Intelligent quality control method and system for medical linear accelerator

Abstract

The invention discloses an intelligent quality control method and system for a medical linear accelerator, and the method comprises the steps: collecting an image in a quality control operation range of the medical linear accelerator in a treatment room; identifying the quality control equipment in the image and the field arrangement information of the quality control equipment through a quality control equipment identification model, comparing with a radiotherapy plan output by a radiotherapy plan system, and determining whether the field arrangement information of the quality control equipment meets the requirement of the radiotherapy plan; and based on dialogue type interaction mode communication, reminding the user of the next operation specification according to the quality control requirements and the quality control steps, and reminding and confirming the user operation according to the comparison information of the quality control system. Based on artificial intelligence and a large language model, the quality control equipment identification model can intelligently identify the types and placement positions of various quality control equipment in the quality control operation, confirm whether the field arrangement information of the quality control equipment meets the requirements of a radiotherapy plan, and provide a basis for orderly and correct quality control work.

Description

Technical Field

[0001] The present invention relates to the technical field of quality control of medical linear accelerators, and in particular to an intelligent quality control method and system for medical linear accelerators. Background Art

[0002] In the radiotherapy industry, the quality control of medical linear accelerators is extremely important. Planned quality control mainly refers to the quality control of the patient's planned dose completed before each patient's treatment. The specific content includes the 2D / 3D dose of the planned dose, which is mainly achieved with the help of ionization chamber flat panel detectors, solid water and other tools. If the equipment, position and other abnormalities are found during the inspection, they must be corrected in time to ensure that the planned performance indicators are within the prescribed allowable error range. Therefore, once the quality control equipment is found to be used incorrectly or the deviation is large, it must be corrected in time.

[0003] However, at present, the quality control of medical linear accelerators is usually performed by physicists. Human operational errors by physicists are inevitable, and the quality control operations also place high demands on physicists.

[0004] In view of this, the present invention patent is proposed. Summary of the invention

[0005] In order to overcome the errors caused by human factors in the use of quality control equipment in the existing medical linear accelerator quality control method and lower the threshold of accelerator quality control technology requirements, the present invention provides an intelligent quality control method and system for a medical linear accelerator. Specifically, the following technical solutions are adopted:

[0006] An intelligent quality control method for a medical linear accelerator, comprising:

[0007] Collect images within the quality control operation range of the medical linear accelerator in the treatment room;

[0008] Identify the quality control equipment and on-site layout information of the quality control equipment in the image by using a quality control equipment identification model, compare them with the radiotherapy plan output by the radiotherapy planning system, and confirm whether the on-site layout information of the quality control equipment meets the requirements of the radiotherapy plan based on the comparison information;

[0009] Communicate in a conversational interactive manner, and according to quality control requirements and quality control steps, remind users of the next step of operation specifications, and remind and confirm user operations based on comparison information.

[0010] As an optional embodiment of the present invention, in an intelligent quality control method for a medical linear accelerator of the present invention, the quality control equipment recognition model is constructed by a target detection deep learning model, which is trained on a large number of labeled quality control equipment placement pictures. The labeled labels include the type of radiation detector, the amount of solid water that can be superimposed on the flat plate, and the position coordinates determined by the laser light and the position coordinates of the actual placement of the quality control equipment.

[0011] As an optional embodiment of the present invention, in an intelligent quality control method for a medical linear accelerator of the present invention, the identifying of the quality control equipment in the image and the on-site arrangement information of the quality control equipment by the quality control equipment recognition model includes:

[0012] Identify the quality control equipment category information, quality control equipment quantity information and quality control equipment placement information in the image through a quality control equipment recognition model;

[0013] The quality control equipment category information includes the category of radiation detectors and the factory information parameters of each category of radiation detectors;

[0014] The quantity information of the quality control equipment includes the stacking quantity of solid water;

[0015] The quality control equipment placement information includes the quality control equipment position placement information determined by the laser light.

[0016] As an optional embodiment of the present invention, in an intelligent quality control method for a medical linear accelerator of the present invention, the quality control equipment and the on-site arrangement information of the quality control equipment in the image are identified by a quality control equipment identification model, and compared with the radiotherapy plan output by the radiotherapy planning system, and confirming whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan includes:

[0017] The type of the radiation detector in the image is identified by the quality control equipment identification model to determine whether the type is consistent with the type of the radiation detector in the quality control information of the radiation therapy plan output by the radiation therapy planning system. If not, the radiation therapy plan requirement is not met.

[0018] Optionally, the quality control equipment recognition model identifies that the type of the radiation detector in the image is cylindrical / flat-plate, and the type of the radiation detector in the quality control information of the radiation therapy plan output by the radiation therapy planning system is flat-plate / cylindrical, which is inconsistent.

