CN111128340A

CN111128340A - Radiotherapy plan generation device, radiotherapy plan generation apparatus, and storage medium

Info

Publication number: CN111128340A
Application number: CN201911357329.2A
Authority: CN
Inventors: 张康; 方磊
Original assignee: Shanghai United Imaging Healthcare Co Ltd
Current assignee: Shanghai United Imaging Healthcare Co Ltd
Priority date: 2019-12-25
Filing date: 2019-12-25
Publication date: 2020-05-08

Abstract

The invention discloses a radiotherapy plan generating device, a radiotherapy plan generating device and a storage medium, wherein the device comprises: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors are caused to perform operations comprising: inputting the image information and the prescription dose information into a preset neural network model to obtain a current operation instruction; controlling a radiation treatment planning module to generate a current radiation treatment plan according to the current operation instruction; if the current radiotherapy plan does not meet the preset cycle ending condition, feeding the current score corresponding to the current radiotherapy plan back to the preset neural network model so as to update the output current operation instruction and return to execute the operation of controlling the radiotherapy plan generating module; and if the current radiotherapy plan meets the preset cycle ending condition, determining the current radiotherapy plan as a target radiotherapy plan, thereby improving the plan generation efficiency and accuracy.

Description

Radiotherapy plan generation device, radiotherapy plan generation apparatus, and storage medium

Technical Field

Embodiments of the present invention relate to the field of radiology, and in particular, to a radiotherapy plan generation apparatus, a radiotherapy plan generation device, and a storage medium.

Background

Radiation therapy is a common treatment for tumors. Before radiation treatment, a radiation treatment plan with better effect is usually required to be made according to patient information, for example, the radiation dose distribution of a tumor region is more uniform, the dose drop around a tumor is larger, the radiation dose of a endangered organ is lower, and the like, so that effective radiation treatment can be performed on a patient based on the better radiation treatment plan, and the treatment effect is improved.

In the prior art, a doctor usually delineates a target region and an organ at risk of a patient image and gives a prescription, and a physicist manually operates a radiotherapy Planning System (TPS) according to the doctor prescription to generate a radiotherapy plan which can satisfy the doctor prescription and has an ideal Dose Volume Histogram (DVH) and Dose distribution through a plurality of operations.

However, the process of a physicist generating a radiation treatment plan by manually operating a radiation therapy planning system is very time consuming. For a typical case, an experienced physicist will take tens of minutes to obtain a radiation treatment plan that meets clinical requirements; for more complex cases, it may take several hours or more, greatly reducing the efficiency of radiation treatment plan generation. And it is difficult for the physicist to determine whether the currently generated radiation treatment plan is the optimal plan based on his own experience, thereby reducing the accuracy of plan generation.

Disclosure of Invention

The embodiment of the invention provides a radiotherapy plan generating device, a radiotherapy plan generating device and a storage medium, so that the generation efficiency and the accuracy of a radiotherapy plan are improved.

In a first aspect, an embodiment of the present invention provides a radiation therapy plan generating apparatus, including:

one or more processors;

a memory for storing one or more programs;

when executed by the one or more processors, cause the one or more processors to perform operations comprising:

inputting image information of a patient and prescription dose information corresponding to the image information into a preset neural network model to obtain a current operation instruction output by the preset neural network model;

controlling a radiation treatment plan module according to the current operation instruction to generate a current radiation treatment plan;

if the fact that the current radiotherapy plan does not meet the preset cycle ending condition is detected, feeding back a current score corresponding to the current radiotherapy plan to the preset neural network model so that the preset neural network model updates an output current operation instruction based on the current score, and returning to execute the operation of controlling the radiotherapy plan generating module according to the current operation instruction based on the updated current operation instruction;

and if the current radiation treatment plan is detected to meet the preset cycle end condition, determining the current radiation treatment plan as the target radiation treatment plan corresponding to the patient.

In a second aspect, an embodiment of the present invention further provides a radiation therapy plan generating apparatus, including:

the current operation instruction obtaining module is used for inputting image information of a patient and prescription dose information corresponding to the image information into a preset neural network model and obtaining a current operation instruction output by the preset neural network model;

the current radiotherapy plan generating module is used for controlling the radiotherapy plan module according to the current operating instruction to generate a current radiotherapy plan;

the current score feedback module is used for feeding back a current score corresponding to the current radiotherapy plan to the preset neural network model if the current radiotherapy plan is detected not to meet a preset cycle ending condition, so that the preset neural network model updates an output current operation instruction based on the current score, and returns to execute the operation of the radiotherapy plan generation module controlled according to the current operation instruction based on the updated current operation instruction;

and the target radiation treatment plan determining module is used for determining the current radiation treatment plan as the target radiation treatment plan corresponding to the patient if the current radiation treatment plan is detected to meet the preset cycle ending condition.

In a third aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the following operations:

According to the embodiment of the invention, a preset neural network model is utilized, so that a current operation instruction can be automatically output based on input image information of a patient and prescription dose information corresponding to the image information, and the current operation instruction is used for controlling a radiotherapy planning module provided with a radiotherapy planning system TPS, so that the radiotherapy planning module generates a current radiotherapy plan under the control operation of the preset neural network model. When the current radiotherapy plan is detected not to meet the preset cycle ending condition, the current radiotherapy plan is not the optimal radiotherapy plan, at the moment, the current score corresponding to the current radiotherapy plan is fed back to the preset neural network model, so that the preset neural network model can update the current operation instruction based on the current score and output the updated current operation instruction, the radiotherapy plan module can update the current radiotherapy plan under the control of the updated current operation instruction until the current radiotherapy plan meets the preset cycle ending condition, the current radiotherapy plan is determined to be the target radiotherapy plan corresponding to the patient, and therefore a very experienced physicist can be simulated by the preset neural network model to automatically operate the radiotherapy plan module to generate a more accurate radiotherapy plan, and human intervention is not needed in the whole process, so that the generation efficiency and the accuracy of the radiation treatment plan are improved.

