sub-field shapes corresponding to the rest of the multiple candidate position combinations.

[0152] Step 1040, determining the value of the loss function based on the estimated sub-field shape and the sample set in the training set.

[0153] In some embodiments, the value of the loss function (also referred to as the loss) can be determined by comparing the processed sample sub-field shapes (e.g., the sub-field shapes determined by the treatment planning system (TPS)) with the estimated sub-field shapes corresponding to the candidate position combination. In some embodiments, the value of the loss function can be determined by comparing all sub-field shapes in the sample set with all output estimated sub-field shapes.

[0154] In some embodiments, when the estimated sub-field shape is compared with the corresponding sub-field shape in the sample set, each leaf position (c P , w P ) can be compared with the leaf position (c T , w T ) for comparison. In some embodiments, (c P , w P ) and (c T , w T ) can be based on (c P , w P ) and (c T , w T ) defined by non-intersecting (i.e., non-overlapping) open regions. For example, if (c T , w T ) indicates a closed leaf pair, while (c P , w P ) represents an open pair of blades, which may introduce greater losses. For another example, if (c P , w P ) and (c T , w T ) both represent a closed pair of leaves, but they are closed in different positions, i.e., w P =w T =0, then it can be considered that (c P , w P ) and (c T , w T ) are equivalent and do not introduce losses.

[0155] In some embodiments, a loss function suitable for the sparsity of the sample set may be selected to determine the loss between the estimated subfield shape and the sample set.

[0156] In some embodiments, the estimated radiation information may be determined based on the estimated sub-field shape. In some embodiments, the estimated radiation information may correspond to each of a plurality of sample control points. In some embodiments, the estimated radiation information may include at least one of an estimated output dose, an estimated dose output rate, an estimated dose distribution in a sample, an estimated dose output for each pulse, and the like, or any combination thereof. In some embodiments, the estimated radiation information may be determined based on the estimated sub-field shape using a treatment planning algorithm or model. In some embodiments, the estimated radiation information may be compared with sample radiation information in a training set (e.g., radiation information determined by a treatment planning system (TPS)). A loss may be determined based on the comparison. The loss may be back-propagated back to the training process to update the intermediate model.

[0157] In step 1050, it may be determined whether a termination condition is satisfied. In some embodiments, the termination condition may be related to at least one of a value of a loss function or an iteration count of an executed iterative process.

[0158] In some embodiments, an exemplary termination condition may be that the value of the loss function is less than a threshold value. Other exemplary termination conditions may include that the iteration has been executed a maximum number of times (or counts), and / or that the difference between the value of the loss function obtained in the previous iteration and the current iteration (or between the values ​​of the loss function within a certain number or count of consecutive iterations) is less than a certain threshold value.

[0159] In response to not meeting the termination condition in step 1050, process 1000 may proceed to step 1060, and further repeat steps 1020-1050 until the termination condition is met. In response to step 1050 determining that the termination condition is met, the iterative process may terminate. Process 1000 may proceed to step 1070, where the intermediate model in the current iteration may be stored and / or output as a subfield shape estimation model.

[0160] In step 1060 , an updated model may be generated by updating the intermediate model based on the value of the loss function.

[0161] During the training process, the training data can be input into the intermediate model so that the intermediate model can learn how to determine, based on new input data (e.g., the geometry of the lesion and / or surrounding tissue, and / or information), a target position combination from a general set of positions and a corresponding target sub-field shape from a general set of distances.

[0162] It should be noted that the above description is provided for illustrative purposes only and is not intended to limit the scope of the present disclosure. For those of ordinary skill in the art, various changes and modifications can be made under the guidance of the present disclosure. However, these changes and modifications do not depart from the scope of the present disclosure.

[0163] Fig.11 140 is a block diagram of a processing device according to some embodiments of the present disclosure. The processing device 140 may include an input acquisition module 1110 , a model acquisition module 1120 , and a shape estimation module 1130 .

[0164] The input acquisition module 1110 may obtain input data related to a target treatment plan for performing radiation therapy on a target using a radiation device. In some embodiments, the target treatment plan may include a plurality of target control points.

[0165] In some embodiments, the input data may include one or more first target images of the target. In some embodiments, the input data may also include one or more second target images of normal tissues around the target, one or more third target images of the target, and / or target radiation information of the target treatment plan. In some embodiments, the first target image may be acquired by scanning the target using an imaging device (e.g., the imaging device 160 of the medical radiation system 100). For example, the first target image may include a single-modality image and / or a multi-modality image. The single-modality image may include, for example, a magnetic resonance (MR) image, a CT image, a PET image, a SPECT image, an ultrasound image, an X-ray image, etc. The multi-modality image may include, for example, a computed tomography-magnetic resonance imaging (MRI-CT) system, a PET-MR image, a SPECT-MR image, a DSA-MR image, a PET-CT image, a SPECT-CT image, etc. In some embodiments, the first target image may be a reconstructed image, for example, a three-dimensional (3D) rendered image, a multi-planar reconstruction (MPR) image, a curved surface reconstruction (CPR) image, etc.

[0166] In some embodiments, the second target image may be a mask image of surrounding normal tissue separated from at least one of the one or more first target images. In some embodiments, the third target image may be a mask image of the target separated from at least one of the one or more first target images. In some embodiments, the target radiation information may include output dose, dose output rate, radiation output per pulse, dose distribution in the target, acceptable dose to normal tissue surrounding the target, minimum dose to the target, etc., or any combination thereof. In some embodiments, the target radiation information may be predicted based on one or more first target images, one or more target sample images, and / or one or more third target images. In some embodiments, a dose prediction model may be used to predict the target radiation information.

[0167] In some embodiments, the type of the target treatment plan may correspond to the sub-field shape estimation model. For example, the number of leaf pairs (or counts) of the MLC corresponding to the target treatment plan may be equal to the number of leaf pairs (or counts) corresponding to the sub-field shape estimation model. For another example, if the general set of positions used to determine the sub-field shape estimation model includes multiple gantry-collimator angle combinations, the target position combination in the target treatment plan may be a gantry-collimator angle combination. For another example, the range of the target position combination in the target treatment plan may be within the range of the candidate position combination in the general set of positions used to determine the sub-field shape estimation model.

[0168] The model acquisition module 1120 may acquire a subfield shape estimation model. In some embodiments, the subfield shape estimation model may be a machine learning model. In some embodiments, the subfield shape estimation model may be a deep learning model. In some embodiments, the subfield shape estimation model may be a deep convolution-deconvolution (e.g., encoder-decoder) network, such as a U-shaped convolutional neural network (U-Net), a V-shaped convolutional neural network (V-Net), a residual network (Res-Net), or a residual dense network (Red-Net).

[0169] In some embodiments, the model acquisition module 1120 may be configured to determine the sub-field shape estimation model by iteratively training an initial model based on training data.

[0170] The shape estimation module 1130 can estimate multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device based on the sub-field shape estimation model and the input data. Each of the multiple target segment shapes can correspond to one of the multiple target position combinations.

[0171] In some embodiments, the sub-field shape estimation model can select some candidate position combinations in the position general set as multiple target position combinations based on input data, and estimate sub-field shapes of the target shape that are each similar to one of the target position combinations from the distance general set. In some embodiments, the sub-field shape estimation model can output multiple candidate sub-field shapes, wherein the sub-field shape corresponding to the target position combination has a true shape, and the remaining sub-field shapes are closed shapes. The output candidate sub-field shape with a true shape can be designated as a target sub-field shape. The candidate position combination corresponding to the sub-field shape can be designated as a target position combination. In some embodiments, the sub-field shape estimation model can directly output multiple target position combinations and corresponding target sub-field shapes.

[0172] It should be noted that the above description is provided for illustrative purposes only and is not intended to limit the scope of the present disclosure. A person skilled in the art may make various changes and modifications under the guidance of the present disclosure. However, these changes and modifications do not depart from the scope of the present disclosure. For example, the processing device 140 may also include a storage module ( Figure 4 140). The storage module may be configured to store data generated during any process performed by any component in the processing device 140. As another example, each component of the processing device 140 may include a storage device. Additionally or alternatively, the components of the processing device 140 may share a common storage device.

[0173] Fig.12 FIG. 1 is a flowchart of an exemplary process for estimating the sub-field shape of an MLC based on a sub-field shape estimation model according to some embodiments of the present disclosure. In some embodiments, Fig.12 One or more steps of the process 1200 shown in FIG. Figure 1 The medical radiation system 100 shown in FIG. Fig.10 The process 1000 shown in FIG. 1 may be stored in a storage device (eg, storage device 150 and / or memory 220) of the medical radiation system 100 in the form of instructions and executed by a processing device 140 (eg, Figure 2 The processor 210 of the computing device 200 shown in Figure 3 ) or one or more modules (e.g., a CPU 340 of the mobile device 300 shown in FIG. Fig.11 The steps of the illustrated process presented below are intended to be illustrative. In some embodiments, process 1200 may be completed with one or more additional steps not described and / or one or more steps not discussed. In addition, Fig.12 The order of the steps of process 1200 illustrated in and described below is not intended to be limiting.

[0174] In step 1210, the processing device 140 (eg, the input acquisition module 1110) may acquire input data related to a target treatment plan for performing radiation therapy on a lesion using a radiation device. In some embodiments, the target treatment plan may include a plurality of target control points.

[0175] In some embodiments, the input data may include one or more first target images of the target. In some embodiments, the input data may also include one or more second target images of normal tissues around the target, one or more third target images of the target, and / or target radiation information of the target treatment plan. In some embodiments, the first target image may be acquired by scanning the target using an imaging device (e.g., the imaging device 160 of the medical radiation system 100). For example, the first target image may include a single-modality image and / or a multi-modality image. The single-modality image may include, for example, a magnetic resonance (MR) image, a CT image, a PET image, a SPECT image, an ultrasound image, an X-ray image, etc. The multi-modality image may include, for example, a computed tomography-magnetic resonance imaging (MRI-CT) system, a PET-MR image, a SPECT-MR image, a DSA-MR image, a PET-CT image, a SPECT-CT image, etc. In some embodiments, the first target image may be a reconstructed image, for example, a three-dimensional (3D) rendered image, a multi-planar reconstruction (MPR) image, a curved surface reconstruction (CPR) image, etc.

[0176] In some embodiments, the second target image may be a mask image of surrounding normal tissue separated from at least one of the one or more first target images. In some embodiments, the third target image may be a mask image of the target separated from at least one of the one or more first target images. In some embodiments, the target radiation information may include output dose, dose output rate, radiation output per pulse, dose distribution in the target, acceptable dose to normal tissue surrounding the target, minimum dose to the target, etc., or any combination thereof. In some embodiments, the target radiation information may be predicted based on one or more first target images, one or more target sample images, and / or one or more third target images. In some embodiments, a dose prediction model may be used to predict the target radiation information.

