WO2024032570A1

WO2024032570A1 - Interventional planning system, method and apparatus, and a storage medium

Info

Publication number: WO2024032570A1
Application number: PCT/CN2023/111597
Authority: WO
Inventors: 汪国强; 张璟; 方伟
Original assignee: 武汉联影智融医疗科技有限公司
Priority date: 2022-08-08
Filing date: 2023-08-07
Publication date: 2024-02-15

Abstract

Embodiments of the present description provide an interventional planning system, method and apparatus, and a storage medium. The system comprises a control module, the control module being used for: acquiring patient data of a target patient; and determining a target parameter on the basis of the patient data.

Description

Intervention planning system, method, device and storage medium

cross reference

This application claims the priority of the Chinese application with application number 202210946070.0 submitted on August 8, 2022, the priority of the Chinese application with application number 202211695278.6 submitted on December 28, 2022, and the priority of the Chinese application with application number 202211695278.6 submitted on December 28, 2022 Priority to the Chinese application with application number 202211695301.1 filed on 2020-01-28, the entire content of which is incorporated herein by reference.

Technical field

This specification relates to the technical field of medical devices, and in particular to an intervention planning system, method, device and storage medium.

Background technique

Percutaneous interventional surgery is a commonly used clinical treatment or detection method. Percutaneous interventional surgery can be used for puncture biopsy, ablation therapy, and TPS radioactive seed implantation. Currently, before an interventional needle enters human tissue, doctors need to plan appropriate target parameters for interventional surgery based on preoperative images and experience. Planning target parameters is difficult. Especially for ablation treatment of lesions with a diameter of 3-5cm, it is necessary to consider the method of combined multiple needle ablation, which brings greater challenges to preoperative planning and also increases the difficulty of the doctor's work.

Therefore, it is expected to propose an interventional planning system, method, device and storage medium that can automatically plan reasonable target parameters to improve doctors' surgical efficiency and reduce work intensity.

Contents of the invention

One or more embodiments of this specification provide an intervention planning system. The interventional planning system includes a control module, which is used to: obtain patient data of a target patient; and determine target parameters based on the patient data.

In some embodiments, the target parameter includes a target puncture path, and determining the target puncture path includes: determining the structural characteristics of the target patient based on the patient data; determining the target target based on the structural characteristics; and determining candidate puncture based on the structural characteristics and the target target. Path set; determine the target puncture path from the candidate puncture path set.

In some embodiments, determining the structural features of the target patient based on the patient data includes: determining a three-dimensional medical image of the target patient based on the patient data; and determining the structural features based on the three-dimensional medical image.

In some embodiments, determining the set of candidate puncture paths based on the structural features and the target target point includes: determining the target target point as the perspective projection center; using the perspective projection center as the radiation source to emit multiple rays outward, based on the structural features, Calculate the judgment value of the clinical strong constraint conditions of the puncture path corresponding to each ray; determine the puncture path whose judgment value of the clinical strong constraint condition satisfies the preset conditions as the candidate puncture path; determine the set of candidate puncture paths as the candidate Collection of puncture paths.

In some embodiments, strong clinical constraints include one or more of the following conditions: the puncture path does not contact and does not penetrate puncture risk structures, the length of the puncture path is less than the preset needle length threshold, and the angle between the puncture path and the target tissue is not The distance between the puncture path and the structure to be punctured is less than the preset angle threshold and the length of the puncture path passing through the structure to be punctured is greater than the preset distance threshold.

In some embodiments, the set of candidate puncture paths includes at least one candidate puncture path, and determining the target puncture path from the set of candidate puncture paths includes: calculating path association information of at least one candidate puncture path; based on the path association of the at least one candidate puncture path Information, the target puncture path is determined through a preset search algorithm.

In some embodiments, the path association information includes one or more of the following information: the distance between the candidate puncture path and the puncture risk structure, the length of the candidate puncture path, and the angle between the candidate puncture path and the target tissue.

In some embodiments, the preset search algorithm includes a quantum annealing algorithm. Based on path association information of at least one candidate puncture path, determining the target puncture path through the preset search algorithm includes: constructing a first function term according to the path association information; based on the first The function term determines the target puncture path through a preset search algorithm.

In some embodiments, the target parameters include a target puncture path, a target stop point position, and a target ablation sphere parameter. Based on the patient data, determining the target parameters includes: determining at least one set of planning parameters based on the patient data; based on at least one set of planning parameters, Determine the target parameters; wherein, determining the target parameters includes: generating an individual set based on the individual generator. The individual set includes multiple individuals, each individual corresponds to a set of planning parameters; performing at least one first iteration update on the individual set until the An iteration completion condition is met; based on the updated individual set, at least one set of intermediate parameters is determined; and based on at least one set of intermediate parameters, the target parameters are determined.

In some embodiments, determining at least one set of planning parameters based on the patient data includes: performing three-dimensional reconstruction on the patient data to obtain a three-dimensional medical image, where the patient data includes CT or MR data of the patient; and determining at least one set of planning parameters based on the three-dimensional medical image. planning parameters.

In some embodiments, determining at least one set of planning parameters based on the three-dimensional medical image includes: performing a preprocessing operation on the three-dimensional medical image. The preprocessing operation includes one or more of region of interest cropping, data point downsampling, and blood vessel coarse and subdivided grading. ; and based on preprocessing The results obtained from the operation determine at least one set of planning parameters.

In some embodiments, each of the at least one first iteration includes: generating at least one new individual based on the individual generator; and adding at least one new individual to the individual set.

In some embodiments, each iteration in at least one first iteration includes: filtering each individual in the individual set based on the individual filter, updating the individual set, and the filtering includes performing a selection operation on the individuals in the individual set. ; Wherein, the selection operation includes: calculating the first evaluation value and the first constraint determination value of each individual in the individual set; and selecting individuals based on the first evaluation value and the first constraint determination value, and determining the updated individual set.

In some embodiments, determining at least one set of intermediate parameters based on the updated set of individuals includes: determining a Pareto front solution based on the updated set of individuals, and using the Pareto front solution as at least one set of intermediate parameters.

In some embodiments, the target parameters include target puncture path, target stop point position and target ablation sphere parameters. Based on the patient data, determining the target parameters includes: determining the number of puncture paths and the ablation sphere parameter range based on the patient data; based on the patient data, puncture Determine at least one set of planning parameters based on the number of paths and the range of ablation sphere parameters; determine at least one set of feasible solutions based on at least one set of planning parameters; and determine target parameters based on at least one set of feasible solutions.

In some embodiments, determining at least one set of feasible solutions based on at least one set of planning parameters includes: generating a planning set based on at least one set of planning parameters, where the planning set includes a plurality of intermediate solutions, each intermediate solution corresponding to a set of planning parameters; Perform at least one round of second iteration optimization on the planning set until the second iteration completion condition is met to obtain the optimal planning set; and determine at least one set of feasible solutions based on the optimal planning set.

In some embodiments, determining the ablation sphere parameter range includes: determining a lesion mask based on the patient data; determining the lesion long axis and the lesion short axis based on the lesion mask; and determining the ablation sphere parameters based on the lesion long axis and the lesion short axis. scope.

In some embodiments, determining at least one set of planning parameters based on the patient data, the number of puncture paths, and the ablation sphere parameter range includes: performing three-dimensional reconstruction of the patient data to obtain a three-dimensional medical image; performing a preprocessing operation on the three-dimensional medical image; and At least one set of planning parameters is determined based on the results obtained from the preprocessing operation, the number of puncture paths, and the ablation sphere parameter range.

In some embodiments, each iteration in at least one second iteration includes: performing a transformation operation on the planning set to obtain a first preset number of new intermediate solutions; and adding the new intermediate solutions to the planning set to obtain the added A set of plans for new intermediate solutions.

In some embodiments, each of the at least one second iteration includes: calculating a second evaluation value and a second constraint determination value of each intermediate solution added to the planning set of the new intermediate solution; and based on the second The evaluation value and the second constraint determination value select an intermediate solution to obtain a new planning set containing a second preset number of intermediate solutions. The second constraint determination value is determined based on the determination value of at least one constraint condition, and the at least one constraint condition includes the length of the device. Whether the requirements are met, the second evaluation value includes puncture path score and ablation conformity rate.

One embodiment of this specification provides an intervention planning method, which method includes: obtaining patient data of a target patient; and determining target parameters based on the patient data.

One embodiment of this specification provides an interventional planning device, which includes an interventional device, a robotic arm, and a processor; the interventional device includes an interventional needle; the robotic arm is used to carry the interventional needle to perform interventional surgery according to target parameters; the processor is used to control the robotic arm, and determining the target parameters. The determination of the target parameters includes: obtaining patient data of the target patient; and determining the target parameters based on the patient data.

One or more embodiments of this specification provide a computer-readable storage medium. The storage medium stores computer instructions. After the computer reads the computer instructions in the storage medium, the computer executes the parameter planning method of the intervention planning system.

In order to solve the problem of how to automatically plan target parameters to improve doctors' surgical efficiency and reduce work difficulty, some embodiments of this specification determine the target parameters based on the patient data of the target patient. On the one hand, the set of candidate puncture paths is determined based on the structural characteristics and target points, and then based on the path association information of at least one candidate puncture path, the target puncture path is determined through a preset search algorithm, which can accurately and effectively plan the target puncture path, improving the Target puncture path planning efficiency. On the other hand, more reasonable target parameters can be obtained by determining at least one set of planning parameters based on patient data, and determining target parameters including target puncture path, target stay point location, and target ablation sphere parameters from at least one set of planning parameters. . When determining at least one set of planning parameters, you can also narrow the range of planning parameters by determining the number of puncture paths and the range of ablation ball parameters, and obtain more reasonable target parameters more quickly, so as to improve the doctor's surgical efficiency and reduce the difficulty of the doctor's work. .

Description of drawings

This specification is further explained by way of example embodiments, which are described in detail by means of the accompanying drawings. These embodiments are not limiting. In these embodiments, the same numbers represent the same structures, where:

Figure 1 is a schematic diagram of an application scenario of an intervention planning system according to some embodiments of this specification;

Figure 2 is an exemplary schematic diagram of an intervention planning device according to some embodiments of this specification;

Figure 3 is an exemplary flow chart of an intervention planning method according to some embodiments of this specification;

Figure 4 is an exemplary flowchart of determining a target puncture path according to some embodiments of this specification;

Figure 5 is an exemplary schematic diagram of a divided area of a light source according to some embodiments of this specification;

Figure 6 is an exemplary schematic diagram of a set of candidate puncture paths according to some embodiments of this specification;

Figure 7 is another exemplary flow chart for determining a target puncture path according to some embodiments of this specification;

Figure 8 is an exemplary flow chart of a quantum annealing algorithm according to some embodiments of this specification;

Figure 9 is an exemplary flowchart of determining target parameters according to some embodiments of this specification;

Figure 10 is an exemplary schematic diagram of planning parameters according to some embodiments of this specification;

Figure 11 is another exemplary flowchart of determining target parameters according to some embodiments of this specification;

Figure 12 is an exemplary flow diagram of an individual filter screening process according to some embodiments of the present specification;

Figure 13 is another exemplary flowchart of determining target parameters according to some embodiments of this specification;

Figure 14 is an exemplary schematic diagram of determining the maximum short axis of a lesion according to some embodiments of this specification;

Figure 15 is an exemplary flowchart of determining planning parameters according to some embodiments of this specification;

Figure 16A is an exemplary schematic diagram of another planning parameter according to some embodiments of the present specification;

Figure 16B is an exemplary schematic diagram of determining the needle entry point set and the target point set according to some embodiments of this specification;

Figure 17 is another exemplary flowchart for determining target parameters according to some embodiments of the present specification;

Figure 18 is a schematic flowchart of selecting an intermediate solution according to some embodiments of this specification;

Figure 19 is an exemplary flowchart of the internal structure of a computer device according to some embodiments of the present specification.

Detailed ways

Interventional surgery on patients is a commonly used clinical treatment or examination method. However, in the process of applying this method, a device capable of performing interventional surgery is required. Some embodiments of this specification illustrate such a device.

Moreover, during the process of performing interventional surgery, it is necessary to plan target parameters such as the target puncture path, target stop point location, and target ablation sphere parameters. Some embodiments of this specification illustrate such methods. These methods are not directly implemented on living human or animal bodies, but on devices that perform interventional surgeries. Its role is not to identify, determine or eliminate the cause or lesion, but to optimize the puncture process, such as improving the doctor's surgical efficiency and reducing work intensity. Wherein, the device for performing interventional surgery may be the interventional planning device described in this specification.

In order to explain the technical solutions of the embodiments of this specification more clearly, the accompanying drawings needed to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some examples or embodiments of this specification. For those of ordinary skill in the art, without exerting any creative efforts, this specification can also be applied to other applications based on these drawings. Other similar scenarios. Unless obvious from the locale or otherwise stated, the same reference numbers in the figures represent the same structure or operation.

It will be understood that the terms "system", "apparatus", "unit" and/or "module" as used herein are a means of distinguishing between different components, elements, parts, portions or assemblies at different levels. However, said words may be replaced by other expressions if they serve the same purpose.

As shown in this specification and claims, words such as "a", "an", "an" and/or "the" do not specifically refer to the singular and may include the plural unless the context clearly indicates an exception. Generally speaking, the terms "comprising" and "comprising" only imply the inclusion of clearly identified steps and elements, and these steps and elements do not constitute an exclusive list. The method or apparatus may also include other steps or elements.

Flowcharts are used in this specification to illustrate operations performed by systems according to embodiments of this specification. It should be understood that preceding or following operations are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. At the same time, you can add other operations to these processes, or remove a step or steps from these processes.

Figure 1 is a schematic diagram of an application scenario of an intervention planning system according to some embodiments of this specification. In some embodiments, the application scenario 100 of the intervention planning system may include an image scanning device 110, a processing device 120, an intervention planning device 130, and a patient 140. In some embodiments, the application scenario 100 of the intervention planning system may also include storage devices, networks and/or user terminals (not shown in the figure).

Image scanning device 110 refers to a device for scanning a patient 140 to obtain patient data. For example, positron emission tomography (PET) equipment, direct digital flat-panel X-ray imaging (Digital Radiography, DR) equipment, magnetic resonance imaging (Magnetic Resonance Imaging, MRI) equipment, and electronic computed tomography used to obtain patient data (Computed Tomography, CT) equipment, etc. In some embodiments, the image scanning device 110 can scan the patient 140 and send the scanned image data to the processing device 120 for three-dimensional reconstruction and subsequent processing to determine target parameters.

The processing device 120 may be used to process the patient data obtained by the image scanning device 110 to determine target parameters. For example, processing device 120 may obtain patient data of the target patient. Further, the processing device 120 may determine the target parameters based on the patient data. In some embodiments, processing device 120 may send target parameters to intervention planning device 130 for ablation treatment.

In some embodiments, processing device 120 may include one or more sub-processing devices (eg, a single-core processing device or a multi-core processing device). By way of example only, processing device 120 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or the like. In some embodiments, processing device 120 may be integrated or installed in interventional planning device 130 (eg, interventional device 131).

Interventional planning device 130 may be used to perform interventional procedures on patient 140 . In some embodiments, the interventional planning device 130 may include an interventional device 131, a robotic arm 132, and a processor (not shown in the figure). Interventional device 131 may include an interventional needle (not shown). In some embodiments, electrodes may be provided on the interventional needle for delivering ablation energy. The robotic arm 132 may be used to carry an interventional needle to perform interventional surgery according to target parameters. The processor is used to control the robotic arm and determine target parameters. The processor may be a part of the processing device 120, or the processor may be two independent components from the processing device 120. For more information about the intervention planning device 130, see Figure 2 and its associated description.

In some embodiments, the application scenario 100 of the intervention planning system may also include some or more other devices, such as storage devices, networks and/or user terminals (not shown in the figure).

Storage devices may be used to store data, instructions, and/or any other information. In some embodiments, the storage device may store data and/or instructions related to the interventional procedure. For example, the storage device may store patient data scanned by image scanning device 110 . For another example, the storage device may also store target parameters determined by the processing device 120 and the like.

The storage device may include one or more storage components, and each storage component may be an independent device or a part of other devices (such as the processing device 120, etc.). In some embodiments, the storage device may be implemented on a cloud platform.

The network may connect various components in the application scenario 100 of the intervention planning system for communication. For example, image scanning device 110 may send patient data over the network to processing device 120 for processing. For another example, the processing device 120 may send the target parameters to the interventional planning device 130 through the network for performing the interventional surgery. In some embodiments, the network may be any one or more of a wired network or a wireless network.

User terminal refers to one or more terminal devices or software used by users. The user using the user terminal can be one or multiple users. For example, a doctor who performs an interventional procedure on the patient 140, or the like. In some embodiments, the doctor who performs the ablation treatment operation can select the target parameters from at least a set of intermediate parameters or feasible solutions determined by the processing device 120 through the user terminal to perform the ablation treatment operation. The user terminal may be one or any combination of a mobile device, a tablet computer, a laptop computer, a desktop computer, and other devices with input and/or output functions.

It should be noted that the application scenario 100 of the ablative hand treatment surgical system is provided for illustrative purposes only and is not intended to limit the scope of the present application. For those of ordinary skill in the art, various modifications or changes can be made based on the description of this specification. However, such changes and modifications would not depart from the scope of the present application.

Figure 2 is an exemplary schematic diagram of an intervention planning device 130 according to some embodiments of the present specification.

In some embodiments, the interventional planning device 130 may include an interventional device 131 , a robotic arm 132 and a processor 133 .

The interventional device 131 refers to a device that can enter the patient's body tissue to perform interventional surgery. When interventional surgery is used for biopsy, the interventional device 131 can be used to puncture and sample the target patient to further detect the lesion. When interventional surgery is used for TPS radioactive seed implantation, the interventional device 131 can be used to puncture the target patient and implant the radioactive seeds into the tumor to achieve precise tumor treatment. When interventional surgery is used for ablation treatment, the interventional device 131 can be used to puncture the target patient and perform ablation treatment on the lesion. In some embodiments, the interventional device 131 is provided with an interventional needle 131-1 for entering the patient's lesion area to perform interventional surgery on the lesion. In some embodiments, the interventional device 131 can perform interventional surgery with target parameters under the control of the processor 133 .