Drawings

Fig. 1 is a schematic structural diagram of a radiation therapy plan generating apparatus according to an embodiment of the present invention;

FIG. 2 is a flow chart of a radiation treatment plan generation process performed by the processor in accordance with an embodiment of the present invention;

FIG. 3 is a flow chart of a training process of a neural network model according to an embodiment of the present invention;

fig. 4 is a flowchart of a radiation treatment plan generating process executed by a processor in a radiation treatment plan generating apparatus according to a second embodiment of the present invention;

fig. 5 is a schematic structural diagram of a radiation therapy plan generating apparatus according to a third embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.

Example one

Fig. 1 is a schematic structural diagram of a radiation therapy plan generating apparatus according to an embodiment of the present invention; fig. 2 is a flow chart of a radiation treatment plan generation process performed by a processor in accordance with an embodiment of the present invention. As shown in fig. 1 and 2, the radiation therapy plan generating apparatus in the present embodiment includes:

one or more processors 110;

a memory 120 for storing one or more programs;

when executed by the one or more processors 110, the one or more programs enable the one or more processors 110 to perform operations S210-S250 to generate an optimal radiation treatment plan.

In FIG. 1, a processor 110 is illustrated; the processor 110 and the memory 120 in the device may be connected by a bus or other means, which is exemplified in fig. 1. The memory 120 is a computer-readable storage medium for storing software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to steps S210-S250 in the embodiments of the present invention. The processor 110 executes various functional applications and data processing in the device, i.e., performs the operations of steps S210-S250, by executing software programs, instructions, and modules stored in the memory 120.

The memory 120 mainly includes a program storage area and a data storage area, wherein the program storage area can store an operating system and an application program required by at least one function; the storage data area may store data created according to use of the device, and the like. Further, the memory 120 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, the memory 120 may further include memory located remotely from the processor 110, which may be connected to the device over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

Referring to fig. 2, the radiation treatment plan generation process performed by the processor 110 specifically includes the following steps:

s210, inputting the image information of the patient and the prescription dose information corresponding to the image information into a preset neural network model, and obtaining a current operation instruction output by the preset neural network model.

The image information may include, but is not limited to, a tomographic image and a region of interest that has been outlined in the tomographic image. The tomographic image may be an image obtained by scanning a portion of a patient with a scanning device and reconstructing the image from the obtained scan data. The scanning device may be, but is not limited to, an X-ray machine (X-ray tomography), a CT device (computed tomography), and an MR device (Magnetic Resonance scanning). The region of interest may include a target region and an organ-at-risk region. The target region may refer to a region of a tumor that requires radiation therapy. An organ-at-risk region may refer to a region that cannot receive an excessively high radiation dose during radiation treatment, such as an eyeball, a spinal cord, a lung, a heart, and the like. The region of interest in the tomographic image in the present embodiment may be manually obtained by a doctor or automatically obtained based on an existing algorithm.

The prescription dose information corresponding to the image information may refer to a prescription dose distribution given by a doctor based on the image information. The preset neural network model may refer to a neural network model for determining an operation instruction. Illustratively, the preset Neural network model may be, but is not limited to, a Convolutional Neural Network (CNN) model. The preset neural network model in this embodiment is a model obtained after training based on sample data, so that the preset neural network model can output a more accurate current operation instruction. The current operating instructions may include, but are not limited to, at least one of: adding a constraint, deleting a constraint, adding an auxiliary structure, deleting an auxiliary structure, increasing a constraint value, decreasing a constraint value, increasing an iteration number, decreasing an iteration number, starting optimization, continuing optimization, increasing a particle beam, decreasing a particle beam, increasing a collimator angle of a particle beam, and decreasing a collimator angle of a particle beam.

Specifically, the image information of the patient and the prescription dose information corresponding to the image information are input into a pre-trained preset neural network model, and the preset neural network model can automatically output a matched current operation instruction based on the image information and the prescription dose information.

And S220, controlling a radiation treatment planning module according to the current operation instruction to generate a current radiation treatment plan.

The radiotherapy planning module may be a module on which a radiotherapy planning system TPS is mounted. The current radiation treatment plan may refer to a radiation treatment plan prepared at the current time. The radiation therapy plan may include a radiation field angle, a radiation field area, a dose volume histogram DVH, a dose distribution, etc. at the time of radiation therapy.

Specifically, the current operation instruction output by the preset neural network model may be input into the radiation therapy planning module, so that the radiation therapy planning module may be controlled to perform automatic operation based on the current operation instruction to generate a corresponding current radiation therapy plan. The present embodiment can automatically operate the radiation therapy planning module to generate the current radiation therapy plan through the current operation instruction output by the preset neural network model, thereby replacing the way in which a physicist manually operates the radiation therapy planning module to generate the radiation therapy plan, so that the generation process of the radiation therapy plan does not need manual participation, the generation efficiency is improved, and the preset neural network model is obtained in advance based on sample data training, so that the automatic operation of the preset neural network model is more accurate, which is equivalent to a very experienced physicist, thereby improving the accuracy of the generation of the radiation therapy plan.

S230, detecting whether the current radiotherapy plan meets a preset cycle ending condition, and if so, executing the operation of S240; if not, the operation of S250 is executed.

The preset end-of-cycle condition may refer to a condition for characterizing the generated radiation therapy plan as an optimal radiation therapy plan. The preset loop ending condition may be preset based on the service requirement and the scenario. For example, the preset loop ending condition may be that the loop operation is ended when the current loop number is equal to the preset loop number; or when the current score corresponding to the current radiation treatment plan is greater than or equal to the preset score, the circulation operation is ended.