[0177] In some embodiments, the type of the target treatment plan may correspond to the sub-field shape estimation model. For example, the number (or count) of leaf pairs of the MLC corresponding to the target treatment plan may be equal to the number (or count) of leaf pairs corresponding to the sub-field shape estimation model. For another example, if the general set of positions used to determine the sub-field shape estimation model includes multiple gantry-collimator angle combinations, the target position combination in the target treatment plan may be a gantry-collimator angle combination. For another example, the range of the target position combination in the target treatment plan may be within the range of the candidate position combination in the general set of positions used to determine the sub-field shape estimation model.

[0178] In step 1220, the processing device 140 (e.g., the model acquisition module 1120) may acquire a subfield shape estimation model. In some embodiments, the subfield shape estimation model may be a machine learning model. In some embodiments, the subfield shape estimation model may be a deep learning model. In some embodiments, the subfield shape estimation model may be a deep convolution-deconvolution (e.g., encoder-decoder) network, such as a U-shaped convolutional neural network (U-Net), a V-shaped convolutional neural network (V-Net), a residual network (Res-Net), or a residual dense network (Red-Net).

[0179] In some embodiments, a sub-field shape estimation model can be obtained based on processes 700-1000. In some embodiments, the sub-field shape estimation model can be considered to have learned how to determine a target position combination from a general set of positions and determine a corresponding target sub-field shape from a general set of distances based on new input data (e.g., the geometry and / or information of the lesion and / or surrounding tissue).

[0180] In step 1230, the processing device 140 (e.g., the shape estimation module 1130) may estimate multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device based on the sub-field shape estimation model and the input data. Each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations. In some embodiments, the multiple target position combinations may be within a position universal set of the sub-field shape estimation model. The universal set may include multiple discrete candidate position combinations. The multiple target sub-field shapes may be within a distance universal set of the sub-field shape estimation model. The distance universal set may include multiple discrete candidate leaf positions.

[0181] In some embodiments, the sub-field shape estimation model can select some candidate position combinations in the position general set as multiple target position combinations based on input data, and estimate sub-field shapes of the target shape that are each similar to one of the target position combinations from the distance general set. In some embodiments, the sub-field shape estimation model can output multiple candidate sub-field shapes, wherein the sub-field shape corresponding to the target position combination has a true shape, and the remaining sub-field shapes are closed shapes. The output candidate sub-field shape with a true shape can be designated as a target sub-field shape. The candidate position combination corresponding to the sub-field shape can be designated as a target position combination. In some embodiments, the sub-field shape estimation model can directly output multiple target position combinations and corresponding target sub-field shapes.

[0182] In some embodiments, radiation information may not be used for training or prediction of sub-field shape. In some embodiments, the sub-field shape estimation model may predict the sub-field shape without dose distribution estimation. In some embodiments, the sub-field shape may be inferred from the relationship between the geometry of the sub-field structure in one or more first target images and the sub-field shape. The addition of information such as radiation information will further provide the sub-field shape estimation model with the possibility of inferring dosimetric relationships without the need for a dose calculation algorithm to estimate, nor the need for additional elements (such as neural networks) to make such inferences more explicitly.

[0183] It should be noted that the above description is provided for illustrative purposes only and is not intended to limit the scope of the present disclosure. For those of ordinary skill in the art, various changes and modifications can be made under the guidance of the present disclosure. However, these changes and modifications do not depart from the scope of the specification.

[0184] The basic concepts have been described above. Obviously, for those of ordinary skill in the art who have read this application, the above invention disclosure is only for example and does not constitute a limitation of this application. Although not explicitly stated here, those of ordinary skill in the art may make various modifications, improvements and amendments to this application. Such modifications, improvements and amendments are suggested in this application, so such modifications, improvements and amendments still belong to the spirit and scope of the exemplary embodiments of this application.

[0185] At the same time, the present application uses specific words to describe the embodiments of the present application. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, some features, structures or characteristics in one or more embodiments of the present application can be appropriately combined.

[0186] In addition, it will be understood by those skilled in the art that various aspects of the present application may be illustrated and described by a number of patentable categories or situations, including any new and useful process, machine, product, or combination of substances, or any new and useful improvement thereof. Therefore, various aspects of the present application may be implemented entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or by combining software and hardware implementations, which are generally collectively referred to herein as "units", "modules" or "systems". In addition, various aspects of the present application may take the form of a computer program product embodied in one or more computer-readable media, which has computer-readable program code embodied thereon.

[0187] A computer readable signal medium may include a propagated data signal containing computer program code, for example, in baseband or as part of a carrier wave. Such propagated signals may be in a variety of forms, including electromagnetic, optical, etc., or any suitable combination. A computer readable signal medium may be any computer readable medium other than a computer readable storage medium, which may be connected to an instruction execution system, device or apparatus to communicate, propagate or transmit a program for use. The program code on a computer readable signal medium may be propagated via any suitable medium, including radio, cable, fiber optic cable, RF, etc., or any combination of the above.

[0188] The computer program code for performing the operations of the present disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages, such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python and other conventional program programming languages, such as "C" programming language, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, dynamic programming languages, such as Python, Ruby and Groovy, or other programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer, partially on a remote computer or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., using an Internet service provider via the Internet) or in a cloud computing environment or as a service, such as software as a service (SaaS).

[0189] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences described in this application, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this application. Although the above disclosure discusses some invention embodiments that are currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and scope of the embodiments of this application. For example, although the implementation of the various components described above can be embodied in a hardware device, it can also be implemented as a pure software solution, for example, installation on an existing server or mobile device.

[0190] Similarly, it should be noted that in order to simplify the description disclosed in this application and thus facilitate the understanding of one or more embodiments of the invention, in the foregoing description of the embodiments of the present application, multiple features are sometimes grouped into one embodiment, figure, or description thereof. However, this method of the present application should not be interpreted as reflecting the intention that the claimed object material to be scanned requires more features than those expressly stated in each claim. On the contrary, the subject matter of the invention should have fewer features than the single embodiment described above.

[0191] In some embodiments, the numbers representing quantities or properties used to describe and claim certain embodiments of the present application should be understood to be modified by the terms "approximately", "approximately" or "substantially" in some cases. For example, unless otherwise stated, "approximately", "approximately" or "substantially" can indicate a ±20% variation of the value described. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which can change according to the required characteristics of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of the present application are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible range.

[0192] All patents, patent applications, patent application publications, and other materials (such as papers, books, specifications, publications, records, things, and / or the like) referred to herein are hereby incorporated by reference in their entirety for all purposes, except any prosecution record related to such documents, any such documents that are inconsistent or conflicting with this document, or any such documents that have the effect of limiting the broad scope of the claims that may or may not be related to this document. For example, if there is any inconsistency or conflict between the description, definition, and / or use of terms associated with any incorporated material and the terminology associated with this document, the description, definition, and / or terminology used in this document shall control.

[0193] Finally, it should be understood that the embodiments described in this application are only used to illustrate the principles of the embodiments of the present application. Other variations may also fall within the scope of the present application. Therefore, as an example and not a limitation, the alternative configurations of the embodiments of the present application may be considered to be consistent with the teachings of the present application. Accordingly, the embodiments of the present application are not limited to the embodiments explicitly introduced and described in the present application.

[0194] Additional Invention Statement

[0195] Invention Statement 1

[0196] A system comprising:

[0197] at least one storage device including a set of instructions; and

[0198] At least one processor is configured to communicate with at least one storage device, wherein when executing the instruction, the at least one processor is configured to instruct the system to perform operations including:

[0199] Acquiring input data related to a target treatment plan for performing radiotherapy on a lesion using a radiation device, the input data including a first target image of the lesion;

[0200] Obtaining a subfield shape estimation model; and

[0201] Based on the sub-field shape estimation model and input data, multiple target position combinations of a target treatment plan and multiple target sub-field shapes of a collimator of a radiation device are estimated, one of the multiple target position combinations indicates the position of the collimator relative to a lesion, and each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations.

[0202] Statement of Invention 2

[0203] The system according to Invention 1, wherein the target treatment plan includes a plurality of control points, and each of a plurality of target position combinations or a plurality of target sub-field shapes corresponds to one of the plurality of control points.

[0204] Statement of Invention 3

[0205] A system according to Invention 1 or 2, wherein the input data includes a second target image of normal tissue surrounding the lesion, a third target image of the lesion, or at least one of target radiation information of a target treatment plan, and the target radiation information includes at least one of an output dose, a dose output rate, a dose per pulse, or a dose distribution in the lesion.

[0206] Statement of Invention 4

[0207] The system according to Invention 3, wherein target radiation information is predicted based on a first target image of the lesion, a second target image of normal tissue surrounding the lesion, and a third target image of the lesion.

[0208] Statement of Invention 5

[0209] A system according to any one of Inventions 1-4, wherein one of the plurality of target position combinations comprises a combination of one or more positions at which one or more components of the radiation device operate.

[0210] Statement of Invention 6

[0211] The system according to Invention 5, wherein one of the plurality of target position combinations includes a rack angle of a rack of the radiation device.

[0212] Statement of Invention 7

[0213] The system according to Invention 5, wherein one of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device and a collimator angle of a collimator of the radiation device.

[0214] Statement of Invention 8

[0215] The system according to Invention 5, wherein one of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device, a collimator angle of a collimator of the radiation device, and a diagnosis table position of the radiation device.

[0216] Statement of Invention 9

[0217] A system according to any one of Inventions 1-8, wherein:

[0218] The collimator includes a plurality of pairs of blades; and

[0219] One of the plurality of target sub-field shapes of the collimator includes leaf positions of each of the plurality of pairs of leaves.

[0220] Statement of Invention 10

[0221] A system according to invention book 9, wherein the blade position of each pair of blades in the plurality of pairs of blades includes a center position of an opening of a pair of blades and an opening width of a pair of blades.

[0222] Statement of Invention 11

[0223] A system according to any one of Inventions 1-10, wherein a plurality of target position combinations are within a plurality of discrete candidate position combinations in a common set of positions.

[0224] Statement of Invention 12

[0225] According to the system of invention 11, the sub-field shape estimation model is obtained by executing a training process, comprising:

[0226] Obtaining a general set of locations including a plurality of candidate location combinations; and

[0227] The initial model is iteratively trained based on the general set of positions to determine the subfield shape estimation model.

[0228] Statement of Invention 13

[0229] According to the system of Invention 12, obtaining a general set of locations including a plurality of candidate location combinations comprises:

[0230] Acquire a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions; and

[0231] A general set of positions is obtained based on a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions.

[0232] Statement of Invention 14

[0233] A system according to invention 12 or 13, wherein the plurality of target sub-field shapes are within a common set of distances comprising a plurality of discrete candidate leaf positions.

