Nursing decision-making method for perioperative period of oral cancer

Abstract

The invention relates to the technical field of medical care, and particularly discloses an oral cancer perioperative nursing decision method, which comprises the steps of image acquisition, model generation, preoperative evaluation, dynamic nursing, risk evaluation and decision support. According to the scheme, the tumor area is extracted from the three-dimensional image data based on the U-Net model, the tumor area is separated from surrounding healthy tissue, complex tumor forms are automatically processed, higher segmentation precision and robustness are provided, and finally the accurate three-dimensional anatomical model is generated; according to the scheme, the individualized data, the real-time physiological indexes and the treatment progress of the patient are obtained, the intelligent optimization algorithm is used for calculating the priority of each nursing measure, the individualized nursing decision is dynamically generated according to the priority, the most suitable nursing scheme is automatically provided for the patient, and individualized requirements of different patients are better met.

Description

Technical Field

[0001] The present invention relates to the field of medical care technology, and in particular to a nursing decision-making method for perioperative period of oral cancer. Background Art

[0002] Oral cancer is one of the most common malignant tumors in the head and neck. Since surgical treatment of oral cancer is high-risk and complex, perioperative care is crucial to the patient's recovery and prognosis. The traditional oral cancer image acquisition process is often affected by noise, resulting in poor ability to handle complex structures when the tumor boundaries are blurred and the shape is irregular, and unable to provide stable segmentation results; traditional nursing methods lack specificity for the perioperative care needs of specific diseases such as oral cancer, resulting in problems such as non-standardized nursing processes, low information integration, and insufficient decision support. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a nursing decision method for perioperative oral cancer. In view of the problem that the traditional oral cancer image acquisition process is often affected by noise, resulting in poor ability to process complex structures when the tumor boundaries are blurred and the morphology is irregular, and stable segmentation results cannot be provided, this solution extracts the tumor area from the three-dimensional image data based on the U-Net model, separates it from the surrounding healthy tissue, automatically processes complex tumor morphology, provides higher segmentation accuracy and robustness, and finally generates an accurate three-dimensional anatomical model; in view of the lack of pertinence of traditional nursing methods for perioperative nursing needs of specific diseases such as oral cancer, resulting in irregular nursing processes, low information integration, and insufficient decision support, this solution obtains the patient's individualized data, real-time physiological indicators and treatment progress, uses an intelligent optimization algorithm to calculate the priority of each nursing measure, and dynamically generates personalized nursing decisions based on the priority, automatically provides the patient with the most suitable nursing plan, and better adapts to the personalized needs of different patients.

[0004] The technical solution adopted by the present invention is as follows: The present invention provides a nursing decision-making method for oral cancer perioperative period, the method comprising the following steps:

[0005] Step S1: image acquisition, performing CT scanning on the neck of an oral cancer patient to obtain three-dimensional image data;

[0006] Step S2: Model generation, using an intelligent iterative algorithm to process the three-dimensional image data to generate a three-dimensional anatomical model;

[0007] Step S3: preoperative evaluation, based on the three-dimensional anatomical model, evaluating the size, location and surrounding key structures of the tumor, generating preoperative information support, and formulating a surgical plan based on the preoperative information support;

[0008] Step S4: Dynamic nursing, obtaining individualized patient data, combining preoperative information support and surgical plan, generating a personalized nursing plan, and dynamically optimizing the personalized nursing plan using a multi-objective optimization method;

[0009] Step S5: risk assessment, analyzing possible complications during surgery based on the personalized nursing plan and generating a risk assessment report;

[0010] Step S6: Decision support, providing decision support for intraoperative and postoperative care based on the risk assessment report.

[0011] Furthermore, in step S2, the model generation specifically includes the following steps:

[0012] Step S21: removing noise by performing convolution operation on the three-dimensional image data using Gaussian filtering, calculating the weighted average value of each pixel, and obtaining a smoothed image;

[0013] Step S22: edge enhancement, using the Sobel operator to calculate the gradient value of the smoothed image, enhancing the high-value area in the smoothed image, and highlighting the edge area to obtain an enhanced image;