Specifically, since the preset neural network model needs to perform several control operations on the radiation therapy planning module, the radiation therapy planning module can generate an optimal radiation therapy plan, and thus after the current radiation therapy plan is generated each time, it needs to detect whether the current radiation therapy plan meets a preset cycle end condition so as to determine whether the current radiation therapy plan is a radiation therapy plan meeting clinical standards.

And S240, determining the current radiation treatment plan as a target radiation treatment plan corresponding to the patient.

Specifically, when it is detected that the current radiation treatment plan satisfies the preset loop end condition, which indicates that the current radiation treatment plan is the optimal radiation treatment plan, the current radiation treatment plan may be determined as the target radiation treatment plan corresponding to the patient.

And S250, feeding back the current score corresponding to the current radiotherapy plan to the preset neural network model so that the preset neural network model updates the output current operation instruction based on the current score, and returns to execute the operation of S220 based on the updated current operation instruction.

Wherein, the current score corresponding to the current radiation therapy plan can be used to measure the quality of the current radiation therapy plan. For example, a higher current score indicates a higher quality current radiation treatment plan and better treatment.

Specifically, the present embodiment may score the current radiation treatment plan based on a preset scoring manner, and determine a corresponding current score. For example, the performance of each evaluation parameter may be analyzed according to the current radiation treatment plan, each evaluation parameter is scored, all scores are weighted and stacked, and the obtained stacking result is determined as the current score corresponding to the current radiation treatment plan. When it is detected that the current radiotherapy plan does not meet the preset cycle end condition, it indicates that the current radiotherapy plan does not meet the radiotherapy plan of the clinical standard, at this time, the current score corresponding to the current radiotherapy plan may be fed back to the preset neural network model, the preset neural network model may re-output a current operation instruction based on the current score, update the current operation instruction, and return to perform the operation of S220 based on the updated current operation instruction, so as to control the radiotherapy plan module again according to the updated current operation instruction, and re-generate the current radiotherapy plan until the generated current radiotherapy plan meets the preset cycle end condition.

According to the technical scheme of the embodiment, by using a preset neural network model, a current operation instruction can be automatically output based on input image information of a patient and prescription dose information corresponding to the image information, and the current operation instruction is used for controlling a radiotherapy planning module provided with a radiotherapy planning system TPS, so that the radiotherapy planning module generates a current radiotherapy plan under the control operation of the preset neural network model. When the current radiotherapy plan is detected not to meet the preset cycle ending condition, the current radiotherapy plan is not the optimal radiotherapy plan, at the moment, the current score corresponding to the current radiotherapy plan is fed back to the preset neural network model, so that the preset neural network model can update the current operation instruction based on the current score and output the updated current operation instruction, the radiotherapy plan module can update the current radiotherapy plan under the control of the updated current operation instruction until the current radiotherapy plan meets the preset cycle ending condition, the current radiotherapy plan is determined to be the target radiotherapy plan corresponding to the patient, and therefore a very experienced physicist can be simulated by the preset neural network model to automatically operate the radiotherapy plan module to generate a more accurate radiotherapy plan, and human intervention is not needed in the whole process, so that the generation efficiency and the accuracy of the radiation treatment plan are improved.

On the basis of the technical scheme, before the preset neural network model is used, the training process of the preset neural network model is also included. The present embodiment may train the preset neural network model based on, but not limited to, a reinforcement learning manner, so that the trained preset neural network model may replace an experienced physicist to do planning work, and an effect of automatically generating a radiotherapy plan is achieved. Fig. 3 is a flow chart of a training process of a neural network model. As shown in fig. 3, the training process of the preset neural network model specifically includes the following steps:

s310, sample image information of a plurality of sample patients and sample prescription dose information corresponding to the sample image information are obtained.

The image information and the prescription dose information of the real patient at the same part can be acquired, and the image information and the prescription dose information of the real patient can be used for generating new image information and prescription dose information through a random algorithm, so that a large amount of sample data can be acquired, and the training precision is improved.

Specifically, when training is based on reinforcement learning, sample image information and sample prescription dose information of a sample patient can be used as sample data, and a high-quality radiotherapy plan made by a physicist does not need to be obtained, so that labor cost can be saved.

In the training process of reinforcement learning, two roles of a preset neural network model agent and an environment exist. The preset neural network model agent can make corresponding operations after acquiring the environment state (possibly only partial information), and the environment state is changed accordingly. The environment gives the pre-defined neural network model agent new status information and a reward or penalty (e.g., current rating) to the pre-defined neural network model agent based on how well the status changes. And the preset neural network model agent performs new operation according to the new environment state, and the environment state is changed again, so that the operation is continuously circulated. The goal of the pre-defined neural network model agent is to maximize the final accumulated reward. The preset neural network model agent can continuously adjust the network parameters of the preset neural network model agent in the training process according to the obtained feedback information until the final accumulated reward is maximized.

Reinforcement learning may be a Markov decision process. The Markov decision process may include a set S and a set a, where the set S is used to represent the environmental state; the set A is used for representing operation instructions which can be output by the preset neural network model agent. P_a(s,s′)＝Prob(s_t+1＝s′|s_t＝s,a_tA) may refer to a probability that the environment state is from the state s to the state s' after the preset neural network model agent outputs the operation instruction a. R_a(s, s ') is that agent outputs operation instruction a after presetting neural network model agent, and then the environmental state is from s to s' so that agentt the reward (or penalty) earned.