[0234] Statement of Invention 15

[0235] According to the system of invention 14, the training process comprises:

[0236] obtaining a distance general set including a plurality of candidate leaf positions; and

[0237] An initial model is iteratively trained based on the distance universal set, and a sub-field shape estimation model is determined, so that the candidate sub-field shape corresponding to each of the multiple candidate position combinations output by the sub-field shape estimation model is in the distance universal set.

[0238] Statement of Invention 16

[0239] A system according to claim 15, wherein the plurality of candidate blade positions comprises a plurality of candidate opening positions and a plurality of candidate opening widths.

[0240] Statement of Invention 17

[0241] The system according to claim 15 or 16, wherein the training process comprises:

[0242] Obtain training data including multiple training sets.

[0243] Statement of Invention 18

[0244] According to the system of claim 17, acquiring training data comprises:

[0245] For one of the multiple training sets,

[0246] Obtain historical treatment plans previously generated based on lesion samples;

[0247] acquiring a first sample image of a lesion sample corresponding to a historical treatment plan;

[0248] Obtaining sample position combinations and corresponding sample sub-field shapes in historical treatment plans; and

[0249] A training set is obtained based on the first sample image, a sample position combination in a historical treatment plan, and a sample sub-field shape.

[0250] Statement of Invention 19

[0251] According to the system of Invention 18, the training set is obtained based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape, comprising:

[0252] Obtaining a processed sample position combination within the position universal set, wherein the processed sample position combination is obtained by processing the sample position combination based on the position universal set;

[0253] Obtaining a processed sample subfield shape located within the distance universal set, wherein the processed sample subfield shape is obtained by processing the sample subfield shape based on the distance universal set;

[0254] Acquire a sample set including a processed sample sub-field shape and a closed sub-field shape, wherein the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination; and

[0255] A training set is obtained by a sample set including a first sample image, a processed sample position combination, and a historical treatment plan.

[0256] Statement of Invention 20

[0257] The system according to claim 19, wherein the training process comprises:

[0258] Initializing the initial model; and

[0259] The subfield shape estimation model is obtained by updating the initialized initial model using an iterative process including multiple iterations, wherein at least one iteration of the multiple iterations of the iterative process includes:

[0260] Get one of multiple training sets;

[0261] By inputting a first sample image of a training set into an intermediate model, an estimated subfield shape corresponding to a plurality of candidate position combinations is generated, wherein the intermediate model is an initial model initialized in a first iteration of multiple iterations of an iterative process or a previously updated model generated in a previous iteration of the iterative process;

[0262] Determine the value of the loss function based on the estimated subfield shape and the sample set in the training set;

[0263] Determine whether termination conditions are met;

[0264] In response to the termination condition not being satisfied,

[0265] Update the intermediate model based on the value of the loss function to generate an updated model; and

[0266] Launch the next iteration; and

[0267] An intermediate model in a last iteration of the multiple iterations of the iterative process is designated as a subfield shape estimation model.

[0268] Statement of Invention 21

[0269] A system according to claim 20, wherein at least one of the plurality of iterations of the iterative process comprises:

[0270] In response to determining that the termination condition is satisfied, the iterative process is terminated.

[0271] Statement of Invention 22

[0272] A system according to invention 20 or 21, wherein the value of the loss function is determined based on the sparsity of the sample set, and the sparsity of the sample set is related to the closed sub-field shape of the sample set.

[0273] Statement of Invention 23

[0274] A system according to any one of Inventions 20-22, wherein the termination condition is related to at least one of the value of the loss function or the number of iterations of the iterative process that have been executed.

[0275] Statement of Invention 24

[0276] A system according to any one of Inventions 20-23, wherein the training set includes a second sample image of normal tissue surrounding the lesion sample, a third sample image of the lesion sample, or sample radiation information of a historical treatment plan, and the sample radiation information includes at least one of a sample output dose, a sample dose output rate, a sample dose per pulse, or a sample dose distribution in the lesion sample.

[0277] Statement of Invention 25

[0278] A system according to Invention 24, wherein the sample radiation information is predicted based on a first sample image of the lesion sample, a second sample image of normal tissue surrounding the lesion sample, and a third sample image of the lesion sample.

[0279] Statement of Invention 26

[0280] A system according to claim 24 or 25, wherein at least one of the multiple iterations of the iterative process comprises:

[0281] An estimated sub-field shape is generated by inputting at least one of the second sample image, the third sample image, or sample radiation information of a training set into the intermediate model.

[0282] Statement of Invention 27

[0283] A system according to claim 24 or 25, wherein at least one of the multiple iterations of the iterative process comprises:

[0284] Determining estimated radiation information based on the estimated sub-field shape;

[0285] comparing the estimated radiation information to the sample radiation information; and

[0286] The intermediate model is updated based on the comparison to generate an updated model.

[0287] Statement of Invention 28

[0288] A system comprising:

[0289] at least one storage device including a set of instructions; and

[0290] At least one processor is configured to communicate with at least one storage device, wherein when executing the instruction, the at least one processor is configured to instruct the system to perform operations including:

[0291] Get the initial model;

[0292] Obtaining training data; and

[0293] An initial model is trained based on training data to obtain a sub-field shape estimation model, wherein the sub-field shape estimation model is configured to estimate multiple target position combinations in a target treatment plan based on input data, wherein the target position combinations are used to perform radiotherapy on the lesion using multiple target sub-field shapes of a radiation device and a collimator of the radiation device, wherein the input data includes a first target image of the lesion, each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, and one of the multiple target position combinations indicates a position of the collimator relative to the lesion.

[0294] Statement of Invention 29

[0295] A system according to Invention 28, wherein the target treatment plan includes a plurality of control points, and each of a plurality of target position combinations or a plurality of target sub-field shapes corresponds to one of the plurality of control points.

[0296] Statement of Invention 30

[0297] A system according to Invention 28 or 29, wherein the input data includes a second target image of normal tissue surrounding the lesion, a third target image of the lesion, or at least one of target radiation information of a target treatment plan, the target radiation information including at least one of an output dose, a dose output rate, a dose per pulse, or a dose distribution in the lesion.

[0298] Statement of Invention 31

[0299] A system according to Invention 30, wherein target radiation information is predicted based on a first target image of the lesion, a second target image of normal tissue surrounding the lesion, and a third target image of the lesion.

[0300] Statement of Invention 32

[0301] A system according to any one of Inventions 28-31, wherein one of the plurality of target position combinations comprises a combination of one or more positions at which one or more components of the radiation device operate.

[0302] Statement of Invention 33

[0303] A system according to claim 32, wherein one of the plurality of target position combinations includes a rack angle of a rack of the radiation device.

[0304] Statement of Invention 34

[0305] A system according to claim 32, wherein one of the plurality of target position combinations comprises a combination of a gantry angle of a gantry of the radiation device and a collimator angle of a collimator of the radiation device.

[0306] Statement of Invention 35

[0307] A system according to Invention 32, wherein one of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device, a collimator angle of a collimator of the radiation device, and a diagnosis table position of the radiation device.

[0308] Statement of Invention 36

[0309] A system according to any one of Inventions 28-35, wherein:

[0310] The collimator includes a plurality of pairs of blades; and

[0311] One of the plurality of target sub-field shapes of the collimator includes leaf positions of each of the plurality of pairs of leaves.

[0312] Statement of Invention 37

[0313] A system according to Invention 36, wherein the blade position of each pair of blades in the plurality of pairs of blades comprises a center position of an opening of the pair of blades and an opening width of the pair of blades.

[0314] Statement of Invention 38

[0315] A system according to any one of Inventions 28-37, wherein a plurality of target location combinations are within a plurality of discrete candidate location combinations in a common set of locations.

[0316] Statement of Invention 39

[0317] According to the system of invention 38, obtaining the initial model comprises:

[0318] Obtaining a general set of locations including a plurality of candidate location combinations; and

[0319] Get an initial model based on a common set of locations.

[0320] Statement of Invention 40

[0321] According to the system of invention 39, obtaining a general set of locations including a plurality of candidate location combinations comprises:

[0322] Acquire a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions; and

[0323] A general set of positions is obtained based on a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions.

[0324] Statement of Invention 41

[0325] A system according to invention 39 or 40, wherein the plurality of target sub-field shapes are within a common set of distances comprising a plurality of discrete candidate leaf positions.

[0326] Statement of Invention 42

[0327] According to the system of invention 41, obtaining the initial model comprises:

[0328] obtaining a distance general set including a plurality of candidate leaf positions; and

[0329] An initial model is acquired based on the distance general set, so that the candidate sub-field shape corresponding to each of the plurality of candidate position combinations output by the sub-field shape estimation model is in the distance general set.

[0330] Statement of Invention 43

[0331] A system according to claim 42, wherein the plurality of candidate blade positions comprises a plurality of candidate opening positions and a plurality of candidate opening widths.

[0332] Statement of Invention 44

[0333] The system according to claim 42 or 43, wherein acquiring training data comprises:

[0334] For one of the multiple training sets,

[0335] Obtain historical treatment plans previously generated based on lesion samples;

[0336] acquiring a first sample image of a lesion sample corresponding to a historical treatment plan;

[0337] Obtaining sample position combinations and corresponding sample sub-field shapes in historical treatment plans; and

[0338] A training set is obtained based on the first sample image, a sample position combination in a historical treatment plan, and a sample sub-field shape.

[0339] Statement of Invention 45

[0340] According to the system of invention 44, the training set is obtained based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape, comprising:

[0341] Obtaining a processed sample position combination within the position universal set, wherein the processed sample position combination is obtained by processing the sample position combination based on the position universal set;

[0342] Obtaining a processed sample subfield shape located within the distance universal set, wherein the processed sample subfield shape is obtained by processing the sample subfield shape based on the distance universal set;

[0343] Acquire a sample set including a processed sample sub-field shape and a closed sub-field shape, wherein the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination; and

[0344] A training set is obtained by a sample set including a first sample image, a processed sample position combination, and a historical treatment plan.

[0345] Statement of Invention 46

[0346] According to the system of invention 45, the initial model is trained based on the training data to obtain the sub-field shape estimation model, comprising:

[0347] Initializing the initial model; and

[0348] The subfield shape estimation model is obtained by updating the initialized initial model using an iterative process including multiple iterations, wherein at least one iteration of the multiple iterations of the iterative process includes:

[0349] Get one of multiple training sets;

[0350] By inputting a first sample image of a training set into an intermediate model, an estimated subfield shape corresponding to a plurality of candidate position combinations is generated, wherein the intermediate model is an initial model initialized in a first iteration of multiple iterations of an iterative process or a previously updated model generated in a previous iteration of the iterative process;

[0351] Determine the value of the loss function based on the estimated subfield shape and the sample set in the training set;

[0352] Determine whether termination conditions are met;

[0353] In response to the termination condition not being satisfied,

[0354] Update the intermediate model based on the value of the loss function to generate an updated model; and

[0355] Launch the next iteration; and

[0356] An intermediate model in a last iteration of the multiple iterations of the iterative process is designated as a subfield shape estimation model.