[0019] As an optional embodiment of the present invention, in an intelligent quality control method for a medical linear accelerator of the present invention, the quality control equipment and the on-site arrangement information of the quality control equipment in the image are identified by a quality control equipment identification model, and compared with the radiotherapy plan output by the radiotherapy planning system, and confirming whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan includes:

[0020] The quality control equipment recognition model is used to identify whether the amount of solid water stacking in the image is consistent with the amount of solid water stacking in the quality control information of the radiotherapy plan output by the radiotherapy planning system. If not, it does not meet the requirements of the radiotherapy plan.

[0021] As an optional embodiment of the present invention, in an intelligent quality control method for a medical linear accelerator of the present invention, the quality control equipment and the on-site arrangement information of the quality control equipment in the image are identified by a quality control equipment identification model, and compared with the radiotherapy plan output by the radiotherapy planning system, and confirming whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan includes:

[0022] Identify, through a quality control equipment recognition model, whether there is quality control equipment position placement information determined by a laser light in the image;

[0023] If the quality control equipment recognition model identifies that there is placement information of the quality control equipment position determined by the laser light in the image, further determine whether the placement coordinates of the quality control equipment determined by the laser light in the image identified by the quality control equipment recognition model are consistent with the current actual placement coordinates of the quality control equipment; if they are inconsistent, the radiotherapy plan requirement is not met; if they are consistent, the radiotherapy plan requirement is met;

[0024] If the quality control equipment recognition model identifies that the image does not contain the quality control equipment position placement information determined by the laser light, then the radiotherapy plan requirements are not met.

[0025] As an optional embodiment of the present invention, an intelligent quality control method for a medical linear accelerator of the present invention includes generating a quality control process suggestion based on a comparison result of comparing the quality control equipment in the image and the on-site arrangement information of the quality control equipment with the radiotherapy plan output by the radiotherapy planning system through a quality control equipment recognition model:

[0026] If the category of the radiation detector in the image identified by the quality control equipment recognition model is inconsistent with the category of the radiation detector in the quality control information of the radiation therapy plan output by the radiation therapy planning system, it is recommended to confirm the category of the current radiation detector and replace it with a radiation detector of the same category as that in the quality control information of the radiation therapy plan.

[0027] As an optional embodiment of the present invention, an intelligent quality control method for a medical linear accelerator of the present invention includes generating a quality control process suggestion based on a comparison result of comparing the quality control equipment in the image and the on-site arrangement information of the quality control equipment with the radiotherapy plan output by the radiotherapy planning system through a quality control equipment recognition model:

[0028] If the amount of solid water stacking in the image identified by the quality control equipment recognition model is inconsistent with the amount of solid water stacking in the quality control information of the radiotherapy plan output by the radiotherapy planning system, it is recommended to prepare and arrange according to the amount of solid water stacking required in the quality control information of the radiotherapy plan.

[0029] As an optional embodiment of the present invention, an intelligent quality control method for a medical linear accelerator of the present invention includes generating a quality control process suggestion based on a comparison result of comparing the quality control equipment in the image and the on-site arrangement information of the quality control equipment with the radiotherapy plan output by the radiotherapy planning system through a quality control equipment recognition model:

[0030] If the quality control equipment recognition model recognizes that there is no placement information of the quality control equipment position determined by the laser light in the image, and the quality control information of the radiotherapy plan output by the radiotherapy planning system requires the laser light to determine the position coordinates, it is recommended to check and confirm whether the positioning of the laser light is normal and whether the recognition ability of the quality control equipment recognition model is normal;

[0031] If the quality control equipment recognition model identifies that the image contains placement information of the quality control equipment position determined by the laser light, and it is further determined that the placement coordinates of the quality control equipment determined by the laser light in the image identified by the quality control equipment recognition model are inconsistent with the current actual placement coordinates of the quality control equipment, it is recommended to manually verify whether the current actual placement coordinates of the quality control equipment meet the quality control information requirements of the radiotherapy plan.

[0032] The present invention also provides an intelligent quality control system using the intelligent quality control method of the medical linear accelerator, comprising:

[0033] Visual image acquisition module, which collects images within the quality control operation range of the medical linear accelerator in the treatment room;

[0034] an intelligent quality control module, which identifies the quality control equipment and the on-site arrangement information of the quality control equipment in the image through a quality control equipment identification model, compares the information with the radiotherapy plan output by the radiotherapy planning system, and confirms whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan;

[0035] The large language model interaction module communicates based on a conversational interaction method. According to the quality control requirements, it reminds users of the next operation specifications according to the quality control steps, and reminds and confirms user operations based on the comparison information of the intelligent quality control system.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] The intelligent quality control method of a medical linear accelerator of the present invention is based on artificial intelligence and a large language model. The quality control equipment recognition model intelligently recognizes various quality control equipment and their placement in the quality control operation, confirms whether the on-site layout information of the quality control equipment meets the requirements of the radiotherapy plan, provides a basis for the orderly and correct conduct of quality control work, improves quality control efficiency, reduces physicist operation errors, and even realizes full-process quality control automation.