[0014] Step S23: Image segmentation, constructing a U-net model, separating the tumor area and surrounding tissue in the enhanced image to obtain a segmented image; the U-net model includes an encoder, a bottleneck layer and a decoder, and restores edge information through a jump connection; the encoder includes a plurality of downsampling layers, each of which is composed of a convolutional layer and a maximum pooling layer, the decoder includes a plurality of upsampling layers, each of which is composed of a transposed convolutional layer and a convolutional layer, and the jump connection adds the feature map of the encoder to the feature map of the corresponding layer of the decoder element by element to restore the edge information; the output layer outputs a single-channel image representing the separation of the tumor area and the non-tumor area;

[0015] Step S24: Image optimization, constructing an optimization function to smooth the segmentation boundary and obtain the final three-dimensional anatomical model. The formula used is as follows: ;

[0016] In the formula, represents the optimization function, represents the segmentation boundary, represents the image gradient, and Weight parameter to control the smoothness and boundary strength of the optimization function.

[0017] Furthermore, in step S4, the dynamic nursing specifically includes the following steps:

[0018] Step S41: Obtain the hospital's electronic medical record system, read the patient's individualized data, and generate a preliminary nursing plan in combination with preoperative information support and surgical plan;

[0019] Step S42: using monitoring equipment to collect the patient's real-time physiological indicators and treatment progress, and dynamically adjust the initial nursing plan;

[0020] Step S43: Calculate the priority of each nursing measure in the preliminary nursing plan according to the patient's individualized data and real-time physiological indicators; the priority calculation is implemented by a particle swarm optimization algorithm, and the priority of each nursing measure is calculated according to the patient's individualized data and real-time physiological indicators. Nursing measures with high priorities are executed first. The formula used is as follows: ;

[0021] In the formula, Indicates the priority of nursing action, Indicates nursing measures, represents the characteristic function related to the nursing measures, Represents the weight of each feature;

[0022] Step S44: Sort and combine nursing measures based on priority to generate personalized nursing decisions; the personalized nursing decisions include the execution order and time schedule of various nursing measures. The generated nursing decisions are pushed to medical staff through the nursing system to guide actual nursing operations. The formula used is as follows: ; ;

[0023] In the formula, Indicates the nursing measures sorting results, Indicates the priority of nursing measures, represents the sorting algorithm, represents the execution time of each nursing measure, Indicates personalized care decisions.

[0024] Furthermore, in step S6, the decision support specifically includes the following steps:

[0025] Step S61: Obtaining risk factors from the risk assessment report, the personalized care plan, and the patient's individualized data;

[0026] Step S62: risk probability assessment, using Bayesian networks to model different risk factors, analyzing whether the current situation matches the risk factors, and calculating the probability and expected severity of the risk factors;

[0027] Step S63: generating decision rules. If the intraoperative risk factor is assessed as high, outputting a drug use adjustment rule; if the postoperative risk factor is assessed as high, recommending infection examination and automatically suggesting whether the antibiotic treatment regimen needs to be adjusted;

[0028] Step S64: Intervention plan, providing nursing plans and intervention schedules during the postoperative recovery process, and dynamically adjusting nursing decisions.

[0029] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0030] (1) In view of the problem that the traditional oral cancer image acquisition process is often affected by noise, resulting in poor ability to process complex structures and inability to provide stable segmentation results when the tumor boundaries are blurred and the morphology is irregular, this solution extracts the tumor area from the 3D image data based on the U-Net model and separates it from the surrounding healthy tissue. It automatically processes complex tumor morphology, provides higher segmentation accuracy and robustness, and ultimately generates an accurate 3D anatomical model.

[0031] (2) Traditional nursing methods lack specificity for perioperative nursing needs of specific diseases such as oral cancer, resulting in irregular nursing processes, low information integration, and insufficient decision support. This solution obtains the patient's individualized data, real-time physiological indicators, and treatment progress, uses an intelligent optimization algorithm to calculate the priority of each nursing measure, and dynamically generates personalized nursing decisions based on the priority, automatically providing patients with the most suitable nursing plan to better meet the personalized needs of different patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A flowchart of a nursing decision-making method for perioperative period of oral cancer proposed by the present invention;

[0033] Figure 2 This is the architecture diagram of the U-net model in Example 3.

