Method and system for diagnosis and surgery matching verification based on graph encoding

Abstract

The application discloses a kind of based on graph coding's diagnosis and surgical matching check method and system, the method includes: obtaining the diagnosis data and surgical plan data of target object;Based on feature extraction algorithm, extract the diagnosis change feature and surgical step feature in the diagnosis data and the surgical plan data;Based on graph network coding algorithm, the diagnosis change feature and surgical step feature are encoded into corresponding node relationship graph;The node relationship graph is input into trained graph neural network to obtain the matching degree parameter of output diagnosis and surgical plan.It can be seen that the present application can realize accurate medical diagnosis and surgery matching evaluation based on graph structure feature association, improve the scientificity and adaptability of surgical plan selection, reduce the surgical risk and treatment deviation caused by mismatch between scheme and diagnosis.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a diagnostic and surgical matching verification method and system based on graph coding. Background Technology

[0002] With the rapid development of precision medicine and intelligent medicine, medical institutions are increasingly emphasizing data-driven methods to improve the scientific rigor and adaptability of surgical plan selection. A key technical issue is how to achieve accurate matching and evaluation of diagnosis and surgical plans to reduce surgical risks and efficacy deviations. Current technologies typically acquire diagnostic and surgical plan data for the target patient, employing simple feature statistics or manual comparison methods to assess the matching degree, and selecting surgical plans based on empirical rules to support clinical decision-making. However, existing solutions lack graph structure encoding of diagnostic variation features and surgical procedure features, as well as dynamic relationship reasoning using graph neural networks. This makes it difficult to accurately output the matching degree parameters between diagnosis and surgical plans. Commonly used linear or isolated analysis strategies cannot capture complex diagnostic-medical relationships, resulting in insufficient adaptability in plan selection. Mismatches between diagnosis and plan can easily lead to surgical risks or efficacy deviations, limiting the scientific nature of surgical decisions and patient prognosis. Therefore, existing technologies have shortcomings that urgently need to be addressed. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a diagnostic and surgical matching verification method and system based on graph coding, which can realize accurate medical diagnosis and surgical matching assessment based on graph structure feature association, improve the scientificity and adaptability of surgical plan selection, and reduce surgical risks and efficacy deviations caused by mismatch between plan and diagnosis.

[0004] To address the aforementioned technical problems, the first aspect of this invention discloses a diagnostic and surgical matching verification method based on graph coding, the method comprising: Acquire diagnostic data and surgical plan data for the target individuals; Based on feature extraction algorithms, diagnostic change features and surgical step features are extracted from the diagnostic data and surgical plan data; Based on the graph network coding algorithm, the diagnostic change features and surgical step features are encoded into corresponding node relationship graphs; The node relationship graph is input into the trained graph neural network to obtain the matching degree parameters of the output diagnosis and surgical plan.

[0005] As an optional implementation, in a first aspect of the invention, the diagnostic data includes medical diagnostic records of the target object at multiple historical time points; the medical diagnostic records include at least two of the following: consultation records, medical examination records, symptom analysis records, disease diagnosis records, and medication recommendations.

[0006] As an optional implementation, in the first aspect of the present invention, the surgical plan data includes at least one of the following: surgical executor, surgical procedure, surgical type, surgical history for the disease, and surgical history execution records.

[0007] As an optional implementation, in the first aspect of the present invention, the step of extracting diagnostic change features and surgical step features from the diagnostic data and the surgical plan data based on a feature extraction algorithm includes: The surgical plan data is input into the trained step prediction model to obtain multiple surgical steps and the key surgical points corresponding to each surgical step. For any two adjacent medical diagnostic records at corresponding historical time points in the diagnostic data, calculate the record similarity between the two medical diagnostic records; Based on the similarity of the records, the diagnostic change characteristics in the diagnostic data are determined.

[0008] As an optional implementation, in the first aspect of the present invention, the step prediction model is a neural network model, which is trained using a training dataset that includes multiple training surgical plan data and corresponding surgical step annotations and surgical point annotations.