In the training process of the preset neural network model, the set S may include, but is not limited to: image information, sample prescription dose information corresponding to the sample image information, a current radiation treatment plan, current constraints, DVH information and dose distribution in the current radiation treatment plan, and a current number of iterations of the apparatus. Set a may include, but is not limited to: adding a constraint, deleting a constraint, adding an auxiliary structure, deleting an auxiliary structure, increasing a constraint value, decreasing a constraint value, increasing an iteration number, decreasing an iteration number, starting optimization, continuing optimization, increasing a particle beam, decreasing a particle beam, increasing a collimator angle of a particle beam, decreasing a collimator angle of a particle beam, and the like.

And S320, inputting the sample image information and the sample prescription dose information into the preset neural network model to obtain the current operation instruction output by the preset neural network model.

In the training phase, the current operation instruction output by the preset neural network model may be at least one of the following: adding a constraint, deleting a constraint, adding an auxiliary structure, deleting an auxiliary structure, increasing a constraint value, decreasing a constraint value, increasing an iteration number, decreasing an iteration number, starting optimization, continuing optimization, increasing a particle beam, decreasing a particle beam, increasing a collimator angle of a particle beam, and decreasing a collimator angle of a particle beam.

And S330, controlling a radiation treatment plan module according to the current operation instruction to generate a current radiation treatment plan.

And S340, determining the current score corresponding to the current radiotherapy plan, and accumulating the current score after each circulation to obtain the current total score.

Specifically, the present embodiment may score the current radiation treatment plan based on a preset scoring manner, and determine a corresponding current score. For example, the performance of each evaluation parameter may be analyzed according to the current radiation treatment plan, each evaluation parameter is scored, all scores are weighted and stacked, and the obtained stacking result is determined as the current score corresponding to the current radiation treatment plan. And adding the current scores corresponding to each current radiation treatment plan obtained in the current cycle to obtain a current total score, so that the current total score can be used as the accumulated reward of the current cycle.

It should be noted that the weight of each evaluation parameter can be set based on the preference of the physicist to learn the preference of the physicist, so that the trained preset neural network model can automatically generate a radiation therapy plan more preferred by the physicist.

S350, detecting whether the current cycle number is smaller than the preset cycle number, and if so, executing the operation of S360; if not, the operation of S370 is performed.

Wherein the preset number of cycles may be preset based on business requirements for characterizing conditions that may generate a radiation treatment plan meeting clinical criteria.

And S360, accumulating the current cycle times by 1, feeding the current score back to the preset neural network model to adjust the network parameters in the preset neural network model, updating the output current operation instruction, and returning to execute the operation of the S330 based on the updated current operation instruction.

Specifically, when the current cycle number is less than the preset cycle number, it indicates that the current cycle number has not ended, at this time, the current cycle number may be added by 1, so as to update and obtain the total cycle number, and the current score is fed back to the preset neural network model, the preset neural network model may automatically adjust the network parameters based on the current score, and according to the adjusted network parameters, re-output the current operation instruction, so as to update the current operation instruction, and by returning to perform the operation of S330, control the radiation treatment planning module again based on the updated current operation instruction, re-generate the current radiation treatment plan, so as to update the current radiation treatment plan, until the current cycle number is equal to the preset cycle number.

S370, detecting whether the current total score is larger than or equal to a preset score value, if so, executing the operation of S380; if not, the operation of S390 is executed.

Specifically, when the current cycle number is equal to the preset cycle number, it indicates that when the sub-cycle has ended, it may be determined whether the current radiation treatment plan is the optimal radiation treatment plan by detecting whether the current total score is greater than or equal to the preset score value.

And S380, determining that the preset neural network model training is finished.

Specifically, when the current total score is greater than or equal to the preset score value, it indicates that the current radiotherapy plan is the best radiotherapy plan meeting the clinical standard, and at this time, it may be determined that the training of the preset neural network model is finished, that is, the preset neural network model may operate and control the radiotherapy planning module like a real experienced physicist, so that the radiotherapy planning module may more quickly generate a high-quality radiotherapy plan meeting the clinical standard.

And S390, clearing the current cycle number and the current total score, feeding the current score back to the preset neural network model to adjust network parameters in the preset neural network model, updating the output current operation instruction, and returning to execute the operation of S330 based on the updated current operation instruction.

Specifically, when the current total score is smaller than the preset score value, it indicates that the currently generated current radiotherapy plan is not the optimal radiotherapy plan meeting the clinical standard, and the preset neural network model needs to be trained continuously, and at this time, the current cycle number and the current total score may be cleared to restart the next cycle until the current total score obtained after the cycle is greater than or equal to the preset score value.

Example two

Fig. 4 is a flowchart of a radiation treatment plan generating process executed by a processor in a radiation treatment plan generating apparatus according to a second embodiment of the present invention, and this embodiment describes in detail a determination process of a current score corresponding to a current radiation treatment plan based on the above embodiment. Wherein explanations of the same or corresponding terms as those of the above-described embodiments are omitted.

Referring to fig. 4, the radiation therapy plan generation process provided by the present embodiment specifically includes the following steps:

s410, inputting the image information of the patient and the prescription dose information corresponding to the image information into a preset neural network model, and obtaining a current operation instruction output by the preset neural network model.

And S420, controlling a radiation treatment planning module according to the current operation instruction to generate a current radiation treatment plan.

And S430, determining a current parameter value corresponding to each preset evaluation parameter according to the current radiotherapy plan.

In which, various preset evaluation parameters for evaluating the quality of the current radiotherapy plan can be preset based on the DVH information of the target area and the organs at risk, the adaptability of the dose field around the target area, the positions of the cold and hot spots, the dose and other factors. For example, for a radiation treatment plan for treating cervical cancer, the preset evaluation parameters may include, but are not limited to: target D95, target D2, maximum dose of target, bladder V40, bladder V30, bladder V20, rectal V40, rectal V30, rectal V20, maximum dose of femoral head, maximum dose of spine, and the like. The target region D95 is the dose achieved by 95% of the volume of the target region. Bladder V40 refers to the volume in the bladder at a dose of greater than 40 Gy.