[0038] The intelligent quality control method for a medical linear accelerator of the present invention is not limited to a certain type of medical linear accelerator or a certain type of quality control equipment, has low requirements on operators, has sufficient reminders and confirmation information for quality control operations in the control room, and can correct errors in a timely manner according to on-site conditions. The intelligent quality control method for a medical linear accelerator of this embodiment requires less equipment, can effectively reduce quality control requirements, and allows technicians and other personnel to perform patient quality control.

[0039] The intelligent quality control method for a medical linear accelerator of the present invention innovatively uses a deep learning model to identify the placement of accelerator quality control equipment obtained through a visual system, and communicates with operators in a conversational interactive manner based on a large language model, making the quality control process fully automatic and intelligent, reducing the risk of erroneous operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flow chart of an intelligent quality control method for a medical linear accelerator according to an embodiment of the present invention;

[0041] Figure 2 A modular diagram of an intelligent quality control system for a medical linear accelerator according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] To make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be described clearly and completely in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them.

[0043] Therefore, the following detailed description of the embodiments of the present invention is not intended to limit the scope of the invention claimed for protection, but merely represents some embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0044] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features and technical solutions in the embodiments may be combined with each other.

[0045] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0046] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper", "lower", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the invention product is usually placed when in use, or the orientation or positional relationship commonly understood by those skilled in the art. Such terms are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0047] See also Figure 1 As shown, an intelligent quality control method for a medical linear accelerator in this embodiment includes:

[0048] Collect images within the quality control operation range of the medical linear accelerator in the treatment room;

[0049] Identify the quality control equipment and on-site layout information of the quality control equipment in the image by using a quality control equipment identification model, compare them with the radiotherapy plan output by the radiotherapy planning system, and confirm whether the on-site layout information of the quality control equipment meets the requirements of the radiotherapy plan based on the comparison information;

[0050] Communicate in a conversational interactive manner, and according to quality control requirements and quality control steps, remind users of the next step of operation specifications, and remind and confirm user operations based on comparison information.

[0051] An intelligent quality control method for a medical linear accelerator in this embodiment is based on artificial intelligence and a large language model. The quality control equipment recognition model intelligently recognizes various quality control equipment and their placement in the quality control operation, confirms whether the on-site layout information of the quality control equipment meets the requirements of the radiotherapy plan, provides a basis for the orderly and correct conduct of quality control work, improves quality control efficiency, reduces physicist operation errors, and even realizes full-process quality control automation.

[0052] The intelligent quality control method for a medical linear accelerator of this embodiment is not limited to a certain type of medical linear accelerator or a certain type of quality control equipment, has low requirements for operators, has sufficient reminders and confirmation information for quality control operations in the control room, and can correct errors in a timely manner according to on-site conditions. The intelligent quality control method for a medical linear accelerator of this embodiment requires less equipment, can effectively reduce quality control requirements, and allows technicians and other personnel to perform patient quality control.

[0053] As an optional implementation of this embodiment, an intelligent quality control method for a medical linear accelerator in this embodiment is based on a multimodal conversational large language model interaction system with the Wenxin large language model as the basic model. It communicates based on a conversational interactive method, and according to the quality control requirements, it reminds the user of the next operation specification according to the established quality control steps, and reminds and confirms the user's operation based on the comparison information of the intelligent quality control system.

[0054] In order to realize the conversational interactive communication in this embodiment, the Wenxin language model is used as the basic model. By mounting the professional terminology of radiotherapy, the user is interacted with in a conversational manner. According to the quality control requirements, the user is reminded of the next operation specification according to the established quality control steps. The user operation is reminded and confirmed based on the comparison information of the intelligent automatic control system. The reminder information includes the type, quantity, and placement of the quality control equipment. The recognition information of the intelligent quality control system and the output plan information of the planning system are input into the large language model in the form of language description.

[0055] Therefore, the intelligent quality control method for a medical linear accelerator in this embodiment innovatively uses a deep learning model to identify the placement of accelerator quality control equipment obtained through a visual system, and interacts with operators based on a large language model, making the quality control process fully automatic and intelligent, reducing the risk of erroneous operations.

[0056] Furthermore, in an intelligent quality control method for a medical linear accelerator of this embodiment, the identifying of the quality control equipment in the image and the on-site arrangement information of the quality control equipment by using the quality control equipment recognition model includes:

[0057] Identify the quality control equipment category information, quality control equipment quantity information and quality control equipment placement information in the image through a quality control equipment recognition model;

[0058] The quality control equipment category information includes the category of radiation detectors (specifically including flat-plate and cylindrical), and the factory information parameters of each category of radiation detectors (specifically including brand);

[0059] The quantity information of the quality control equipment includes the stacking quantity of solid water;

[0060] The quality control equipment placement information includes the quality control equipment position placement information determined by the laser light.

[0061] The intelligent quality control method of a medical linear accelerator of the present embodiment identifies the type and quantity of quality control equipment information based on images within the quality control operation range of the medical linear accelerator in the treatment room collected by the pre-established quality control equipment recognition model; confirms the approximate placement position (within 3mm error) and gives placement requirements based on information such as laser lights; compares the above information with the radiotherapy plan output by the radiotherapy planning system (CT can identify whether it is solid water and confirm whether it is of what type), confirms whether it is the quality control equipment, quantity, placement, etc. required for the quality control operation, and specifically includes:

[0062] In the intelligent quality control method for a medical linear accelerator of the present embodiment, the quality control equipment and the on-site arrangement information of the quality control equipment in the image are identified by the quality control equipment identification model, and compared with the radiotherapy plan output by the radiotherapy planning system to confirm whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan, including:

[0063] The type of the radiation detector in the image is identified by the quality control equipment identification model to determine whether the type is consistent with the type of the radiation detector in the quality control information of the radiation therapy plan output by the radiation therapy planning system. If not, the radiation therapy plan requirement is not met.

[0064] Optionally, the quality control equipment recognition model identifies that the type of the radiation detector in the image is cylindrical / flat-plate, and the type of the radiation detector in the quality control information of the radiation therapy plan output by the radiation therapy planning system is flat-plate / cylindrical, which is inconsistent.

[0065] In the intelligent quality control method for a medical linear accelerator of the present embodiment, the quality control equipment and the on-site arrangement information of the quality control equipment in the image are identified by the quality control equipment identification model, and compared with the radiotherapy plan output by the radiotherapy planning system to confirm whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan, including:

[0066] The quality control equipment recognition model is used to identify whether the amount of solid water stacking in the image is consistent with the amount of solid water stacking in the quality control information of the radiotherapy plan output by the radiotherapy planning system. If not, it does not meet the requirements of the radiotherapy plan.

[0067] In the intelligent quality control method for a medical linear accelerator of the present embodiment, the quality control equipment and the on-site arrangement information of the quality control equipment in the image are identified by the quality control equipment identification model, and compared with the radiotherapy plan output by the radiotherapy planning system to confirm whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan, including:

[0068] Identify, through a quality control equipment recognition model, whether there is quality control equipment position placement information determined by a laser light in the image;

[0069] If the quality control equipment recognition model identifies that there is placement information of the quality control equipment position determined by the laser light in the image, further determine whether the placement coordinates of the quality control equipment determined by the laser light in the image identified by the quality control equipment recognition model are consistent with the current actual placement coordinates of the quality control equipment; if they are inconsistent, the radiotherapy plan requirement is not met; if they are consistent, the radiotherapy plan requirement is met;

[0070] If the quality control equipment recognition model identifies that the image does not contain the quality control equipment position placement information determined by the laser light, then the radiotherapy plan requirements are not met.

[0071] As a specific example of this embodiment, for a quality control operation, the quality control equipment recognition model identifies whether it is a 2D flat plate or 3D (divided into flat plate + phantom, or cylinder, or cross); if it is a 2D flat plate, determine whether it is a dedicated phantom or solid water; solid water needs to determine the stacking quantity. The quality control equipment recognition model determines the specific model of the quality control equipment based on the image. The quality control equipment recognition model determines the parameters, size, height, etc. based on the number of types determined, and makes a judgment with the plan information in TPS (including CT) to ensure consistency.

[0072] Furthermore, an intelligent quality control method for a medical linear accelerator in this embodiment includes generating a quality control process suggestion based on a comparison result of comparing the quality control equipment in the image and the on-site arrangement information of the quality control equipment with the radiotherapy plan output by the radiotherapy planning system through a quality control equipment recognition model:

[0073] If the category of the radiation detector in the image identified by the quality control equipment recognition model is inconsistent with the category of the radiation detector in the quality control information of the radiation therapy plan output by the radiation therapy planning system, it is recommended to confirm the category of the current radiation detector and replace it with a radiation detector of the same category as that in the quality control information of the radiation therapy plan.