[0357] Statement of Invention 47

[0358] A system according to claim 46, wherein at least one of the plurality of iterations of the iterative process comprises:

[0359] In response to determining that the termination condition is satisfied, the iterative process is terminated.

[0360] Statement of Invention 48

[0361] A system according to invention 46 or 47, wherein the value of the loss function is determined based on the sparsity of the sample set, and the sparsity of the sample set is related to the closed sub-field shape of the sample set.

[0362] Statement of Invention 49

[0363] A system according to any one of Inventions 46-48, wherein the termination condition is related to at least one of the value of the loss function or the number of iterations of the iterative process that have been performed.

[0364] Statement of Invention 50

[0365] A system according to any one of Inventions 46-49, wherein the training set includes a second sample image of normal tissue surrounding the lesion sample, a third sample image of the lesion sample, or sample radiation information of a historical treatment plan, the sample radiation information including at least one of a sample output dose, a sample dose output rate, a sample dose per pulse, or a sample dose distribution in the lesion sample.

[0366] Statement of Invention 51

[0367] A system according to Invention 50, wherein the sample radiation information is predicted based on a first sample image of the lesion sample, a second sample image of normal tissue surrounding the lesion sample, and a third sample image of the lesion sample.

[0368] Statement of Invention 52

[0369] A system according to claim 50 or 51, wherein at least one of the multiple iterations of the iterative process comprises:

[0370] An estimated sub-field shape is generated by inputting at least one of the second sample image, the third sample image, or sample radiation information of a training set into the intermediate model.

[0371] Statement of Invention 53

[0372] A system according to claim 52, wherein at least one of the multiple iterations of the iterative process comprises:

[0373] Determining estimated radiation information based on the estimated sub-field shape;

[0374] comparing the estimated radiation information to the sample radiation information; and

[0375] The intermediate model is updated based on the comparison to generate an updated model.

[0376] Statement of Invention 54

[0377] A method implemented on a machine including one or more processors and one or more storage devices, comprising:

[0378] Acquiring input data related to a target treatment plan for performing radiotherapy on a lesion using a radiation device, the input data including a first target image of the lesion;

[0379] Obtaining a subfield shape estimation model; and

[0380] Based on the sub-field shape estimation model and input data, multiple target position combinations of a target treatment plan and multiple target sub-field shapes of a collimator of a radiation device are estimated, one of the multiple target position combinations indicates the position of the collimator relative to a lesion, and each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations.

[0381] Statement of Invention 55

[0382] A method according to Invention 54, wherein the target treatment plan includes a plurality of control points, and each of a plurality of target position combinations or a plurality of target sub-field shapes corresponds to one of the plurality of control points.

[0383] Statement of Invention 56

[0384] A method according to Invention 54 or 55, wherein the input data includes a second target image of normal tissue surrounding the lesion, a third target image of the lesion, or at least one of target radiation information of the target treatment plan, the target radiation information including at least one of output dose, dose output rate, dose per pulse, or dose distribution in the lesion.

[0385] Statement of Invention 57

[0386] The method according to Invention 56, wherein the target radiation information is predicted based on a first target image of the lesion, a second target image of normal tissue surrounding the lesion, and a third target image of the lesion.

[0387] Statement of Invention 58

[0388] A method according to any one of Inventions 54-57, wherein one of the plurality of target position combinations comprises a combination of one or more positions at which one or more components of the radiation device operate.

[0389] Statement of Invention 59

[0390] A method according to Invention 58, wherein one of the plurality of target position combinations includes a rack angle of a rack of the radiation device.

[0391] Statement of Invention 60

[0392] The method according to Invention 58, wherein one of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device and a collimator angle of a collimator of the radiation device.

[0393] Statement of Invention 61

[0394] The method according to Invention 58, wherein one of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device, a collimator angle of a collimator of the radiation device, and a diagnosis table position of the radiation device.

[0395] Statement of Invention 62

[0396] The method according to any one of Inventions 54-61, wherein:

[0397] The collimator includes a plurality of pairs of blades; and

[0398] One of the plurality of target sub-field shapes of the collimator includes leaf positions of each of the plurality of pairs of leaves.

[0399] Statement of Invention 63

[0400] According to the method described in Invention Book 62, the blade position of each pair of blades in the multiple pairs of blades includes the center position of the opening of a pair of blades and the opening width of the pair of blades.

[0401] Statement of Invention 64

[0402] A method according to any one of Inventions 54-63, wherein a plurality of target position combinations are within a plurality of discrete candidate position combinations in a common set of positions.

[0403] Statement of Invention 65

[0404] The method according to claim 64, wherein the subfield shape estimation model is obtained by executing a training process, comprising:

[0405] Obtaining a general set of locations including a plurality of candidate location combinations; and

[0406] The initial model is iteratively trained based on the general set of positions to determine the subfield shape estimation model.

[0407] Statement of Invention 66

[0408] The method according to Invention 65, wherein obtaining a general set of locations including a plurality of candidate location combinations comprises:

[0409] Acquire a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions; and

[0410] A general set of positions is obtained based on a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions.

[0411] Statement of Invention 67

[0412] A method according to invention 65 or 66, wherein the plurality of target sub-field shapes are within a common set of distances comprising a plurality of discrete candidate leaf positions.

[0413] Statement of Invention 68

[0414] The method according to claim 67, wherein the training process comprises:

[0415] obtaining a distance general set including a plurality of candidate leaf positions; and

[0416] An initial model is iteratively trained based on the distance universal set, and a sub-field shape estimation model is determined, so that the candidate sub-field shape corresponding to each of the multiple candidate position combinations output by the sub-field shape estimation model is in the distance universal set.

[0417] Statement of Invention 69

[0418] A method according to Invention 68, wherein the plurality of candidate blade positions comprises a plurality of candidate opening positions and a plurality of candidate opening widths.

[0419] Statement of Invention 70

[0420] The method according to claim 68 or 69, wherein the training process comprises:

[0421] Obtain training data including multiple training sets.

[0422] Statement of Invention 71

[0423] The method according to claim 70, wherein obtaining training data comprises:

[0424] For one of the multiple training sets,

[0425] Obtain historical treatment plans previously generated based on lesion samples;

[0426] acquiring a first sample image of a lesion sample corresponding to a historical treatment plan;

[0427] Obtaining sample position combinations and corresponding sample sub-field shapes in historical treatment plans; and

[0428] A training set is obtained based on the first sample image, a sample position combination in a historical treatment plan, and a sample sub-field shape.

[0429] Statement of Invention 72

[0430] The method according to Invention 71, wherein the training set is obtained based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape, comprising:

[0431] Obtaining a processed sample position combination within the position universal set, wherein the processed sample position combination is obtained by processing the sample position combination based on the position universal set;

[0432] Obtaining a processed sample subfield shape located within the distance universal set, wherein the processed sample subfield shape is obtained by processing the sample subfield shape based on the distance universal set;

[0433] Acquire a sample set including a processed sample sub-field shape and a closed sub-field shape, wherein the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination; and

[0434] A training set is obtained by a sample set including a first sample image, a processed sample position combination, and a historical treatment plan.

[0435] Statement of Invention 73

[0436] The method according to claim 72, wherein the training process comprises:

[0437] Initializing the initial model; and

[0438] The subfield shape estimation model is obtained by updating the initialized initial model using an iterative process including multiple iterations, wherein at least one iteration of the multiple iterations of the iterative process includes:

[0439] Get one of multiple training sets;

[0440] By inputting a first sample image of a training set into an intermediate model, an estimated subfield shape corresponding to a plurality of candidate position combinations is generated, wherein the intermediate model is an initial model initialized in a first iteration of multiple iterations of an iterative process or a previously updated model generated in a previous iteration of the iterative process;

[0441] Determine the value of the loss function based on the estimated subfield shape and the sample set in the training set;

[0442] Determine whether termination conditions are met;

[0443] In response to the termination condition not being satisfied,

[0444] Update the intermediate model based on the value of the loss function to generate an updated model; and

[0445] Launch the next iteration; and

[0446] An intermediate model in a last iteration of the multiple iterations of the iterative process is designated as a subfield shape estimation model.

[0447] Statement of Invention 74

[0448] A method according to claim 73, wherein at least one of the multiple iterations of the iterative process comprises:

[0449] In response to determining that the termination condition is satisfied, the iterative process is terminated.

[0450] Statement of Invention 75

[0451] A method according to Invention 73 or 74, wherein the value of the loss function is determined based on the sparsity of the sample set, and the sparsity of the sample set is related to the closed sub-field shape of the sample set.

[0452] Statement of Invention 76

[0453] A method according to any one of Inventions 73-75, wherein the termination condition is related to at least one of the value of the loss function or the number of iterations of the iterative process that have been performed.

[0454] Statement of Invention 77

[0455] A method according to any one of Inventions 73-76, wherein the training set includes a second sample image of normal tissue surrounding the lesion sample, a third sample image of the lesion sample, or sample radiation information of a historical treatment plan, and the sample radiation information includes at least one of a sample output dose, a sample dose output rate, a sample dose per pulse, or a sample dose distribution in the lesion sample.

[0456] Statement of Invention 78

[0457] The method according to Invention 77, wherein the sample radiation information is predicted based on a first sample image of the lesion sample, a second sample image of normal tissue surrounding the lesion sample, and a third sample image of the lesion sample.

[0458] Statement of Invention 79

[0459] The method according to claim 77 or 78, wherein at least one of the multiple iterations of the iterative process comprises:

[0460] An estimated sub-field shape is generated by inputting at least one of the second sample image, the third sample image, or sample radiation information of a training set into the intermediate model.

[0461] Statement of Invention 80

[0462] The method according to claim 77 or 78, wherein at least one of the multiple iterations of the iterative process comprises:

[0463] Determining estimated radiation information based on the estimated sub-field shape;

[0464] comparing the estimated radiation information to the sample radiation information; and

[0465] The intermediate model is updated based on the comparison to generate an updated model.

[0466] Statement of Invention 81

[0467] A method implemented on a machine including one or more processors and one or more storage devices, comprising:

[0468] Get the initial model;

[0469] Obtaining training data; and

[0470] An initial model is trained based on training data to obtain a sub-field shape estimation model, wherein the sub-field shape estimation model is configured to estimate multiple target position combinations in a target treatment plan based on input data, wherein the target position combinations are used to perform radiotherapy on the lesion using multiple target sub-field shapes of a radiation device and a collimator of the radiation device, wherein the input data includes a first target image of the lesion, each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, and one of the multiple target position combinations indicates a position of the collimator relative to the lesion.