Specifically, the size of the current parameter value corresponding to each preset evaluation parameter may be determined according to the DVH information and dose distribution information in the current radiation treatment plan.

And S440, determining a parameter grading result corresponding to the preset evaluation parameter according to the current parameter value.

Wherein the parameter scoring result can be used to characterize the performance of the current radiation treatment plan at each of the preset evaluation parameters. For example, a higher score in the parameter score results indicates a better performance of the current radiation treatment plan on the preset evaluation parameters.

Exemplarily, S440 may include: when the current parameter value is smaller than the first standard parameter value, determining a parameter grading result corresponding to the preset evaluation parameter according to the current parameter value and the first standard parameter value; when the current parameter value is greater than or equal to the first standard parameter value and less than or equal to the second standard parameter value, determining the preset highest score value as a parameter scoring result corresponding to the preset evaluation parameter; and when the current parameter value is larger than the second standard parameter value, determining a parameter grading result corresponding to the preset evaluation parameter according to the current parameter value and the second standard parameter value.

The first standard parameter value and the second standard parameter value may be used to represent a preset parameter value range of an acceptable preset evaluation parameter. The second standard parameter value is greater than the first standard parameter value. The preset highest score value may refer to a highest score value achievable by the preset evaluation parameter. For example, the highest score value may be set to 1.

Specifically, when the current parameter value is between the first standard parameter value and the second standard parameter value, it indicates that the current parameter value of the preset evaluation parameter meets the preset clinical standard, and at this time, the preset highest score value may be determined as the parameter scoring result corresponding to the preset evaluation parameter. When the current parameter value is smaller than the first standard parameter value or larger than the second standard parameter value, it indicates that the current parameter value does not meet the preset clinical standard, and at this time, a corresponding parameter scoring result may be determined based on the first standard parameter value or the second standard parameter value. For example, the greater the distance between the current parameter value and the first standard parameter value or the second standard parameter value, the lower the quality of the parameter is, and the smaller the score value in the determined parameter scoring result is.

Illustratively, when the preset evaluation parameter is the target D95 and the dose of the target D95 is desired to be at 6000cGy and 6100cGy, i.e., the first standard parameter value is 6000cGy and the second standard parameter value is 6100cGy, then the current parameter value X corresponding to D95 can be calculated by the following formula based on the current parameter value X corresponding to D95₁Determining a parameter scoring result F corresponding to the target area D95₁(X₁)：

S450, determining an accumulated scoring result corresponding to the current radiotherapy plan according to the weight value corresponding to each preset evaluation parameter and the parameter scoring result.

Specifically, the scoring results of the parameters may be weighted and superimposed based on the weighted values corresponding to the preset evaluation parameters, and the obtained superimposed result is determined as the cumulative scoring result corresponding to the current radiotherapy plan, so that the current radiotherapy plan may be individually scored based on each preset evaluation parameter, so as to respectively evaluate the performance of the current radiotherapy plan on each preset evaluation parameter.

And S460, determining a penalty scoring result corresponding to the current radiotherapy plan according to the target preset evaluation parameter in the preset evaluation parameters.

The target preset evaluation parameters can be selected from the preset evaluation parameters and used for comprehensively measuring the quality of the current radiotherapy plan so as to score the current radiotherapy plan as a whole. The target preset evaluation parameter may include at least two preset evaluation parameters. The penalty score results may be used to characterize the degree of quality of the current radiation treatment plan as a whole.

Specifically, the corresponding target preset evaluation parameter may be selected from the preset evaluation parameters for characterization by considering a difference of the parameter scoring results of the preset evaluation parameters, or a uniformity of the dose distribution. And determining a penalty scoring result corresponding to the current radiation treatment plan according to the performance condition of the target preset evaluation parameter, so that the condition that the parameter scoring results of the preset evaluation parameters have large difference or the condition that the dose distribution condition is not uniform can be punished.

Exemplarily, S460 may include: determining a first grading result corresponding to the current radiotherapy plan according to a maximum parameter grading result and a minimum parameter grading result in the parameter grading results corresponding to the preset evaluation parameters; and/or determining a second grading result corresponding to the current radiotherapy plan according to a first current parameter value and a second current parameter value which respectively correspond to a preset evaluation parameter pair used for representing the dose distribution condition in each preset evaluation parameter; and determining a penalty scoring result corresponding to the current radiation treatment plan according to the first scoring result and/or the second scoring result.

And the preset evaluation parameter corresponding to the maximum parameter scoring result is the best-performing preset evaluation parameter. The preset evaluation parameter corresponding to the minimum parameter scoring result is the preset evaluation parameter with the worst performance. The first scoring result may be a difference reflecting how well each of the preset evaluation parameters performs. For example, the larger the difference between the best performing preset evaluation parameter and the worst preset evaluation parameter, the higher the score of the first scoring result. The second scoring result may reflect non-uniformity of dose distribution. The more uneven the dose distribution, the higher the score of the second scoring result.

Specifically, if only the first scoring result is determined, the first scoring result may be determined as a penalty scoring result corresponding to the current radiation treatment plan. If only the second scoring result is determined, the second scoring result may be determined as a penalty scoring result corresponding to the current radiation treatment plan. If the first scoring result and the second scoring result are determined, the sum of the first scoring result and the second scoring result can be determined as a penalty scoring result corresponding to the current radiation treatment plan.

And S470, determining the current score corresponding to the current radiation treatment plan according to the accumulated scoring result and the penalty scoring result.