[0009] As an optional implementation, in a first aspect of the invention, determining the diagnostic change features in the diagnostic data based on the record similarity includes: Calculate the average record similarity between each medical diagnostic record and all other adjacent medical diagnostic records to obtain the record specificity corresponding to each medical diagnostic record; The medical diagnostic records with a specificity greater than a preset specificity threshold are selected to obtain at least one specific diagnostic record; For each specific diagnostic record, records among all medical diagnostic records whose similarity to the specific diagnostic record is greater than a preset similarity threshold are identified as the relevant diagnostic records corresponding to the specific diagnostic record.

[0010] As an optional implementation, in the first aspect of the present invention, the graph network coding algorithm for encoding the diagnostic variation features and surgical step features into a corresponding node relationship graph includes: Each of the surgical steps described is defined as a surgical node; The surgical key points corresponding to each surgical step are determined as the node data corresponding to the surgical node. For each specific diagnostic record, the specific diagnostic record is determined as a record node, and the record data corresponding to the specific diagnostic record and all the related diagnostic records is determined as the node data corresponding to the record node; For any surgical node and any recording node, the node data corresponding to the surgical node and the recording node are respectively input into the trained feature association model to output the node connection features between the surgical node and the recording node; The network, which includes all the surgical nodes and the recording nodes, as well as the corresponding node connection features, is output as a node relationship graph.

[0011] As an optional implementation, in a first aspect of the present invention, the graph neural network includes an inter-node matching prediction model and an overall matching calculation model; the inter-node matching prediction model is used to calculate and output the node matching degree between any surgical node and any recording node; the overall matching calculation model is used to perform the following steps: For each node matching degree, determine the record data corresponding to all the relevant diagnostic records in the node data corresponding to the record node with that matching degree; The diagnostic overlap is obtained by calculating the average of the data similarity between each pair of the record data corresponding to all the relevant diagnostic records; Calculate the average of the matching scores of all other nodes corresponding to the record node corresponding to the matching score of the node, and obtain the record matching score; The surgical matching degree is obtained by calculating the average of the matching degrees of all other nodes corresponding to the surgical node corresponding to the matching degree of the node. Calculate the product of the diagnostic overlap, the record matching degree, and the surgical matching degree to obtain the calculation weight corresponding to the matching degree of the node; Based on the calculated weights, the matching degrees of all the nodes are weighted and averaged to obtain the matching degree parameters of the output diagnosis and surgical plan.

[0012] A second aspect of this invention discloses a graph-coding-based diagnostic and surgical matching verification system, the system comprising: The acquisition module is used to acquire diagnostic data and surgical plan data of the target object; The extraction module is used to extract diagnostic change features and surgical step features from the diagnostic data and surgical plan data based on a feature extraction algorithm. The encoding module is used to encode the diagnostic change features and surgical step features into a corresponding node relationship graph based on a graph network encoding algorithm; The matching module is used to input the node relationship graph into the trained graph neural network to obtain the matching degree parameters of the output diagnosis and surgical plan.

[0013] As an optional implementation, in a second aspect of the invention, the diagnostic data includes medical diagnostic records of the target object at multiple historical time points; the medical diagnostic records include at least two of the following: consultation records, medical examination records, symptom analysis records, disease diagnosis records, and medication recommendations.

[0014] As an optional implementation, in a second aspect of the invention, the surgical protocol data includes at least one of the following: surgical executor, surgical procedure, surgical type, surgical history for the specific condition, and surgical history execution records.

[0015] As an optional implementation, in a second aspect of the invention, the extraction module extracts diagnostic change features and surgical step features from the diagnostic data and surgical plan data based on a feature extraction algorithm, including the following specific methods: The surgical plan data is input into the trained step prediction model to obtain multiple surgical steps and the key surgical points corresponding to each surgical step. For any two adjacent medical diagnostic records at corresponding historical time points in the diagnostic data, calculate the record similarity between the two medical diagnostic records; Based on the similarity of the records, the diagnostic change characteristics in the diagnostic data are determined.

[0016] As an optional implementation, in the second aspect of the present invention, the step prediction model is a neural network model, which is trained using a training dataset that includes multiple training surgical plan data and corresponding surgical step annotations and surgical point annotations.