[0034] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0035] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0036] Example 1, see Figure 1 The present invention provides a nursing decision-making method for oral cancer perioperative period, the method comprising the following steps:

[0037] Step S1: image acquisition, performing CT scanning on the neck of an oral cancer patient to obtain three-dimensional image data;

[0038] Step S2: Model generation, using an intelligent iterative algorithm to process the three-dimensional image data to generate a three-dimensional anatomical model;

[0039] Step S3: preoperative evaluation, based on the three-dimensional anatomical model, evaluating the size, location and surrounding key structures of the tumor, generating preoperative information support, and formulating a surgical plan based on the preoperative information support, wherein the preoperative information support includes: the three-dimensional size and volume of the tumor, the spatial relationship between the tumor and surrounding key structures, and the recommended location and size of the surgical incision;

[0040] Step S4: Dynamic nursing, obtaining individualized patient data through the electronic medical record system, combining preoperative information support and surgical plan, generating a personalized nursing plan, and dynamically optimizing the personalized nursing plan using a multi-objective optimization method;

[0041] Step S5: risk assessment, analyzing possible complications during surgery based on the personalized nursing plan and generating a risk assessment report;

[0042] Step S6: Decision support, providing decision support for intraoperative and postoperative care based on the risk assessment report.

[0043] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, the neck of an oral cancer patient is scanned using an MX16EVO CT scanner, and the scanning parameters are: tube voltage 120kV, tube current 150mA; 80mL of non-ionic contrast agent is injected to obtain scanning data of the arterial phase and the venous phase; the obtained scanning data is input into the CT three-dimensional reconstruction software for three-dimensional reconstruction to generate three-dimensional image data.

[0044] Example 3, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S2, the model is generated, which specifically includes the following steps:

[0045] Step S21: noise removal, using Gaussian filtering to perform convolution operation on the three-dimensional image data, calculating the weighted average of each pixel to obtain a smoothed image; Gaussian filtering parameters are set to: standard deviation of 1.0, filter kernel size of 5×5, by weighted averaging the 5×5 neighborhood pixel values ​​around each pixel point in the three-dimensional image data, to calculate each pixel value in the smoothed image;

[0046] Step S22: Edge enhancement, using the Sobel operator to calculate the gradient value of the smoothed image, enhance the high-value area in the smoothed image, and highlight the edge area to obtain an enhanced image; the Sobel operator includes gradient templates in the horizontal direction and the vertical direction, and the gradient value of each pixel in the horizontal direction and the vertical direction is calculated by performing a convolution operation on the smoothed image. The formula used is as follows: ;

[0047] In the formula, represents the final gradient value, Represents the gradient value in the horizontal direction, Indicates the gradient value in the vertical direction;

[0048] Step S23: image segmentation, constructing a U-net model, separating the tumor area and surrounding tissue in the enhanced image to obtain a segmented image; the input enhanced image size is H×W×C, and the output segmented image size is H×W×1;

[0049] The U-net model includes an encoder, a bottleneck layer and a decoder, and restores edge information through jump connections; the encoder includes 4 downsampling layers, which are composed of convolutional layers and maximum pooling layers, and the structure of each downsampling layer is:

[0050] Layer 1: Convolutional layer: 3x3 convolution, 16 filters; Max pooling: 2x2, stride 2;

[0051] Layer 2: Convolutional layer: 3x3 convolution, 32 filters; Max pooling: 2x2, stride 2;

[0052] Layer 3: Convolutional layer: 3x3 convolution, 64 filters; Max pooling: 2x2, stride 2;

[0053] Layer 4: Convolutional layer: 3x3 convolution, 128 filters; Max pooling: 2x2, stride 2;

[0054] The bottleneck layer is located between the encoder and the decoder and includes a 3x3 convolutional layer with 256 filters;

[0055] The decoder includes 4 upsampling layers, each of which consists of a transposed convolution layer and a convolution layer, and the structures are:

[0056] Layer 1: Transposed convolution: 2x2, stride 2, 128 filters; skip connection to encoder layer 4 feature map; Convolution layer: 3x3, 128 filters;

[0057] Layer 2: Transposed convolution: 2x2, stride 2, 64 filters; skip connection with the feature map of encoder layer 3; Convolution layer: 3x3, 64 filters;

[0058] Layer 3: Transposed convolution: 2x2, stride 2, 32 filters; skip connection with the feature map of encoder layer 2; Convolution layer: 3x3, 32 filters;

[0059] Layer 4: Transposed convolution: 2x2, stride 2, 16 filters; skip connection with the feature map of encoder layer 1; Convolution layer: 3x3, 16 filters;

[0060] The output layer outputs a single-channel image representing the separation of tumor areas from non-tumor areas;

[0061] Step S24: Image optimization, constructing an optimization function to smooth the segmentation boundary and obtain the final three-dimensional anatomical model. The formula used is as follows: ;

[0062] In the formula, represents the optimization function, represents the segmentation boundary, represents the image gradient, and Weight parameter to control the smoothness and boundary strength of the optimization function.

[0063] By performing the operations, the traditional oral cancer image acquisition process is often affected by noise, resulting in poor ability to process complex structures and inability to provide stable segmentation results when the tumor boundaries are blurred and the morphology is irregular. This solution extracts the tumor area from the three-dimensional image data based on the U-Net model, separates it from the surrounding healthy tissue, automatically processes complex tumor morphology, provides higher segmentation accuracy and robustness, and finally generates an accurate three-dimensional anatomical model.

[0064] Example 4, see Figure 1 This embodiment is based on the above embodiment. In step S4, dynamic nursing specifically includes the following steps:

[0065] Step S41: Obtain the hospital's electronic medical record system, read the patient's individualized data, which includes the patient's age, gender, weight, medical history, and pathology report, and generate a preliminary nursing plan in combination with preoperative information support and surgical plan. The preliminary nursing plan includes the patient's rehabilitation plan, nutritional support, and psychological support;

[0066] Step S42: using monitoring equipment to collect the patient's real-time physiological indicators and treatment progress, and dynamically adjust the initial nursing plan; the real-time physiological indicators include heart rate, blood pressure, and blood oxygen saturation, and the treatment progress includes postoperative recovery and the occurrence of complications;

[0067] Step S43: Calculate the priority of each nursing measure in the preliminary nursing plan according to the patient's individualized data and real-time physiological indicators; the priority calculation is implemented by a particle swarm optimization algorithm, and the priority of each nursing measure is calculated according to the patient's individualized data and real-time physiological indicators. Nursing measures with high priorities are executed first. The formula used is as follows: ;

[0068] In the formula, Indicates the priority of nursing action, Indicates nursing measures, represents the characteristic function related to the nursing measures, Represents the weight of each feature;

[0069] Step S44: Sort and combine nursing measures based on priority to generate personalized nursing decisions; the personalized nursing decisions include the execution order and time schedule of various nursing measures. The generated nursing decisions are pushed to medical staff through the nursing system to guide actual nursing operations. The formula used is as follows: ; ;

[0070] In the formula, Indicates the nursing measures sorting results, Indicates the priority of nursing measures, represents the sorting algorithm, represents the execution time of each nursing measure, Indicates personalized care decisions.

[0071] By executing the above operations, traditional nursing methods lack specificity for perioperative nursing needs of specific diseases such as oral cancer, resulting in irregular nursing processes, low information integration, and insufficient decision support. This solution obtains the patient's individualized data, real-time physiological indicators, and treatment progress, uses an intelligent optimization algorithm to calculate the priority of each nursing measure, and dynamically generates personalized nursing decisions based on the priority, automatically providing patients with the most suitable nursing plans to better meet the personalized needs of different patients.

[0072] Example 5, see Figure 1This embodiment is based on the above embodiment. In step S5, the risk assessment report specifically includes: possible types of intraoperative complications and their probability of occurrence, including infection, nerve damage, and bleeding; the impact of complications on the patient's postoperative recovery, including the impact of complications on the patient's recovery time, quality of life, and subsequent treatment; prevention and response measures for each complication, including preoperative preparation, intraoperative operation, and postoperative care measures.

[0073] Example 6, see Figure 1 This embodiment is based on the above embodiment. In step S6, decision support specifically includes the following steps:

[0074] Step S61: Obtaining risk factors from the risk assessment report, the personalized care plan, and the patient's individualized data;

[0075] Step S62: risk probability assessment, using Bayesian networks to model different risk factors, analyzing whether the current situation matches the risk factors, and calculating the probability and expected severity of the risk factors;

[0076] Step S63: generating decision rules. If the intraoperative risk factor is assessed as high, outputting a drug use adjustment rule; if the postoperative risk factor is assessed as high, recommending infection examination and automatically suggesting whether the antibiotic treatment regimen needs to be adjusted;

[0077] Step S64: Intervention plan, providing nursing plans and intervention schedules during the postoperative recovery process, and dynamically adjusting nursing decisions.