[0471] Statement of Invention 82

[0472] A method according to Invention 81, wherein the target treatment plan includes a plurality of control points, and each of a plurality of target position combinations or a plurality of target sub-field shapes corresponds to one of the plurality of control points.

[0473] Statement of Invention 83

[0474] A method according to Invention 81 or 82, wherein the input data includes a second target image of normal tissue surrounding the lesion, a third target image of the lesion, or at least one of target radiation information of the target treatment plan, the target radiation information including at least one of output dose, dose output rate, dose per pulse, or dose distribution in the lesion.

[0475] Statement of Invention 84

[0476] A method according to Invention 83, wherein target radiation information is predicted based on a first target image of the lesion, a second target image of normal tissue surrounding the lesion, and a third target image of the lesion.

[0477] Statement of Invention 85

[0478] A method according to any one of Inventions 81-84, wherein one of the multiple target position combinations includes a combination of one or more positions at which one or more components of the radiation device operate.

[0479] Statement of Invention 86

[0480] A method according to Invention 85, wherein one of the plurality of target position combinations includes a rack angle of a rack of the radiation device.

[0481] Statement of Invention 87

[0482] The method according to Invention 85, wherein one of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device and a collimator angle of a collimator of the radiation device.

[0483] Statement of Invention 88

[0484] The method according to Invention 85, wherein one of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device, a collimator angle of a collimator of the radiation device, and a diagnosis table position of the radiation device.

[0485] Statement of Invention 89

[0486] The method according to any one of Inventions 81-88, wherein:

[0487] The collimator includes a plurality of pairs of blades; and

[0488] One of the plurality of target sub-field shapes of the collimator includes leaf positions of each of the plurality of pairs of leaves.

[0489] Statement of Invention 90

[0490] According to the method described in Invention Book 89, the blade position of each pair of blades in the multiple pairs of blades includes the center position of the opening of a pair of blades and the opening width of the pair of blades.

[0491] Statement of Invention 91

[0492] A method according to any one of Inventions 81-90, wherein a plurality of target position combinations are within a plurality of discrete candidate position combinations in a common set of positions.

[0493] Statement of Invention 92

[0494] The method according to claim 91, wherein the sub-field shape estimation model is obtained by executing a training process, comprising:

[0495] Obtaining a general set of locations including a plurality of candidate location combinations; and

[0496] The initial model is iteratively trained based on the general set of positions to determine the subfield shape estimation model.

[0497] Statement of Invention 93

[0498] The method according to Invention 92, wherein obtaining a general set of locations including a plurality of candidate location combinations comprises:

[0499] Acquire a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions; and

[0500] A general set of positions is obtained based on a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions.

[0501] Statement of Invention 94

[0502] A method according to invention 92 or 93, wherein the plurality of target sub-field shapes are within a universal set of distances comprising a plurality of discrete candidate leaf positions.

[0503] Statement of Invention 95

[0504] The method according to claim 94, wherein the training process comprises:

[0505] obtaining a distance general set including a plurality of candidate leaf positions; and

[0506] An initial model is iteratively trained based on the distance universal set, and a sub-field shape estimation model is determined, so that the candidate sub-field shape corresponding to each of the multiple candidate position combinations output by the sub-field shape estimation model is in the distance universal set.

[0507] Statement of Invention 96

[0508] A method according to Invention 95, wherein the plurality of candidate blade positions comprises a plurality of candidate opening positions and a plurality of candidate opening widths.

[0509] Statement of Invention 97

[0510] The method according to claim 95 or 96, wherein obtaining training data comprises:

[0511] For one of the multiple training sets,

[0512] Obtain historical treatment plans previously generated based on lesion samples;

[0513] acquiring a first sample image of a lesion sample corresponding to a historical treatment plan;

[0514] Obtaining sample position combinations and corresponding sample sub-field shapes in historical treatment plans; and

[0515] A training set is obtained based on the first sample image, a sample position combination in a historical treatment plan, and a sample sub-field shape.

[0516] Statement of Invention 98

[0517] The method according to Invention 97, wherein the training set is obtained based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape, comprising:

[0518] Obtaining a processed sample position combination within the position universal set, wherein the processed sample position combination is obtained by processing the sample position combination based on the position universal set;

[0519] Obtaining a processed sample subfield shape located within the distance universal set, wherein the processed sample subfield shape is obtained by processing the sample subfield shape based on the distance universal set;

[0520] Acquire a sample set including a processed sample sub-field shape and a closed sub-field shape, wherein the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination; and

[0521] A training set is obtained by a sample set including a first sample image, a processed sample position combination, and a historical treatment plan.

[0522] Statement of Invention 99

[0523] The method according to claim 98, wherein the training process comprises:

[0524] Initializing the initial model; and

[0525] The subfield shape estimation model is obtained by updating the initialized initial model using an iterative process including multiple iterations, wherein at least one iteration of the multiple iterations of the iterative process includes:

[0526] Get one of multiple training sets;

[0527] By inputting a first sample image of a training set into an intermediate model, an estimated subfield shape corresponding to a plurality of candidate position combinations is generated, wherein the intermediate model is an initial model initialized in a first iteration of multiple iterations of an iterative process or a previously updated model generated in a previous iteration of the iterative process;

[0528] Determine the value of the loss function based on the estimated subfield shape and the sample set in the training set;

[0529] Determine whether termination conditions are met;

[0530] In response to the termination condition not being satisfied,

[0531] Update the intermediate model based on the value of the loss function to generate an updated model; and

[0532] Launch the next iteration; and

[0533] An intermediate model in a last iteration of the multiple iterations of the iterative process is designated as a subfield shape estimation model.

[0534] Statement of Invention 100

[0535] A method according to claim 99, wherein at least one of the multiple iterations of the iterative process comprises:

[0536] In response to determining that the termination condition is satisfied, the iterative process is terminated.

[0537] Statement of Invention 101

[0538] A method according to Invention 99 or 100, wherein the value of the loss function is determined based on the sparsity of the sample set, and the sparsity of the sample set is related to the closed sub-field shape of the sample set.

[0539] Statement of Invention 102

[0540] A method according to any one of Inventions 99-101, wherein the termination condition is related to at least one of the value of the loss function or the number of iterations of the iterative process that have been performed.

[0541] Statement of Invention 103

[0542] A method according to any one of Inventions 99-102, wherein the training set includes a second sample image of normal tissue surrounding the lesion sample, a third sample image of the lesion sample, or sample radiation information of a historical treatment plan, and the sample radiation information includes at least one of a sample output dose, a sample dose output rate, a sample dose per pulse, or a sample dose distribution in the lesion sample.

[0543] Statement of Invention 104

[0544] The method according to Invention 103, wherein the sample radiation information is predicted based on a first sample image of the lesion sample, a second sample image of normal tissue surrounding the lesion sample, and a third sample image of the lesion sample.

[0545] Statement of Invention 105

[0546] The method according to claim 103 or 104, wherein at least one of the multiple iterations of the iterative process comprises:

[0547] An estimated sub-field shape is generated by inputting at least one of the second sample image, the third sample image, or sample radiation information of a training set into the intermediate model.

[0548] Statement of Invention 106

[0549] The method according to claim 103 or 104, wherein at least one of the multiple iterations of the iterative process comprises:

[0550] Determining estimated radiation information based on the estimated sub-field shape;

[0551] comparing the estimated radiation information to the sample radiation information; and

[0552] The intermediate model is updated based on the comparison to generate an updated model.

[0553] Statement of Invention 107

[0554] A system comprising:

[0555] An input acquisition module, used to acquire input data related to a target treatment plan for performing radiotherapy on a lesion using a radiation device, the input data including a first target image of the lesion;

[0556] A model acquisition module, used to acquire a sub-field shape estimation model; and

[0557] A shape estimation module is used to estimate multiple target position combinations of a target treatment plan and multiple target sub-field shapes of a collimator of a radiation device based on a sub-field shape estimation model and input data, wherein one of the multiple target position combinations indicates a position of the collimator relative to a lesion, and each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations.

[0558] Statement of Invention 108

[0559] A system according to Invention 107, wherein the target treatment plan includes a plurality of control points, and each of a plurality of target position combinations or a plurality of target sub-field shapes corresponds to one of the plurality of control points.

[0560] Statement of Invention 109

[0561] A system according to invention 107 or 108, wherein the input data includes a second target image of normal tissue surrounding the lesion, a third target image of the lesion, or at least one of target radiation information of the target treatment plan, the target radiation information including at least one of output dose, dose output rate, dose per pulse, or dose distribution in the lesion.

[0562] Statement of Invention 110

[0563] A system according to Invention 109, wherein target radiation information is predicted based on a first target image of the lesion, a second target image of normal tissue surrounding the lesion, and a third target image of the lesion.

[0564] Statement of Invention 111

[0565] A system according to any of inventions 107-110, wherein one of the plurality of target position combinations comprises a combination of one or more positions at which one or more components of the radiation device operate.

[0566] Statement of Invention 112

[0567] A system according to invention 111, wherein one of the plurality of target position combinations includes a rack angle of a rack of the radiation device.

[0568] Statement of Invention 113

[0569] A system according to invention 111, wherein one of the plurality of target position combinations comprises a combination of a gantry angle of a gantry of the radiation device and a collimator angle of a collimator of the radiation device.

[0570] Statement of Invention 114

[0571] A system according to Invention 111, wherein one of the plurality of target position combinations comprises a combination of a gantry angle of a gantry of the radiation device, a collimator angle of a collimator of the radiation device, and a diagnosis table position of the radiation device.

[0572] Statement of Invention 115

[0573] A system according to any one of Inventions 107-114, wherein:

[0574] The collimator includes a plurality of pairs of blades; and

[0575] One of the plurality of target sub-field shapes of the collimator includes leaf positions of each of the plurality of pairs of leaves.

[0576] Statement of Invention 116

[0577] A system according to invention book 115, wherein the blade position of each pair of blades in the plurality of pairs of blades includes a center position of an opening of a pair of blades and an opening width of a pair of blades.

[0578] Statement of Invention 117

[0579] A system according to any one of Inventions 107-116, wherein a plurality of target location combinations are within a plurality of discrete candidate location combinations of a common set of locations.

[0580] Statement of Invention 118

[0581] The system according to Invention 117, wherein the sub-field shape estimation model is obtained by executing a training process, comprising:

[0582] Obtaining a general set of locations including a plurality of candidate location combinations; and

[0583] The initial model is iteratively trained based on the general set of positions to determine the subfield shape estimation model.

[0584] Statement of Invention 119

[0585] According to the system of invention 118, obtaining a general set of locations including a plurality of candidate location combinations comprises:

[0586] Acquire a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions; and

[0587] A general set of positions is obtained based on a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions.