[0017] As an optional implementation, in a second aspect of the invention, the extraction module determines the specific method by which it determines the diagnostic change features in the diagnostic data based on the record similarity, including: Calculate the average record similarity between each medical diagnostic record and all other adjacent medical diagnostic records to obtain the record specificity corresponding to each medical diagnostic record; The medical diagnostic records with a specificity greater than a preset specificity threshold are selected to obtain at least one specific diagnostic record; For each specific diagnostic record, records among all medical diagnostic records whose similarity to the specific diagnostic record is greater than a preset similarity threshold are identified as the relevant diagnostic records corresponding to the specific diagnostic record.

[0018] As an optional implementation, in a second aspect of the invention, the encoding module encodes the diagnostic change features and surgical step features into a corresponding node relationship graph based on a graph network encoding algorithm, including: Each of the surgical steps described is defined as a surgical node; The surgical key points corresponding to each surgical step are determined as the node data corresponding to the surgical node. For each specific diagnostic record, the specific diagnostic record is determined as a record node, and the record data corresponding to the specific diagnostic record and all the related diagnostic records is determined as the node data corresponding to the record node; For any surgical node and any recording node, the node data corresponding to the surgical node and the recording node are respectively input into the trained feature association model to output the node connection features between the surgical node and the recording node; The network, which includes all the surgical nodes and the recording nodes, as well as the corresponding node connection features, is output as a node relationship graph.

[0019] As an optional implementation, in a second aspect of the invention, the graph neural network includes an inter-node matching prediction model and an overall matching calculation model; the inter-node matching prediction model is used to calculate and output the node matching degree between any surgical node and any recording node; the overall matching calculation model is used to perform the following steps: For each node matching degree, determine the record data corresponding to all the relevant diagnostic records in the node data corresponding to the record node with that matching degree; The diagnostic overlap is obtained by calculating the average of the data similarity between each pair of the record data corresponding to all the relevant diagnostic records; Calculate the average of the matching scores of all other nodes corresponding to the record node corresponding to the matching score of the node, and obtain the record matching score; The surgical matching degree is obtained by calculating the average of the matching degrees of all other nodes corresponding to the surgical node corresponding to the matching degree of the node. Calculate the product of the diagnostic overlap, the record matching degree, and the surgical matching degree to obtain the calculation weight corresponding to the matching degree of the node; Based on the calculated weights, the matching degrees of all the nodes are weighted and averaged to obtain the matching degree parameters of the output diagnosis and surgical plan.

[0020] A third aspect of the present invention discloses another diagnostic and surgical matching verification system based on graph coding, the system comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute some or all of the steps in the graph coding-based diagnostic and surgical matching verification method disclosed in the first aspect of the present invention.

[0021] The fourth aspect of the present invention discloses a computer storage medium storing computer instructions, which, when invoked, are used to execute some or all of the steps in the graph coding-based diagnostic and surgical matching verification method disclosed in the first aspect of the present invention.

[0022] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: This invention acquires diagnostic data and surgical plan data of the target object and extracts diagnostic change features and surgical step features. It uses a graph network coding algorithm to construct a node relationship graph, inputs it into a graph neural network, and outputs a matching parameter between the diagnosis and the surgical plan. This enables accurate medical diagnosis and surgery matching assessment based on graph structure feature association, improves the scientificity and adaptability of surgical plan selection, and reduces surgical risks and efficacy deviations caused by mismatch between the plan and the diagnosis. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a graph-coded diagnostic and surgical matching verification method disclosed in an embodiment of the present invention.

[0025] Figure 2 This is a schematic diagram of a graph-coded diagnostic and surgical matching verification system disclosed in an embodiment of the present invention.