Specifically, the difference between the cumulative score result and the penalty score result may be determined as the current score corresponding to the current radiation treatment plan. Compared with the method that the accumulated scoring result is directly determined as the current scoring, namely, only the performance condition of each preset evaluation factor is considered independently, the embodiment can consider the association among the preset evaluation factors, take all the preset evaluation parameters into consideration, avoid the condition that the importance of a certain preset evaluation parameter is neglected due to low weight, enable the current scoring to be more consistent with the real condition, and improve the robustness and the reference value of the scoring mode.

S480, detecting whether the current radiotherapy plan meets a preset cycle ending condition, and if so, executing the operation of S490; if not, the operation of S491 is executed.

It should be noted that the execution sequence of steps S430 to S470 is not limited in this embodiment. If the preset loop end condition is a condition set based on the current rating, the operations of steps S430-S470 may be performed before step S480. If the preset loop end condition is not a condition set based on the current score, the operations of steps S430-S470 may be performed after step S480.

And S490, determining the current radiation treatment plan as a target radiation treatment plan corresponding to the patient.

S491, feeding back the current score corresponding to the current radiotherapy plan to the preset neural network model, so that the preset neural network model updates the output current operation instruction based on the current score, and returns to execute the operation of S420 based on the updated current operation instruction.

According to the technical scheme, the penalty scoring result corresponding to the current radiotherapy plan is determined integrally, the cumulative scoring result and the penalty scoring result are used for determining the current score corresponding to the current radiotherapy plan together, and compared with the method of determining the current score only by using the cumulative scoring result, namely, only considering the performance condition of each preset evaluation factor separately, the embodiment can consider the association among the preset evaluation factors, give consideration to each preset evaluation parameter, avoid the condition that the importance of a certain preset evaluation parameter is neglected due to low weight, enable the current score to be more consistent with the real condition, and improve the robustness and the reference value of the scoring mode.

On the basis of the above technical solution, determining a first scoring result corresponding to the current radiotherapy plan according to a maximum parameter scoring result and a minimum parameter scoring result in the parameter scoring results corresponding to each preset evaluation parameter may include:

determining a first difference value between the maximum parameter scoring result and the minimum parameter scoring result according to the maximum parameter scoring result and the minimum parameter scoring result in the parameter scoring results corresponding to the preset evaluation parameters; if the first difference is smaller than or equal to a first preset difference, determining a first scoring result corresponding to the current radiotherapy plan by using a preset lowest score value; and if the first difference is larger than a first preset difference, determining a first grading result corresponding to the current radiation treatment plan according to the first difference.

Wherein the first predetermined difference may be a maximum value of the first difference that is acceptable to the physician. The preset minimum score value may be a preset minimum score value, such as 0, based on traffic demand. If the first difference is smaller, the difference of the performance of each preset evaluation parameter is smaller, that is, the first scoring result corresponding to the current radiation treatment plan is smaller. Specifically, when the first difference is greater than a first preset difference, a first scoring result corresponding to the current radiation treatment plan may be determined according to the following formula:

E⁽¹⁾＝ρ₁(_max-F_min-₃)²；F_max-F_min>A₃

wherein E is⁽¹⁾Is a first scoring result; f_maxMeans the maximum parameter score result; f_minMeans minimum parameter score result value; a. the₃Is a first preset difference; rho₁Is a first predetermined coefficient. For example, for generating a radiation treatment plan for treating cervical cancer, the first preset difference A₃May be set to 0.2, a first preset coefficient ρ₁May be set to 40. Therefore, the method can be used for treating the blood pressure change.

On the basis of the above technical solution, determining a second scoring result corresponding to the current radiation treatment plan according to a first current parameter value and a second current parameter value respectively corresponding to a preset evaluation parameter pair used for characterizing a dose distribution condition in each preset evaluation parameter may include:

determining a second difference value between the first current parameter value and the second current parameter value according to a first current parameter value and a second current parameter value which respectively correspond to a preset evaluation parameter pair used for representing the dose distribution condition in each preset evaluation parameter; if the second difference is smaller than or equal to a second preset difference, determining a second scoring result corresponding to the current radiotherapy plan by using the preset lowest score value; and if the second difference is larger than a second preset difference, determining a second grading result corresponding to the current radiation treatment plan according to the second difference.

Wherein the second predetermined difference may be a maximum value of the second difference that is acceptable to the physician. The preset evaluation parameter pairs for characterizing the dose distribution in each preset evaluation parameter may be set as two preset evaluation parameters of the target zone D95 and the target zone D2. If the difference between the dose value corresponding to the target area D95 and the dose value corresponding to the target area D2 is smaller, it indicates that the dose distribution of the target area is more uniform, i.e. the second scoring result is smaller. Specifically, when the second difference is greater than the second preset difference, the second scoring result corresponding to the current radiation treatment plan may be determined according to the following formula:

>A₄

wherein E is⁽²⁾Refers to the second scoring result; d is a second difference; a. the₄Is a second predetermined difference; rho₂Is a second predetermined coefficient; rho₃Is a third predetermined coefficient. For generating a radiation treatment plan for treating cervical cancer, the second preset difference A₄May be set to 100; second predetermined coefficient ρ₂May be set to 10; third predetermined coefficient ρ₃May be set to 100.

The following is an embodiment of a radiation therapy plan generation device provided in an embodiment of the present invention, which belongs to the same inventive concept as the radiation therapy plan generation apparatuses of the above embodiments, and reference may be made to the embodiments of the radiation therapy plan generation apparatus for details that are not described in detail in the embodiments of the radiation therapy plan generation device.