[0588] Statement of Invention 120

[0589] A system according to invention 118 or 119, wherein the plurality of target sub-field shapes are within a common set of distances comprising a plurality of discrete candidate leaf positions.

[0590] Statement of Invention 121

[0591] The system according to claim 120, wherein the training process comprises:

[0592] obtaining a distance general set including a plurality of candidate leaf positions; and

[0593] An initial model is iteratively trained based on the distance universal set, and a sub-field shape estimation model is determined, so that the candidate sub-field shape corresponding to each of the multiple candidate position combinations output by the sub-field shape estimation model is in the distance universal set.

[0594] Statement of Invention 122

[0595] A system according to invention 121, wherein the plurality of candidate blade positions comprises a plurality of candidate opening positions and a plurality of candidate opening widths.

[0596] Statement of Invention 123

[0597] The system according to claim 121 or 122, wherein the training process comprises:

[0598] Obtain training data including multiple training sets.

[0599] Statement of Invention 124

[0600] According to the system of invention 123, obtaining training data comprises:

[0601] For one of the multiple training sets,

[0602] Obtain historical treatment plans previously generated based on lesion samples;

[0603] acquiring a first sample image of a lesion sample corresponding to a historical treatment plan;

[0604] Obtaining sample position combinations and corresponding sample sub-field shapes in historical treatment plans; and

[0605] A training set is obtained based on the first sample image, a sample position combination in a historical treatment plan, and a sample sub-field shape.

[0606] Statement of Invention 125

[0607] According to the system of invention 124, the training set is obtained based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape, comprising:

[0608] Obtaining a processed sample position combination within the position universal set, wherein the processed sample position combination is obtained by processing the sample position combination based on the position universal set;

[0609] Obtaining a processed sample subfield shape located within the distance universal set, wherein the processed sample subfield shape is obtained by processing the sample subfield shape based on the distance universal set;

[0610] Acquire a sample set including a processed sample sub-field shape and a closed sub-field shape, wherein the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination; and

[0611] A training set is obtained by a sample set including a first sample image, a processed sample position combination, and a historical treatment plan.

[0612] Statement of Invention 126

[0613] A system according to claim 125, wherein the training process comprises:

[0614] Initializing the initial model; and

[0615] The subfield shape estimation model is obtained by updating the initialized initial model using an iterative process including multiple iterations, wherein at least one iteration of the multiple iterations of the iterative process includes:

[0616] Get one of multiple training sets;

[0617] By inputting a first sample image of a training set into an intermediate model, an estimated subfield shape corresponding to a plurality of candidate position combinations is generated, wherein the intermediate model is an initial model initialized in a first iteration of multiple iterations of an iterative process or a previously updated model generated in a previous iteration of the iterative process;

[0618] Determine the value of the loss function based on the estimated subfield shape and the sample set in the training set;

[0619] Determine whether termination conditions are met;

[0620] In response to the termination condition not being satisfied,

[0621] Update the intermediate model based on the value of the loss function to generate an updated model; and

[0622] Launch the next iteration; and

[0623] An intermediate model in a last iteration of the multiple iterations of the iterative process is designated as a subfield shape estimation model.

[0624] Statement of Invention 127

[0625] A system according to claim 126, wherein at least one of the plurality of iterations of the iterative process comprises:

[0626] In response to determining that the termination condition is satisfied, the iterative process is terminated.

[0627] Statement of Invention 128

[0628] A system according to invention 126 or 127, wherein the value of the loss function is determined based on the sparsity of the sample set, and the sparsity of the sample set is related to the closed sub-field shape of the sample set.

[0629] Statement of Invention 129

[0630] A system according to any one of Inventions 126-128, wherein the termination condition is related to at least one of the value of the loss function or the number of iterations of the iterative process that have been executed.

[0631] Statement of Invention 130

[0632] A system according to any one of Inventions 126-129, wherein the training set includes a second sample image of normal tissue surrounding the lesion sample, a third sample image of the lesion sample, or sample radiation information of a historical treatment plan, the sample radiation information including at least one of a sample output dose, a sample dose output rate, a sample dose per pulse, or a sample dose distribution in the lesion sample.

[0633] Statement of Invention 131

[0634] A system according to Invention 130, wherein sample radiation information is predicted based on a first sample image of the lesion sample, a second sample image of normal tissue surrounding the lesion sample, and a third sample image of the lesion sample.

[0635] Invention Chen according to 132

[0636] The system of invention 130 or 131, wherein at least one of the plurality of iterations of the iterative process comprises:

[0637] An estimated sub-field shape is generated by inputting at least one of the second sample image, the third sample image, or sample radiation information of a training set into the intermediate model.

[0638] Statement of Invention 133

[0639] A system according to claim 130 or 131, wherein at least one of the plurality of iterations of the iterative process comprises:

[0640] Determining estimated radiation information based on the estimated sub-field shape;

[0641] comparing the estimated radiation information to the sample radiation information; and

[0642] The intermediate model is updated based on the comparison to generate an updated model.

[0643] Statement of Invention 134

[0644] A system includes a model acquisition module configured to:

[0645] Get the initial model;

[0646] Obtaining training data; and

[0647] An initial model is trained based on training data to obtain a sub-field shape estimation model, wherein the sub-field shape estimation model is configured to estimate multiple target position combinations in a target treatment plan based on input data, wherein the target position combinations are used to perform radiotherapy on the lesion using multiple target sub-field shapes of a radiation device and a collimator of the radiation device, wherein the input data includes a first target image of the lesion, each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, and one of the multiple target position combinations indicates a position of the collimator relative to the lesion.

[0648] Statement of Invention 135

[0649] A system according to Invention 134, wherein the target treatment plan includes a plurality of control points, and each of a plurality of target position combinations or a plurality of target sub-field shapes corresponds to one of the plurality of control points.

[0650] Statement of Invention 136

[0651] A system according to invention 134 or 135, wherein the input data includes a second target image of normal tissue surrounding the lesion, a third target image of the lesion, or at least one of target radiation information of the target treatment plan, the target radiation information including at least one of output dose, dose output rate, dose per pulse, or dose distribution in the lesion.

[0652] Statement of Invention 137

[0653] A system according to Invention 136, wherein target radiation information is predicted based on a first target image of the lesion, a second target image of normal tissue surrounding the lesion, and a third target image of the lesion.

[0654] Statement of Invention 138

[0655] A system according to any of inventions 134-137, wherein one of the plurality of target position combinations comprises a combination of one or more positions at which one or more components of the radiation device operate.

[0656] Statement of Invention 139

[0657] A system according to invention 138, wherein one of the multiple target position combinations includes a rack angle of a rack of the radiation device.

[0658] Statement of Invention 140

[0659] A system according to invention 138, wherein one of the plurality of target position combinations comprises a combination of a gantry angle of a gantry of the radiation device and a collimator angle of a collimator of the radiation device.

[0660] Statement of Invention 141

[0661] A system according to Invention 138, wherein one of the plurality of target position combinations comprises a combination of a gantry angle of a gantry of the radiation device, a collimator angle of a collimator of the radiation device, and a diagnosis table position of the radiation device.

[0662] Statement of Invention 142

[0663] A system according to any one of Inventions 134-141, wherein:

[0664] The collimator includes a plurality of pairs of blades; and

[0665] One of the plurality of target sub-field shapes of the collimator includes leaf positions of each of the plurality of pairs of leaves.

[0666] Statement of Invention 143

[0667] A system according to invention 142, wherein the blade position of each pair of blades in the plurality of pairs of blades includes a center position of an opening of the pair of blades and an opening width of the pair of blades.

[0668] Statement of Invention 144

[0669] A system according to any one of Inventions 141-143, wherein a plurality of target location combinations are within a plurality of discrete candidate location combinations of a common set of locations.

[0670] Statement of Invention 145

[0671] According to the system of invention 144, the sub-field shape estimation model is obtained by executing a training process, comprising:

[0672] Obtaining a general set of locations including a plurality of candidate location combinations; and

[0673] The initial model is iteratively trained based on the general set of positions to determine the subfield shape estimation model.

[0674] Statement of Invention 146

[0675] According to the system of invention 145, obtaining a general set of locations including a plurality of candidate location combinations comprises:

[0676] Acquire a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions; and

[0677] A general set of positions is obtained based on a plurality of candidate gantry angles, a plurality of candidate collimator angles, or a plurality of candidate table positions.

[0678] Statement of Invention 147

[0679] A system according to invention 145 or 146, wherein the plurality of target sub-field shapes are within a common set of distances comprising a plurality of discrete candidate leaf positions.

[0680] Statement of Invention 148

[0681] A system according to claim 147, wherein the training process comprises:

[0682] obtaining a distance general set including a plurality of candidate leaf positions; and

[0683] An initial model is iteratively trained based on the distance universal set, and a sub-field shape estimation model is determined, so that the candidate sub-field shape corresponding to each of the multiple candidate position combinations output by the sub-field shape estimation model is in the distance universal set.

[0684] Statement of Invention 149

[0685] A system according to invention 148, wherein the plurality of candidate blade positions comprises a plurality of candidate opening positions and a plurality of candidate opening widths.

[0686] Statement of Invention 150

[0687] According to the system of inventions 148-149, obtaining training data comprises:

[0688] For one of the multiple training sets,

[0689] Obtain historical treatment plans previously generated based on lesion samples;

[0690] acquiring a first sample image of a lesion sample corresponding to a historical treatment plan;

[0691] Obtaining sample position combinations and corresponding sample sub-field shapes in historical treatment plans; and

[0692] A training set is obtained based on the first sample image, a sample position combination in a historical treatment plan, and a sample sub-field shape.

[0693] Statement of Invention 151

[0694] According to the system of invention 150, the training set is obtained based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape, comprising:

[0695] Obtaining a processed sample position combination within the position universal set, wherein the processed sample position combination is obtained by processing the sample position combination based on the position universal set;

[0696] Obtaining a processed sample subfield shape located within the distance universal set, wherein the processed sample subfield shape is obtained by processing the sample subfield shape based on the distance universal set;

[0697] Acquire a sample set including a processed sample sub-field shape and a closed sub-field shape, wherein the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination; and

[0698] A training set is obtained by a sample set including a first sample image, a processed sample position combination, and a historical treatment plan.