[0026] Figure 3 This is a schematic diagram of another graph-coded diagnostic and surgical matching verification system disclosed in an embodiment of the present invention. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0029] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0030] This invention discloses a diagnostic and surgical matching verification method and system based on graph coding. By acquiring diagnostic data and surgical plan data of the target object and extracting diagnostic change features and surgical step features, a node relationship graph is constructed using a graph network coding algorithm. The graph network is then input into a graph neural network to output diagnostic and surgical plan matching parameters. This enables accurate medical diagnosis and surgical matching assessment based on graph structure feature association, improving the scientific rigor and adaptability of surgical plan selection, and reducing surgical risks and efficacy deviations caused by mismatch between the plan and diagnosis. Detailed explanations follow.

[0031] Example 1 Please see Figure 1 , Figure 1 This is a flowchart illustrating a graph-coding-based diagnostic and surgical matching verification method disclosed in an embodiment of the present invention. Figure 1 The described graph-coding-based diagnostic and surgical matching verification method can be applied to data processing systems / data processing equipment / data processing servers (including local processing servers or cloud processing servers). Figure 1 As shown, the graph-coding-based diagnostic and surgical matching verification method may include the following operations: 101. Obtain diagnostic data and surgical plan data for the target subject.

[0032] Optionally, diagnostic data may include medical diagnostic records of the target subject at multiple historical time points.

[0033] Optionally, medical diagnostic records may include at least two of the following: consultation records, medical examination records, symptom analysis records, disease diagnosis records, and medication recommendations.

[0034] Optionally, the surgical protocol data may include at least one of the following: surgeon, surgical procedure, surgical type, surgical history for the condition, and surgical history execution records.

[0035] Optionally, the diagnostic data may include medical records, imaging reports, laboratory test results, or multiple follow-up diagnostic conclusions, and the present invention does not limit this.

[0036] Optionally, the surgical plan data can be a surgical plan, surgical path description, step list, or intraoperative record; this invention does not impose any limitations.

[0037] 102. Based on feature extraction algorithms, extract diagnostic change features and surgical step features from diagnostic data and surgical plan data. Optionally, the feature extraction algorithm can be a hybrid extraction of BERT Chinese Medical Edition entity recognition + temporal difference or rule + deep learning, and this invention does not limit it.

[0038] 103. Based on graph network coding algorithm, the diagnostic change features and surgical step features are encoded into corresponding node relationship graphs. Optionally, the graph network encoding algorithm can be heterogeneous graph construction + relation embedding, which is not limited in this invention.

[0039] 104. Input the node relationship graph into the trained graph neural network to obtain the matching degree parameters of the output diagnosis and surgical plan.

[0040] Optionally, the matching degree parameter can be a continuous matching score or matching level of 0-1, which is not limited in this invention.

[0041] As can be seen, the above-described embodiments of the invention acquire diagnostic data and surgical plan data of the target object and extract diagnostic change features and surgical step features. They then use a graph network coding algorithm to construct a node relationship graph and input a graph neural network to output diagnostic and surgical plan matching parameters. This enables accurate medical diagnosis and surgical matching assessment based on graph structure feature association, improves the scientificity and adaptability of surgical plan selection, and reduces surgical risks and efficacy deviations caused by mismatch between the plan and the diagnosis.

[0042] As an optional embodiment, the step described above, extracting diagnostic change features and surgical step features from the diagnostic data and surgical plan data based on a feature extraction algorithm, includes: The surgical plan data is input into the trained step prediction model to obtain multiple surgical steps and the key points of each surgical step. For any two adjacent medical diagnostic records at corresponding historical time points in the diagnostic data, calculate the record similarity between the two medical diagnostic records; Based on record similarity, diagnostic variation characteristics in diagnostic data are determined.

[0043] Optionally, the step prediction model is a neural network model, which is trained using a training dataset that includes multiple training surgical plans and corresponding surgical step annotations and surgical point annotations.

[0044] Optionally, the step prediction model can be a 4-layer Transformer decoder (12 attention heads, 768 hidden dimensions), trained for 100 training epochs on 200,000 surgical protocol labeled data, with a step prediction F1 score of 0.94. This invention does not limit the model.

[0045] Optionally, the key points of the surgery may include critical instruments, precautions, risk points, or time estimates, which are not limited in this invention.

[0046] Optionally, the record similarity can be obtained using cosine similarity after Sentence-BERT encoding in the medical field; this invention does not impose any limitation on this method.

[0047] Optionally, the diagnostic change feature is a pair of records with similarity below a threshold and a description of their differences; this invention is not limited to this.

[0048] As can be seen, through the above optional embodiments, by inputting surgical plan data into the step prediction model and outputting surgical steps and key points, and identifying specific and related diagnostic records based on the similarity of adjacent records in diagnostic data, the accurate quantitative extraction of diagnostic change features is achieved, improving the accuracy and pertinence of diagnostic dynamic feature identification, providing high-quality change features for node relationship graph construction, and reducing the risk of matching evaluation deviation caused by misjudgment of change features.

[0049] As an optional embodiment, the step above, determining the diagnostic change characteristics in the diagnostic data based on record similarity, includes: Calculate the average record similarity between each medical diagnostic record and all other adjacent medical diagnostic records to obtain the record specificity corresponding to each medical diagnostic record; Medical diagnostic records with a specificity greater than a preset specificity threshold are selected to obtain at least one specific diagnostic record; For each specific diagnostic record, records among all medical diagnostic records whose similarity to that specific diagnostic record is greater than a preset similarity threshold are identified as the relevant diagnostic records corresponding to that specific diagnostic record.

[0050] Optionally, the relevant diagnostic records form a contextual cluster of specific changes, which is not limited in this invention.

[0051] As can be seen, through the above optional embodiments, by calculating the average similarity between medical diagnostic records as the record specificity to screen specific diagnostic records, and determining related diagnostic records based on the similarity threshold, accurate diagnostic change feature localization based on specificity and correlation analysis is achieved, improving the sensitivity and reliability of feature extraction, and reducing the risk of change feature distortion caused by interference from non-specific records.

[0052] As an optional embodiment, the step described above, encoding diagnostic variation features and surgical procedure features into a corresponding node relationship graph based on a graph network coding algorithm, includes: Each surgical step is defined as a surgical node; The key surgical points corresponding to each surgical step are identified as the node data corresponding to the corresponding surgical node. For each specific diagnostic record, the specific diagnostic record is defined as a record node, and the record data corresponding to the specific diagnostic record and all related diagnostic records is defined as the node data corresponding to the record node. For any surgical node and any recording node, the node data corresponding to the surgical node and the recording node are respectively input into the trained feature association model to output the node connection features between the surgical node and the recording node. The network, which includes all surgical nodes and recording nodes, as well as their corresponding node connection features, is output as a node relationship graph.

[0053] Optionally, the feature association model can be a dual-tower Siamese network (each tower has 3 fully connected layers, 512-256-128), trained with contrast loss, and trained for 80 training epochs on 150,000 diagnosis-surgery association pairs, with an association prediction AUC of 0.96. This invention does not limit the model.

[0054] Optionally, the node relationship graph is a heterogeneous graph, and the edges between surgical nodes and recording nodes have feature vectors with node connection characteristics. This invention does not limit this.

[0055] As can be seen, through the above optional embodiments, by constructing surgical nodes with surgical steps and key points, and constructing record nodes with specific and related diagnostic records, and using the feature association model to output node connection features to generate node relationship graphs, a structured graph representation of diagnostic changes and surgical steps is realized, which improves the topological capture capability and visualization expression of the relationship between features, and reduces the risk of loss of matching information caused by linear feature association.

[0056] As an optional embodiment, in the above steps, the graph neural network includes an inter-node matching prediction model and an overall matching calculation model; the inter-node matching prediction model is used to calculate and output the node matching degree between any surgical node and any recording node; the overall matching calculation model is used to perform the following steps: For each node matching degree, determine the record data corresponding to all relevant diagnostic records in the node data corresponding to the node matching degree; The diagnostic overlap is obtained by calculating the average of the pairwise similarity between the data of all relevant diagnostic records. Calculate the average of the matching scores of all other nodes corresponding to the matching score of the node, and obtain the record matching score; Calculate the average of the matching scores of all other nodes corresponding to the surgical node corresponding to the matching score of this node, and obtain the surgical matching score; Calculate the product of diagnostic overlap, record matching, and surgical matching to obtain the calculation weight corresponding to the matching degree of the node; Based on the calculated weights, the matching degree of all nodes is weighted and averaged to obtain the matching degree parameters of the output diagnosis and surgical plan.

[0057] Optionally, the node matching prediction model can be a 3-layer heterogeneous graph attention network (GAT), with 8 attention heads per layer and 256 hidden dimensions, trained on 120,000 node pairs. This invention does not impose any limitations.

[0058] Optionally, the weighted summation average can be used to aggregate local matching into global matching, but this invention does not limit this.

[0059] As can be seen, through the above optional embodiments, the matching degree of a single node is calculated by the node matching prediction model, and the overall matching calculation model obtains the final matching degree parameter by weighted summation of the matching degrees of all nodes based on the product of diagnostic overlap, record matching degree, and surgical matching degree. This achieves accurate diagnostic surgical plan matching quantification based on local matching and global weighting, improves the depth and objectivity of matching evaluation, and reduces the risk of overall matching deviation caused by a single matching dimension.

[0060] Example 2 Please see Figure 2 , Figure 2This is a schematic diagram of a graph-coding-based diagnostic and surgical matching verification system disclosed in an embodiment of the present invention. Figure 2 The described graph-coding-based diagnostic and surgical matching verification system can be applied to data processing systems / data processing equipment / data processing servers (including local processing servers or cloud processing servers). For example... Figure 2 As shown, the graph-coding-based diagnostic and surgical matching verification system may include: The acquisition module 201 is used to acquire diagnostic data and surgical plan data of the target object.

[0061] Extraction module 202 is used to extract diagnostic change features and surgical step features from diagnostic data and surgical plan data based on feature extraction algorithms. The encoding module 203 is used to encode diagnostic change features and surgical step features into corresponding node relationship graphs based on graph network encoding algorithms. The matching module 204 is used to input the node relationship graph into the trained graph neural network to obtain the matching degree parameters of the output diagnosis and surgical plan.

[0062] As can be seen, the above-described embodiments of the invention acquire diagnostic data and surgical plan data of the target object and extract diagnostic change features and surgical step features. They then use a graph network coding algorithm to construct a node relationship graph and input a graph neural network to output diagnostic and surgical plan matching parameters. This enables accurate medical diagnosis and surgical matching assessment based on graph structure feature association, improves the scientificity and adaptability of surgical plan selection, and reduces surgical risks and efficacy deviations caused by mismatch between the plan and the diagnosis.

[0063] As an optional embodiment, the diagnostic data includes medical diagnostic records of the target subject at multiple historical time points; the medical diagnostic records include at least two of the following: consultation records, medical examination records, symptom analysis records, disease diagnosis records, and medication recommendations.

[0064] As can be seen, the above optional embodiments limit the content of the diagnostic data, enabling the solution to perform matching calculations based on more comprehensive diagnostic record data, assisting in the realization of accurate medical diagnosis and surgery matching assessment based on graph structure feature association, and improving the scientificity and adaptability of surgical plan selection.

[0065] As an optional embodiment, the surgical protocol data includes at least one of the following: surgical executor, surgical procedure, surgical type, surgical history for the condition, and surgical history execution records.

[0066] As can be seen, the above optional embodiments limit the data content of the surgical plan data, so that the solution can perform matching calculations based on more comprehensive surgical plan data, assist in realizing accurate medical diagnosis and surgical matching evaluation based on graph structure feature association, and improve the scientificity and adaptability of surgical plan selection.

[0067] As an optional embodiment, the extraction module, based on a feature extraction algorithm, extracts diagnostic change features and surgical step features from diagnostic data and surgical plan data in the following specific ways: The surgical plan data is input into the trained step prediction model to obtain multiple surgical steps and the key points of each surgical step. For any two adjacent medical diagnostic records at corresponding historical time points in the diagnostic data, calculate the record similarity between the two medical diagnostic records; Based on record similarity, diagnostic variation characteristics in diagnostic data are determined.