EXAMPLE III

Fig. 5 is a schematic structural diagram of a radiation therapy plan generating device according to a third embodiment of the present invention, which is applicable to automatically generating an optimal radiation therapy plan, and the device specifically includes: a current operating instruction obtaining module 510, a current radiation treatment plan generating module 520, a current score feedback module 530, and a target radiation treatment plan determining module 540.

The current operation instruction obtaining module 510 is configured to input image information of a patient and prescription dose information corresponding to the image information into a preset neural network model, and obtain a current operation instruction output by the preset neural network model; a current radiation treatment plan generating module 520, configured to control the radiation treatment plan module according to the current operation instruction, and generate a current radiation treatment plan; a current score feedback module 530, configured to, if it is detected that the current radiotherapy plan does not meet the preset cycle end condition, feed back a current score corresponding to the current radiotherapy plan to the preset neural network model, so that the preset neural network model updates an output current operation instruction based on the current score, and returns to execute an operation of controlling the radiotherapy plan generation module according to the current operation instruction based on the updated current operation instruction; a target radiation treatment plan determining module 540, configured to determine the current radiation treatment plan as a target radiation treatment plan corresponding to the patient if it is detected that the current radiation treatment plan meets a preset cycle end condition.

Optionally, the image information includes the tomogram and a region of interest that has been outlined in the tomogram.

Optionally, the current operation instruction comprises at least one of:

adding a constraint, deleting a constraint, adding an auxiliary structure, deleting an auxiliary structure, increasing a constraint value, decreasing a constraint value, increasing an iteration number, decreasing an iteration number, starting optimization, continuing optimization, increasing a particle beam, decreasing a particle beam, increasing a collimator angle of a particle beam, and decreasing a collimator angle of a particle beam.

Optionally, the apparatus further comprises: the training module is specifically configured to:

before using a preset neural network model, obtaining sample image information of a plurality of sample patients and sample prescription dose information corresponding to the sample image information; inputting the sample image information and the sample prescription dose information into a preset neural network model to obtain a current operation instruction output by the preset neural network model; controlling a radiation treatment plan module according to the current operation instruction to generate a current radiation treatment plan; determining a current score corresponding to the current radiotherapy plan, and accumulating the current scores after each circulation to obtain a current total score; when the current cycle number is smaller than the preset cycle number, feeding the current score back to the preset neural network model to adjust network parameters in the preset neural network model, updating the output current operation instruction, returning to execute the operation of controlling the radiotherapy plan generation module according to the current operation instruction based on the updated current operation instruction, and accumulating the current cycle number by 1; when the current cycle number is equal to the preset cycle number, detecting whether the current total score is greater than or equal to a preset score value; if so, determining that the preset neural network model is trained to be finished; and if not, resetting the current cycle number and the current total score, feeding the current score back to the preset neural network model to adjust network parameters in the preset neural network model, updating the output current operation instruction, and returning to execute the operation of controlling the radiotherapy plan generation module according to the current operation instruction based on the updated current operation instruction.

Optionally, the apparatus further comprises:

the current parameter value determining module is used for determining a current parameter value corresponding to each preset evaluation parameter according to the current radiotherapy plan before feeding back the current score corresponding to the current radiotherapy plan to the preset neural network model;

the parameter scoring result determining module is used for determining a parameter scoring result corresponding to the preset evaluation parameter according to the current parameter value;

the cumulative scoring result determining module is used for determining a cumulative scoring result corresponding to the current radiotherapy plan according to the weight value corresponding to each preset evaluation parameter and the parameter scoring result;

the penalty scoring result determining module is used for determining a penalty scoring result corresponding to the current radiotherapy plan according to the target preset evaluation parameter in the preset evaluation parameters;

and the current score determining module is used for determining the current score corresponding to the current radiation treatment plan according to the accumulated scoring result and the punishment scoring result.

Optionally, the parameter scoring result determining module is specifically configured to:

when the current parameter value is smaller than the first standard parameter value, determining a parameter grading result corresponding to the preset evaluation parameter according to the current parameter value and the first standard parameter value; when the current parameter value is greater than or equal to the first standard parameter value and less than or equal to the second standard parameter value, determining the preset highest score value as a parameter scoring result corresponding to the preset evaluation parameter; and when the current parameter value is larger than the second standard parameter value, determining a parameter grading result corresponding to the preset evaluation parameter according to the current parameter value and the second standard parameter value.

Optionally, the penalty scoring result determining module includes:

the first scoring result determining unit is used for determining a first scoring result corresponding to the current radiotherapy plan according to a maximum parameter scoring result and a minimum parameter scoring result in the parameter scoring results corresponding to the preset evaluation parameters; and/or the presence of a gas in the gas,

the second grading result determining unit is used for determining a second grading result corresponding to the current radiotherapy plan according to a first current parameter value and a second current parameter value which are respectively corresponding to preset evaluation parameters used for representing the dose distribution condition in each preset evaluation parameter;

and the penalty scoring result determining unit is used for determining a penalty scoring result corresponding to the current radiation treatment plan according to the first scoring result and/or the second scoring result.

Optionally, the first scoring result determining unit is specifically configured to:

Optionally, the second scoring result determining unit is specifically configured to:

The radiotherapy plan generating device provided by the embodiment of the invention can execute the radiotherapy plan generating process provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects for executing the radiotherapy plan generating process.

Example four

The present embodiments provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a radiation treatment plan generation process as provided by any of the embodiments of the present invention, which may include the operations of:

if the fact that the current radiotherapy plan does not meet the preset circulation ending condition is detected, feeding back a current score corresponding to the current radiotherapy plan to a preset neural network model so that the preset neural network model updates an output current operation instruction based on the current score, and returning to execute the operation of controlling a radiotherapy plan generating module according to the current operation instruction based on the updated current operation instruction;

and if the current radiation treatment plan is detected to meet the preset cycle end condition, determining the current radiation treatment plan as a target radiation treatment plan corresponding to the patient.

Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. The computer-readable storage medium may be, for example but not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may 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 the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

It will be understood by those skilled in the art that the modules or steps of the invention described above may be implemented by a general purpose computing device, they may be centralized on a single computing device or distributed across a network of computing devices, and optionally they may be implemented by program code executable by a computing device, such that it may be stored in a memory device and executed by a computing device, or it may be separately fabricated into various integrated circuit modules, or it may be fabricated by fabricating a plurality of modules or steps thereof into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.

It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A radiation therapy plan generation apparatus, characterized in that the apparatus comprises:

one or more processors;

a memory for storing one or more programs;

2. The apparatus of claim 1, wherein the imagery information comprises a tomogram and a region of interest delineated in the tomogram.

3. The apparatus of claim 1, wherein the current operating instruction comprises at least one of:

4. The apparatus according to claim 1, further comprising a training process of the preset neural network model before using the preset neural network model, specifically comprising:

acquiring sample image information of a plurality of sample patients and sample prescription dose information corresponding to the sample image information;

inputting the sample image information and the sample prescription dose information into a preset neural network model to obtain a current operation instruction output by the preset neural network model;

determining a current score corresponding to the current radiotherapy plan, and accumulating the current scores after each circulation to obtain a current total score;

when the current cycle number is smaller than the preset cycle number, feeding back the current score to the preset neural network model to adjust network parameters in the preset neural network model, updating the output current operation instruction, returning to execute the operation of controlling the radiotherapy plan generating module according to the current operation instruction based on the updated current operation instruction, and accumulating the current cycle number by 1;

when the current cycle number is equal to the preset cycle number, detecting whether the current total score is greater than or equal to a preset score value;

if so, determining that the training of the preset neural network model is finished;

and if not, resetting the current cycle number and the current total score, feeding the current score back to the preset neural network model to adjust network parameters in the preset neural network model, updating the output current operation instruction, and returning to execute the operation of controlling the radiotherapy plan generation module according to the current operation instruction based on the updated current operation instruction.

5. The apparatus according to any one of claims 1-4, further comprising, before feeding back the current score corresponding to the current radiation therapy plan to the pre-set neural network model:

determining a current parameter value corresponding to each preset evaluation parameter according to the current radiotherapy plan;

determining a parameter grading result corresponding to the preset evaluation parameter according to the current parameter value;

determining an accumulated scoring result corresponding to the current radiotherapy plan according to the weight value corresponding to each preset evaluation parameter and the parameter scoring result;

determining a penalty scoring result corresponding to the current radiotherapy plan according to a target preset evaluation parameter in each preset evaluation parameter;

and determining the current score corresponding to the current radiation treatment plan according to the accumulated scoring result and the punishment scoring result.

6. The apparatus according to claim 5, wherein determining a parameter scoring result corresponding to the preset evaluation parameter according to the current parameter value includes:

when the current parameter value is smaller than a first standard parameter value, determining a parameter grading result corresponding to the preset evaluation parameter according to the current parameter value and the first standard parameter value;

when the current parameter value is greater than or equal to a first standard parameter value and less than or equal to a second standard parameter value, determining a preset highest score value as a parameter scoring result corresponding to the preset evaluation parameter;

and when the current parameter value is larger than a second standard parameter value, determining a parameter grading result corresponding to the preset evaluation parameter according to the current parameter value and the second standard parameter value.

7. The apparatus of claim 5, wherein determining a penalty score corresponding to the current radiation treatment plan according to a target preset evaluation parameter of the preset evaluation parameters comprises:

determining a first scoring result corresponding to the current radiotherapy plan according to a maximum parameter scoring result and a minimum parameter scoring result in the parameter scoring results corresponding to the preset evaluation parameters; and/or the presence of a gas in the gas,

determining a second grading result corresponding to the current radiotherapy plan according to a first current parameter value and a second current parameter value which respectively correspond to a preset evaluation parameter pair used for representing the dose distribution condition in each preset evaluation parameter;

and determining a penalty scoring result corresponding to the current radiation treatment plan according to the first scoring result and/or the second scoring result.

8. The apparatus of claim 7, wherein determining a first scoring result corresponding to the current radiation treatment plan according to a maximum parameter scoring result and a minimum parameter scoring result of the parameter scoring results corresponding to each of the preset evaluation parameters comprises:

determining a first difference value between the maximum parameter scoring result and the minimum parameter scoring result according to the maximum parameter scoring result and the minimum parameter scoring result in the parameter scoring results corresponding to the preset evaluation parameters;

if the first difference is smaller than or equal to a first preset difference, determining a first scoring result corresponding to the current radiotherapy plan by using a preset lowest score value;

and if the first difference is larger than a first preset difference, determining a first grading result corresponding to the current radiotherapy plan according to the first difference.

9. The apparatus of claim 7, wherein determining a second scoring result corresponding to the current radiation treatment plan according to a first current parameter value and a second current parameter value respectively corresponding to a preset evaluation parameter pair for characterizing a dose distribution condition in the preset evaluation parameters comprises:

determining a second difference value between a first current parameter value and a second current parameter value according to a first current parameter value and a second current parameter value which respectively correspond to a preset evaluation parameter pair used for representing the dose distribution condition in each preset evaluation parameter;

if the second difference is smaller than or equal to a second preset difference, determining a second scoring result corresponding to the current radiotherapy plan by using a preset lowest score value;

and if the second difference is larger than a second preset difference, determining a second grading result corresponding to the current radiotherapy plan according to the second difference.

10. A radiation therapy plan generation apparatus, comprising:

11. A computer-readable storage medium, on which a computer program is stored, the program, when executed by a processor, performing operations comprising: