A gynecological tumor multi-disciplinary joint operation path planning system and method

Abstract

The application belongs to the technical field of path planning, and discloses a gynecological tumor multidisciplinary joint operation path planning system and method. First joint operation path planning features are obtained by processing first lesion image data collected from a gynecological tumor operation site, first physiological function parameters for evaluating a patient state, and first operation state parameters of an operating physician. A multidisciplinary joint operation path planning model is constructed by using the first joint operation path planning features and a corresponding first tumor operation planning method. Second joint operation path planning features of a target to be operated are input into the model to obtain a second tumor operation planning method. By combining image and physiological parameters such as electroencephalogram for reference, a tumor operation path planning route suggestion is obtained to assist a long-time operating physician in path selection, reduce physician fatigue and operation comfort, and accurately perform operation path management to improve operation completion rate.

Description

Technical Field

[0001] This invention belongs to the field of path planning technology, and in particular relates to a multidisciplinary surgical path planning system and method for gynecological tumors. Background Technology

[0002] Image-guided surgery has become a hot research and application area in the field of tumor surgery, accelerating the development of precision, minimally invasive, and intelligent tumor treatment. Image-based surgical path planning methods directly affect the human-computer interaction performance of tumor surgery.

[0003] Advance surgical pathway planning is widely used in clinical oncology. Early surgical pathway planning methods mainly involved manual or semi-automatic interactive operations by physicians through software interfaces. These methods and interfaces exhibited significant engineering characteristics, making them difficult to integrate with physicians' existing operating habits, resulting in low efficiency and hindering the widespread adoption of assisted surgical procedures.

[0004] Statistical shape modeling and deep learning methods offer potential solutions for exploring automated and intelligent planning techniques for tumor surgical pathways. Existing research on automated tumor surgical pathway planning based on preoperative 3D CT images and deep neural networks has preliminarily demonstrated the potential advantages and application feasibility of intelligent technologies in the field of automated planning of multidisciplinary-assisted tumor surgical pathways.

[0005] Based on the above analysis, the problems and shortcomings of the existing technology are as follows:

[0006] In existing gynecological tumor surgical pathway planning, only the patient's condition is considered. Furthermore, with limited and inaccurate tumor image data, how can surgical pathway planning be designed to assist surgeons in planning tumor surgeries? How can we combine high-accuracy images with physiological parameters such as EEG to derive surgical pathway recommendations? This would assist surgeons selecting pathways during long surgeries, reduce surgeon fatigue and comfort, and enable precise surgical pathway management, thereby improving surgical success rates.

[0007] The question is how to use machine learning methods to adaptively acquire tumor surgical pathways, and how to simultaneously handle different states of patients and doctors to comprehensively acquire unified features to reduce the doctor's workload in making judgments when generating surgical pathways, thereby generating the optimal tumor surgical pathway and enabling precise tumor surgical pathway treatment for different patients. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention proposes a method and system for multidisciplinary surgical pathway planning in gynecological tumors.

[0009] In a first aspect of the present invention, a method for multidisciplinary surgical pathway planning in gynecologic oncology is provided, the method comprising:

[0010] Surgical path planning features were obtained by collecting image data of the first lesion from the surgical site of gynecological tumors, assessing the patient's physiological function parameters, and processing the surgeon's surgical status parameters.

[0011] And collect tumor surgery planning methods determined by the surgeons;

[0012] A first multidisciplinary joint surgical path planning model was constructed using the surgical path planning features and the tumor surgical planning method.

[0013] Furthermore, before acquiring the surgical path planning features, the first lesion image data is used to generate second lesion image data by using an improved generative adversarial network model based on the physiological function parameters. Based on the second lesion image data, the physiological function parameters, and the surgical state parameters, joint path planning features are generated. A second multidisciplinary joint surgical path planning model is constructed using the joint surgical path planning features and the tumor surgical planning method.

[0014] Furthermore, the lesion image data is obtained through CT scan, and the obtained lesion image data is used to calculate and filter target lesion data to obtain first lesion image data or second lesion image data.

[0015] Furthermore, the physiological function parameters are calculated based on the patient's tumor size, blood flow velocity at the tumor location, blood oxygen saturation percentage, and respiratory rate.

[0016] Furthermore, the surgical status parameters are obtained by processing the doctor's blood flow rate, blood oxygen saturation, respiratory rate ratio, and brain electrophysiological parameters.

[0017] Furthermore, an improved generative adversarial network model based on physiological function parameters is used to generate lesion image data by optimizing its discriminator.

[0018] Furthermore, the first multidisciplinary joint surgical pathway planning model or the second multidisciplinary joint surgical pathway planning model adopts the SVM model.

[0019] It also provides a multidisciplinary surgical pathway planning system for gynecological tumors, including a tumor lesion image data collection module, a patient status assessment module, a physician status assessment module, a surgical pathway planning feature processing module, a generative adversarial network image generation module, a multidisciplinary surgical pathway planning model construction module, and a multidisciplinary surgical pathway planning module.

[0020] The tumor lesion image data collection module is used to collect image data of the first lesion from the surgical site of gynecological tumors.

[0021] The patient status assessment module is used to assess the physiological functional parameters of the patient's status.

[0022] The physician status assessment module is used to collect surgical status parameters of the surgeon performing the operation.

[0023] The surgical path planning feature processing module is used to receive the first lesion image data, the physiological function parameters, and the surgical state parameters, and process them to obtain surgical path planning features; it is also used to receive the second lesion image data, the physiological function parameters, and the surgical state parameters, and process them to obtain combined surgical path planning features.

[0024] The generative adversarial network image generation module: generates second lesion image data from the first lesion image data using an improved generative adversarial network model based on the physiological function parameters;

[0025] The multidisciplinary joint surgical pathway planning model construction module is used to construct a first multidisciplinary joint surgical pathway planning model based on the surgical pathway planning features and the tumor surgical planning method, and is also used to construct a second multidisciplinary joint surgical pathway planning model based on the joint surgical pathway planning features and the tumor surgical planning method.

[0026] The multidisciplinary surgical pathway planning module: uses the first or second multidisciplinary surgical pathway planning model in the multidisciplinary surgical pathway planning model construction module to generate a multidisciplinary surgical pathway planning method to assist doctors in planning tumor surgical pathways.

[0027] Furthermore, an improved generative adversarial network model based on physiological function parameters is used to generate lesion image data by optimizing its discriminator.

[0028] This invention, building upon existing automated intelligent selection assistance technology for gynecological tumor surgical pathway planning, considers not only the patient's condition but also the surgeon's condition. Furthermore, when the amount of tumor image data is limited or inaccurate, it utilizes an improved GAN image data generation model to adaptively modify the discriminator, enabling faster acquisition of accurate disease image data for tumor surgical pathway planning and design, thus assisting surgeons in this process. This invention also combines the generation of highly accurate images with physiological parameters such as EEG to acquire model input features, training a more accurate model to provide tumor surgical pathway planning suggestions. This assists surgeons performing long surgeries in selecting pathways, reducing surgeon fatigue and surgical comfort, and enabling precise surgical pathway management, thereby improving surgical completion rates.

[0029] By employing a support vector machine classifier method for adaptive acquisition of tumor surgical pathways and simultaneously handling different states of patients and doctors, a unified feature acquisition method is used to reduce the workload of doctors in making judgments during the generation of surgical pathways, thereby generating the optimal tumor surgical pathway. This allows for precise tumor surgical pathway treatment for different patients, which is what distinguishes this invention from existing technologies.

[0030] Further embodiments and improvements of the present invention will be described in conjunction with the accompanying drawings and specific examples. Attached Figure Description

[0031] Figure 1 This is a flowchart of the gynecologic oncology multidisciplinary surgical pathway planning method of the present invention;

[0032] Figure 2 This is a schematic diagram of the gynecologic oncology multidisciplinary surgical pathway planning system of the present invention;

[0033] Figure 3 This is a schematic diagram of the GAN function in an embodiment of the present invention;

[0034] Figure 4 This is a schematic diagram of the SVM model in the embodiments of the present invention;

[0035] Figure 5 This is a schematic diagram of an electronic device structure for implementing the method of the present invention in an embodiment of the present invention. Detailed Implementation

[0036] The invention will now be further described in conjunction with the accompanying drawings and specific embodiments. The images acquired in this invention are all color images. In addition to acquiring image data using CT equipment, color ultrasound equipment can also be used for acquisition, but it is not limited to CT equipment.

[0037] In a first aspect of the invention, a method for multidisciplinary surgical pathway planning in gynecologic oncology is provided, the method comprising:

[0038] Surgical path planning features were obtained by collecting image data of the first lesion from the surgical site of gynecological tumors, assessing the patient's physiological function parameters, and processing the surgeon's surgical status parameters.

[0039] And collect tumor surgery planning methods determined by the surgeons;

[0040] A first multidisciplinary joint surgical path planning model was constructed using the surgical path planning features and the tumor surgical planning method.

[0041] Furthermore, before acquiring the surgical path planning features, the first lesion image data is used to generate second lesion image data by using an improved generative adversarial network model based on the physiological function parameters. Based on the second lesion image data, the physiological function parameters, and the surgical state parameters, joint path planning features are generated. A second multidisciplinary joint surgical path planning model is constructed using the joint surgical path planning features and the tumor surgical planning method.

[0042] Furthermore, the lesion image data is obtained through CT scans, and the obtained lesion image data is used to calculate and filter target lesion data to obtain either first lesion image data or second lesion image data. The formula for calculating and filtering target lesion data is as follows:

[0043]

[0044] This embodiment uses the Canny edge detection algorithm to obtain the operator threshold. Since the Canny edge detection algorithm can detect image edges and has good resistance to noise, the threshold is determined by the operators in the Canny edge detection algorithm. Specifically, the maximum value of the calculated horizontal and vertical gradient values ​​is taken, and the target image pixels are filtered based on values ​​above (including the maximum gradient value) to obtain the final threshold. This can better remove image interference outside the target lesion and obtain image data of the first lesion or the second lesion, where... The values ​​obtained after processing image data based on grayscale values ​​are filtered using operators, i.e., the image data of the first lesion or the image data of the second lesion. , Represents the gradient operator, This is the convolution operator. , , The image grayscale coefficient is used. , , Both can be 0.333. , , These are the pixel values ​​of the red, green, and blue channels of the image. This represents the maximum value of the gradient in the horizontal and vertical directions, where k is the number of pixel values ​​greater than the maximum value. Indicates greater than or equal to The sum of pixel values.

[0045] Furthermore, the physiological function parameters are calculated based on the patient's tumor size, blood flow velocity at the tumor location, blood oxygen saturation percentage, and respiratory rate. The calculation formula is as follows:

[0046]

[0047] Because the tumor condition and the patient's overall condition must be carefully considered during surgery, physiological parameters are calculated by taking into account tumor size, blood flow velocity at the tumor location, blood oxygen saturation percentage, and respiratory rate. In the formula... Physiological function parameters, The value is set according to the tumor size and TNM staging, with a range of 0-3 based on the staging. These represent blood flow velocity and blood oxygen saturation at the tumor site, respectively. These are the average blood flow velocity and average blood oxygen saturation in the human body, respectively. For the patient's respiratory rate, This represents the normal respiratory rate for a woman.

[0048] Furthermore, the surgical status parameters are obtained by processing the doctor's blood flow rate, blood oxygen saturation, respiratory rate ratio, and brain electrophysiological parameters, and the calculation formula is as follows:

[0049]

[0050] Since electroencephalogram (EEG) parameters often differ from other parameters by more than three orders of magnitude, the ln function is used for correction. In the formula... This represents the correction factor for electroencephalographic parameters. This is the average EEG reading, in microvolts. This indicates the surgeon's surgical status parameters. For the doctor's blood oxygen saturation, For the doctor's blood flow rate, For the doctor's respiratory rate, The normal breathing rate for men and women is given. To avoid the EEG parameters being too small, the value of k in the formula is 0.15-0.25.

[0051] Furthermore, an improved generative adversarial network model based on physiological function parameters is used to generate lesion image data by optimizing its discriminator. The discriminator's calculation formula is as follows:

[0052]

[0053] In the formula, The minimum classification accuracy of the discriminator. The image distribution of the first lesion. The distribution of lesion images generated by the improved generative adversarial network model. Physiological function parameters, The physiological function parameters of the patient at the smallest tumor size are used. Since the discriminator in the existing GAN still has general characteristics in judging the authenticity of data, it does not take into account the impact of the differences in the samples on the discriminator. In this invention, based on the tumor scenario, the ratio of the patient's current physiological function parameters to the patient's initial physiological function parameters is used to correct the generated lesion image distribution data, thereby improving the accuracy of data generation and carrying out multidisciplinary joint planning for subsequent classification and surgical path acquisition.

[0054] Furthermore, the first multidisciplinary joint surgical pathway planning model or the second multidisciplinary joint surgical pathway planning model adopts the SVM model.

[0055] Furthermore, the calculation formula for the SVM model is as follows:

[0056]

[0057] in , These are the normal vector and intercept of the hyperplane, respectively, obtained by training using the surgical path planning features or the combined surgical path planning features and the tumor surgical planning method. The output is a tumor surgery planning method. This refers to either the surgical path planning feature or the combined surgical path planning feature.