[0699] Statement of Invention 152

[0700] A system according to claim 151, wherein the training process comprises:

[0701] Initializing the initial model; and

[0702] The subfield shape estimation model is obtained by updating the initialized initial model using an iterative process including multiple iterations, wherein at least one iteration of the multiple iterations of the iterative process includes:

[0703] Get one of multiple training sets;

[0704] By inputting a first sample image of a training set into an intermediate model, an estimated subfield shape corresponding to a plurality of candidate position combinations is generated, wherein the intermediate model is an initial model initialized in a first iteration of multiple iterations of an iterative process or a previously updated model generated in a previous iteration of the iterative process;

[0705] Determine the value of the loss function based on the estimated subfield shape and the sample set in the training set;

[0706] Determine whether termination conditions are met;

[0707] In response to the termination condition not being satisfied,

[0708] Update the intermediate model based on the value of the loss function to generate an updated model; and

[0709] Launch the next iteration; and

[0710] An intermediate model in a last iteration of the multiple iterations of the iterative process is designated as a subfield shape estimation model.

[0711] Statement of Invention 153

[0712] A system according to claim 152, wherein at least one of the plurality of iterations of the iterative process comprises:

[0713] In response to determining that the termination condition is satisfied, the iterative process is terminated.

[0714] Statement of Invention 154

[0715] A system according to invention 152 or 153, wherein the value of the loss function is determined based on the sparsity of the sample set, and the sparsity of the sample set is related to the closed sub-field shape of the sample set.

[0716] Statement of Invention 155

[0717] A system according to any one of Inventions 152-154, wherein the termination condition is related to at least one of the value of the loss function or the number of iterations of the iterative process that have been performed.

[0718] Statement of Invention 156

[0719] A system according to any one of Inventions 152-155, wherein the training set includes a second sample image of normal tissue surrounding the lesion sample, a third sample image of the lesion sample, or sample radiation information of a historical treatment plan, and the sample radiation information includes at least one of a sample output dose, a sample dose output rate, a sample dose per pulse, or a sample dose distribution in the lesion sample.

[0720] Statement of Invention 157

[0721] A system according to Invention 156, wherein sample radiation information is predicted based on a first sample image of the lesion sample, a second sample image of normal tissue surrounding the lesion sample, and a third sample image of the lesion sample.

[0722] Statement of Invention 158

[0723] A system according to claim 156 or 157, wherein at least one of the multiple iterations of the iterative process comprises:

[0724] An estimated sub-field shape is generated by inputting at least one of the second sample image, the third sample image, or sample radiation information of a training set into the intermediate model.

[0725] Statement of Invention 159

[0726] A system according to claim 156 or 157, wherein at least one of the multiple iterations of the iterative process comprises:

[0727] Determining estimated radiation information based on the estimated sub-field shape;

[0728] comparing the estimated radiation information to the sample radiation information; and

[0729] The intermediate model is updated based on the comparison to generate an updated model.

[0730] Statement of Invention 160

[0731] A non-transitory computer readable medium comprising at least one set of instructions, wherein when executed by one or more processors of a computing device, the at least one set of instructions causes the computing device to perform a method comprising:

[0732] Acquiring input data related to a target treatment plan for performing radiotherapy on a lesion using a radiation device, the input data including a first target image of the lesion;

[0733] Obtaining a subfield shape estimation model; and

[0734] Based on the sub-field shape estimation model and input data, multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device are estimated, each of the multiple target position combinations indicates the position of the collimator relative to the lesion, and each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations.

[0735] Statement of Invention 161

[0736] A non-transitory computer-readable medium comprising at least one set of instructions, wherein when executed by one or more processors of a computing device, the at least one set of instructions causes the computing device to perform a method comprising:

[0737] Get the initial model;

[0738] Obtaining training data; and

[0739] An initial model is trained based on training data to obtain a sub-field shape estimation model, wherein the sub-field shape estimation model is configured to estimate multiple target position combinations in a target treatment plan based on input data, wherein the target position combinations are used to perform radiotherapy on the lesion using multiple target sub-field shapes of a radiation device and a collimator of the radiation device, wherein the input data includes a first target image of the lesion, each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, and one of the multiple target position combinations indicates a position of the collimator relative to the lesion.

Claims

1. A system comprising: at least one storage device including a set of instructions; as well as At least one processor is configured to communicate with the at least one storage device, wherein when executing the instruction, the at least one processor is configured to instruct the system to perform operations including: Acquiring input data related to a target treatment plan for performing radiotherapy on a lesion using a radiation device, the input data comprising a first target image of the lesion; Obtaining a subfield shape estimation model; and Based on the sub-field shape estimation model and the input data, multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device are estimated, one of the multiple target position combinations indicates the position of the collimator relative to the lesion, and each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, wherein the sub-field shape estimation model is configured to determine the target position combinations and the target sub-field shapes.

2. The system according to claim 1, characterized in that The target treatment plan includes a plurality of control points, and each of the plurality of target position combinations or the plurality of target sub-field shapes corresponds to one of the plurality of control points.

3. The system according to claim 1, characterized in that The input data includes at least one of a second target image of normal tissue surrounding the lesion, a third target image of the lesion, or target radiation information of the target treatment plan, and the target radiation information includes at least one of an output dose, a dose output rate, a dose per pulse, or a dose distribution in the lesion.

4. The system according to claim 3, characterized in that The target radiation information is predicted based on the first target image of the lesion, the second target image of normal tissue surrounding the lesion, and the third target image of the lesion.

5. The system according to any one of claims 1 to 4, characterized in that: One of the plurality of target position combinations comprises a combination of one or more positions at which one or more components of the radiation device operate.

6. The system according to claim 5, characterized in that One of the plurality of target position combinations includes a rack angle of a rack of the radiation device.

7. The system according to claim 5, characterized in that One of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device and a collimator angle of a collimator of the radiation device.

8. The system according to claim 5, characterized in that One of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device, a collimator angle of a collimator of the radiation device, and a diagnosis table position of the radiation device.

9. The system according to any one of claims 1 to 4, characterized in that: The collimator comprises a plurality of pairs of blades; and One of the plurality of target sub-field shapes of the collimator includes a leaf position of each pair of leaves in the plurality of pairs of leaves.

10. The system according to claim 9, characterized in that The blade position of each pair of blades in the plurality of pairs of blades includes a center position of an opening of the pair of blades and an opening width of the pair of blades.

11. The system according to any one of claims 1 to 4, characterized in that: The plurality of target position combinations are within a plurality of discrete candidate position combinations of the general set of positions.

12. The system according to claim 11, characterized in that The sub-field shape estimation model is obtained by executing a training process, including: Acquire the general set of locations including the plurality of candidate location combinations; and An initial model is iteratively trained based on the general set of positions to determine the sub-field shape estimation model.

13. The system according to claim 12, characterized in that Acquiring the general set of positions including the plurality of candidate position combinations, comprising: Acquire multiple candidate gantry angles, multiple candidate collimator angles, or multiple candidate table positions; and The general set of positions is obtained based on the multiple candidate gantry angles, the multiple candidate collimator angles, or the multiple candidate table positions.

14. The system according to claim 12, characterized in that The plurality of target sub-field shapes are within a universal set of distances that includes a plurality of discrete candidate leaf positions.

15. The system according to claim 14, characterized in that The training process includes: obtaining the universal set of distances including the plurality of candidate blade positions; and The initial model is iteratively trained based on the distance universal set to determine the sub-field shape estimation model, so that the candidate sub-field shape corresponding to each of the multiple candidate position combinations output by the sub-field shape estimation model is in the distance universal set.

16. The system according to claim 15, characterized in that The plurality of candidate blade positions include a plurality of candidate opening positions and a plurality of candidate opening widths.

17. The system according to claim 15, characterized in that The training process includes: Obtain training data including multiple training sets.

18. The system according to claim 17, characterized in that Acquiring the training data includes: For one of the multiple training sets, Obtain historical treatment plans previously generated based on lesion samples; Acquire a first sample image of the lesion sample corresponding to the historical treatment plan; Obtaining a sample position combination and a corresponding sample sub-field shape in the historical treatment plan; and The training set is acquired based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape.

19. The system according to claim 18, characterized in that Acquiring the training set based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape, comprises: Acquire a processed sample position combination located in the position universal set, wherein the processed sample position combination is obtained by processing the sample position combination based on the position universal set; Acquire a processed sample subfield shape located in the distance universal set, wherein the processed sample subfield shape is obtained by processing the sample subfield shape based on the distance universal set; Acquire a sample set including the processed sample sub-field shape and a closed sub-field shape, wherein the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination; and The training set is obtained through the sample set including the first sample image, the processed sample position combination and the historical treatment plan.

20. The system according to claim 19, characterized in that The training process includes: Initializing the initial model; and The subfield shape estimation model is obtained by updating the initialized initial model using an iterative process including multiple iterations, wherein at least one iteration of the multiple iterations of the iterative process includes: Obtain one of the multiple training sets; Generating estimated subfield shapes corresponding to the plurality of candidate position combinations by inputting the first sample image of the training set into an intermediate model, wherein the intermediate model is the initial model initialized in the first iteration of the plurality of iterations of the iterative process or a previously updated model generated in a previous iteration of the iterative process; Determining a value of a loss function based on the estimated subfield shape and a sample set in the training set; Determine whether termination conditions are met; In response to not satisfying the termination condition, Updating the intermediate model based on the value of the loss function to generate an updated model; and Launch the next iteration; and The intermediate model in a last iteration of the plurality of iterations of the iterative process is designated as the sub-field shape estimation model.

21. The system according to claim 20, characterized in that At least one of the plurality of iterations of the iterative process comprises: In response to determining that the termination condition is satisfied, the iterative process is terminated.

22. The system according to claim 20, characterized in that The value of the loss function is determined based on the sparsity of the sample set, and the sparsity of the sample set is related to the closed sub-field shape of the sample set.

23. The system according to claim 20, characterized in that The termination condition is related to at least one of a value of the loss function or a number of iterations of the iterative process that have been performed.

24. The system according to claim 20, characterized in that The training set includes a second sample image of normal tissue surrounding the lesion sample, a third sample image of the lesion sample, or sample radiation information of the historical treatment plan, wherein the sample radiation information includes at least one of a sample output dose, a sample dose output rate, a sample dose per pulse, or a sample dose distribution in the lesion sample.

25. The system according to claim 24, characterized in that The sample radiation information is predicted based on the first sample image of the lesion sample, the second sample image of normal tissue around the lesion sample, and the third sample image of the lesion sample.

26. The system according to claim 24, characterized in that At least one of the plurality of iterations of the iterative process comprises: The estimated sub-field shape is generated by inputting at least one of the second sample image, the third sample image, or the sample radiation information of the training set into the intermediate model.

27. The system according to claim 24, characterized in that At least one of the plurality of iterations of the iterative process comprises: Determining estimated radiation information based on the estimated sub-field shape; comparing the estimated radiation information with the sample radiation information; and The intermediate model is updated based on the comparison to generate the updated model.