Taking ablation treatment as an example, the interventional needle 131-1 may include a thermal interventional needle, a cryointerventional needle, a chemical interventional needle and other devices. In some embodiments, one or more interventional needles 131-1 may be provided on the interventional device 131. The needle tip of the interventional needle 131-1 is provided with an ablation electrode, which can deliver ablation energy to the lesion according to the ablation power for ablation treatment. The ablation energy means that when acting on the lesion, it can cause destructive damage to the cells of the lesion and cause them to die. The ablation power is controlled by the processor 133. In some embodiments, the interventional device 131 can perform thermal ablation treatment (eg, microwave ablation, radiofrequency ablation, etc.), cold ablation treatment, chemical ablation, etc., on the lesion area. Among them, thermal ablation treatment means that the electrode of the interventional needle 131-1 locally generates a high temperature reaching 70 degrees, cauterizing the lesion and stimulating necrosis; cold ablation treatment means releasing argon and helium gas at the electrode, and the electrode reaches a low temperature of -185 degrees. , causing the water in the cells of the lesion to crystallize, and then rewarming to cause necrosis of the lesion; chemical ablation refers to injecting some absolute alcohol and other chemical drugs into the lesion through the interventional needle 131-1 to cause the lesion to achieve necrosis. The interventional equipment 131 suitable for different ablation treatment methods can be configured according to specific needs.

The robotic arm 132 may be used to carry the interventional needle 131-1 to perform interventional surgery according to target parameters. The robotic arm 132 refers to an instrument that provides support for the interventional needle 131-1 to perform interventional surgery. For example, the robotic arm 132 may include an operating surgical robot, a positioning surgical robot, and the like.

The processor 133 may be used to control the robotic arm 132 to carry the interventional needle 131-1 to perform interventional surgery according to target parameters. For example, the processor 133 may control the robotic arm 132 to carry the interventional needle 131-1 into the human tissue at a target needle entry point in target parameters. For another example, taking ablation treatment as an example, the processor 133 may determine the number of interventional needles 131-1 used by the interventional device 131 based on the number of target puncture paths in the target parameters. For another example, taking ablation treatment as an example, the processor 133 can control the robotic arm 132 to carry the interventional needle 131-1 into the human tissue from the target needle entry point, and perform ablation treatment according to the target ablation sphere parameters at the target stay point.

In some embodiments, the processor 133 can automatically control the robotic arm 132 to carry the interventional needle 131-1 to perform interventional surgery according to target parameters. In some embodiments, the processor 133 can navigate the doctor based on the target parameters, and the doctor controls the robotic arm 132 to carry the interventional needle 131-1 according to the navigation to perform the interventional surgery according to the target parameters. Among them, navigation can display target parameters based on three-dimensional medical images and implement The position of the interventional needle 131-1 and/or the progress of the operation during the interventional operation are displayed. The user may input relevant instructions instructing the robotic arm 132 to perform interventional surgery on the target patient via the user terminal. In some embodiments, the processor 133 can also perform access surgery based on other methods, for example, direct operation by a doctor based on navigation.

Figure 3 is an exemplary flowchart of an intervention planning method according to some embodiments of the present specification. As shown in Figure 3, process 300 includes the following steps. In some embodiments, one or more operations of the process 300 shown in FIG. 3 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 300 shown in FIG. 3 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

Step 310: Obtain patient data of the target patient.

The target patient is the person who is to undergo interventional surgery.

Patient data refers to data that reflects the lesion status of the target patient. For example, patient data may include PET image data, DR image data, MRI image data, CT image data, etc.

In some embodiments, processing device 120 may obtain patient data in a variety of ways. For example, after the image scanning device 110 scans a patient, the scanned patient data can be directly sent to the processing device 120 through the network. For another example, the image scanning device 110 scans the patient and stores the patient data in the storage device. The processing device 120 can actively read the patient data in the storage device when needed (such as when the user issues relevant instructions through the user terminal).

Step 320: Determine target parameters based on patient data.

The target parameters refer to parameters used by the interventional planning device to perform the interventional surgery. For example, processing device 120 may control parameters for interventional device 131 to perform ablation at target parameters.

In some embodiments, the target parameters may include a target puncture path. In some embodiments, when interventional surgery is used for ablation treatment, the target parameters may also include target stop point positions and target ablation sphere parameters. The target puncture path can be determined by the target needle entry point and the target target point.

The target puncture path refers to the trajectory of the interventional needle of the interventional device entering human tissue. For example, as shown in Figure 10, the target puncture path can be {P _i1j1 , P _i2j2 }, where P _i1j1 can be determined by the target needle entry point P _i1 and the target target point P _j1 , and P _i2j2 can be determined by the target needle entry point P. _i2 and target target point P _j2 are determined.

The target stop point position refers to the stop point position of the interventional needle electrode during the puncture process of the interventional needle along the target puncture path. For example, as shown in Figure 10, the target stay point position on the target puncture path P _i1j1 is {T _1p1 , T _2p1 }, and the target stay point position on the target puncture path P _i2j2 is T _1p2 .

The target ablation sphere parameters refer to the size data of the ablation area of the interventional needle during ablation treatment surgery. For example, when the ablation sphere is an ellipsoid, the ablation sphere parameters can be the size of the major and minor axes of the ablation sphere. Since the thermal diffusion ability is the same in all directions and is circular in the direction perpendicular to the axis, the parameters of the ablation sphere can be expressed as a × a × c, where a is the short axis length of the ablation sphere ellipsoid, and c is the ablation sphere ellipsoid. The length of the major axis. The ablation sphere parameter range refers to the respective value ranges of the long and short axes of the ablation sphere. Different ablation power and ablation time correspond to different ablation sphere sizes. The greater the ablation power, the longer the ablation time, and the larger the ablation sphere. For example, as shown in Figure 10, the target ablation sphere parameter can be R (for example, if the target ablation sphere is an ellipsoid, R can be an ellipsoid of 36×36×42 mm, where 36 is the short axis length of the ellipsoid, 42 is the length of the major axis of the ellipsoid).

In some embodiments, the target parameters may include a target puncture path, the processing device 120 may determine structural characteristics of the target patient based on the patient data, and determine the target target point based on the structural characteristics. Next, the processing device 120 may determine a set of candidate puncture paths based on the structural features and the target target point, and then determine the target puncture path from the set of candidate puncture paths. After the target puncture path is determined, the processor can control the robotic arm to carry the interventional needle to puncture the target patient according to the target puncture path. For example, when the interventional surgery is used for needle biopsy, the processing device 120 can control the robotic arm to carry the interventional needle to puncture the target patient and extract the tissue according to the target puncture path. For example, when interventional surgery is used for TPS radioactive particle implantation, the processing device 120 can control the robotic arm to carry the interventional needle, puncture the target patient according to the target puncture path, and implant multiple particles at multiple stay points. The resulting multiple doses are superimposed to form a larger therapeutic range. For another example, when an interventional surgery is used for ablation treatment, the processing device 120 can control the robotic arm to carry the interventional needle to puncture the target patient according to the target puncture path, and perform ablation treatment with the target ablation sphere parameters at the target stop point. Among them, the target stop point position and target ablation sphere parameters during the puncture process can be set by the doctor. The target stop point position and target ablation sphere parameters during the puncture process can also be obtained through other methods, for example, further determined based on the method in Figure 9 or Figure 13. For more information on determining the target puncture path, see Figures 4-8 and their related descriptions.

In some embodiments, when interventional surgery is used for ablation treatment, the target parameters may include a target puncture path, a target stop point position, and a target ablation sphere parameter. The processing device 120 may determine at least one set of planning parameters based on the patient data, and based on At least one set of planning parameters to determine the target parameters. For example, the processing device 120 can generate an individual set based on the individual generator. The individual set includes multiple individuals, each individual corresponds to a set of planning parameters, and performs at least one first iteration update on the individual set until the first iteration. The completion condition is met. Next, the processing device 120 may determine at least one set of intermediate parameters based on the updated set of individuals, and determine the target parameters based on the at least one set of intermediate parameters. For more information on the above-mentioned determination of target parameters, please refer to Figures 9 to 12 and their related descriptions.

In some embodiments, when interventional surgery is used for ablation treatment, the target parameters may include the target puncture path and the target stay point. To set and target ablation sphere parameters, the processing device 120 may determine the number of puncture paths and the ablation sphere parameter range based on the patient data, and determine at least one set of planning parameters based on the patient data, the number of puncture paths, and the ablation sphere parameter range. Next, the processing device 120 may determine at least one set of feasible solutions based on at least one set of planning parameters, and determine the target parameters based on at least one set of feasible solutions. For more information on the above-mentioned determination of target parameters, please refer to Figures 13 to 19 and their related descriptions.

In some embodiments of this specification, by obtaining the patient data of the target patient and then determining the target parameters, more reasonable target parameters can be obtained to improve the doctor's surgical efficiency and reduce the difficulty of the doctor's work.

It should be noted that the above description of process 300 is only for example and illustration, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process 300 under the guidance of this description. However, such modifications and changes remain within the scope of this specification.

When interventional surgery is used for puncture biopsy, the processing device 120 may adopt the process of FIG. 4 to determine the target puncture path. When the lesion is small in size, the processing device 120 can use a target puncture path to implant TPS radioactive seeds, and the processing device 120 can adopt the process in Figure 4 to determine the target puncture path. When the size of the lesion is small and the number of target puncture paths is 1, the ablation can be completed. The processing device 120 can adopt the process of FIG. 4 to determine the target puncture path. See Figure 13 and its related description for more details on determining the number of target puncture paths.

Figure 4 is an exemplary flowchart of determining a target puncture path according to some embodiments of this specification. As shown in Figure 4, process 400 includes the following steps. In some embodiments, one or more operations of the process 400 shown in FIG. 4 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 400 shown in FIG. 4 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

In some embodiments, the target parameters include a target puncture path.

In some embodiments, the processing device 120 can acquire a surgical area to be punctured (or a target object) constructed of several voxels, and determine a candidate needle entry point based on the target target point in the structure to be punctured inside the surgical area to be punctured. Set; several voxels include surface voxels and internal voxels. The voxels corresponding to each candidate needle entry point in the candidate needle entry point set belong to the surface voxels, and there are corresponding candidate punctures between each candidate needle entry point and the target target point. Path; determine the path association information corresponding to each candidate puncture path; the path association information includes the separation distance between the candidate puncture path and the puncture risk structure inside the surgical area to be punctured, and the path length of the candidate puncture path; combine the separation distance and path length , and the preset path search parameters, construct a Hamiltonian (or path selection function); according to the Hamiltonian, search for the target puncture path from each candidate puncture path, and use the candidate needle entry point corresponding to the target puncture path as Target entry point. Wherein, the path search parameters include iteration parameters used to indicate the path search calculation.

Step 410: Determine the structural characteristics of the target patient based on the patient data.

Structural features refer to the features of the surgical area to be punctured surrounded by the surface contour of the structure to be punctured. The surgical area to be punctured may include surface voxels and internal voxels, and the structure to be punctured refers to the organ to which the target patient's lesion belongs (for example, taking a liver tumor as an example, the structure to be punctured may refer to the liver of the target patient). For example, as shown in FIG. 5 , the structural features may include features of the structure to be punctured 510 , surface voxels 520 , target tissue 530 , puncture risk structure 540 and other structures.

In some embodiments, the processing device 120 may determine a three-dimensional medical image of the target patient based on the patient data, and determine structural features based on the three-dimensional medical image.

Three-dimensional medical imaging refers to three-dimensional rendered images used to represent surface voxels and internal voxels. Three-dimensional medical images can be used to display the local anatomical scene of the target patient (for example, the abdominal anatomical scene) to clarify the anatomical structure and spatial location of the target patient's lesions. Among them, surface voxels can be used to characterize the epidermal contour, and internal voxels can be used to characterize internal anatomical structures.

In some embodiments, the processing device 120 can perform three-dimensional reconstruction of the surgical area to be punctured based on the patient data of the target patient and match the spatial coordinates to obtain a three-dimensional model of the target patient constructed from voxels, that is, a three-dimensional medical image.

In some embodiments, the processing device 120 can distinguish the boundaries between different tissues and organs based on the grayscale features of different tissues in the patient data, and then segment the tissues, organs, etc. in the three-dimensional medical image to determine the structural features. In some embodiments, the tissues and organs that need to be segmented may include structures to be punctured, skin, lesions, bones, tissues adjacent to the structures to be punctured, blood vessels inside the structures to be punctured, etc. In some embodiments, the above-mentioned segmentation method may include automatic segmentation and interactive editing segmentation. Among them, automatic segmentation may include but is not limited to threshold-based segmentation, machine learning-based segmentation, and deep learning-based segmentation methods. Interactive editing segmentation may include but is not limited to region growing, flood filling, slice interpolation, and interactive rendering.

Step 420: Determine the target target based on the structural characteristics.

The target point refers to the target point of the lesion on the structure to be punctured. For example, the target point may be the center of mass of the lesion. In some embodiments, the target target point can also be other points on the lesion. For example, the target target point can be the geometric center of the lesion.

In some embodiments, the processing device 120 may determine the centroid of the lesion on the structure to be punctured based on the structural characteristics, and use the determined centroid as the target target.

Step 430: Determine a set of candidate puncture paths based on structural features and target points.

In some embodiments, the processing device 120 can obtain the patient data of the target patient and generate a target rendering corresponding to the target patient. Image; the target rendering image is used to represent the surgical area to be punctured based on patient data; the target point in the structure to be punctured is the perspective projection center, and the puncture risk structure is the perspective projection object, and a perspective projection model is established; the perspective projection model is used , determine the set of candidate needle entry points. Among them, the method of determining the set of candidate needle entry points using a perspective projection model may include: the processing device 120 may use a perspective projection model to obtain the light collision results corresponding to each target source light; the target source light is based on the target target point, and The light is scattered in all directions at the location of the structure to be punctured. The light collision results are used to characterize the collision results of the target source light and puncture risk structures and/or surface voxels; according to the strong clinical constraints (or preset partition conditions) ), based on the light collision results corresponding to the light rays of each target point, determine the feasible needle entry voxels from the surface voxels of the surgical area to be punctured, as candidate needle entry points; the pre-clinical strong constraints are for the surgical area to be punctured Whether the surface voxels meet the constraints required for puncture, the candidate puncture path corresponding to each candidate needle entry point is obtained based on the target source light path corresponding to each feasible needle entry voxel; a set of candidate needle entry points is obtained based on each candidate needle entry point . Strong clinical constraints may include any one or more of the following: the candidate puncture path does not contact and does not penetrate the puncture risk structure, the path length of the candidate puncture path is less than the preset needle length threshold, and the distance length of the candidate puncture path through the structure to be punctured is greater than Default distance threshold.

The set of candidate puncture paths refers to a set composed of paths that can be used as puncture paths. In some embodiments, the set of candidate puncture paths may include at least one candidate puncture path. Candidate puncture paths can be obtained by removing obviously unreasonable paths through strong clinical constraints. The candidate puncture path may be composed of at least one candidate needle entry point and a target target point. The voxel corresponding to each of the at least one candidate needle entry point belongs to the surface voxel.

In some embodiments, the processing device 120 can determine the perspective projection center based on the target target, and use the perspective projection center as the source to emit multiple rays outwards, and calculate the clinical strength of the puncture path corresponding to each ray based on the structural characteristics. The judgment value of the constraint condition. Further, the processing device 120 may determine puncture paths whose judgment values of strong clinical constraints satisfy preset conditions as candidate puncture paths, and determine a set of candidate puncture paths as a candidate puncture path set.

The perspective projection center refers to the center of the perspective projection model. For example, the center of the perspective projection can be the target target point. In some embodiments, processing device 120 may determine the target target point as the perspective projection center.

Strong clinical constraints refer to constraints that can determine whether the puncture path corresponding to each ray meets the puncture requirements. Among them, the puncture path corresponding to each ray refers to the puncture path that coincides with each ray.

In some embodiments, strong clinical constraints may include one or more of the following: the puncture path does not contact and does not penetrate puncture risk structures, the length of the candidate puncture path is less than the preset needle length threshold, and the angle between the puncture path and the target tissue It is not less than the preset angle threshold and the distance length of the puncture path through the structure to be punctured is greater than the preset distance threshold. The following can be explained by taking Figure 5 as an example.

In some embodiments, strong clinical constraints may include that the puncture path does not contact or penetrate puncture risk structures, that is, the puncture path needs to avoid contacting or penetrating inaccessible risk structures. Among them, risk structures can include thicker blood vessels, important organs, bones, etc. In some embodiments, the structural features can reflect the conditions of key parts such as organs and bones in the surgical area to be punctured, and the processing device 120 can determine whether the puncture path contacts and/or penetrates puncture risk structures based on the structural features. For example, as shown in Figure 5, the target point O can be used as the perspective projection center, and the target point O can be used as the radiation source to emit multiple rays (for example, OP ₁ , OP ₂ , OP ₃ , OP ₄ , OP ₅ , OP ₆ ), 540 is the puncture risk structure. In some embodiments, processing device 120 may consider the intersection of ray OP ₆ that strikes the puncture risk mechanism and surface voxel 520 as a no-no point. Further, the processing device 120 may mark the body surface area behind the puncture risk structure 540 as a prohibited needle entry area, and determine a line connecting a point on the prohibited needle entry area and the target point O as a prohibited puncture path.

In some embodiments, the clinically strong constraint may include that the length of the puncture path is less than a preset needle length threshold. Among them, according to clinical standards, the preset needle length threshold can be 10cm~15cm. Taking the interventional needle length as 15cm as an example, according to clinical standards, 15cm can be selected as the preset needle length threshold. In some embodiments, the processing device 120 may calculate the length of the puncture path, determine the puncture path whose length is greater than or equal to the preset needle length threshold as the prohibited puncture path, and determine the intersection of the prohibited puncture path and the surface voxel as the prohibited puncture path. into the needle area.

In some embodiments, the strong clinical constraint may include that the angle between the puncture path and the target tissue is not less than a preset angle threshold. Among them, the target tissue refers to the outermost membranous structure of the structure to be punctured. For example, when the structure to be punctured is the liver and the target tissue is the liver capsule, the angle threshold can be 10° to 30° according to clinical standards. For another example, when the structure to be punctured is the spleen, the target tissue is the splenic capsule. The angle between the puncture path and the target tissue refers to the angle between the puncture path and the normal vector of the target tissue surface. For example, taking FIG. 5 as an example, the angle between the puncture path OP ₅ and the target tissue 530 is the angle θ between the puncture path OP ₅ and the normal vector of the surface of the target tissue 530 .

It should be noted that if the lesion is inside the structure to be punctured, it is necessary to consider the strong clinical constraint that the angle between the puncture path and the target tissue is not less than the preset angle threshold; if the lesion is not inside the structure to be punctured, such as on the surface of the structure to be punctured, (for example, in the target tissue), the strong clinical constraints may not be considered. In some embodiments, if the structure to be punctured is the heart, lungs, etc., the strong constraint angle condition may not be considered. If the structure to be punctured is the liver or spleen, the strong constraint angle condition may be considered.

For example, if the structure to be punctured is the liver and the target tissue is the liver capsule, if the angle between the puncture path and the liver capsule is large, the interventional needle will more easily penetrate the liver parenchyma during the puncture process and produce smaller stress deformation. According to clinical standards, θ>20° can be selected as the angle threshold.

In some embodiments, the processing device 120 calculates the angle between the ray direction and the normal vector of the target tissue surface, and then for rays greater than the angle threshold, all fluoroscopic body surface areas behind the ray can be calibrated as needle-inhibited areas, and Align the points on the prohibited needle area with the target The line connecting the target point O is determined as the prohibited puncture path. In some embodiments, the processing device 120 may also record the obtained angle value θ in each voxel of the feasible needle insertion area on the body surface (ie, the voxel corresponding to the candidate needle insertion point) to facilitate subsequent stage calculation processing.

In some embodiments, the strong clinical constraint may include that the distance length of the puncture path through the structure to be punctured is greater than a preset distance threshold. Among them, according to clinical standards, the preset distance threshold can be 3 mm to 8 mm. For example, as shown in FIG. 5 , the distance between the puncture path OP ₅ and the structure to be punctured is the length between OH. Taking the structure to be punctured as the liver as an example, according to clinical standards, 5 mm can be selected as the preset distance threshold.

In some embodiments, the processing device 120 can mark all the fluoroscopic body surface areas behind the rays whose puncture path passes through the structure to be punctured and whose distance length is less than or equal to the preset distance threshold as the needle-inhibited area, and place the needle-inhibited areas on the rays. The line connecting the point and the target point O is determined as the prohibited puncture path.

The judgment value of a strong clinical constraint refers to a numerical value or letter that can reflect whether the strong clinical constraint is satisfied. For example, when a certain strong clinical constraint condition is satisfied, the judgment value of the strong clinical constraint condition is assigned a value of 0; otherwise, the value assigned is -1. In some embodiments, the processing device 120 may calculate the sum of the determination values of different clinical strong constraint conditions, and use the summed value as the clinical strong constraint condition.

In some embodiments, the processing device may determine a puncture path whose determination value of the strong clinical constraint satisfies the preset condition as a candidate puncture path. Among them, the preset conditions may include that the judgment value of the clinical strong constraint condition is 0, the judgment value of the clinical strong constraint condition is not greater than 1, the judgment value of the clinical strong constraint condition is not greater than 2, etc. In some embodiments, the processing device can calculate the union of the prohibited needle entry areas calibrated by multiple strong clinical constraints, and then obtain the feasible needle entry point area on the body surface (as represented by the feasible needle entry point area 610 in Figure 6 (shown), the feasible needle entry point area includes the screened candidate needle entry points. Further, the processing device 120 may determine a line connecting the candidate needle entry point and the target target point O as the candidate puncture path.

In some embodiments of this specification, by using a perspective projection model and based on strong clinical constraints, candidate puncture paths whose judgment values of clinical strong constraints meet preset conditions are determined, providing data support for subsequent puncture path planning.

Step 440: Determine the target puncture path from the set of candidate puncture paths.

In some embodiments, the processing device 120 may calculate path association information of at least one candidate puncture path, and determine the target puncture path through a preset search algorithm based on the path association information of the at least one candidate puncture path. For more information on determining the target puncture path through the preset search algorithm, see Figure 7 and its related description.

In some embodiments of this specification, by determining the structural characteristics and target points of the target patient, a candidate puncture path whose judgment value satisfies the preset conditions for strong clinical constraints can be determined, so that a reasonable target puncture can be obtained automatically and efficiently. path.

It should be noted that the above description of process 400 is only for example and illustration, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process 400 under the guidance of this description. However, such modifications and changes remain within the scope of this specification.

Figure 7 is another exemplary flowchart of determining a target puncture path according to some embodiments of this specification. As shown in Figure 7, process 700 includes the following steps. In some embodiments, one or more operations of the process 700 shown in FIG. 7 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 700 shown in FIG. 7 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

Step 710: Calculate path association information of at least one candidate puncture path. The set of candidate puncture paths may include at least one candidate puncture path.

Path related information refers to information that can reflect whether the candidate puncture path meets the puncture requirements. In some embodiments, the path association information may include one or more of the following information: the distance between the candidate puncture path and the puncture risk structure, the length of the candidate puncture path, and the angle between the candidate puncture path and the target tissue. For more information about the distance between the candidate puncture path and the puncture risk structure, the length of the candidate puncture path, and the angle between the candidate puncture path and the target tissue, see Figure 4 and its related description.

In some embodiments, in order to facilitate calculation, considering that although the feasible needle insertion area is a three-dimensional voxel, it only has one layer of body surface elements (ie, surface voxels), the skin needle insertion points can be uniformly sampled based on the light source emission angle. Among them, L is the maximum length of the original body surface, W is the maximum width of the original body surface, and each surface voxel P _ij is a candidate needle entry point. Considering that the original body surface elements are irregular polygons, the prohibition of needle entry can be The surface voxels corresponding to the area are filled in the matrix with 0. Each candidate needle entry point and the target target point can be combined into a candidate puncture path, that is, the surface voxel P _ij corresponding to each candidate needle entry point can have a candidate puncture path.

The greater the distance between the candidate puncture path and the puncture risk structure, the probability of instrument injury or therapeutic injury during the puncture process will be reduced accordingly, and the safety of puncture will be improved accordingly. In some embodiments, the candidate puncture path passes through multiple voxels, and the processing device 120 can use a classic computer to calculate the distances between the multiple voxels passed by the candidate puncture path and all puncture risk structures using a three-dimensional distance transform (Distance Transform, DTF) algorithm, and recorded in each voxel. Then the distances recorded in each voxel are averaged, and the minimum value of the average distance of each voxel is determined as the distance between the candidate puncture path and the puncture risk structure, which can be represented by D _dangerous . The larger the D _dangerous value, the better the path. When solving multi-objectives, the smaller the objective function is defined, the better the result. In order to solve the objectives in a minimizing way, the candidate path is The inverse function operation is performed on the distance between the puncture path and the puncture risk structure to obtain D _ij =1/D _dangerous .

The shorter the length of the candidate puncture path, the fewer key structures need to be considered during the puncture process, and the smaller the chance of cancer cells spreading along the puncture path with the interventional needle. In some embodiments, the processing device can scale the number of passing voxels into the actual path length, which is recorded as d _ij .

In some embodiments, the processing device 120 can retrieve the included angle value θ in the voxel of each candidate needle insertion point when calculating strong clinical constraints, and use the included angle value θ as the candidate corresponding to the candidate needle insertion point. The angle between the puncture path and the target tissue is recorded as θ _ij . For more information about the angle value θ, see Figure 4 and its related description.

Step 720: Determine the target puncture path through a preset search algorithm based on the path association information of at least one candidate puncture path.

In some embodiments, the processing device 120 can construct the first function term according to the separation distance, path length, feature angle, and preset weight information according to the preset selection conditions; the preset selection condition is for the separation distance, path length, The constraint conditions of the characteristic angle; according to the preset path search parameters, the first function term is used to construct the Hamiltonian based on the quantum annealing algorithm; the path search parameters include iteration parameters used to indicate the path search calculation. The method of constructing the first function term may include: the processing device 120 may obtain preset weight information; the preset weight information includes a first weight coefficient for distance, a second weight coefficient for path length, and a characteristic angle. The third weight coefficient; according to the preset selection conditions, the first function term is obtained based on the distance and the first weight coefficient, the path length and the second weight coefficient, the characteristic angle and the third weight coefficient.

In some embodiments, the preset algorithm may include a quantum annealing algorithm. The quantum annealing algorithm is a quantum algorithm applied to combinatorial optimization problems. It finds the optimal solution by utilizing the inherent characteristics of quantum systems, such as quantum superposition states and quantum entanglement states. The quantum annealing algorithm generally consists of two parts: one part is called Hamiltonian potential energy, which maps the objective function to be optimized into a potential field applied to the quantum system, that is, the objective function is regarded as the potential energy part of the Hamiltonian of the quantum system. ; The other part is called Hamiltonian kinetic energy, which is usually introduced as a kinetic energy term with controllable amplitude as a perturbation to the system (for example, it can be imagined as a perturbation of a transverse magnetic field). When a perturbation exists, the system will gradually evolve along the direction of smaller gradient in the solution space, and even directly "pass through" the part with higher potential energy through the quantum tunneling effect. With the gradual iteration, the energy of the quantum system will become lower and lower (that is, the temperature will decrease), and finally it will converge to the ground state of the system. This process is equivalent to obtaining the global optimum of the objective function.

In some embodiments, the processing device 120 may construct a first function term according to the path association information, then construct a second function term, and construct a Hamiltonian based on the first function term, the second function term and the coupling coefficient, thereby based on the Hamiltonian Quantity, the target puncture path is determined through the quantum annealing algorithm. Among them, the first function term can characterize the potential energy part of the Hamiltonian in the quantum annealing algorithm, and the second function term can characterize the kinetic energy part of the Hamiltonian in the quantum annealing algorithm.

In some embodiments, the processing device 120 may use formula (1) to construct the first function term of the candidate puncture path corresponding to the surface voxel P _ij .

in, is the first function term, R _ij is the distance between the normalized candidate puncture path and the puncture risk structure, L _ij is the length of the normalized candidate puncture path, A _ij is the distance between the normalized candidate puncture path and the puncture risk structure. For the angle between the target tissue, w ₁ is the weight of the distance between the candidate puncture path and the puncture risk structure, w ₂ is the weight of the length of the candidate puncture path, and w ₃ is the weight of the angle between the candidate puncture path and the target tissue. Among them, w ₁ +w ₂ +w ₃ =1 to ensure that the weight of each group is normalized.

Since the distance D _ij between the candidate puncture path and the puncture risk structure of the path correlation information, the length d _ij of the candidate puncture path, and the angle θ _ij between the candidate puncture path and the target tissue are different in dimensions, formula (2)-Eq. (4) Normalize the above correlation information to determine the R _ij , L _ij and A _ij of the candidate puncture path corresponding to the surface voxel P _ij .

Among them, R _ij is the distance between the normalized candidate puncture path and the puncture risk structure, L _ij is the length of the normalized candidate puncture path, and A _ij is the distance between the normalized candidate puncture path and the target tissue. Angle, D _ij is the distance between the candidate puncture path and the puncture risk structure, d _ij is the length of the candidate puncture path, θ _ij is the angle between the candidate puncture path and the target tissue, λ ₁ is the normalization parameter of D _ij , λ ₂ is the normalization parameter of d _ij , λ ₃ is the normalization parameter of θ _ij , D _min and D _max are respectively the minimum and maximum values of the distances between all candidate puncture paths and puncture risk structures, d _min and d _max are respectively the minimum value and the maximum value among the lengths of all candidate puncture paths, θ _min and θ _max are respectively the minimum value and the maximum value among the angles between all candidate puncture paths and the target tissue.

In some embodiments, the processing device 120 may determine the candidate puncture corresponding to the surface voxel P _ij based on equations (2) to (4) R _ij , L _ij and A _ij of the path, and then substitute R _ij , L _ij and A _ij into formula (1) to obtain the first function term of the candidate puncture path corresponding to the surface voxel P _ij

In some embodiments, considering the second function term (ie, the kinetic energy part of the Hamiltonian) for providing the perturbation, the processing device 120 can construct the second candidate puncture path corresponding to the surface voxel P _ij using equation (5). function term.
H _k =Γ ₀ e ^-1/T (5)

Among them, Γ ₀ is a constant term, e is a natural number, and T is the temperature of the iterative process.

In some embodiments, the temperature T of the iterative process can be determined using formula (6), which can be a linear function.

Among them, T(s) is the temperature when the current iteration number of the outer loop is s, T ₀ is the initial temperature during the iteration process, s is the current iteration number of the outer loop, and M is the preset maximum number of iterations of the outer loop. The meaning of the outer loop is to jump out of the currently selected initial needle entry point, randomly update another initial needle entry point on the body surface contour (that is, any other candidate needle entry point), and re-search for the optimal path at the current location, once N iterations need to be executed in the outer loop to find the optimal needle entry point that meets the conditions. During the process of searching for the two-dimensional needle entry point, a new solution P′ can be obtained based on random step sizes. If the new solution is smaller than the current solution, the new solution can be updated as the criterion. Otherwise, if it is larger than the current solution , you can judge whether to accept the new solution according to exp(-ΔH/T)>random(0,1). As the number of iterations of the outer loop increases or decreases, the temperature gradually drops from T ₀ to zero, and H _k also drops accordingly. . As the temperature decreases, the probability of acceptance will become lower and lower, that is, the probability of "passing through" the potential barriers on both sides of the local optimal solution (ie, the optimal path analysis result) due to the quantum tunneling effect becomes lower and lower. For more information on outer loop, inner loop, new solution, etc., see Figure 8 and its related descriptions.

In some embodiments, the processing device 120 may use formula (7) to construct the Hamiltonian of the candidate puncture path corresponding to the surface voxel P _ij .

in, is the first function term (that is, the potential energy part of the Hamiltonian), which describes the candidate puncture path to be optimized and can be determined based on formula (1); J _T is the coupling coefficient, which satisfies J _T > 0 and is usually a constant, Its meaning is to adjust the energy ratio of the potential energy term and the kinetic energy term; H _k is the second function term (that is, the kinetic energy part of the Hamiltonian). It represents the disturbance to the entire system and can be described by a function that changes linearly with the number of iterations or by describing the thermal balance. The function composition of the state can be determined based on formula (5).

In some embodiments, the processing device may calculate the Hamiltonian of all candidate puncture paths based on formula (7), and determine the candidate puncture path corresponding to the smallest Hamiltonian as the target puncture path.

In some embodiments, the processing device 120 may determine the target puncture path through a quantum annealing algorithm based on the Hamiltonian. For more information on determining the target puncture path through the quantum annealing algorithm, see Figure 8 and its related description.

In some embodiments of this specification, the first function term is constructed according to the path association information, and then the second function term is constructed, and the Hamiltonian is constructed based on the first function term, the second function term and the coupling coefficient, and then based on the Hamiltonian Quantity, the target puncture path is determined through the quantum annealing algorithm, which improves the algorithm efficiency and the probability of finding the optimal solution.

In some embodiments of this specification, by calculating path association information of at least one candidate puncture path, and determining the target puncture path through a preset search algorithm based on the path association information of at least one candidate puncture path, automatic identification of the target puncture path is achieved. Quantitative planning can accurately and effectively plan the target puncture path, improving the efficiency of target puncture path planning.

It should be noted that the above description of process 700 is only for example and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 700 under the guidance of this specification. However, such modifications and changes remain within the scope of this specification.

Figure 8 is an exemplary flowchart of a quantum annealing algorithm according to some embodiments of this specification. As shown in Figure 8, process 800 includes the following steps. In some embodiments, one or more operations of the process 800 shown in FIG. 8 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 800 shown in FIG. 8 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

Step 8010: Initialize various parameters and the number of outer loop iterations.

In some embodiments, the user can initially set the weight w ₁ for the distance between the candidate puncture path and the puncture risk structure, the weight w ₂ for the length of the candidate puncture path, the weight w ₃ for the angle between the candidate puncture path and the target tissue, and the step size. α, coupling coefficient J _T , initial temperature T ₀ during the iteration process, maximum number of iterations of the outer loop M, maximum number of iterations of the inner loop N; and set the number of iterations of the outer loop s to 1.

In some embodiments, the above parameters and the number of iterations can also be preset fixed configurations without user input.

In some embodiments, the processing device 120 may perform at least one round of outer loop by executing steps 8020 to 8130 until s>M. During each outer loop, the processing device may perform at least one round of inner loop by executing steps 8030 to 8110 until s>N.

Step 8020: Randomly initialize the needle entry point and the number of inner loop iterations.

In some embodiments, each time the outer loop starts, the processing device 120 can randomly initialize the needle entry point, determine the initial needle entry point for this round of outer loop iterations, and set the number of inner loop iterations t to 1.

In some embodiments, the processing device 120 may randomly select any candidate needle insertion point as the initial needle insertion point.

In some embodiments, when the number of external circulations is greater than or equal to the preset threshold of the number of external circulations, the processing device 120 can determine the needle insertion point mass distribution, and divide the candidate needle insertion point set into at least one sample according to the needle insertion point quality distribution. area. Then, one of the sampling areas is selected as the target sampling area from at least one sampling area with unequal probability, and the initial entry point of this round of outer loop iteration is determined from the target sampling area. Among them, the preset external circulation times threshold can be set manually, for example, 100 times, 200 times, 300 times, etc. In some embodiments, when the number of external circulation times is greater than or equal to the preset threshold number of external circulation times, the processing device 120 may use the above method to determine the initial needle insertion point each time it determines the initial needle insertion point. In some embodiments, when the number of external loops is greater than or equal to the preset threshold of the number of external loops, the processing device 120 may determine the number of external loop iterations using the above method to determine the initial needle entry point according to the preset rules. Among them, the preset rules may include interval selection, random selection, etc. For example, if the preset outer loop number threshold is 200 times and the maximum number of outer loop iterations is 500, the processing device 120 can use the above method at the 200th, 210th, 220th... 490th and 500th times. Determine the initial needle entry point.

The mass distribution of needle entry points may include needle entry points and their corresponding point evaluation values for each iteration of the outer loop and/or inner loop iterations.

Point evaluation value refers to letters, numerical values, etc. that can evaluate the quality of the needle entry point. For example, the processing device 120 may use the value of the first function term corresponding to the needle insertion point as the point evaluation value. The lower the point evaluation value, the better the needle insertion point.

In some embodiments, the processing device 120 may calculate the value of the first function term corresponding to the needle entry point, and based on the value of the first function term, determine that the value of the first function term corresponding to the needle entry point is the value of the first function term corresponding to the needle entry point after completing the outer loop and /or the ranking value among the values of the first function term corresponding to the needle entry point of each iteration in the inner loop iteration. Further, based on the ranking value, the point evaluation value corresponding to the needle entry point is determined. For example, the processing device 120 may calculate the value of the first function term corresponding to the needle entry point based on formula (1), and calculate the value of the first function term corresponding to the needle entry point of each iteration of the completed outer loop and/or inner loop iteration. The values are arranged from small to large to determine the ranking value of the value of the first function item corresponding to the needle entry point. The lower the ranking value, the better the needle entry point and the higher the point evaluation value.

The sampling area refers to multiple areas where needle entry points with similar point evaluation values are divided together based on the characteristics of needle entry point mass distribution. For example, the sampling areas can be 2, 3, 4, etc.

In some embodiments, the processing device 120 may divide the sampling area using a random division method. Specifically, the processing device 120 can randomly divide the sampling area, and sequentially evaluate whether each division can meet the preset requirements. If the preset requirements can be met, the sampling area divided this time will be used as the final sampling area; otherwise, the sampling area will be randomly divided again. The preset requirement may be that the variance of the value of the first function term corresponding to the needle entry point in each sampling area is less than the preset variance value.

In some embodiments, the processing device 120 may also use other methods to divide the sampling area, such as clustering.

The target sampling area refers to the final selected sampling area.

In some embodiments, the processing device 120 may select one of the sampling areas from at least one sampling area as the target sampling area with non-equal probability. For example, for each sampling area in at least one sampling area, the processing device 120 may calculate the average point evaluation value of the needle entry point of all completed outer loop and/or inner loop iterations of each iteration in the sampling area. , and use this evaluation value as the average point evaluation value. Next, the processing device 120 may determine the probability of selecting each sampling area as the target sampling area based on the proportion of the average point evaluation value of each sampling area in the at least one sampling area. For example, if the average point evaluation value of sampling area 1: the average point evaluation value of sampling area 2: the average point evaluation value of sampling area 3 = 1:2:1, then the probability of selecting sampling area 1 as the target sampling area for The probability of selecting sampling area 2 as the target sampling area is The probability of selecting sampling area 3 as the target sampling area is

In some embodiments, the processing device 120 may randomly select any candidate needle entry point from the target sampling area as the initial needle entry point.

In some embodiments of this specification, by determining the mass distribution of needle entry points, the set of candidate needle entry points is divided into at least one sampling area, and one of the sampling areas is selected as the target sampling area with unequal probability, and is sampled from the target The needle entry point of this round of outer loop iteration is determined in the area, and a better needle entry point of this round of outer loop iteration can be selected, which is beneficial to subsequent determination of the optimal target puncture path. Since one of the sampling areas is selected as the target sampling area with unequal probability, rather than one of the sampling areas as the target sampling area, falling into a local optimum is avoided.

Step 8030, randomly move the needle entry point with a step size α to generate a new solution P'.

In some embodiments, if the previous needle entry point is P _ij , the needle entry point P _ij is randomly moved with a step size α to generate a new needle entry point Pi _±α,j or _Pi,j±α . It should be noted that the new needle entry point needs to be one of the candidate needle entry points. Correspondingly, the solution of the previous iteration is P, and the new solution generated is P'.

In some embodiments, the step size α can be any positive integer. For example, α can be a smaller value such as 1, 2, 3, etc.

In some embodiments, the processing device 120 may randomly select a positive integer within 10 as the step size α.

In some embodiments, the processor 120 may determine the average curvature of the skin surface where the preset neighborhood of the current needle insertion point position is located, and determine the step size α of the loop iteration within this round based on the average curvature. The preset neighborhood refers to the set of voxel points whose distance from the needle entry point is no greater than the preset threshold. The preset threshold may be a preset threshold.

The average curvature is a physical quantity that reflects the curvature of the skin surface where the preset neighborhood is located. In some embodiments, the processing device 120 may obtain the curvature radius of each surface voxel on the skin surface where the preset neighborhood is located, calculate the average of all curvature radii, and use the reciprocal of the average of the curvature radii as the average curvature. The greater the average curvature, the smaller the average curvature radius, and the greater the undulations of the skin surface where the preset neighborhood is located, then the two adjacent voxel points on the skin surface may have a greater difference when used as the needle insertion point. Therefore, The step size α needs to be reduced.

In some embodiments, the processing device may determine the step size of the loop iteration within this round based on the average curvature of the skin surface where the preset neighborhood of the current needle entry point location is located. For example, the greater the average curvature of the skin surface where the preset neighborhood of the current needle entry point is located, the smaller the path point step size α can be.

In some embodiments of this specification, the step size of the inner loop iteration of this round is determined based on the average curvature of the skin surface where the preset neighborhood of the current needle entry point is located. Taking into account the difference in surrounding voxels, it can be used This step size is more reasonable.

Step 8040, calculate the first function term H _p (P').

In some embodiments, the processing device 120 may calculate the first function term H _p (P') of the new solution P' based on formula (1).

Step 8050, determine whether the system energy is reduced.

In some embodiments, the processing device 120 can compare the Hamiltonian H(P) of the solution P in the previous iteration with the Hamiltonian H(P') of the new solution P', and determine H(P') Whether it decreases relative to H(P). In some embodiments, the processing device 120 may compare the first function term H _p (P) of the solution P in the previous iteration with the first function term H _p (P') of the new solution P', and determine H Whether _p (P') decreases relative to H _p (P).

In some embodiments, in response to a decrease in system energy, processing device 120 may perform step 8070. In some embodiments, in response to the system energy not being reduced, processing device 120 may perform step 8060.

Step 8060, determine whether exp(-△H/T)>random(0, 1).

In some embodiments, the processing device can determine whether the Metropolis criterion is met, that is, whether exp(-ΔH/T)>random(0, 1) is true. Among them, △H is the difference between the Hamiltonian H(P') of the new solution P' and the Hamiltonian H(P _best ) of the current optimal solution P _best , T is the temperature of the iterative process, random (0, 1 ) is a random number of (0, 1). exp(-△H/T)>random(0,1) can represent the condition for accepting new solutions that perform poorly. As the number of iterations of the outer loop increases or decreases, the temperature gradually decreases from T ₀ to zero, and the second function term H _k also decreases. As the temperature decreases, the probability of its acceptance will become lower and lower, that is, the probability of "passing through" the potential barriers on both sides of the local optimal solution due to the quantum tunneling effect will become lower and lower.

In some embodiments, in response to exp(-ΔH/T)>random(0,1), processing device 120 may perform step 8070. In some embodiments, processing device 120 may perform step 8100 in response to exp(-ΔH/T)≤random(0,1).

Step 8070, P=P′.

In some embodiments, in response to the determination result of step 8050 or step 8060 being yes, record P=P′, and enter step 8080.

Step 8080, determine whether H(P')<H(P _best ).

In some embodiments, the processing device 120 may determine whether the Hamiltonian H(P′) of the solution P′ is smaller than the Hamiltonian H(P _best ) of the current optimal solution P _best .

In some embodiments, in response to H(P') < H(P _best ), processing device 120 may perform step 8090. In some embodiments, processing device 120 may perform step 8100 in response to H(P′)≥H(P _best ).

Step 8090, P _best =P′.

In some embodiments, in response to the determination result of step 8080 being yes, record P _best =P, and enter step 8100.

Step 8100, determine whether t≤N.

In some embodiments, in response to t≤N, processing device 120 may perform step 8110. In some embodiments, in response to t>N, processing device 120 may perform step 8120.

Step 8110, t=t+1.

In some embodiments, in response to the judgment result of step 8100 being yes, that is, t≤N, let t=t+1, and jump to step 8030 to enter the next inner loop.

Step 8120, determine whether s≤M.

In some embodiments, in response to s≤M, processing device 120 may perform step 8130. In some embodiments, in response to s>M, processing device 120 may perform step 8140.

Step 8130, s=s+1.

In some embodiments, in response to the determination result of step 8120 being yes, that is, s≤M, let s=s+1, and adjust to step 8020, Enter the next outer loop.

Step 8140: Determine the target puncture path.

In some embodiments, in response to the determination result of step 8120 being negative, the outer loop ends, and the processing device may determine the candidate puncture path corresponding to the optimal solution P _best as the target puncture path.

In some embodiments of this specification, the optimal solution is determined through a quantum annealing algorithm, and the candidate puncture path corresponding to the optimal solution is determined as the target puncture path, which can accurately and effectively plan the target puncture path and improve the efficiency of target puncture path planning. . Since target puncture path planning is a clinical multi-constraint optimization problem, the problem is a non-deterministic polynomial problem, that is, there is a definite answer, but the time complexity of obtaining the solution increases exponentially. Classic computers have calculation problems due to their own performance limitations. If the time is too long or the optimal solution cannot be reached, the technical solution of this embodiment can be run on a quantum annealing machine, such as mapping the Hamiltonian to the real qubits of the quantum annealing machine, so that The puncture path planning is more efficient and can achieve fast and accurate preoperative puncture path planning.

It should be noted that the above description of process 800 is only for example and illustration, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 800 under the guidance of this specification. However, such modifications and changes remain within the scope of this specification.

In some embodiments, taking any candidate needle entry point as the target search point, the processing device can calculate the path analysis result corresponding to the target search point according to the Hamiltonian based on the candidate puncture path corresponding to any candidate needle entry point; The optimal path analysis result is searched from multiple path analysis results, and the candidate puncture path corresponding to the optimal path analysis result is used as the target puncture path. For example, the processing device 120 may calculate the Hamiltonian corresponding to all candidate puncture paths, and determine the candidate puncture path with the smallest Hamiltonian as the target puncture path.

In some embodiments, the processing device 120 may determine the target puncture path through other preset search algorithms. For example, genetic algorithm, simulated annealing algorithm, etc.

In some embodiments, when interventional surgery is used for ablation treatment, the target parameters may include target puncture path, target stop point location, and target ablation sphere parameters. The processing device 120 may adopt the process of FIG. 9 or FIG. 13 to determine the target parameters.

Figure 9 is an exemplary flowchart of determining target parameters according to some embodiments of the present specification. As shown in Figure 9, process 900 includes the following steps. In some embodiments, one or more operations of the process 900 shown in FIG. 9 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 900 shown in FIG. 9 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

Step 910: Generate at least one set of planning parameters based on the patient data.

Planning parameters refer to a set of parameters that can be selected as an interventional device to perform an ablation treatment procedure. In some embodiments, the planning parameters may include the number of puncture paths n, the needle entry point set {P _i1 , _Pi2 , ..., P _in } and the interventional needle target point set {P _j1 , P _j2 , ..., P _jn } The puncture path consisting of {P _i1j1 , P _i2j2 ,..., P _injn }, at least one stop point position on each puncture path {T _1p1 , T _2p1 ,..., T _kpn }, optional ablation ball parameters {R ₁ , R ₂ ,…, R _s } etc. Different ablation power and ablation time correspond to different ablation sphere parameters. The greater the ablation power, the longer the ablation time, and the larger the ablation sphere corresponding to the ablation sphere parameters.

Among them, P _injn represents the nth path composed of the needle entry point P _in and the target point of the interventional needle P _jn ; T _kpn represents the kth stop point position on the nth path; R _s represents the sth power-time The corresponding ablation sphere parameters (for example, if the ablation sphere is an ellipsoid, R _s can be an ellipsoid with a length of 36 × 36 × 42 mm in its major and minor axes), s represents the number of different ablation sphere parameters for selection. In some embodiments, the ablation sphere may be in the shape of an ellipsoid, a sphere, or the like. In some embodiments, ablation ball parameters corresponding to different stop point positions on the same puncture path may be the same.

For example, as shown in Figure 10, the dotted area is the lesion, and a set of planning parameters may include the number of puncture paths n=2, the set of needle entry points of the puncture path as {P _i1 , P _i2 }, and the set of target points of the interventional needle as { The puncture path {P _i1j1 , P _i2j2 } composed of P _j1 , P _j2 }, the puncture path P _i1j1 has a stay point position {T _1p1 , T _2p1 }, the puncture path P _i2j2 has a stay point position T _1p2 , optional Ablation sphere parameters {R ₁ , R ₂ , R ₃ }.

In some embodiments, the processing device 120 can perform three-dimensional reconstruction of patient data to obtain a three-dimensional medical image, and determine structural features based on the three-dimensional medical image, thereby determining at least one set of planning parameters based on the three-dimensional medical image. Regarding three-dimensional medical imaging, see Figure 4 and its related description for more information on structural characteristics.

In some embodiments, the processing device 120 may determine planning parameters based on the three-dimensional medical images in various ways. For example, the processing device 120 may use a path composed of any two points in the area of interest as a puncture path, randomly select at least one point on each puncture path as at least one stop point position on each puncture path, and select Select ablation sphere parameters to determine the ablation sphere parameters. For another example, doctors can manually input planning parameters based on three-dimensional medical images. For another example, the processing device 120 may use actual parameters corresponding to three-dimensional medical images similar to the three-dimensional medical images in the historical database as planning parameters.

In some embodiments, the processing device 120 can also perform a preprocessing operation on the three-dimensional medical image, where the preprocessing operation includes one or more of region of interest cropping, data point downsampling, and blood vessel coarse subdivision grading; and then based on the preprocessing The results obtained from the operation determine at least one set of planning parameters.

Region-of-interest cropping refers to cropping three-dimensional medical images based on the location of the lesion, thereby limiting the effective range of the needle entry point. For example, the processing device 120 can select upper and lower adjacent Lmm (for example, 10mm, 20mm, 30mm, etc.) slices for needle entry point planning based on the target area bounding box or the target area centroid position. For another example, the processing device 120 may crop a 1/4 area of the skin contour bounding box as the effective needle entry point area based on the skin contour bounding box and the center position of the target point. For another example, the processing device 120 may use the above two methods to crop the area of interest.

Data point downsampling can sparse the higher-precision data points in three-dimensional medical images, thereby sparse points in the lesion area and skin points. Methods for downsampling data points include but are not limited to resampling, interpolation, etc.

Blood vessel thickness grading refers to grading blood vessels in three-dimensional medical images according to their diameter. Blood vessels of different thicknesses have different effects on ablation. The blood vessel thickness classification allows the path planning to be processed according to the results of different blood vessel thickness classifications to avoid the puncture path passing through thicker blood vessels and the ablation field causing damage to thicker blood vessels. In some embodiments, the processing device 120 can distinguish blood vessels in three-dimensional medical images, construct blood vessel tree branches through graph theory, and calculate blood vessel segment diameters through blood vessel center lines to complete blood vessel thickness classification. In some embodiments, the results of the blood vessel thickness classification can be expressed in grades based on the diameter of the blood vessel segment. For example, the blood vessel diameter can be divided into level 1, level 2, level 3, etc. from small to large according to the diameter of the blood vessel.

In some embodiments, the processing device 120 can determine the needle entry point set and the target point set in the planning parameters based on the downsampling results of data points in the cropped region of interest obtained by the preprocessing operation (for example, the downsampling results are obtained by The data points are used as needle entry point set, target point set, etc.), you can filter out the puncture paths that will pass through thicker blood vessels based on the blood vessel thickness classification results, and set the ablation range composed of optional ablation sphere parameters so as not to affect thicker blood vessels. within the range of blood vessel damage.

In some embodiments of this specification, by preprocessing three-dimensional medical images, the pressure brought by the determination process of target parameters on the performance of the processing equipment can be alleviated, while the efficiency of optimizing the ablation process can be improved.

Step 920: Determine target parameters based on at least one set of planning parameters.

In some embodiments, the target parameters may include target puncture path, target stop point location, and target ablation sphere parameters. For more information on target parameters, target puncture path, target stop point location, and target ablation sphere parameters, see Figure 3 and its related descriptions.

In some embodiments, the processing device 120 may generate a set of individuals based on the individual generator. Perform at least one iterative update on the individual collection until the first iteration completion condition is met. Based on the updated individual set, at least one set of intermediate parameters is determined, and then based on at least one set of intermediate parameters, the target parameters are determined. For more information on determining target parameters, see Figure 11 and its associated description.

In some embodiments of this specification, by determining the target parameters from the planning parameters, more reasonable target parameters can be obtained to improve the doctor's surgical efficiency and ease the doctor's work difficulty.

It should be noted that the above description of process 900 is only for example and illustration, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 900 under the guidance of this specification. However, such modifications and changes remain within the scope of this specification. For example, one or more of the preprocessing operations in step 910 may be omitted, or all of them may be omitted.

In some embodiments, the processing device 120 may first determine the target puncture path based on the process of FIG. 4 , and then determine the target parameters other than the target puncture path (ie, the target stay point position and the target ablation sphere parameters) based on the process of FIG. 9 . In step 910, since the target puncture path has been determined, the processing device 120 can generate at least one set of planning parameters only by changing the stay point position and the ablation ball parameters.

Figure 11 is another exemplary flowchart for determining target parameters illustrated in some embodiments of the present specification. As shown in Figure 11, process 1100 includes the following steps. In some embodiments, one or more operations of the process 1100 shown in FIG. 11 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 1100 shown in FIG. 11 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

Step 1110: Generate an individual set based on the individual generator.

An individual refers to a specific subject corresponding to the planning parameters. For example, each individual can correspond to a set of planning parameters, that is, each individual can include a set of puncture paths, ablation sphere parameters, stay point locations and other parameters.

In some embodiments, the individual generator can generate at least one individual based on certain rules and calculate individual attributes of each of the at least one individual. The individual attributes may include the individual's first evaluation value, first constraint determination value, etc. For more details about the first evaluation value and the first constraint determination value, please refer to step 1130 and its related description.

An individual collection refers to a collection of multiple individuals. For example, a set of multiple planning parameters.

In some embodiments, processing device 120 may generate a set of individuals based on the individual generator in a variety of ways. Just as an example, for example, the individual generator can randomly select the needle entry point and the target point from the needle entry point set and the target point set to pair up to form at least one puncture path, and determine the stop point position according to certain intervals on each puncture path. (such as at the midpoint, trisection point, etc. of the intersection between the puncture path and the lesion area), and randomly select the ablation sphere parameters from the optional ablation sphere parameters, and add at least one path with a stay point position and ablation sphere parameters. Randomly combine to form multiple individuals containing at least one puncture path, and the multiple individuals constitute an individual set. For another example, the individual generator can also use the actual parameter corresponding to at least one patient data whose similarity is higher than the similarity threshold in the historical database as an individual to construct an individual set. Among them, the similarity is the similarity between the three-dimensional medical image corresponding to the patient data and the three-dimensional medical image corresponding to the patient data in the historical database. To a certain extent, the similarity threshold can be set manually.

Step 1120: Generate at least one new individual based on the individual generator, and add at least one new individual to the individual set.

A new individual refers to an individual whose at least one parameter is different from any of the original individuals in the individual set. For example, a new individual whose needle entry point is different from all the original individuals.

In some embodiments, the number of new individuals may be the same as the number of individuals in the original individual set, or may be different from the number of individuals in the original individual set.

In some embodiments, the individual generator can generate new individuals in the same manner as generating individuals in the individual set described in step 1110 above, which will not be described again here. In some embodiments, the individual generator can also generate a new individual by changing at least one parameter of the individuals in the collection (eg, needle entry point, target point, number of stay points, stay point locations, ablation sphere parameters, etc.) . For example, the individual generator can reduce the ablation sphere parameters of at least one individual by 20% to generate at least one new individual.

In some embodiments, the individual generator can also pair individuals with the same number of original puncture paths in the individual set, and use the paired individuals to generate new individuals through crossover operations, etc., where the crossover operation can include a single At least one of point crossover, multi-point crossover, uniform crossover, polynomial crossover, etc.

In some embodiments, the individual generator can generate a third individual with a higher probability and use the third individual as a new individual. Among them, the third individual is an individual of the same dimension that has the highest proportion in the current individual set.

The third individual refers to the individual with the same latitude as the dimension with the highest proportion in the current individual set. For example, if the proportion of individuals of a certain dimension in the current individual set is 60%, and the individuals of other dimensions are less than 60%, then the dimensions of the third individual are the same as the above dimensions.

Dimension refers to the number of certain parameters in the planning parameters corresponding to the individual. For example, the number of puncture paths, the number of stopping points on each puncture path, etc. in the individual planning parameters.

The same dimension means that the dimensions of individuals are exactly the same. For example, the number of puncture paths for individual A is 2, and the number of stay points on each puncture path is 3. The number of puncture paths for individual B is also 2, and the number of stay points on each puncture path is also 3. Then the individual A and individual B have the same dimensions.

In some embodiments of this specification, the individual generator generates the third individual with a higher probability, which can stabilize the dimensions of the individuals in the individual set, improve the efficiency of iterative updates, and at the same time promote the individual generator to generate a higher probability. The probability of generating a new individual based on the current set of individuals.

In some embodiments, in response to satisfying the preset conditions, the individual generator may also generate new individuals based on the current individual set with a higher probability.

Preset conditions refer to the requirements that a collection of individuals needs to meet. For example, the preset condition may be that the number of iterative updates of the individual collection is greater than the iteration number threshold (such as 500 times, etc.). For another example, the preset condition may also be that the proportion of individuals in a certain dimension in the individual collection is higher than a proportion threshold (such as 80%).

In some embodiments, in response to satisfying the preset conditions, the individual generator can generate a new individual based on at least one current individual through a certain evolutionary method. For example, new individuals are generated through evolutionary methods such as genetic mutation and particle swarm, rather than randomly generating new individuals. When the individual set meets the preset conditions, the dimensions of the individuals in the individual set tend to be the same. At this time, the individual corresponding to the dimension with the largest proportion in the individual set has a greater probability of being a better individual. The individual generator can evolve to generate new individuals based on individuals of this dimension with a higher probability.

In some embodiments, the individual generator can generate new individuals based on current individuals with a higher probability through various methods. For example, the individual generator can pair individuals corresponding to the highest proportion of dimensions, and then exchange at least one parameter (such as needle entry point, ablation sphere parameters, etc.) of the paired individuals to form two new individuals. For another example, the individual generator can also allow the new individual generated after at least one parameter exchange to randomly change at least one of its parameters (such as the number of stay points, the location of the stay points, etc.) according to a certain probability to generate a new individual. For another example, the individual generator can directly randomly change at least one parameter of the individual corresponding to the dimension with the highest proportion (such as the number of stay points, the location of the stay points, etc.) to generate a new individual.

In some embodiments of this specification, the individual generator generates a new individual based on the current individual with a higher probability, so that the generated new individual can retain some relatively reliable parameters that still survive after multiple iterative updates, defining The variation range of parameters during iterative update keeps the quality of the individual collection relatively stable during iterative update and obtains high-quality individuals faster.

In some embodiments, the individual generator can add new individuals to the individual set based on certain rules. For example, an individual generator can add all new individuals generated to the individual collection. For another example, the individual generator can also randomly select new individuals with the same number as the number of individuals in the current individual set to join the individual set.

In some embodiments, the processing device 120 may add individuals whose first constraint determination value is greater than the determination value threshold (eg, -1, -2, -3, etc.) to the individual set based on the individual generator. Among them, the judgment value threshold can be set manually. For more information about the first constraint determination value, please refer to step 1140 and its related description.

In some embodiments, the processing device 120 may filter individuals in the individual set based on the individual filter to update the individual set. The screening may include performing the selection operations of steps 1130 to 1140 on the individuals in the individual collection.

Step 1130: Calculate the first evaluation value and the first constraint determination value of each individual in the individual set.

The first constraint judgment value refers to a numerical value or letter that can reflect the degree of compliance of an individual with the constraint conditions. For example, the first constraint determination value can be represented by a numerical value between 1-10, or letters a-f, or a star rating. The larger the value, the greater the dictionary sorting, or the higher the star rating, the higher the degree of compliance.

In some embodiments, the first constraint decision value may be determined based on the decision value of at least one constraint condition. Among them, at least one constraint may include whether the instrument length meets the requirements.

Constraints refer to conditions that restrict the selection of planning parameters. For example, the length of the interventional needle will limit the length of the puncture path and thus the location of the needle entry point and target point.

In some embodiments, the constraints may include whether the instrument length meets the requirements, the puncture does not penetrate the risk structure, the ablation field completely covers the lesion, the ablation electrode distance is reasonable and the angle between the puncture paths is reasonable, etc.

The judgment value of a constraint condition refers to a numerical value or letter that can reflect whether the constraint condition is satisfied. For example, when a certain constraint is satisfied, the judgment value is assigned a value of 0, and when the constraint is not satisfied, the judgment value is assigned a value of -1.

Whether the instrument length meets the required constraints refers to the limitation of the instrument length on the length of the puncture path. The length of the puncture path cannot exceed the length of the instrument, otherwise the interventional needle may not be able to reach the target point and/or the stop point, resulting in the inability to perform ablation treatment on the lesion area. For example, when the instrument length is 80mm, the length of the puncture path cannot exceed 80mm, otherwise the instrument length constraint cannot be met. In some embodiments, the instrument length may be the length of the interventional needle.

In some embodiments, the processing device 120 can calculate the length of the path P injn composed of the needle entry point P _in and the target point P _jn of each puncture path in the individual based on the individual generator, and compare the _{length of P injn} _with the instrument length. Compare to determine whether the instrument length constraint is met. When the instrument length constraint is satisfied, the judgment value of the constraint is 0; when the instrument length constraint is not satisfied, the judgment value is -1.

The constraint that the puncture does not penetrate the risk structure refers to the restriction of the puncture path by the dangerous tissue. Dangerous tissues can include thicker blood vessels, vital organs, etc. When an interventional needle enters human tissue according to the puncture path and performs ablation treatment on the lesion, it will cause damage to healthy tissues other than the lesion. When dangerous tissue is damaged, it may threaten the patient's life.

In some embodiments, the processing device 120 may characterize the structure to determine whether at least one puncture path of the individual punctures but does not penetrate the risk structure. See Figure 4 and its associated description for more information on structural features. When the puncture path does not interfere with dangerous tissues in the three-dimensional medical image, the constraint is satisfied, and the judgment value is 0. When the puncture path interferes with at least one dangerous tissue, the constraint is not satisfied, and the judgment value is - 1.

The constraint that the ablation area completely covers the lesion is the restriction on the scope of the ablation field. When the ablation field completely covers the lesion area, the lesion can be completely ablated and treated. In some embodiments, the processing device 120 can combine the individual's puncture path, stop point location, and ablation sphere parameters with three-dimensional medical images to determine whether all ablation field ranges corresponding to the individual completely encompass the lesion area. When the ablation field range completely surrounds the lesion area, the constraint that the ablation area completely covers the lesion is satisfied. The corresponding judgment value is 0, otherwise, the corresponding judgment value is -1.

The constraint on the rationality of the ablation electrode distance is the restriction on the ablation electrode distance between multiple puncture paths. When multiple interventional needles perform ablation treatment simultaneously along the puncture path, if the electrodes in the ablation electrode segment are too close, arcs may occur, posing a threat to patient safety.

In some embodiments, if the number of puncture paths is greater than or equal to 2, the processing device 120 can calculate the shortest distance between the ablation electrode segments of different puncture paths in the individual. When the shortest distance is greater than the distance threshold, the puncture path ablation electrode distance is reasonable. The constraint condition is satisfied, and the judgment value is assigned a value of 0. On the contrary, the constraint condition of the reasonable distance of the ablation electrode on the puncture path is not satisfied, and the judgment value is assigned a value of -1. The shortest distance refers to the shortest distance between at least two ablation electrode segments corresponding to different puncture paths. The distance threshold can be set according to the power parameters and usage specifications of the interventional device.

The constraints on the rationality of the safety of the angles between puncture paths refer to the constraints on the angles of the puncture paths under multiple puncture paths. When multiple interventional needles perform ablation treatment simultaneously along the puncture path, if the angle between the ablation electrode segments is too large, arcs may easily occur, posing a threat to patient safety.

In some embodiments, the processing device 120 can project different puncture paths in an individual onto the same plane and calculate the included angle of the projection. When the included angle is less than the included angle threshold, the constraints on the rationality of the angle between the puncture paths are safe. is satisfied, the corresponding judgment value is assigned to 0, otherwise, the judgment value is assigned to -1.

In some embodiments, the processing device 120 may calculate the sum of the determination values of different constraint conditions, and use the sum of the determination values of the constraint conditions as the first constraint determination value.

The first evaluation value refers to a numerical value that can reflect the quality of an individual.

In some embodiments, the first evaluation value may include the number of puncture paths and the ablation conformation rate.

Due to factors such as the size and location of the lesion, one puncture path may not be able to achieve complete ablation, and multiple combined needle ablation methods need to be considered. The number of puncture paths refers to the number of interventional needle paths included in an individual. When an interventional needle enters the human body, it will cause more or less damage to the human body. The smaller the number of individual puncture paths, the smaller the damage to the patient's body, which can be used to evaluate the individual's pros and cons. For example, if the number of puncture paths for individual A is 2 and the number of puncture paths for individual B is 3, then individual A will cause less damage to the patient's body due to the smaller number of puncture paths.

The ablation conformity rate refers to the percentage of the lesion volume to the total planned ablation volume under complete ablation conditions. For example, under complete ablation conditions, The lesion volume is 1.5 cubic centimeters and the total planned ablation volume is 2 cubic centimeters, so the ablation conformity rate is 75%. Among them, the total planned ablation volume is the volume of the union of the planned ablation volumes corresponding to all stay point positions. For example, as shown in Figure 10, the planned ablation volume corresponding to the stay point position T _1p1 is 2 cubic centimeters, the planned ablation volume corresponding to the stay point position T _2p1 is 2.5 cubic centimeters, and the planned ablation volume corresponding to the stay point position T _1p2 is 2 cubic centimeters. The planned ablation volumes corresponding to the stay point positions T _1p1 and T _2p1 have an overlap of 1 cubic centimeter (i.e., the diagonal line part). The planned ablation volumes corresponding to the stay point position T _1p2 have no overlap, so the total planned ablation volume is 2+2.5- 1+2=5.5 cubic centimeters. The greater the ablation conformity rate, the smaller the damage to surrounding tissue caused by the ablation treatment surgery, and the more it meets the treatment expectations.

Step 1140: Select individuals based on the first evaluation value and the first constraint determination value, and determine an updated individual set.

In some embodiments, when the number of new individuals is the same as the number of individuals in the original individual set, the individual filter can randomly select two individuals from the individual set added to the new individual as two matching individuals, And select one of the individuals based on the first evaluation value and the first constraint judgment value, then randomly select two individuals from the remaining individual set and select one of them, and so on, all the selected individuals are collected as an update The subsequent collection of individuals. In some embodiments, the processing device 120 may calculate the first evaluation value and the first constraint determination value of the individuals in the individual set, select the individual based on the first evaluation value and the first constraint determination value, and determine the updated individual set. For more information about the individual filter selecting individuals based on the first evaluation value and the first constraint determination value, see Figure 12 and its related description.

Step 1150: Determine whether the first iteration completion condition is met. In response to the first iteration completion condition being satisfied, the processing device may perform step 1160; in response to the first iteration completion condition being not satisfied, the processing device 120 may perform step 1120.

The first iteration completion condition refers to the condition that needs to be met for the individual collection to stop iterative update. For example, the number of iterative updates reaches the maximum number threshold (such as 500 times, 1200 times, 2000 times, etc.). For another example, the proportion of individuals of a certain dimension in the individual set exceeds the maximum threshold (such as 80%, 85%, 90%, etc.). For another example, the proportion difference of the individuals with the highest proportion in the individual set is less than the minimum threshold (such as 1%, 2%, 3%, etc.).

In some embodiments, while the processing device 120 iteratively updates the individual set, it counts each iterative update. For example, when the first update of the individual set is completed, the count is 1; when the second update of the individual set is completed, the count is 2. When the count exceeds the maximum times threshold, it is determined that the first iteration completion condition is met. In some embodiments, after each iteration update is completed, the processing device 120 can calculate the proportion of individuals in each dimension in the current individual set. When the proportion of individuals in a certain dimension exceeds the maximum threshold, it is determined that the first iteration completion condition is met. . In some embodiments, after each iterative update is completed, the processing device 120 can calculate the difference between the individual proportion of the dimension with the highest proportion and the proportion when the last iterative update is completed. When the proportion difference is less than the minimum threshold, It is determined that the first iteration completion condition is met.

In some embodiments, in response to the first iteration completion condition being satisfied, the processing device 120 may determine the target parameter by performing step 1160 . In some embodiments, in response to the first iteration completion condition being not satisfied, processing device 120 may continue to step 1120 .

Step 1160: Determine at least one set of intermediate parameters based on the updated individual set.

The intermediate parameters refer to a better set of planning parameters for selection.

In some embodiments, the processing device 120 may determine a Pareto front solution based on the updated set of individuals, using the Pareto front solution as at least one set of intermediate parameters.

The Pareto front solution refers to a set of individuals that are not completely dominated by other individuals. In some embodiments, the processing device 120 may compare the first evaluation value of any individual in the individual set with the first evaluation values of other individuals to see whether there is any indicator in the first evaluation value of some other individual ( For example, the number of puncture paths, ablation conformity rate) are better than any one individual, and the first evaluation value in response to the absence of some other individual is better than any one individual (that is, the number of puncture paths for any one individual) , at least one of the ablation conformal rates is better than other individuals in the individual set), then any individual is determined to be one of the Pareto front solutions. For example, if the number of puncture paths for individual A is 3 and the ablation conformity rate is 90%, and there is no other individual in the individual set with a number of puncture paths less than 3 and an ablation conformity rate higher than 90%, then there is no The first evaluation value of some other individuals is better than that of individual A, then individual A is determined to be one of the Pareto front solutions.

In some embodiments of this specification, multiple individuals in the Pareto front solution are used as intermediate parameters for selection, which allows users to select the most appropriate target parameters according to their own surgical habits and improves the universality of the puncture path planning scheme. sex.

Step 1170: Determine target parameters based on at least one set of intermediate parameters.

In some embodiments, the processing device 120 may use the intermediate parameter with the highest ablation conformation rate among at least one set of intermediate parameters as the target parameter.

In some embodiments, the processing device 120 may send at least one set of intermediate parameters to the user terminal used by the doctor for the doctor to independently select, and use the intermediate parameters independently selected by the doctor as target parameters.

In some embodiments, when the processing device 120 sends at least one set of intermediate parameters to the doctor, the doctor may also choose to automatically recommend the target parameters. When the doctor chooses to automatically recommend target parameters, the processing device 120 can calculate the damage value F of at least one set of intermediate parameters, and recommend the intermediate parameter with the smallest damage value F to the user as the target parameter. Among them, the damage value F can be calculated using formula (8):
F＝a*nb*η+c*L (8)

Among them, n is the number of puncture paths, eta is the ablation conformation rate, L is the total length of all puncture paths in the same set of intermediate parameters; a is a larger constant, used to make the damage value as close as possible to the number of puncture paths. , the smaller the number of puncture paths, the smaller the damage to healthy tissue, the smaller the damage value F, and the easier it is to be used as a target parameter; b is a parameter that can balance the number of puncture paths and the ablation conformity rate. When puncture When the number of paths is the same, the smaller the ablation conformation rate, the greater the damage value F, and the greater the damage to healthy tissue, which will not be used as a target parameter; the coefficient c is used when the number of puncture paths is the same and the ablation conformation rate is similar. Optimize the control parameters under the conditions, and optimize the path with a shorter total puncture path to reduce patient injury and intraoperative risks. By automatically recommending target parameters, the optimal intermediate parameters are directly obtained, which improves the doctor's surgical efficiency.