[0058] A multidisciplinary surgical pathway planning system for gynecological tumors is also provided, including a tumor lesion image data collection module, a patient status assessment module, a physician status assessment module, a surgical pathway planning feature processing module, a generative adversarial network image generation module, a multidisciplinary surgical pathway planning model construction module, and a multidisciplinary surgical pathway planning module.

[0059] The tumor lesion image data collection module is used to collect image data of the first lesion from the surgical site of gynecological tumors.

[0060] The patient status assessment module is used to assess the physiological functional parameters of the patient's status.

[0061] The physician status assessment module is used to collect surgical status parameters of the surgeon performing the operation.

[0062] The surgical path planning feature processing module is used to receive the first lesion image data, the physiological function parameters, and the surgical state parameters, and process them to obtain surgical path planning features; it is also used to receive the second lesion image data, the physiological function parameters, and the surgical state parameters, and process them to obtain combined surgical path planning features.

[0063] The generative adversarial network image generation module: generates second lesion image data from the first lesion image data using an improved generative adversarial network model based on the physiological function parameters;

[0064] The multidisciplinary joint surgical pathway planning model construction module is used to construct a first multidisciplinary joint surgical pathway planning model based on the surgical pathway planning features and the tumor surgical planning method, and is also used to construct a second multidisciplinary joint surgical pathway planning model based on the joint surgical pathway planning features and the tumor surgical planning method.

[0065] The multidisciplinary surgical pathway planning module: uses the first or second multidisciplinary surgical pathway planning model in the multidisciplinary surgical pathway planning model construction module to generate a multidisciplinary surgical pathway planning method to assist doctors in planning tumor surgical pathways.

[0066] Furthermore, an improved generative adversarial network model based on physiological function parameters is used to generate lesion image data by optimizing its discriminator. The discriminator's calculation formula is as follows:

[0067] In the formula, The minimum classification accuracy of the discriminator. The image distribution of the first lesion. The distribution of lesion images generated by the improved generative adversarial network model. Physiological function parameters, The physiological function parameters of the patient at the smallest tumor size are used. Since the discriminator in the existing GAN still has general characteristics in judging the authenticity of data, it does not take into account the impact of the differences in the samples on the discriminator. In this invention, based on the tumor scenario, the ratio of the patient's current physiological function parameters to the patient's initial physiological function parameters is used to correct the generated lesion image distribution data, thereby improving the accuracy of data generation and carrying out multidisciplinary joint planning for subsequent classification and surgical path acquisition.

[0068] This invention, building upon existing automated intelligent selection assistance technology for gynecological tumor surgical pathway planning, considers not only the patient's condition but also the surgeon's condition. Furthermore, when the amount of tumor image data is limited or inaccurate, it utilizes an improved GAN image data generation model to adaptively modify the discriminator, enabling faster acquisition of accurate disease image data for tumor surgical pathway planning and design, thus assisting surgeons in this process. This invention also combines the generation of highly accurate images with physiological parameters such as EEG to acquire model input features, training a more accurate model to provide tumor surgical pathway planning suggestions. This assists surgeons performing long surgeries in selecting pathways, reducing surgeon fatigue and surgical comfort, and enabling precise surgical pathway management, thereby improving surgical completion rates.

[0069] By employing a support vector machine classifier method for adaptive acquisition of tumor surgical pathways and simultaneously handling different states of patients and doctors, a unified feature acquisition method is used to reduce the workload of doctors in making judgments during the generation of surgical pathways, thereby generating the optimal tumor surgical pathway. This allows for precise tumor surgical pathway treatment for different patients, which is what distinguishes this invention from existing technologies.

[0070] For any module structures not specifically defined in this invention, the existing technical specifications shall prevail. The existing technical specifications mentioned in the foregoing background and specific embodiments sections are considered part of this invention and are used to understand the meaning of certain technical features or parameters. The scope of protection of this invention is determined by the actual contents of the claims.

Claims

1. A method for multidisciplinary surgical pathway planning in gynecological tumors, characterized in that, The method includes: Surgical path planning features were obtained by collecting image data of the first lesion from the surgical site of gynecological tumors, assessing the patient's physiological function parameters, and processing the surgeon's surgical status parameters. And collect tumor surgery planning methods determined by the surgeons; A first multidisciplinary joint surgical path planning model was constructed using the surgical path planning features and the tumor surgical planning method. Before obtaining the surgical path planning features, the first lesion image data is used to generate second lesion image data by using an improved generative adversarial network model based on the physiological function parameters. Then, a joint path planning feature is generated based on the second lesion image data, the physiological function parameters, and the surgical state parameters. A second multidisciplinary joint surgical path planning model is constructed using the joint path planning feature and the tumor surgical planning method.