[0068] As can be seen, through the above optional embodiments, by inputting surgical plan data into the step prediction model and outputting surgical steps and key points, and identifying specific and related diagnostic records based on the similarity of adjacent records in diagnostic data, the accurate quantitative extraction of diagnostic change features is achieved, improving the accuracy and pertinence of diagnostic dynamic feature identification, providing high-quality change features for node relationship graph construction, and reducing the risk of matching evaluation deviation caused by misjudgment of change features.

[0069] As an optional embodiment, the step prediction model is a neural network model, which is trained on a training dataset that includes multiple training surgical plan data and corresponding surgical step annotations and surgical point annotations.

[0070] As can be seen, the model details of the step prediction model are limited through the above optional embodiments, so as to more accurately predict multiple surgical steps corresponding to the surgical plan. This enables the solution to perform matching calculations based on more accurate surgical information, assists in realizing accurate medical diagnosis and surgical matching evaluation based on graph structure feature association, and improves the scientificity and adaptability of surgical plan selection.

[0071] As an optional embodiment, the extraction module determines the specific method by which it identifies diagnostic change features in the diagnostic data based on record similarity, including: Calculate the average record similarity between each medical diagnostic record and all other adjacent medical diagnostic records to obtain the record specificity corresponding to each medical diagnostic record; Medical diagnostic records with a specificity greater than a preset specificity threshold are selected to obtain at least one specific diagnostic record; For each specific diagnostic record, records among all medical diagnostic records whose similarity to that specific diagnostic record is greater than a preset similarity threshold are identified as the relevant diagnostic records corresponding to that specific diagnostic record.

[0072] As can be seen, through the above optional embodiments, by calculating the average similarity between medical diagnostic records as the record specificity to screen specific diagnostic records, and determining related diagnostic records based on the similarity threshold, accurate diagnostic change feature localization based on specificity and correlation analysis is achieved, improving the sensitivity and reliability of feature extraction, and reducing the risk of change feature distortion caused by interference from non-specific records.

[0073] As an optional embodiment, the encoding module, based on a graph network encoding algorithm, encodes diagnostic variation features and surgical step features into corresponding node relationship graphs in the following specific ways: Each surgical step is defined as a surgical node; The key surgical points corresponding to each surgical step are identified as the node data corresponding to the corresponding surgical node. For each specific diagnostic record, the specific diagnostic record is defined as a record node, and the record data corresponding to the specific diagnostic record and all related diagnostic records is defined as the node data corresponding to the record node. For any surgical node and any recording node, the node data corresponding to the surgical node and the recording node are respectively input into the trained feature association model to output the node connection features between the surgical node and the recording node. The network, which includes all surgical nodes and recording nodes, as well as their corresponding node connection features, is output as a node relationship graph.

[0074] As can be seen, through the above optional embodiments, by constructing surgical nodes with surgical steps and key points, and constructing record nodes with specific and related diagnostic records, and using the feature association model to output node connection features to generate node relationship graphs, a structured graph representation of diagnostic changes and surgical steps is realized, which improves the topological capture capability and visualization expression of the relationship between features, and reduces the risk of loss of matching information caused by linear feature association.

[0075] As an optional embodiment, the graph neural network includes an inter-node matching prediction model and an overall matching calculation model; the inter-node matching prediction model is used to calculate and output the node matching degree between any surgical node and any recording node; the overall matching calculation model is used to perform the following steps: For each node matching degree, determine the record data corresponding to all relevant diagnostic records in the node data corresponding to the node matching degree; The diagnostic overlap is obtained by calculating the average of the pairwise similarity between the data of all relevant diagnostic records. Calculate the average of the matching scores of all other nodes corresponding to the matching score of the node, and obtain the record matching score; Calculate the average of the matching scores of all other nodes corresponding to the surgical node corresponding to the matching score of this node, and obtain the surgical matching score; Calculate the product of diagnostic overlap, record matching, and surgical matching to obtain the calculation weight corresponding to the matching degree of the node; Based on the calculated weights, the matching degree of all nodes is weighted and averaged to obtain the matching degree parameters of the output diagnosis and surgical plan.