[0078] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0079] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

[0080] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A nursing decision-making method for perioperative period of oral cancer, characterized by: The method comprises the following steps: Step S1: image acquisition, performing CT scanning on the neck of an oral cancer patient to obtain three-dimensional image data; Step S2: Model generation, using an intelligent iterative algorithm to process the three-dimensional image data to generate a three-dimensional anatomical model; Step S3: preoperative evaluation, based on the three-dimensional anatomical model, evaluating the size, location and surrounding key structures of the tumor, generating preoperative information support, and formulating a surgical plan based on the preoperative information support; Step S4: Dynamic nursing, obtaining individualized patient data, combining preoperative information support and surgical plan, generating a personalized nursing plan, and dynamically optimizing the personalized nursing plan using a multi-objective optimization method; Step S5: risk assessment, analyzing possible complications during surgery based on the personalized nursing plan and generating a risk assessment report; Step S6: Decision support, providing decision support for intraoperative and postoperative care based on the risk assessment report.

2. The method for oral cancer perioperative nursing decision-making according to claim 1, characterized in that: In step S2, the model is generated, including the following steps: Step S21: removing noise by performing convolution operation on the three-dimensional image data using Gaussian filtering, calculating the weighted average value of each pixel, and obtaining a smoothed image; Step S22: edge enhancement, using the Sobel operator to calculate the gradient value of the smoothed image, enhancing the high-value area in the smoothed image, and highlighting the edge area to obtain an enhanced image; Step S23: Image segmentation, constructing a U-net model, separating the tumor area and surrounding tissue in the enhanced image to obtain a segmented image; the U-net model includes an encoder, a bottleneck layer and a decoder, and restores edge information through a jump connection; the encoder includes a plurality of downsampling layers, each of which is composed of a convolutional layer and a maximum pooling layer, the decoder includes a plurality of upsampling layers, each of which is composed of a transposed convolutional layer and a convolutional layer, and the jump connection adds the feature map of the encoder to the feature map of the corresponding layer of the decoder element by element to restore the edge information; the output layer outputs a single-channel image representing the separation of the tumor area and the non-tumor area; Step S24: Image optimization, constructing an optimization function to smooth the segmentation boundary and obtain the final three-dimensional anatomical model. The formula used is as follows: ； In the formula, represents the optimization function, represents the segmentation boundary, represents the image gradient, and Weight parameter to control the smoothness and boundary strength of the optimization function.

3. The oral cancer perioperative nursing decision-making method according to claim 1, characterized in that: In step S4, the dynamic nursing comprises the following steps: Step S41: Obtain the hospital's electronic medical record system, read the patient's individualized data, and generate a preliminary nursing plan in combination with preoperative information support and surgical plan; Step S42: using monitoring equipment to collect the patient's real-time physiological indicators and treatment progress, and dynamically adjust the initial nursing plan; Step S43: Calculate the priority of each nursing measure in the preliminary nursing plan according to the patient's individualized data and real-time physiological indicators; the priority calculation is implemented by a particle swarm optimization algorithm, and the priority of each nursing measure is calculated according to the patient's individualized data and real-time physiological indicators. Nursing measures with high priorities are executed first. The formula used is as follows: ； In the formula, Indicates the priority of nursing action, Indicates nursing measures, represents the characteristic function related to the nursing measures, Represents the weight of each feature; Step S44: Sort and combine nursing measures based on priority to generate personalized nursing decisions; the personalized nursing decisions include the execution order and time schedule of various nursing measures. The generated nursing decisions are pushed to medical staff through the nursing system to guide actual nursing operations. The formula used is as follows: ； ； In the formula, Indicates the nursing measures sorting results, Indicates the priority of nursing measures, represents the sorting algorithm, represents the execution time of each nursing measure, Indicates personalized care decisions.

4. The oral cancer perioperative nursing decision-making method according to claim 1, characterized in that: In step S6, the decision support comprises the following steps: Step S61: Obtaining risk factors from the risk assessment report, the personalized care plan, and the patient's individualized data; Step S62: risk probability assessment, using Bayesian networks to model different risk factors, analyzing whether the current situation matches the risk factors, and calculating the probability and expected severity of the risk factors; Step S63: generating decision rules. If the intraoperative risk factor is assessed as high, outputting a drug use adjustment rule; if the postoperative risk factor is assessed as high, recommending infection examination and automatically suggesting whether the antibiotic treatment regimen needs to be adjusted; Step S64: Intervention plan, providing nursing plans and intervention schedules during the postoperative recovery process, and dynamically adjusting nursing decisions.

Images