[0074] The intelligent quality control method for a medical linear accelerator of this embodiment includes generating a quality control process suggestion based on a comparison result of comparing the quality control equipment in the image and the on-site arrangement information of the quality control equipment with the radiotherapy plan output by the radiotherapy planning system through a quality control equipment recognition model:

[0075] If the amount of solid water stacking in the image identified by the quality control equipment recognition model is inconsistent with the amount of solid water stacking in the quality control information of the radiotherapy plan output by the radiotherapy planning system, it is recommended to prepare and arrange according to the amount of solid water stacking required in the quality control information of the radiotherapy plan.

[0076] The intelligent quality control method for a medical linear accelerator of this embodiment includes generating a quality control process suggestion based on a comparison result of comparing the quality control equipment in the image and the on-site arrangement information of the quality control equipment with the radiotherapy plan output by the radiotherapy planning system through a quality control equipment recognition model:

[0077] If the quality control equipment recognition model recognizes that there is no placement information of the quality control equipment position determined by the laser light in the image, and the quality control information of the radiotherapy plan output by the radiotherapy planning system requires the laser light to determine the position coordinates, it is recommended to check and confirm whether the positioning of the laser light is normal and whether the recognition ability of the quality control equipment recognition model is normal;

[0078] If the quality control equipment recognition model identifies that the image contains placement information of the quality control equipment position determined by the laser light, and it is further determined that the placement coordinates of the quality control equipment determined by the laser light in the image identified by the quality control equipment recognition model are inconsistent with the current actual placement coordinates of the quality control equipment, it is recommended to manually verify whether the current actual placement coordinates of the quality control equipment meet the quality control information requirements of the radiotherapy plan.

[0079] As a specific example of this embodiment:

[0080] The first case: the quality control information identified by the quality control equipment recognition model: the quality control equipment is cylindrical, and the coordinates determined by the laser light are consistent with the equipment placement coordinates; in the quality control information of the radiotherapy plan: the quality control equipment is flat, with 3 layers of solid water superimposed, and the coordinates are determined by the laser light; it can be seen that the quality control equipment identified by the quality control equipment recognition model is different from the quality control equipment in the quality control information of the radiotherapy plan.

[0081] In the first case, we compared the QC equipment information identified by the QC equipment identification model with the QC information of the radiotherapy plan and found obvious differences. The following is an analysis of these differences and the subsequent QC process recommendations:

[0082] Variance Analysis

[0083] 1. Shape of quality control equipment: The quality control equipment identified by the quality control equipment recognition model is cylindrical, while the quality control equipment required by the planning system is flat. The difference in shape may affect the measurement and evaluation of dose distribution.

[0084] 2. Number of solid water layers: The planning system requires the superposition of 3 layers of solid water, but the information recognized by the quality control equipment recognition model does not mention the use or layer teaching of solid water. The use of solid water is crucial for simulating human tissue and accurately evaluating dose distribution.

[0085] 3. Laser light to determine coordinates: Although both mentioned laser light to determine coordinates, this information itself does not directly reflect the problem, because laser light is usually used for positioning, and its accuracy has an impact on the accuracy of treatment, but it is not the main difference in this case.

[0086] Quality control process suggestions

[0087] 1. Confirm the device type:

[0088] Immediately confirm the correct type of quality control equipment with the planning system provider or production management team. If the identification system does require the use of a flat-shaped device instead of a cylindrical device, the equipment will need to be replaced.

[0089] Check for any misunderstandings or miscommunication leading to inconsistent information.

[0090] 2. Prepare solid water:

[0091] If the identification system requires the stacking of 3 layers of solid water, it is necessary to prepare a corresponding amount of solid water and ensure that its quality meets the standards. The solid water should be placed on parallel poles according to the identification requirements and ensure that it is stable and without deviation.

[0092] 3. Repositioning and repositioning:

[0093] Use the laser light to reposition the parallel plate and solid water to ensure that they are consistent with the coordinates in the planning system.

[0094] Adjust the position and angle of the treatment bed so that the equipment is accurately aimed at the treatment area.

[0095] 4. Irradiation and dose collection:

[0096] After all preparations are completed, the radiotherapy equipment is started to irradiate the parallel plates and solid water.

[0097] Dose data are collected using dose measurement equipment and compared with the expected dose in the planning system.

[0098] The second situation: the quality control information identified by the quality control equipment recognition model: the quality control equipment is in the shape of a flat plate, with two layers of solid water superimposed on it, and the coordinates determined by the laser light are not recognized; in the quality control information of the radiotherapy plan: the quality control equipment is in the shape of a flat plate, with two layers of solid water superimposed on it, and the coordinates determined by the laser light; it can be seen that the quality control equipment recognition model recognizes that the laser light is not turned on.

[0099] In the second case, we compared the quality control equipment information identified by the quality control equipment identification model with the quality control information of the radiotherapy plan and found that the two were consistent in the shape of the quality control equipment and the number of solid water layers, but the information identified by the quality control equipment identification model lacked the part for determining the coordinates of the laser light. In view of this difference, we can take the following quality control process suggestions:

[0100] Quality control process suggestions

[0101] 1. Confirm the laser light position:

[0102] First, it is necessary to confirm that the laser light has been properly turned on and is in working condition. If the laser light is not turned on or is faulty, it should be repaired or replaced immediately. Use the laser light to reposition the position of the quality control equipment (flat plate and solid water) to ensure that it is consistent with the coordinates in the planning system, which usually involves adjusting the position and angle of the treatment bed so that the laser line is accurately aligned with the marked point.

[0103] 2. Check the camera recognition capability:

[0104] Evaluate whether the camera system has the ability to identify the laser light positioning coordinates. If the camera system itself does not support this function, or the function fails due to a malfunction, you need to consider upgrading or replacing the camera system.

[0105] If the camera system supports this feature but fails to recognize it correctly, check the camera settings, lighting conditions, image processing algorithms and other factors to ensure that they can accurately capture and recognize the laser light positioning coordinates.

[0106] 3. Manually verify the coordinates:

[0107] In the case where the camera cannot identify the laser light to determine the coordinates, manual verification can be used. Use other measuring tools (such as rulers, angle rulers, etc.) to measure the relative position relationship between the quality control equipment and the treatment bed to ensure that they match the coordinates in the planning system. Although this method is more cumbersome and may have certain errors, it is an effective alternative when the camera system cannot work properly.

[0108] The third situation: the quality control information identified by the quality control equipment recognition model: the quality control equipment is cylindrical, and the coordinates determined by the laser light are inconsistent with the equipment placement coordinates; in the quality control information of the radiotherapy plan: the quality control equipment is cylindrical, and the coordinates are determined by the laser light; then the position determined by the laser light is different from the equipment placement position.

[0109] In the third case, we are faced with the problem that the quality control equipment information identified by the quality control equipment recognition model is consistent with the quality control information of the radiotherapy plan in terms of equipment shape, but there is a discrepancy between the laser light determination coordinates and the equipment placement coordinates. In view of this difference, we can take the following quality control process suggestions:

[0110] Quality control process suggestions

[0111] 1. Confirm the consistency of equipment shape:

[0112] First, the shape of the quality control equipment identified by the quality control equipment identification model (cylindrical) is consistent with the shape required by the planning system. Since this has been confirmed, we can proceed to the next step.

[0113] 2. Check the laser light coordinates:

[0114] Carefully compare the coordinates of the laser light identified by the quality control equipment recognition model with the coordinates required in the planning system. Use measuring tools (such as rulers, angle rulers, etc.) or the calibration tools that come with the treatment device to verify the accuracy of the coordinates.

[0115] If the coordinates are inconsistent, you need to determine which link has a problem. It may be that the laser light itself is not positioned accurately, or the device is offset when it is placed.

[0116] 3. Adjust the device position:

[0117] If the device placement is inconsistent with the coordinates determined by the laser, the device position needs to be adjusted to match the coordinates in the planning system. This usually involves moving the treatment bed or adjusting the position of the quality control equipment.

[0118] During the adjustment process, the laser light should be used as a reference to ensure that the equipment is accurately aligned.

[0119] 4. Re-verify coordinates:

[0120] After adjusting the position of the device, use the laser light again to determine the coordinates and verify that they are consistent with the coordinates in the planning system.

[0121] If inconsistencies persist, further inspection of the laser light system or treatment device calibration status is required.

[0122] As an optional implementation of this embodiment, in an intelligent quality control method for a medical linear accelerator of this embodiment, the quality control equipment recognition model is constructed by a target detection deep learning model, which is trained on a large number of labeled quality control equipment placement pictures. The labeled labels include the type of radiation detector, the amount of solid water that can be superimposed on the flat plate, and the position coordinates determined by the laser light and the position coordinates of the actual placement of the quality control equipment.

[0123] See also Figure 2 As shown, this embodiment also provides an intelligent quality control system using the intelligent quality control method of the medical linear accelerator, including:

[0124] Visual image acquisition module, which collects images within the quality control operation range of the medical linear accelerator in the treatment room;

[0125] The intelligent quality control module identifies the quality control equipment and the on-site layout information of the quality control equipment in the image through a quality control equipment recognition model, compares the information with the radiotherapy plan output by the radiotherapy planning system, and confirms whether the on-site layout information of the quality control equipment meets the requirements of the radiotherapy plan.

[0126] An intelligent quality control system for a medical linear accelerator in this embodiment is based on artificial intelligence and a large language model. The intelligent quality control module intelligently identifies various quality control equipment and their placement in the quality control operation through a quality control equipment recognition model, confirms whether the on-site layout information of the quality control equipment meets the requirements of the radiotherapy plan, provides a basis for the orderly and correct conduct of quality control work, improves quality control efficiency, reduces physicists' operating errors, and even realizes full-process quality control automation.

[0127] The intelligent quality control system for a medical linear accelerator of the present embodiment is not limited to a certain type of medical linear accelerator or a certain type of quality control equipment, has low requirements for operators, has sufficient reminders and confirmation information for quality control operations in the control room, and can correct errors in a timely manner according to on-site conditions. The intelligent quality control system for a medical linear accelerator of the present embodiment requires less equipment, can effectively reduce quality control requirements, and allows technicians and other personnel to perform patient quality control.

[0128] Furthermore, this embodiment is an intelligent quality control system using the medical linear accelerator, including: a large language model interaction module, which is based on the Wenxin large language model as a basic model, communicates based on a conversational interactive method, and reminds the user of the next operation specification according to the quality control requirements and the established quality control steps, and reminds and confirms the user's operation according to the comparison information of the intelligent quality control system.

[0129] Furthermore, the large language model interaction module is based on the Wenxin large language model as the basic model. It interacts with users in a dialogic manner by mounting radiotherapy professional terminology. According to the quality control requirements, it reminds users of the next operation specifications according to the established quality control steps, and reminds and confirms user operations based on the comparison information of the intelligent automatic control system. The reminder information includes the type, quantity, and placement of quality control equipment. The intelligent quality control system recognition information and the planning system output plan information are input into the large language model in the form of language description.

[0130] The above embodiments are only used to illustrate the present invention and are not intended to limit the technical solutions described in the present invention. Although the present invention has been described in detail with reference to the above embodiments, the present invention is not limited to the above specific implementation methods. Therefore, any modification or equivalent replacement of the present invention; and all technical solutions and improvements thereof that do not depart from the spirit and scope of the invention are included in the scope of the claims of the present invention.

Claims

1. An intelligent quality control method for a medical linear accelerator, characterized in that: include: Collect images within the quality control operation range of the medical linear accelerator in the treatment room; Identify the quality control equipment and on-site layout information of the quality control equipment in the image by using a quality control equipment identification model, compare them with the radiotherapy plan output by the radiotherapy planning system, and confirm whether the on-site layout information of the quality control equipment meets the requirements of the radiotherapy plan based on the comparison information; Communicate in a conversational interactive manner, and according to quality control requirements and quality control steps, remind users of the next step of operation specifications, and remind and confirm user operations based on the comparison information of the quality control system.

2. The intelligent quality control method for a medical linear accelerator according to claim 1, characterized in that: The quality control equipment recognition model is constructed by a target detection deep learning model, which is trained on a large number of labeled quality control equipment placement pictures. The labeled labels include the type of radiation detector, the amount of solid water that can be superimposed on the flat plate, and the position coordinates determined by the laser light and the position coordinates of the actual placement of the quality control equipment.

3. The intelligent quality control method for a medical linear accelerator according to claim 1, characterized in that: The identifying of the quality control equipment in the image and the on-site arrangement information of the quality control equipment by using the quality control equipment identification model includes: Identify the quality control equipment category information, quality control equipment quantity information and quality control equipment placement information in the image through a quality control equipment recognition model; The quality control equipment category information includes the category of radiation detectors and the factory information parameters of each category of radiation detectors; The quantity information of the quality control equipment includes the stacking quantity of solid water; The quality control equipment placement information includes the quality control equipment position placement information determined by the laser light.

4. The intelligent quality control method for a medical linear accelerator according to claim 3, characterized in that: The identifying of the quality control equipment and the on-site arrangement information of the quality control equipment in the image by the quality control equipment identification model, and comparing the information with the radiotherapy plan output by the radiotherapy planning system to confirm whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan includes: Identify the type of the radiation detector in the image through the quality control equipment recognition model, and determine whether the type is consistent with the type of the radiation detector in the quality control information of the radiation therapy plan output by the radiation therapy planning system. If not, the radiation therapy plan requirement is not met. Optionally, the quality control equipment recognition model identifies that the type of the radiation detector in the image is cylindrical / flat-plate, and the type of the radiation detector in the quality control information of the radiation therapy plan output by the radiation therapy planning system is flat-plate / cylindrical, which is inconsistent.