In some embodiments of this specification, the individual generator generates individual sets and new individuals, and the Pareto front solution is used as at least one set of intermediate parameters, which can expand the scope of searching for optimal individuals and screen out relatively small global ones. Optimal solution, thereby avoiding the final confirmed target parameters are only local optimal parameters.

It should be noted that the above description of process 1100 is only for example and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process 1100 under the guidance of this description. However, such modifications and changes remain within the scope of this specification. For example, when performing iterative update, you can choose one round of iterative update, that is, iteratively update the individual collection by generating 10 new individuals each time, for a total of 10 iterative updates, and merge them to directly generate 100 at a time. The individual filter is directly generated from Determine at least one set of intermediate parameters among 100 individuals to achieve one-time iterative update.

Figure 12 is an exemplary flow diagram of an individual filter screening process according to some embodiments of the present specification. As shown in Figure 12, process 1200 includes the following steps. In some embodiments, one or more operations of the process 1200 shown in FIG. 12 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 1200 shown in FIG. 12 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

Step 1210: Calculate the first evaluation value and the first constraint determination value WA, WB of the two matching individuals _A and _B.

In some embodiments, when the individual generator calculates the individual attributes of each of at least one individual, the individual filter can directly call the first of the two matching individuals A and B in the individual attributes calculated by the individual generator. Each index in the evaluation value (for example, the number of puncture paths, ablation conformity rate) and the first constraint determination value W _A of individual A and the first constraint determination value WB of individual _B are sufficient. In some embodiments, when the individual generator does not calculate the individual attributes of each of the at least one individual, the individual filter may calculate each indicator in the first evaluation value of the two matching individuals A and B and the individual A's first evaluation value. A constraint determination value W _A and the first constraint determination value _WB of individual B.

Step 1220, determine whether W _A < _WB .

In some embodiments, in response to _WA < _WB , processing device 120 may perform step 1260 of selecting individual B. In some embodiments, in response to W _A ≥ _WB , processing device 120 may proceed to step 1230 .

Step 1230, determine whether _WA = _WB .

In some embodiments, in response to W _A = _WB , processing device 120 may proceed to step 1240 . In some embodiments, in response to WA>WB, processing device 120 may perform step 1280 of selecting individual A.

Step 1240, determine whether individual A completely dominates individual B.

In some embodiments, in response to individual A completely dominating individual B, processing device 120 may perform step 1280 of selecting individual A. In some embodiments, in response to individual A not fully dominating individual B, processing device 120 may proceed to step 1250.

Complete dominance means that the first evaluation value of an individual is better than that of another individual. For example, when the number of puncture paths for individual A is 2 and the ablation conformity rate is 95%, and the number of puncture paths for individual B is 3 and the ablation conformity rate is 85%, the number of puncture paths and the ablation conformity rate for individual A are both equal. is better than individual B, then individual A is said to completely dominate individual B.

Step 1250, determine whether individual B completely dominates individual A.

In some embodiments, in response to individual B not completely dominating individual A, processing device 120 may perform step 1270 of randomly selecting. In some embodiments, in response to individual B completely dominating individual A, processing device 120 may perform step 1260 of selecting individual B.

Step 1260, select individual B.

Step 1270, random selection.

In some embodiments, after step 1250 determines, in response to individual B not completely dominating individual A, that is, when two matching individuals do not completely dominate each other (that is, individual A does not completely dominate individual B, and individual B does not completely dominate individual A), the processing device 120 may determine the updated individual set based on the number of individuals that the two matching individuals each fully dominate in the individual set.

Matching individuals refer to two individuals randomly selected from the individual set to which the new individual is added. For more information about pairwise matching, please refer to step 1140 and its related description.

In some embodiments, the processing device 120 may determine the number of individuals that are completely dominated by each of the two matching individuals by determining the relationship between the first evaluation values of the two matching individuals and other individuals in the individual set, and assign the respective The set of individuals is updated as the proportion of the number of fully dominated individuals to the total number of individuals dominated by the two matched individuals as the probability that the two matched individuals are retained. For example, individual C and individual D are two matching individuals, and one of them needs to be selected as the individual in the updated individual set. Assume that individual C completely dominates 10 individuals in the current individual set, and individual D is in the current individual set. If 5 individuals are completely dominated, the probability of individual C being retained is 10/15=2/3, so the probability of individual D being retained is 1/3.

In some embodiments of this specification, by determining the updated individual set based on the number of fully dominated individuals, better individuals can be retained with a higher probability for the next iterative update, ensuring the success of multiple rounds of iterative updates. Progress and quality.

Step 1280, select individual A.

In some embodiments of this specification, by updating the individual set based on the first constraint determination value and the first evaluation value through the individual filter, individuals with better effects and the lowest damage to patients can be screened out to enter the next round of iterative updates. , prompting iterative updates to obtain the target parameters most suitable for the patient's condition faster, thereby improving the doctor's surgical efficiency and reducing the difficulty of the doctor's work.

It should be noted that the above description of process 1200 is only for example and illustration, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process 1200 under the guidance of this description. However, such modifications and changes remain within the scope of this specification.

In some embodiments, steps 1220 to 1230 may be omitted, that is, the processing device 120 may determine whether individual A completely dominates individual B based only on the first evaluation values of the two matching individuals A and B. In response to individual A completely dominating individual B, the processing device 120 may perform step 1280, that is, selecting individual A; in response to individual B completely dominating individual A, the processing device 120 may perform step 1260, that is, selecting individual B; in response to individual A and individual B are not completely dominated by each other, and the processing device 120 can perform step 1270, that is, perform random selection. When making a selection based only on the first evaluation values of the two matched individuals A and B, the processing device 120 can determine whether the length of the instrument meets the requirements, the puncture does not penetrate the risk structure, the ablation field range completely covers the lesion, and the ablation electrode is used in the pre-processing stage. Constraints such as the rationality of distance and the safety of angles between puncture paths restrict the selection range of planning parameters to ensure that individuals in the individual set satisfy the constraints.

Figure 13 is another exemplary flowchart for determining target parameters according to some embodiments of the present specification. As shown in Figure 13, process 1300 includes the following steps. In some embodiments, one or more operations of the process 1300 shown in FIG. 13 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 1300 shown in FIG. 13 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

Step 1310: Determine the number of puncture paths and the parameter range of the ablation sphere based on the patient data.

In some embodiments, the processing device 120 may determine the lesion volume based on the patient data and determine the number of puncture paths based on the lesion volume. The number of puncture paths is the same as the number of needle insertion points.

Lesion volume refers to the volume of the lesion area. For example, the volume of a patient's tumor.

In some embodiments, the processing device 120 may identify voxels of the lesion area based on the patient data and calculate the lesion volume based on the voxels of the lesion area. For example, the processing device 120 may determine the volume of the lesion area based on the number of voxel points in the target lesion area in the three-dimensional medical image of the patient data and the voxel equivalent value. Among them, the voxel equivalent value is the actual physical size corresponding to a voxel point. For more information on three-dimensional reconstruction, see Figure 5 and its related description.

In some embodiments, the processing device 120 may determine the number of puncture paths based on the relationship between the lesion volume and the volume threshold, where the volume threshold is a preset lesion volume. The volume threshold can be determined based on manual experience or historical data. For example, the volume threshold can include the maximum lesion volume that can be ablated by a single needle ablation in historical data, etc. In some embodiments, there may be multiple volume thresholds, and different volume thresholds correspond to different numbers of puncture paths. For example, the volume thresholds can be V ₁ , V ₂ , V ₃ ,..., when the lesion volume V satisfies 0＜V＜V ₁ , the number of puncture paths is 1; when the lesion volume V satisfies V ₁ ≤V＜V When ₂ , the number of puncture paths is 2; when the lesion volume V satisfies V ₂ ≤ V < V ₃ , the number of puncture paths is 3.

In some embodiments of this specification, by fully considering the volume of the lesion to determine the preferred number of puncture paths, the time for determining target parameters can be reduced and the efficiency of puncture path planning can be improved.

In some embodiments, the processing device 120 can determine the lesion mask based on the patient data, determine the long axis of the lesion and the short axis of the lesion based on the lesion mask, and then determine the ablation sphere parameter range based on the long axis of the lesion and the short axis of the lesion.

The lesion mask refers to the lesion area segmented from the patient data. For example, tumor regions segmented from patient data. In some embodiments, the lesion mask may be three-dimensional.

In some embodiments, the processing device 120 can segment the lesion area from the patient data to generate a lesion mask. For example, the processing device 120 may set voxel values in areas other than lesions in the patient data to 0, thereby generating a lesion mask.

In some embodiments, patient data can be processed using a mask extraction model to obtain a lesion mask. For example, patient data can be input into a mask extraction model, and the mask extraction model outputs a lesion mask. The mask extraction model may be a graph neural network (Graph Neural Network, GNN), a convolutional neural network (Convolutional Neural Networks, CNN), or a deep neural network (Deep Neural Networks, DNN), etc. The patient data in historical data can be used as training data to train the mask extraction model, so that the mask extraction model can output the lesion mask based on the patient data. The labels corresponding to the training data can be determined by manually annotated lesion masks.

The long axis of the lesion refers to the longest line segment connecting any two points on the boundary of the lesion area.

In some embodiments, the processing device 120 can obtain the peripheral boundary points of the lesion according to the lesion mask to form a set E. Next, the processing device 120 can randomly select two points A ₁ and A ₂ from the set E. A 1 and _{A 2} _are connected to form a space line segment, and calculate the length L ₁ of the line segment; fix point A ₁ and update A ₂ from E. point, form a new line segment, calculate the length of the line segment L ₂ ; repeat the above steps, update the A ₂ point, calculate the length of the line segment A ₁ A ₂ , until A ₂ has completely traversed the points in the set E; update A ₁ from the set E point, repeat the above steps until A ₁ has completely traversed the points in the set E; use the maximum value among the line segment lengths L ₁ , L ₂ ,..., L _n as the length l of the long axis of the lesion.

The short axis of the lesion refers to the longest line segment in the smallest projected area of the lesion area.

In some embodiments, as shown in Figure 14, the processing device 120 can obtain the peripheral boundary points of the lesion according to the lesion mask to form a set F; use a distance field calculation method to calculate the center point Y of the lesion mask (for example, the lesion center of mass, geometric center, etc.); treat Y as a radiation source and emit rays in all directions; take a ray I in a certain direction and calculate the projection H of the lesion on the plane G perpendicular to the ray I; calculate the minimum circumference of the projection H The diameter D _I of the circle C is such that the circle can just completely cover the projection of the lesion; repeat the above steps to obtain a set D of diameters of the smallest circles in all directions, and take the minimum value in the set D as the length d of the short axis of the lesion.

In some embodiments, the processing device 120 may determine the ablation sphere parameter range based on the long axis l of the lesion and the short axis d of the lesion according to the number of puncture paths n. For example, if the ablation ball parameters on multiple puncture paths are the same, when the number of puncture paths n=1, the value range of a is d<a<l; when the number of puncture paths n>1, the value range of a is For example, when n=1, if a>l, then the length of the short axis of the ablation sphere is greater than the long axis of the lesion, the ablation conformity rate is too small, and the damage to the surrounding tissue is greater; if a<d, then the length of the ablation sphere is The length of the short axis will inevitably not completely cover the lesion area, resulting in incomplete ablation. Therefore a needs to satisfy the above conditions.

In some embodiments, if the ablation ball parameters on multiple puncture paths are different. When the number of puncture paths is n≥1, the parameter range of ablation balls on multiple puncture paths can all be d<a<l. For example, when the number of determined puncture paths is 2 (satisfying n≥1), the short axis length of the ablation spherical ellipsoid of the first puncture path is a ₁ , and the short axis length of the ablation spherical ellipsoid of the second puncture path is is a ₂ , a ₁ and a ₂ can be different, the ablation sphere parameter range of the first puncture path can be d<a ₁ <l, and the ablation sphere parameter range of the second puncture path can be d<a ₂ <l.

In some embodiments of this specification, the range of ablation sphere parameters is limited by the long axis of the lesion and the short axis of the lesion, and the selection range of the ablation sphere parameters can be compressed to speed up the subsequent determination of target parameters through iterative optimization.

Step 1320: Determine at least one set of planning parameters based on the patient data, the number of puncture paths, and the ablation sphere parameter range.

It should be noted that, for the convenience of subsequent description, the identification of the needle insertion point and the target point in Figure 10 has been adjusted in Figure 16A. The needle insertion point P _i1 is adjusted to the needle insertion point P (i ₁ , j ₁ ). Point P _i2 is adjusted to needle insertion point P(i ₂ , j ₂ ), target point P _j1 is adjusted to target point Q ₁ , and target point P _j2 is adjusted to target point Q ₂ .

Planning parameters refer to a set of parameters that can be selected as an interventional device to perform an ablation treatment procedure. In some embodiments, the planning parameters include the needle entry point set {P(i ₁ , j ₁ ), P(i ₂ , j ₂ ), ..., P(i _n , j _n )} and the target point set {Q ₁ , Q ₂ ,..., Q _n } composed of puncture path P(i _n , j _n )-Q _n , system ablation ball parameters {R ₁ , R ₂ ,..., R _s }, each The stop point position T _kpn on the puncture path. Among them, each of the ablation sphere parameters available for selection by the system meets the ablation sphere parameter range requirements, and the value of n is determined based on the number of puncture paths (for example, when the number of puncture paths is 2, n=1, 2); stay point pn in position T _kpn is the n-th puncture path, and k is the k-th stop point position on the puncture path. For example, as shown in Figure 16A, the ablation sphere parameters on multiple puncture paths are the same, the dotted line area is the lesion, the number of puncture paths determined based on patient data is 2, and the ablation sphere parameter range is A set of planning parameters can include the needle entry point set {P(i ₁ , j ₁ ), P(i ₂ , j ₂ )} and the target point set {Q ₁ , Q ₂ } to form two puncture paths P(i ₁ , j ₁ )-Q ₁ , P(i ₂ , j ₂ )-Q ₂ , there are stay point positions T _1p1 , T _2p1 on the puncture path P(i ₁ , j ₁ )-Q ₁ , the puncture path P(i ₂ , there is a stay point position T _2p1 on j ₂ )-Q ₂ , and the optional ablation ball parameter is R. Among them, the length range of the minor axis a of R is: The size of the major axis c of R is determined based on the corresponding minor axis. For example, the size of the major axis c corresponding to different minor axes a can be determined by querying a comparison table. Among them, the comparison table can be obtained from the interventional equipment manufacturer.

In some embodiments, ablation ball parameters at different stop point locations on the same puncture path may be the same. For example, as shown in Figure 16A, the ablation ball parameters at the stay point positions T _1p1 and T _2p1 on the puncture path P(i ₁ , j ₁ )-Q ₁ may be the same.

In some embodiments, the processing device 120 may determine the needle entry point set {P(i ₁ , j ₁ ), P(i ₂ , j ₂ ), ..., P(in , j _n ) based on the skin area corresponding to the _lesion . )}. In some embodiments, the processing device 120 may use a set of points in the lesion area as a target point set {Q ₁ , Q ₂ , ..., Q _n }. In some embodiments, the processing device 120 can connect at least one line segment formed by connecting the point P( _in , _jn ) in the needle insertion point set and the point _Qn in the target point set as at least one puncture path P( _in ,j _n )-Q _n , and set the possible stay point position T _kpn and ablation ball parameter R on the puncture path. For example, set the stay point position equidistantly at the intersection of the puncture path and the lesion area, and randomly select the ablation ball parameter R . See Figure 15 and its associated description for more information on determining planning parameters.

Step 1330: Determine at least one set of feasible solutions based on at least one set of planning parameters.

In some embodiments, the processing device 120 can generate a planning set based on at least one set of planning parameters. The planning set includes multiple intermediate solutions, each intermediate solution corresponding to a set of planning parameters; and perform at least one round of iterative optimization on the planning set until The second iteration completion condition is met, and the optimal planning set is obtained; based on the optimal planning set, at least one set of feasible solutions is determined. See Figures 17-18 and their associated descriptions for more information on determining at least one set of feasible solutions.

Step 1340: Determine target parameters based on at least one set of feasible solutions.

In some embodiments, processing device 120 may determine target parameters based on at least one set of feasible solutions. For more information on determining target parameters, see Figures 17-18 and their related descriptions.

In some embodiments of this specification, at least one set of planning parameters is determined by determining the number of puncture paths and the parameter range of the ablation sphere, And by determining the target parameters based on at least one set of planning parameters, the range of planning parameters can be narrowed, and more reasonable target parameters can be obtained more quickly to improve the doctor's surgical efficiency and reduce the difficulty of the doctor's work.

It should be noted that the above description of process 1300 is only for example and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 1300 under the guidance of this description. However, such modifications and changes remain within the scope of this specification.

In some embodiments, the processing device 120 may first determine the target puncture path based on the process of FIG. 4 , and then determine the target parameters other than the target puncture path (ie, the target stop point position and the target ablation sphere parameters) based on the process of FIG. 13 . In step 1320, since the target puncture path has been determined, the processing device 120 can generate at least one set of planning parameters only by changing the stay point position and the ablation ball parameters.

In some embodiments, the user can manually draw the target puncture path, and then determine the target parameters other than the target puncture path (ie, the target stop point location and the target ablation sphere parameters) based on the process of FIG. 13 . In step 1320, since the target puncture path has been determined, the processing device 120 can generate at least one set of planning parameters only by changing the stay point position and the ablation ball parameters.

Figure 15 is an exemplary flowchart of determining planning parameters according to some embodiments of the present specification. As shown in Figure 15, process 1500 includes the following steps. In some embodiments, one or more operations of the process 1500 shown in FIG. 15 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 1500 shown in FIG. 15 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

Step 1510: Perform three-dimensional reconstruction on the patient data to obtain a three-dimensional medical image.

In some embodiments, the patient data may include CT or MR data of the patient. See Figure 3 and its associated description for more information on patient data.

In some embodiments, the processing device 120 can perform three-dimensional reconstruction of patient data to obtain a three-dimensional medical image, and determine structural features based on the three-dimensional medical image. For more information on three-dimensional medical images and structural features, see Figure 4 and its related descriptions.

Step 1520: Perform preprocessing operations on the three-dimensional reconstruction results.