2. The method for multidisciplinary surgical pathway planning in gynecological tumors as described in claim 1, characterized in that: The lesion image data is obtained through CT scans. The obtained lesion image data is then filtered to obtain either first lesion image data or second lesion image data. The formula for calculating and filtering the target lesion data is as follows: In the formula The values ​​obtained after processing image data based on grayscale values ​​are filtered using operators, i.e., the image data of the first lesion or the image data of the second lesion. , Represents the gradient operator, This is the convolution operator. , , The image grayscale coefficient is... , , These are the pixel values ​​of the red, green, and blue channels of the image. This represents the maximum value of the gradient in the horizontal and vertical directions, where k is the number of pixel values ​​greater than the maximum value. Indicates greater than or equal to The sum of pixel values.

3. The method for multidisciplinary surgical pathway planning in gynecological tumors as described in claim 1, characterized in that: The physiological function parameters are calculated based on the patient's tumor size, blood flow velocity at the tumor location, blood oxygen saturation percentage, and respiratory rate. The calculation formula is as follows: In the formula Physiological function parameters, The value is set according to the tumor size and TNM staging, with a range of 0-3 based on the staging. , These represent blood flow velocity and blood oxygen saturation at the tumor site, respectively. , These are the average blood flow velocity and average blood oxygen saturation in the human body, respectively. For the patient's respiratory rate, This represents the normal respiratory rate for a woman.

4. The method for multidisciplinary surgical pathway planning in gynecological tumors as described in claim 1, characterized in that: The surgical status parameters are obtained by processing the doctor's blood flow rate, blood oxygen saturation, respiratory rate ratio, and brain electrophysiological parameters. The calculation formula is as follows: In the formula This represents the correction factor for electroencephalographic parameters. This is the average EEG reading, in microvolts. This indicates the doctor's surgical status parameters. For the doctor's blood oxygen saturation, For the doctor's blood flow rate, For the doctor's respiratory rate, The normal respiratory rate for both men and women is represented by the value of k, which ranges from 0.15 to 0.

25.

5. The method for multidisciplinary surgical pathway planning in gynecological tumors as described in claim 1, characterized in that: An improved generative adversarial network model based on physiological function parameters was used to generate lesion image data by optimizing its discriminator.

6. The method for multidisciplinary surgical pathway planning in gynecological tumors as described in claim 5, characterized in that: The first multidisciplinary joint surgical pathway planning model or the second multidisciplinary joint surgical pathway planning model adopts the SVM model.

7. A multidisciplinary surgical pathway planning system for gynecological tumors, comprising a tumor lesion image data collection module, a patient status assessment module, a physician status assessment module, a surgical pathway planning feature processing module, a generative adversarial network image generation module, a multidisciplinary surgical pathway planning model construction module, and a multidisciplinary surgical pathway planning module, characterized in that: The tumor lesion image data collection module is used to collect image data of the first lesion from the surgical site of gynecological tumors. The patient status assessment module is used to assess the physiological functional parameters of the patient's status. The physician status assessment module is used to collect surgical status parameters of the surgeon performing the operation. The surgical path planning feature processing module is used to receive the first lesion image data, the physiological function parameters, and the surgical state parameters, and process them to obtain surgical path planning features. It is also used to receive the second lesion image data, the physiological function parameters, and the surgical status parameters, and process them to obtain combined surgical path planning features; The generative adversarial network image generation module: generates second lesion image data from the first lesion image data using an improved generative adversarial network model based on the physiological function parameters; The multidisciplinary joint surgical pathway planning model construction module is used to construct a first multidisciplinary joint surgical pathway planning model based on the surgical pathway planning features and the tumor surgical planning method, and is also used to construct a second multidisciplinary joint surgical pathway planning model based on the joint surgical pathway planning features and the tumor surgical planning method. The multidisciplinary surgical pathway planning module generates a multidisciplinary surgical pathway planning method using either the first or second multidisciplinary surgical pathway planning model in the multidisciplinary surgical pathway planning model construction module, thereby assisting doctors in planning tumor surgical pathways.

8. The gynecological oncology multidisciplinary surgical pathway planning system as described in claim 7, characterized in that: An improved generative adversarial network model based on physiological function parameters was used to generate lesion image data by optimizing its discriminator.

Images