5. The intelligent quality control method for a medical linear accelerator according to claim 3 or 4, characterized in that: The identifying of the quality control equipment and the on-site arrangement information of the quality control equipment in the image by the quality control equipment identification model, and comparing the information with the radiotherapy plan output by the radiotherapy planning system to confirm whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan includes: The quality control equipment recognition model is used to identify whether the amount of solid water stacking in the image is consistent with the amount of solid water stacking in the quality control information of the radiotherapy plan output by the radiotherapy planning system. If not, it does not meet the requirements of the radiotherapy plan.

6. An intelligent quality control method for a medical linear accelerator according to any one of claims 3 to 5, characterized in that: The identifying of the quality control equipment and the on-site arrangement information of the quality control equipment in the image by the quality control equipment identification model, and comparing the information with the radiotherapy plan output by the radiotherapy planning system to confirm whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan includes: Identify, through a quality control equipment recognition model, whether there is quality control equipment position placement information determined by a laser light in the image; If the quality control equipment recognition model identifies that there is placement information of the quality control equipment position determined by the laser light in the image, further determine whether the placement coordinates of the quality control equipment determined by the laser light in the image identified by the quality control equipment recognition model are consistent with the current actual placement coordinates of the quality control equipment; if they are inconsistent, the radiotherapy plan requirement is not met; if they are consistent, the radiotherapy plan requirement is met; If the quality control equipment recognition model identifies that the image does not contain the quality control equipment position placement information determined by the laser light, then the radiotherapy plan requirements are not met.

7. The intelligent quality control method for a medical linear accelerator according to claim 6, characterized in that: The method includes generating a quality control process suggestion based on a comparison result of comparing the quality control equipment in the image and the on-site arrangement information of the quality control equipment with the radiotherapy plan output by the radiotherapy planning system through the quality control equipment recognition model: If the category of the radiation detector in the image identified by the quality control equipment recognition model is inconsistent with the category of the radiation detector in the quality control information of the radiation therapy plan output by the radiation therapy planning system, it is recommended to confirm the category of the current radiation detector and replace it with a radiation detector of the same category as that in the quality control information of the radiation therapy plan.

8. The intelligent quality control method for a medical linear accelerator according to claim 6, characterized in that: The method includes generating a quality control process suggestion based on a comparison result of comparing the quality control equipment in the image and the on-site arrangement information of the quality control equipment with the radiotherapy plan output by the radiotherapy planning system through the quality control equipment recognition model: If the amount of solid water stacking in the image identified by the quality control equipment recognition model is inconsistent with the amount of solid water stacking in the quality control information of the radiotherapy plan output by the radiotherapy planning system, it is recommended to prepare and arrange according to the amount of solid water stacking required in the quality control information of the radiotherapy plan.

9. The intelligent quality control method for a medical linear accelerator according to claim 6, characterized in that: The method includes generating a quality control process suggestion based on a comparison result of comparing the quality control equipment in the image and the on-site arrangement information of the quality control equipment with the radiotherapy plan output by the radiotherapy planning system through the quality control equipment recognition model: If the quality control equipment recognition model recognizes that there is no placement information of the quality control equipment position determined by the laser light in the image, and the quality control information of the radiotherapy plan output by the radiotherapy planning system requires the laser light to determine the position coordinates, it is recommended to check and confirm whether the positioning of the laser light is normal and whether the recognition ability of the quality control equipment recognition model is normal; If the quality control equipment recognition model identifies that the image contains placement information of the quality control equipment position determined by the laser light, and it is further determined that the placement coordinates of the quality control equipment determined by the laser light in the image identified by the quality control equipment recognition model are inconsistent with the current actual placement coordinates of the quality control equipment, it is recommended to manually verify whether the current actual placement coordinates of the quality control equipment meet the quality control information requirements of the radiotherapy plan.

10. An intelligent quality control system using the intelligent quality control method for a medical linear accelerator according to any one of claims 1 to 9, characterized in that: include: Visual image acquisition module, which collects images within the quality control operation range of the medical linear accelerator in the treatment room; an intelligent quality control module, which identifies the quality control equipment and the on-site arrangement information of the quality control equipment in the image through a quality control equipment identification model, compares the information with the radiotherapy plan output by the radiotherapy planning system, and confirms whether the on-site arrangement information of the quality control equipment meets the requirements of the radiotherapy plan; The large language model interaction module communicates based on a conversational interaction method. According to the quality control requirements and quality control steps, it reminds users of the next operation specifications, and reminds and confirms user operations based on the comparison information of the quality control system of the intelligent quality control module.

Images