In some embodiments, the processing device 120 can also perform preprocessing operations on the results of the three-dimensional reconstruction, where the preprocessing operations include one or more of region of interest cropping, data point downsampling, and blood vessel coarse subdivision grading. See Figure 9 and its associated description for more information on preprocessing operations.

Step 1530: Determine at least one set of planning parameters based on the results of the preprocessing operation, the number of puncture paths, and the ablation sphere parameter range.

In some embodiments, the processing device 120 may determine the needle entry point set based on results obtained from the preprocessing operation. For example, Figure 16B is a schematic diagram of a region of interest of a lesion in the lungs. The dotted area in the cross-sectional view of the patient's lungs is the patient's lesion, point Q is a point in the lesion area, and point Z is the surrounding body of the region of interest. The center point of a circle of skin outline, select the skin corresponding to the 1/4 area of interest that contains the most lesion areas as the needle entry point planning area, use the center point Z of the skin outline as the radiation source, and move towards the 1/4 area of interest. The side boundary emits a number of M rays uniformly (for example, the dotted line with an arrow in the figure), calculates the intersection point of the ray and the skin contour, and stores the intersection points in a counterclockwise order (for example, starting from the first point on the right in a counterclockwise direction) They are stored in sequence as P(1,1), P(2,1), P(3,1),...P(M,1). Among them, the center point Z can be determined by the center of mass solution. Then, in this Arrange N slices evenly in the vertical direction of the boundary. Repeat the above steps for different slices to find intersection points. An N×M matrix of needle entry points can be constructed, so that N×M needle entry point sets can be confirmed. For example, P(i, j) in the figure can be the coordinates of a needle entry point, where i is the i-th counterclockwise ray, and j represents the j-th layer.

In some embodiments, the processing device 120 may determine the target point set based on the results of downsampling the data points. For example, as shown in Figure 16B, the processing device can use the downsampled data points in the lesion area as target points _Qn , and multiple target points constitute a target point set { _Q1 , _Q2 ,..., _Qn }.

In some embodiments, the processing device 120 may determine the puncture path based on the results of the blood vessel coarse subdivision. For example, the processing device 120 can pair the points in the needle entry point concentration and the points in the target point concentration to form multiple line segments, and combine the line segments of the blood vessels that do not pass through the preset thickness level (for example, the blood vessel thickness level is level 2 and above). The puncture path as a planning parameter. In some embodiments, the processing device 120 may equidistantly set possible stay point locations based on the portion of the puncture path that intersects the lesion area. As shown in Figure 16A, the needle entry point set {P(i ₁ , j ₁ ), P(i ₂ , j ₂ )} and the target point set {Q ₁ , Q ₂ } form two puncture paths P( i ₁ , j ₁ )-Q ₁ , P(i ₂ , j ₂ )-Q ₂ , there are stay point positions T _1p1 , T _2p1 on the puncture path P(i ₁ , j ₁ )-Q ₁ , the puncture path P( There is a stay point position T _2p1 on i ₂ , j ₂ )-Q ₂ . In some embodiments, the processing device 120 can be based on selectable ablation sphere parameters and ensure that the ablation sphere parameters are within the above-determined ablation sphere parameter range while not causing damage to blood vessels of a preset thickness level.

In some embodiments of this specification, three-dimensional reconstruction is performed based on patient data to obtain three-dimensional medical images, and preprocessing operations are performed on the three-dimensional medical images to determine planning parameters. The planning parameters can be adaptively reduced according to different lesion conditions. scope, regulation The delineation parameters are more suitable for the lesion conditions of different patients and improve the efficiency of subsequent iterative optimization.

Figure 17 is an exemplary flowchart of determining target puncture parameters according to some embodiments of the present specification. As shown in Figure 17, process 1700 includes the following steps. In some embodiments, one or more operations of the process 1700 shown in FIG. 17 may be implemented in the application scenario 100 of the intervention planning system shown in FIG. 1 . For example, the process 1700 shown in FIG. 17 may be stored in a storage device in the form of instructions and called and/or executed by the processing device 120 .

Step 1710: Generate a planning set based on at least one set of planning parameters.

In some embodiments, the planning set includes a plurality of intermediate solutions, each intermediate solution corresponding to a set of planning parameters.

A planning set is a set composed of multiple intermediate solutions. Among them, the intermediate solution is the specific subject corresponding to the planning parameters. In some embodiments, the properties of the intermediate solution include a second constraint determination value and a second evaluation value. For more information about the second constraint determination value and the second evaluation value, please refer to step 1740 and its related description.

In some embodiments, the processing device 120 may randomly select intermediate solutions corresponding to a certain number (for example, 20 groups, 30 groups, etc.) of planning parameters to form a planning set based on at least one set of planning parameters. In some embodiments, the processing device 120 may select an intermediate solution from at least one set of planning parameters as a planning set based on a tournament selection strategy. For example, the processing device 120 may randomly select m (eg, 5, 10, etc.) intermediate solutions from at least one set of planning parameters, and use the optimal intermediate solution as the optimal intermediate solution based on the second evaluation value and the second constraint determination value. The planning set is repeated N times until the number of planning sets meets the quantity requirements, where the quantity requirements can be set in advance.

In some embodiments, the processing device 120 may perform step 1720 to perform at least one round of iterative optimization on the planning set until the second iteration completion condition is satisfied, thereby obtaining the optimal planning set.

Step 1720: Perform a transformation operation on the planning set to obtain a first preset number of new intermediate solutions.

The transformation operation refers to the operation of generating a new intermediate solution based on the intermediate solution in the planning set. For example, operations such as crossover operations and mutation operations are performed on the parameters of the intermediate solution to generate a new intermediate solution.

In some embodiments, the processing device 120 can pair the intermediate solutions in the planning set (for example, randomly pair them), and perform a crossover operation on the intermediate solutions in the two paired planning sets to obtain two new intermediate solutions. . Among them, the crossover operation refers to exchanging at least one parameter among the planning parameters corresponding to the paired intermediate solutions through single-point crossover, multi-point crossover, uniform crossover, arithmetic crossover, etc., to generate two new intermediate solutions. planning parameters. For example, at least one of the needle entry points, target points, ablation sphere parameters, etc. of two paired intermediate solutions is exchanged with each other to generate planning parameters corresponding to two new intermediate solutions.

In some embodiments, the processing device 120 can perform a mutation operation on the intermediate solutions in the planning set to obtain a new intermediate solution. The mutation operation refers to making certain changes to at least one parameter in the planning parameters corresponding to the intermediate solution in the planning set to obtain the planning parameters corresponding to the new intermediate solution. For example, at least one of the needle entry point, stay point position, target point, ablation sphere parameters, etc. of the intermediate solution in the planning set is changed.

In some embodiments, the processing device 120 may perform a mutation operation on at least one of the new intermediate solutions obtained by the crossover operation to obtain a mutated new intermediate solution.

In some embodiments, as shown in FIG. 16B , the processing device 120 may construct a set of confirmed N×M needle insertion points and determine the coordinates of the needle insertion points as (i, j). Next, the processing device 120 can calculate the minimum bounding box of the lesion through the patient data of the lesion (eg, CT, MR data, etc.) and determine the coordinates (x, y, z) of the lesion target in the XYZ direction. Further, the processing device 120 may sort the ablation spheres from small to large based on the ablation sphere parameter size, and set the ablation sphere encoding of each size to (0, s).

In some embodiments, when the number of puncture paths is 1, the processing device 120 may determine a set of planning parameters as (i, j, x, y, z, s), that is, there are 6 variables. In some embodiments, when the number of puncture paths is 2, the processing device 120 may determine a set of planning parameters as (i ₁ , j ₁ , x ₁ , y ₁ , z ₁ , s ₁ , i ₂ , j ₂ , x ₂ , y ₂ , z ₂ , s ₂ ), that is, there are 12 variables. It should be noted that, in order to simplify the complexity of problem design, there is no need to consider the situation that there are different stop point locations on the puncture path. The optimal stop point position that satisfies complete ablation can be determined based on the ablation sphere size and lesion volume without treating the stop point position as a variable.

In some embodiments, the processing device may use coding and interleaving methods to determine the new intermediate solution. For example, for two parent intermediate solutions, the processing device 120 may adjust the integer variable to a binary value encoding, and use an integer crossover operator to generate two offspring intermediate solutions. The integer values of the intermediate solution of the child and the intermediate solution of the parent remain consistent and will not generate points that deviate too far from the position of the current intermediate solution of the parent, thus achieving an iterative convergence effect.

In some embodiments, the processing device may adopt a mutation method to determine a new intermediate solution. For example, for a parent intermediate solution, the processing device 120 can adjust the integer variable to a binary value encoding, and randomly invert certain bits of the binary encoding by 0 and 1 to generate a new binary value, thereby forming the descendant intermediate solution. untie. Each variable in the planning problem is an integer or floating point real number type, and the polynomial mutation method is used to generate new individuals.

In some embodiments, the processing device 120 may perform transformation operations on the planning set in other ways to obtain a new intermediate solution, for example, perform transformation operations through polynomial crossover, polynomial mutation, etc. to obtain a new intermediate solution.

A new intermediate solution refers to an intermediate solution in which at least one of the planning parameters is different from the original intermediate solution in the planning set. For example, by crossing The new intermediate solution obtained by operation and mutation operation.

In some embodiments, the processing device 120 may select two intermediate solutions from the planning set based on the second constraint determination value and perform the crossover operation in the above manner to generate two new intermediate solutions. In some embodiments, the processing device 120 may select an intermediate solution from the planning set based on the second constraint determination value and perform a mutation operation in the above manner to generate a new intermediate solution. In some embodiments, the processing device 120 can perform crossover operations and mutation operations simultaneously to generate a new intermediate solution.

In some embodiments, the processing device 120 can also evolve to generate a new intermediate solution through other methods (such as particle swarm algorithm, etc.).

The first preset number refers to the number of new intermediate solutions set in advance. In some embodiments, the first preset number may be the same as the number of intermediate solutions in the planning set, or may be different from the number of intermediate solutions in the planning set.

Meeting the conditions means that the planning parameters corresponding to the new intermediate solution satisfy the conditions. In some embodiments, meeting the conditions includes that the needle entry point, target point, and puncture path meet the restriction range after the pretreatment operation, and the ablation sphere parameters satisfy the ablation sphere parameter range, wherein the restriction range includes that the needle entry point does not exceed the area of interest cropping The obtained 1/4 area of interest and puncture path do not pass through thick blood vessels, etc.

Step 1730: Add the new intermediate solution to the planning set to obtain a planning set with the new intermediate solution added.

Step 1740: Calculate the second evaluation value and the second constraint determination value of each intermediate solution added to the planning set of the new intermediate solution.

The second evaluation value refers to the value obtained by evaluating the quality of the intermediate solution. In some embodiments, the second evaluation value may be at least one of puncture path score and ablation conformity rate.

The puncture path score is to rate the quality of the puncture path.

In some embodiments, the processing device 120 may calculate the score S of the puncture path through formula (9):
S＝w ₁ ×L+w ₂ ×h+w ₃ ×c (9)

Among them, L is the length of the puncture path, h is the shortest distance between the puncture path and the dangerous organ, c is the number of slices spanned by the puncture path, w ₁ and w ₃ are coefficients less than 0, and w ₂ is a coefficient greater than 0. The specific values of w ₁ , w ₂ , and w ₃ can be set according to actual needs. When the puncture path L is shorter, the distance h between the puncture path and the dangerous organ is larger, and the number of slices c spanned by the puncture path is smaller, the harm to the patient is smaller, and it is easier to guide and observe the needle path during the operation. Therefore, the puncture path score S is positively correlated with h and negatively correlated with L and c. The higher the puncture path score, the better the puncture path.

In some embodiments, under complete ablation conditions, the processing device 120 may calculate the ablation conformity rate based on the lesion volume and the planned total ablation volume. Among them, the total planned ablation volume is the volume of the union of the planned ablation volumes corresponding to all stay point positions. See Figure 11 and its associated description for more information on ablation conformity rate.

In some embodiments, the processing device 120 may calculate the sum of the determination values of different constraint conditions, and use the sum of the determination values of the constraint conditions as the second constraint determination value. The second constraint determination value is similar to the first constraint determination value, and will not be described again here.

In some embodiments, the second evaluation value and the second constraint determination value may be determined in other ways, for example, the second evaluation value is determined based on the lesion coverage or the puncture path length, and the second constraint determination value is determined based on the ablation conformity rate, There are no restrictions here.

Step 1750: Select an intermediate solution based on the second evaluation value and the second constraint determination value, and obtain a new planning set containing a second preset number of intermediate solutions.

In some embodiments, the processing device 120 may select the intermediate solution based on the second evaluation value and the second constraint decision value of the intermediate solution being high. For more information on selecting an intermediate solution based on the second evaluation value and the second constraint decision value, see Figure 18 and its related description.

The second preset number refers to the number of new planning sets after selecting the intermediate solution. The second preset number is the same as the number of intermediate solutions in the planning set.

In some embodiments of this specification, selecting an intermediate solution through the second evaluation value and the second constraint determination value can ensure the feasibility of the planning parameters corresponding to the intermediate solution in each generation of planning set.

Step 1760: Determine whether the second iteration completion condition is met.

The second iteration completion condition refers to the condition used to determine whether the iteration optimization is completed. For example, the number of iterations reaches the maximum value, the second evaluation value reaches the preset evaluation threshold, etc.

In some embodiments, the processing device 120 may count each round of iterative optimization, starting from 0, and adding 1 to the count for each iteration. When the count reaches the maximum number of iterative optimization times, it is determined that the second iteration completion condition is met. In some embodiments, the processing device 120 may also compare the second evaluation value of the intermediate solution in the planning set after each round of iterative optimization with the preset evaluation threshold, when the second evaluation value is greater than or equal to the preset evaluation threshold. , determine that the second iteration completion condition is met.

In response to the second iteration completion condition being not satisfied, the processing device 120 may use the new planning set as the planning set, and continue to iteratively update the planning set by performing step 1720 until the second iteration completion condition is satisfied. In response to the second iteration completion condition being satisfied, processing device 120 may perform step 1770.

Step 1770: Obtain the optimal planning set, and determine at least one set of feasible solutions based on the optimal planning set.

The optimal planning set refers to the planning set after the second iteration completion condition is satisfied. For example, the planning set generated after the last iteration of optimization, the planning set containing intermediate solutions whose second evaluation value exceeds the preset evaluation threshold, etc.

In some embodiments, the processing device 120 may regard the planning set that satisfies the second iteration completion condition as the optimal planning set.

In some embodiments, the processing device 120 may determine the Pareto front solution based on the set of optimal plans as at least one set of feasible solutions. The process of determining feasible solutions is similar to the process of determining intermediate parameters. For more information on determining intermediate parameters, see Figure 11 and its related description.

Step 1780: Determine target parameters based on at least one set of feasible solutions.

In some embodiments, the processing device 120 may use the feasible solution with the highest ablation conformation rate among at least one set of feasible solutions as the target parameter. In some embodiments, the processing device 120 may send at least one set of feasible solutions to the user terminal used by the doctor for the doctor to independently select, and use the feasible solutions independently selected by the doctor as target parameters.

In some embodiments, the processing device 120 may determine the objective function values of at least one set of feasible solutions based on the puncture path score and the ablation conformation rate, and determine the target puncture parameters based on the objective function values.

The objective function value refers to the numerical value used to select the target puncture parameters. For example, values related to puncture path score and ablation conformity rate.

In some embodiments, the processing device 120 may calculate the objective function value F _max by formula (10):
F _max =z ₁ ×S+z ₂ ×η (10)

Among them, S is the puncture path score, eta is the ablation conformity rate, z ₁ and z ₂ are parameters greater than 0, which are used to balance the puncture path score and the ablation conformity rate.

In some embodiments, the processing device 120 may determine the puncture parameter with the largest objective function value as the target puncture parameter.

In some embodiments of this specification, the target parameter is determined through the objective function value, taking into account not only the ablation effect, but also the degree of harm to the patient, which can ensure that the obtained target parameter is the most balanced and suitable second evaluation value in all aspects. Parameters of the current lesion status.

In some embodiments of this specification, by performing multiple rounds of iterative optimization on the planning set to determine the optimal planning set, the optimal planning set can be searched globally to avoid the intermediate solutions in the optimal planning set from tending to the local optimum; through the optimal The optimal planning set determines the target parameters, and the most suitable target parameters can be obtained.

It should be noted that the above description of process 1700 is only for example and illustration, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 1700 under the guidance of this specification. However, such modifications and changes remain within the scope of this specification.

Figure 18 is a schematic flowchart of selecting an intermediate solution according to some embodiments of this specification.

In some embodiments, the processing device 120 may select an intermediate solution based on the second evaluation value and the second constraint determination value to obtain a new planning set including a second preset number of intermediate solutions. For example, the processing device 120 may hierarchize the planning set to which the new intermediate solution is added based on the second evaluation value and the second constraint determination value, and then determine the congestion distance of each intermediate solution based on the second evaluation value and the second constraint determination value. , thereby selecting a new planning set.

In some embodiments, the processing device 120 can divide the planning set into several layers according to the dominance relationship. The first layer is the Pareto front solution of the planning set, and the second layer is the result of removing the first layer of intermediate solutions from the planning set. The third level of the Pareto front solution obtained is the Pareto front solution obtained after removing the intermediate solutions of the first and second levels from the planning set, and so on. For example, as shown in FIG. 18 , the processing device 120 may divide the planning set to which a new intermediate solution is added into the first layer, the second layer, the third layer, etc. according to the above method.

In some embodiments, the processing device 120 may determine the selection function based on the second evaluation value and the second constraint decision value, and determine the crowding distance based on the selection function. For example, processing device 120 may determine selection function f via equation (11):
f＝m ₁ ×S+m ₂ ×η+m ₃ ×W (11)

Among them, S is the puncture path score, eta is the ablation conformity rate, W is the second constraint judgment value, m ₁ , m ₂ and m ₃ are all parameters greater than 0, which are used to balance the puncture path score, ablation conformity rate and The second constraint decision value.