28. A system comprising: at least one storage device including a set of instructions; as well as At least one processor is configured to communicate with the at least one storage device, wherein when executing the instruction, the at least one processor is configured to instruct the system to perform operations including: Get the initial model; Obtaining training data; and The initial model is trained based on the training data to obtain a sub-field shape estimation model, wherein the sub-field shape estimation model is configured to estimate a plurality of target position combinations in a target treatment plan based on input data, wherein the target position combinations are used to perform radiotherapy on a lesion using a radiation device and a plurality of target sub-field shapes of a collimator of the radiation device, wherein the input data includes a first target image of the lesion, each of the plurality of target sub-field shapes corresponds to one of the plurality of target position combinations, and one of the plurality of target position combinations indicates a position of the collimator relative to the lesion, wherein the sub-field shape estimation model is configured to determine the target position combinations and the target sub-field shapes.

29. The system according to claim 28, characterized in that The target treatment plan includes a plurality of control points, and each of the plurality of target position combinations or the plurality of target sub-field shapes corresponds to one of the plurality of control points.

30. The system according to claim 28, characterized in that The input data includes at least one of a second target image of normal tissue surrounding the lesion, a third target image of the lesion, or target radiation information of the target treatment plan, and the target radiation information includes at least one of an output dose, a dose output rate, a dose per pulse, or a dose distribution in the lesion.

31. The system according to claim 30, characterized in that The target radiation information is predicted based on the first target image of the lesion, the second target image of normal tissue surrounding the lesion, and the third target image of the lesion.

32. The system according to any one of claims 28 to 31, characterized in that One of the plurality of target position combinations comprises a combination of one or more positions at which one or more components of the radiation device operate.

33. The system according to claim 32, characterized in that One of the plurality of target position combinations includes a rack angle of a rack of the radiation device.

34. The system according to claim 32, characterized in that One of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device and a collimator angle of a collimator of the radiation device.

35. The system according to claim 32, characterized in that One of the plurality of target position combinations includes a combination of a gantry angle of a gantry of the radiation device, a collimator angle of a collimator of the radiation device, and a diagnosis table position of the radiation device.

36. The system according to any one of claims 28 to 31, characterized in that The collimator comprises a plurality of pairs of blades; and One of the plurality of target sub-field shapes of the collimator includes a leaf position of each pair of leaves in the plurality of pairs of leaves.

37. The system according to claim 36, characterized in that The blade position of each pair of blades in the plurality of pairs of blades includes a center position of an opening of the pair of blades and an opening width of the pair of blades.

38. The system according to any one of claims 28 to 31, characterized in that The plurality of target position combinations are within a plurality of discrete candidate position combinations of the general set of positions.

39. The system according to claim 38, characterized in that Getting the initial model includes: Acquire the general set of locations including the plurality of candidate location combinations; and An initial model is obtained based on the general set of locations.

40. The system according to claim 39, characterized in that Acquiring the general set of positions including the plurality of candidate position combinations, comprising: Acquire multiple candidate gantry angles, multiple candidate collimator angles, or multiple candidate table positions; and The general set of positions is obtained based on the multiple candidate gantry angles, the multiple candidate collimator angles, or the multiple candidate table positions.

41. The system according to claim 39, characterized in that The plurality of target sub-field shapes are within a universal set of distances that includes a plurality of discrete candidate leaf positions.

42. The system according to claim 41, characterized in that Getting the initial model includes: obtaining the universal set of distances including the plurality of candidate blade positions; and An initial model is acquired based on the distance universal set, so that the candidate sub-field shapes corresponding to each of the plurality of candidate position combinations output by the sub-field shape estimation model are all in the distance universal set.

43. The system according to claim 42, characterized in that The plurality of candidate blade positions include a plurality of candidate opening positions and a plurality of candidate opening widths.

44. The system according to claim 42, characterized in that Acquiring the training data includes: For one of the multiple training sets, Obtain historical treatment plans previously generated based on lesion samples; Acquire a first sample image of the lesion sample corresponding to the historical treatment plan; Obtaining a sample position combination and a corresponding sample sub-field shape in the historical treatment plan; and The training set is acquired based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape.

45. The system according to claim 44, characterized in that Acquiring the training set based on the first sample image, the sample position combination in the historical treatment plan, and the sample sub-field shape, comprises: Acquire a processed sample position combination located in the position universal set, wherein the processed sample position combination is obtained by processing the sample position combination based on the position universal set; Acquire a processed sample subfield shape located in the distance universal set, wherein the processed sample subfield shape is obtained by processing the sample subfield shape based on the distance universal set; Acquire a sample set including the processed sample sub-field shape and a closed sub-field shape, wherein the candidate position combination corresponding to the closed sub-field shape does not include the processed sample position combination; and The training set is obtained through the sample set including the first sample image, the processed sample position combination and the historical treatment plan.

46. ​​The system according to claim 45, characterized in that The subfield shape estimation model is obtained by training the initial model based on the training data, including: Initializing the initial model; and The subfield shape estimation model is obtained by updating the initialized initial model using an iterative process including multiple iterations, wherein at least one iteration of the multiple iterations of the iterative process includes: Obtain one of the multiple training sets; Generating estimated subfield shapes corresponding to the plurality of candidate position combinations by inputting the first sample image of the training set into an intermediate model, wherein the intermediate model is the initial model initialized in the first iteration of the plurality of iterations of the iterative process or a previously updated model generated in a previous iteration of the iterative process; Determining a value of a loss function based on the estimated subfield shape and a sample set in the training set; Determine whether termination conditions are met; In response to not satisfying the termination condition, Updating the intermediate model based on the value of the loss function to generate an updated model; and Launch the next iteration; and The intermediate model in a last iteration of the plurality of iterations of the iterative process is designated as the sub-field shape estimation model.

47. The system according to claim 46, characterized in that At least one of the plurality of iterations of the iterative process comprises: In response to determining that the termination condition is satisfied, the iterative process is terminated.

48. The system according to claim 46, characterized in that The value of the loss function is determined based on the sparsity of the sample set, and the sparsity of the sample set is related to the closed sub-field shape of the sample set.

49. The system according to claim 46, characterized in that The termination condition is related to at least one of a value of the loss function or a number of iterations of the iterative process that have been performed.

50. The system of claim 46, wherein: The training set includes a second sample image of normal tissue surrounding the lesion sample, a third sample image of the lesion sample, or sample radiation information of the historical treatment plan, wherein the sample radiation information includes at least one of a sample output dose, a sample dose output rate, a sample dose per pulse, or a sample dose distribution in the lesion sample.

51. The system according to claim 50, characterized in that The sample radiation information is predicted based on the first sample image of the lesion sample, the second sample image of normal tissue around the lesion sample, and the third sample image of the lesion sample.

52. The system according to claim 50, characterized in that At least one of the plurality of iterations of the iterative process comprises: The estimated sub-field shape is generated by inputting at least one of the second sample image, the third sample image, or the sample radiation information of the training set into the intermediate model.

53. The system according to claim 52, characterized in that At least one of the plurality of iterations of the iterative process comprises: Determining estimated radiation information based on the estimated sub-field shape; comparing the estimated radiation information with the sample radiation information; and The intermediate model is updated based on the comparison to generate the updated model.

54. A method implemented on a machine comprising one or more processors and one or more storage devices, comprising: Acquiring input data related to a target treatment plan for performing radiotherapy on a lesion using a radiation device, the input data comprising a first target image of the lesion; Obtaining a subfield shape estimation model; as well as Based on the sub-field shape estimation model and the input data, multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device are estimated, one of the multiple target position combinations indicates the position of the collimator relative to the lesion, and each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, wherein the sub-field shape estimation model is configured to determine the target position combinations and the target sub-field shapes.

55. A method implemented on a machine comprising one or more processors and one or more storage devices, comprising: Get the initial model; Get training data; as well as By training the initial model based on the training data, a sub-field shape estimation model is obtained, wherein the sub-field shape estimation model is configured to estimate multiple target position combinations in a target treatment plan based on input data, wherein the target position combinations are used to perform radiotherapy on a lesion using a radiation device and multiple target sub-field shapes of a collimator of the radiation device, wherein the input data includes a first target image of the lesion, each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, and one of the multiple target position combinations indicates a position of the collimator relative to the lesion, wherein the sub-field shape estimation model is configured to determine the target position combinations and the target sub-field shapes.

56. A system comprising: An input acquisition module, used to acquire input data related to a target treatment plan for performing radiotherapy on a lesion using a radiation device, wherein the input data includes a first target image of the lesion; A model acquisition module, used to acquire a sub-field shape estimation model; and A shape estimation module is used to estimate multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device based on the sub-field shape estimation model and the input data, one of the multiple target position combinations indicates the position of the collimator relative to the lesion, and each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, wherein the sub-field shape estimation model is configured to determine the target position combinations and the target sub-field shapes.

57. A system comprising a model acquisition module, wherein the model acquisition module is used to: Get the initial model; Obtaining training data; and The sub-field shape estimation model is obtained by training the initial model based on the training data, wherein the sub-field shape estimation model is configured to estimate a plurality of target position combinations in a target treatment plan based on input data, wherein the target position combinations are used to perform radiotherapy on a lesion using a radiation device and a plurality of target sub-field shapes of a collimator of the radiation device, wherein the input data includes a first target image of the lesion, each of the plurality of target sub-field shapes corresponds to one of the plurality of target position combinations, and one of the plurality of target position combinations indicates a position of the collimator relative to the lesion, wherein The sub-field shape estimation model is configured to determine the target position combination and the target sub-field shape.

58. A non-transitory computer readable medium comprising at least one set of instructions, wherein: When executed by one or more processors of a computing device, the at least one set of instructions causes the computing device to perform a method comprising: Acquiring input data related to a target treatment plan for performing radiotherapy on a lesion using a radiation device, the input data comprising a first target image of the lesion; Obtaining a subfield shape estimation model; and Based on the sub-field shape estimation model and the input data, multiple target position combinations of the target treatment plan and multiple target sub-field shapes of the collimator of the radiation device are estimated, one of the multiple target position combinations indicates the position of the collimator relative to the lesion, and each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, wherein the sub-field shape estimation model is configured to determine the target position combinations and the target sub-field shapes.

59. A non-transitory computer readable medium comprising at least one set of instructions, wherein: When executed by one or more processors of a computing device, the at least one set of instructions causes the computing device to perform a method comprising: Get the initial model; Obtaining training data; and By training the initial model based on the training data, a sub-field shape estimation model is obtained, wherein the sub-field shape estimation model is configured to estimate multiple target position combinations in a target treatment plan based on input data, wherein the target position combinations are used to perform radiotherapy on a lesion using a radiation device and multiple target sub-field shapes of a collimator of the radiation device, wherein the input data includes a first target image of the lesion, each of the multiple target sub-field shapes corresponds to one of the multiple target position combinations, and one of the multiple target position combinations indicates a position of the collimator relative to the lesion, wherein the sub-field shape estimation model is configured to determine the target position combinations and the target sub-field shapes.

Images