Then, the processing device 120 can calculate the selection function value of each intermediate solution in the planning set added to the new intermediate solution, and sort all the intermediate solutions in ascending order based on the selection function value of each intermediate solution, ranking first, that is, The crowding distance of the intermediate solution with the smallest function value and the crowding distance of the last intermediate solution with the largest function value are set to infinity, that is, for each selection function, the boundary intermediate solution (the intermediate solution with the largest and smallest selection function values) ) and the corresponding crowding distance is set to infinity. The crowding distance of the non-boundary i-th intermediate solution can be calculated by formula (12):

Among them, f is the selection function, i is the serial number of the intermediate solution arranged in ascending order, f _max is the maximum value of the selection function values corresponding to all intermediate solutions in the planning set that adds the new intermediate solution, and f _min is the value of the new intermediate solution added The minimum value of the selection function values corresponding to all intermediate solutions in the planning set of solutions, d is the crowding distance.

In some embodiments, the processing device 120 may first add the intermediate solutions in the first layer to the new planning set. If the number of intermediate solutions in the first layer is less than the second preset number, then add the intermediate solutions in the second layer. The new planning set is deduced in sequence until the x-th layer is added and the number of new planning sets is greater than or equal to the second preset number. When the x-th layer is added, the number of new planning sets is equal to the second preset number, the x-th layer and the previous intermediate solutions are used as the new planning set; when the x-th layer is added, the number of new planning sets is greater than When the second preset number is reached, the congestion distance of the xth layer is sorted in descending order, and new planning sets are added from front to back until the number of new planning sets is equal to the 2. Default quantity. For example, as shown in Figure 9, assume that the number of planning sets is 50, the number of new intermediate solutions is 50, the number of the first level divided by the planning set adding the new intermediate solution is 25, the number of the second level is 10, and the number of the second level is 10. The number of three layers is 30... Add the intermediate solution of the first layer to the new planning set. The number of intermediate solutions of the new planning set is 25, which does not reach the second preset number (that is, the number of planning sets is 50); then The intermediate solutions of the second level are also added to the new planning set. The number of intermediate solutions in the new planning set is 35, which does not reach the second preset number. Then the intermediate solutions of the third level are also added to the new planning set. The new planning set has The number of intermediate solutions in the planning set is 65, which exceeds the second preset number. Therefore, 50-25-10=15 intermediate solutions need to be selected in the third layer to add to the new planning set. The specific selection method is: calculate the crowding distance of each intermediate solution in the third layer, arrange the crowding distance in descending order, take the first 15 intermediate solutions and add them to the new planning set, the last 15 in the third layer and the first Layers other than the first, second and third layers are not superior enough and will be eliminated. Since the larger the crowding distance is, the smaller the crowding degree is. Therefore, choosing an intermediate solution with a larger crowding distance can maintain the diversity of the planning set and avoid falling into local optimality.

In some embodiments of this specification, an intermediate solution is selected based on the second evaluation value and the second constraint judgment value, and the intermediate solution with better effect and lowest damage to the patient can be selected to enter the next round of iterative update, thereby promoting the iterative update. Quickly obtain the target parameters that are most suitable for the patient's condition, thereby improving the doctor's surgical efficiency and reducing the difficulty of the doctor's work.

In some embodiments, the processing device 120 only filters the intermediate solution through the second evaluation value to determine the new planning set to enter the next round of iteration, that is, the selection function f in formula (11) may not include the last term m ₃ ×W. For example only, the processing device 120 may calculate the dominance relationship of the intermediate solution based on the second evaluation value and stratify the planning set, and based on the stratification result and the crowding distance calculated by the selection function f excluding the m ₃ ×W term, by The intermediate solutions are screened to determine the new planning set in the aforementioned process. If only the second evaluation value is used to filter the intermediate solution to determine a new planning set, the processing device 120 can determine the length of the instrument, the puncture not penetrating the risk structure, the ablation field range completely covering the lesion, the rationality of the ablation electrode distance, and the puncture path in the pre-processing stage. Constraints such as the rationality of angle safety constrain the selection range of planning parameters to ensure that the intermediate solutions in the planning set satisfy the constraints.

In some embodiments, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in Figure 19. The computer device includes a processor, memory, communication interface, display screen and input device connected through a system bus. Wherein, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes non-volatile storage media and internal memory. The non-volatile storage medium stores operating systems and computer programs. This internal memory provides an environment for the execution of operating systems and computer programs in non-volatile storage media. The communication interface of the computer device is used for wired or wireless communication with external terminals. The wireless mode can be implemented through WIFI, mobile cellular network, NFC (Near Field Communication) or other technologies. The computer program implements a puncture path planning method when executed by the processor. The display screen of the computer device may be a liquid crystal display or an electronic ink display. The input device of the computer device may be a touch layer covered on the display screen, or may be a button, trackball or touch pad provided on the computer device shell. , it can also be an external keyboard, trackpad or mouse, etc.

Those skilled in the art can understand that the structure shown in Figure 19 is only a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer equipment to which the solution of the present application is applied. The specific computer equipment may May include more or fewer parts than shown, or combine certain parts, or have a different arrangement of parts.

In some embodiments, the computer-readable storage medium stores computer instructions. After the computer reads the computer instructions in the storage medium, the computer runs the parameter planning method of the intervention planning system.

The beneficial effects that may be brought about by the embodiments of this specification include but are not limited to: (1) By obtaining the patient data of the target patient and then determining the target parameters, more reasonable target parameters can be obtained to improve the doctor's surgical efficiency and reduce the difficulty of the doctor's work. . (2) By determining the structural characteristics and target points of the target patient, the candidate puncture path whose judgment value satisfies the preset conditions for strong clinical constraints can be determined, and a reasonable target puncture path can be obtained automatically and efficiently. (3) By calculating the path association information of at least one candidate puncture path, and determining the target puncture path through a preset search algorithm based on the path association information of at least one candidate puncture path, automatic planning of the target puncture path is realized, which can accurately and effectively Planning the target puncture path improves the efficiency of target puncture path planning. (4) The optimal solution is determined through the quantum annealing algorithm, and the candidate puncture path corresponding to the optimal solution is determined as the target puncture path, which can accurately and effectively plan the target puncture path and improve the efficiency of target puncture path planning. Since target puncture path planning is a clinical multi-constraint optimization problem, the problem is a non-deterministic polynomial problem, that is, there is a definite answer, but the time complexity of obtaining the solution increases exponentially. Classic computers have calculation problems due to their own performance limitations. If the time is too long or the optimal solution cannot be reached, the technical solution of this embodiment can be run on a quantum annealing machine, such as mapping the Hamiltonian to the real qubits of the quantum annealing machine, so that The puncture path planning is more efficient and can achieve fast and accurate preoperative puncture path planning. (5) By generating individual sets and new individuals through the individual generator, and using the Pareto front solution as at least one set of intermediate parameters, the scope of searching for the optimal individual can be expanded, and the better global solution can be screened out, thereby avoiding the final The confirmed target parameters are only the local optimal parameters. (6) Through the individual filter to update the individual set based on the constraint judgment value and evaluation value, the individuals with better effects and the lowest damage to the patient can be screened out to enter the next round of iterative update, prompting the iterative update to obtain the most suitable patient more quickly target parameters of the situation. (7) By determining the number of puncture paths and the range of ablation sphere parameters, determining at least one set of planning parameters, and determining the target parameters based on at least one set of planning parameters, the range of planning parameters can be narrowed and more reasonable target parameters can be obtained more quickly to improve Improve the doctor's surgical efficiency and reduce the difficulty of the doctor's work. (8) Perform three-dimensional reconstruction based on patient data to obtain three-dimensional medical images, and perform preprocessing operations on the three-dimensional medical images to determine planning parameters. The range of planning parameters can be adaptively narrowed according to different lesion conditions, so that the planning parameters The number is more suitable for the lesion conditions of different patients and improves the efficiency of subsequent iterative optimization. (9) Automatically determine and crop the area of interest, and discretize data points through downsampling, which can reduce the computing pressure on the device. (10) By performing multiple rounds of iterative optimization on the planning set to determine the optimal planning set, the optimal planning set can be searched globally to avoid individuals in the optimal planning set tending to the local optimum; the target parameters are determined through the optimal planning set, The most suitable target parameters can be obtained. (11) Selecting individuals based on evaluation values and constraint judgment values can select individuals with better effects and the least damage to patients to enter the next round of iterative updates, prompting the iterative updates to obtain the target parameters most suitable for the patient's situation faster, thereby improving It improves the doctor's surgical efficiency and reduces the difficulty of the doctor's work.

The basic concepts have been described above. It is obvious to those skilled in the art that the above detailed disclosure is only an example and does not constitute a limitation of this specification. Although not explicitly stated herein, various modifications, improvements, and corrections may be made to this specification by those skilled in the art. Such modifications, improvements, and corrections are suggested in this specification, and therefore such modifications, improvements, and corrections remain within the spirit and scope of the exemplary embodiments of this specification.

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

In addition, unless explicitly stated in the claims, the order of the processing elements and sequences, the use of numbers and letters, or the use of other names in this specification are not intended to limit the order of the processes and methods in this specification. Although the foregoing disclosure discusses by various examples some embodiments of the invention that are presently considered useful, it is to be understood that such details are for purposes of illustration only and that the appended claims are not limited to the disclosed embodiments. To the contrary, rights The claims are intended to cover all modifications and equivalent combinations consistent with the spirit and scope of the embodiments of this specification. For example, although the system components described above can be implemented through hardware devices, they can also be implemented through software-only solutions, such as installing the described system on an existing server or mobile device.

Similarly, it should be noted that, in order to simplify the expression disclosed in this specification and thereby help understand one or more embodiments of the invention, in the previous description of the embodiments of this specification, multiple features are sometimes combined into one embodiment. accompanying drawings or descriptions thereof. However, this method of disclosure does not imply that the subject matter of the description requires more features than are mentioned in the claims. In fact, embodiments may have less than all features of a single disclosed embodiment.

In some embodiments, numbers are used to describe the quantities of components and properties. It should be understood that such numbers used to describe the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Grooming. Unless otherwise stated, "about," "approximately," or "substantially" means that the stated number is allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximations that may vary depending on the desired features of the individual embodiment. In some embodiments, numerical parameters should account for the specified number of significant digits and use general digit preservation methods. Although the numerical ranges and parameters used to identify the breadth of ranges in some embodiments of this specification are approximations, in specific embodiments, such numerical values are set as accurately as is feasible.

Each patent, patent application, patent application publication and other material, such as articles, books, instructions, publications, documents, etc. cited in this specification is hereby incorporated by reference into this specification in its entirety. Application history documents that are inconsistent with or conflict with the content of this specification are excluded, as are documents (currently or later appended to this specification) that limit the broadest scope of the claims in this specification. It should be noted that if there is any inconsistency or conflict between the descriptions, definitions, and/or the use of terms in the accompanying materials of this manual and the content described in this manual, the descriptions, definitions, and/or the use of terms in this manual shall prevail. .

Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Accordingly, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to those expressly introduced and described in this specification.

Claims

An intervention planning system, characterized in that it includes a control module, and the control module is used for:

Obtain patient data for target patients; and

Based on the patient data, the target parameters are determined.
The system according to claim 1, wherein the target parameter includes a target puncture path, and determining the target puncture path includes:

determining structural characteristics of the target patient based on the patient data;

Based on the structural characteristics, determine the target target;

Based on the structural characteristics and the target target, determine a set of candidate puncture paths;

From the set of candidate puncture paths, a target puncture path is determined.
The system of claim 2, wherein determining the structural characteristics of the target patient based on the patient data includes:

Based on the patient data, determine a three-dimensional medical image of the target patient;

Based on the three-dimensional medical image, the structural characteristics are determined.
The system according to any one of claims 2-3, wherein determining a set of candidate puncture paths based on the structural characteristics and the target target includes:

Determine the target target point as the perspective projection center;

Use the perspective projection center as the source to emit multiple rays outwards, and based on the structural characteristics, calculate the judgment value of the strong clinical constraints of the puncture path corresponding to each ray;

Determine the puncture path whose judgment value of the strong clinical constraint conditions satisfies the preset conditions as the candidate puncture path;

A set of the candidate puncture paths is determined as the candidate puncture path set.
The system according to claim 4, wherein the strong clinical constraints include one or more of the following conditions:

The puncture path does not contact and does not penetrate puncture risk structures, the length of the puncture path is less than the preset needle length threshold, the angle between the puncture path and the target tissue is not less than the preset angle threshold, and the puncture path passes through the preset angle threshold. The distance length of the puncture structure is greater than the preset distance threshold.
The system according to any one of claims 2-3, wherein the set of candidate puncture paths includes at least one candidate puncture path, and determining the target puncture path from the set of candidate puncture paths includes:

Calculate path association information of the at least one candidate puncture path;

Based on the path association information of the at least one candidate puncture path, the target puncture path is determined through a preset search algorithm.
The system according to claim 6, wherein the path association information includes one or more of the following information:

The distance between the candidate puncture path and the puncture risk structure, the length of the candidate puncture path, and the angle between the candidate puncture path and the target tissue.
The system according to claim 6, wherein determining the target puncture path through a preset search algorithm based on the path association information of the at least one candidate puncture path includes:

Construct a first function term according to the path association information;

Based on the first function term, the target puncture path is determined through the preset search algorithm.
The system according to claim 1, wherein the target parameters include target puncture path, target stop point position and target ablation sphere parameters, and determining the target parameters based on the patient data includes:

determining at least one set of planning parameters based on the patient data;

Determine target parameters based on the at least one set of planning parameters;

Among them, the determined target parameters include:

Based on the individual generator, generate an individual set, the individual set includes multiple individuals, each of the individuals corresponds to a set of planning parameters;

Perform at least one round of first iteration update on the individual set until the first iteration completion condition is met;

determining the at least one set of intermediate parameters based on the updated set of individuals; and

The target parameters are determined based on the at least one set of intermediate parameters.
The system of claim 9, wherein determining at least one set of planning parameters based on patient data includes:

Perform three-dimensional reconstruction on the patient data to obtain a three-dimensional medical image, where the patient data includes the patient's CT or MR data; and

The at least one set of planning parameters is determined based on the three-dimensional medical image.
The system of claim 10, wherein determining the at least one set of planning parameters based on the three-dimensional medical image includes:

Perform preprocessing operations on the three-dimensional medical image, the preprocessing operations including one or more of region of interest cropping, data point downsampling, and blood vessel coarse subdivision; and

The at least one set of planning parameters is determined based on results obtained from the preprocessing operation.
The system according to any one of claims 9-11, wherein each iteration in the at least one first iteration includes:

Generate at least one new individual based on the individual generator; and

Add the at least one new individual to the individual collection.
The system of claim 12, wherein each of the at least one first iteration includes:

Based on the individual filter, filter each individual in the individual collection and update the individual collection, where the screening includes selecting individuals in the individual collection;

Wherein, the selection operation includes:

Calculating a first evaluation value and a first constraint determination value for each individual in the individual set; and

Select individuals based on the first evaluation value and the first constraint determination value, and determine the updated individual set.
The system of claim 9, wherein determining the at least one set of intermediate parameters based on the updated individual set includes:

A Pareto front solution is determined based on the updated individual set, and the Pareto front solution is used as the at least one set of intermediate parameters.
The system according to claim 1, wherein the target parameters include a target puncture path, a target stop point position and a target ablation sphere parameter, and determining the target parameters based on the patient data includes:

Determine the number of puncture paths and the range of ablation ball parameters based on patient data;

Determine at least one set of planning parameters based on the patient data, the number of puncture paths, and the ablation sphere parameter range;

determining at least one set of feasible solutions based on the at least one set of planning parameters; and

Based on the at least one set of feasible solutions, the target parameters are determined.
The system of claim 15, wherein determining at least one set of feasible solutions based on the at least one set of planning parameters includes:

Generate a planning set based on the at least one set of planning parameters, the planning set including a plurality of intermediate solutions, each of the intermediate solutions corresponding to a set of planning parameters;

Perform at least one round of second iteration optimization on the planning set until the second iteration completion condition is satisfied, and obtain the optimal planning set; and

Based on the set of optimal plans, at least one set of feasible solutions is determined.
The system according to any one of claims 15-16, wherein determining the ablation sphere parameter range includes:

Based on the patient data, determine a lesion mask;

Based on the lesion mask, determine the long axis of the lesion and the short axis of the lesion; and

Based on the long axis of the lesion and the short axis of the lesion, the parameter range of the ablation sphere is determined.
The system according to any one of claims 15-16, wherein determining at least one set of planning parameters based on the patient data, the number of puncture paths, and the ablation sphere parameter range includes:

Perform three-dimensional reconstruction on the patient data to obtain three-dimensional medical images;

Perform preprocessing operations on the three-dimensional medical images; and

The at least one set of planning parameters is determined based on the results obtained from the preprocessing operation, the number of puncture paths, and the ablation sphere parameter range.
The system of claim 16, wherein each of the at least one second iteration includes:

Perform a transformation operation on the planning set to obtain a first preset number of new intermediate solutions; and

The new intermediate solution is added to the planning set to obtain a planning set to which the new intermediate solution is added.
The system of claim 19, wherein each of the at least one second iteration includes:

Calculate the second evaluation value and the second constraint determination value of each intermediate solution in the planning set to which the new intermediate solution is added; and

Select an intermediate solution based on the second evaluation value and the second constraint decision value, and obtain a new planning set including a second preset number of intermediate solutions. The second constraint decision value is based on the decision value of at least one constraint condition. It is determined that the at least one constraint condition includes whether the instrument length meets the requirements, and the second evaluation value includes puncture path score and ablation conformity rate.
An intervention planning method, characterized in that the method includes:

Obtain patient data of target patients;

Based on the patient data, target parameters are determined.
An interventional planning device, characterized by including interventional equipment, a robotic arm and a processor;

The interventional device includes an interventional needle;

The robotic arm is used to carry the interventional needle to perform interventional surgery according to target parameters;

The processor is used to control the robotic arm and determine the target parameters. The determination of the target parameters includes:

Obtain patient data for target patients; and

Based on the patient data, the target parameters are determined.
A computer-readable storage medium, characterized in that the storage medium stores computer instructions. After the computer reads the computer instructions in the storage medium, the computer executes the intervention planning method as claimed in claim 21.