[0076] As can be seen, through the above optional embodiments, the matching degree of a single node is calculated by the node matching prediction model, and the overall matching calculation model obtains the final matching degree parameter by weighted summation of the matching degrees of all nodes based on the product of diagnostic overlap, record matching degree, and surgical matching degree. This achieves accurate diagnostic surgical plan matching quantification based on local matching and global weighting, improves the depth and objectivity of matching evaluation, and reduces the risk of overall matching deviation caused by a single matching dimension.

[0077] Example 3 Please see Figure 3 , Figure 3 This is another diagnostic and surgical matching verification system based on graph coding disclosed in the embodiments of the present invention. Figure 3 The described graph-coding-based diagnostic and surgical matching verification system is applied in a data processing system / data processing equipment / data processing server (wherein, the server includes a local processing server or a cloud processing server). Figure 3 As shown, the graph-coding-based diagnostic and surgical matching verification system may include: Memory 301 storing executable program code; Processor 302 coupled to memory 301; The processor 302 calls the executable program code stored in the memory 301 to execute the steps of the graph coding-based diagnostic and surgical matching verification method described in Embodiment 1.

[0078] Example 4 This invention discloses a computer read storage medium that stores a computer program for electronic data interchange, wherein the computer program causes a computer to execute the steps of the graph coding-based diagnostic and surgical matching verification method described in Embodiment 1.

[0079] Example 5 This invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform the steps of the graph coding-based diagnostic and surgical matching verification method described in Embodiment 1.

[0080] The foregoing has described specific embodiments of this specification; other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily have to follow the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0081] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0082] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware components.

[0083] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0084] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0085] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0086] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0087] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0088] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0089] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0090] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0091] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0092] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0093] Finally, it should be noted that the diagnostic and surgical matching verification method and system based on graph coding disclosed in the embodiments of the present invention are merely preferred embodiments of the present invention, and are only used to illustrate the technical solutions of the present invention, not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A diagnostic and surgical matching verification method based on graph coding, characterized in that, The method includes: Acquire diagnostic data and surgical plan data for the target individuals; Based on feature extraction algorithms, diagnostic change features and surgical step features are extracted from the diagnostic data and surgical plan data; Based on the graph network coding algorithm, the diagnostic change features and surgical step features are encoded into corresponding node relationship graphs; The node relationship graph is input into the trained graph neural network to obtain the matching degree parameters of the output diagnosis and surgical plan.

2. The diagnostic and surgical matching verification method based on graph coding according to claim 1, characterized in that, The diagnostic data includes medical diagnostic records of the target subject at multiple historical time points; the medical diagnostic records include at least two of the following: consultation records, medical examination records, symptom analysis records, disease diagnosis records, and medication recommendations.

3. The diagnostic and surgical matching verification method based on graph coding according to claim 1, characterized in that, The surgical protocol data includes at least one of the following: surgeon, surgical procedure, surgical type, surgical history for the specific condition, and surgical history execution records.

4. The diagnostic and surgical matching verification method based on graph coding according to claim 2, characterized in that, The feature extraction algorithm extracts diagnostic change features and surgical step features from the diagnostic data and surgical plan data, including: The surgical plan data is input into the trained step prediction model to obtain multiple surgical steps and the key surgical points corresponding to each surgical step. For any two adjacent medical diagnostic records at corresponding historical time points in the diagnostic data, calculate the record similarity between the two medical diagnostic records; Based on the similarity of the records, the diagnostic change characteristics in the diagnostic data are determined.

5. The diagnostic and surgical matching verification method based on graph coding according to claim 4, characterized in that, The step prediction model is a neural network model, which is trained using a training dataset that includes multiple training surgical plan data and corresponding surgical step annotations and surgical point annotations.

6. The diagnostic and surgical matching verification method based on graph coding according to claim 4, characterized in that, The step of determining the diagnostic change features in the diagnostic data based on the record similarity includes: Calculate the average record similarity between each medical diagnostic record and all other adjacent medical diagnostic records to obtain the record specificity corresponding to each medical diagnostic record; The medical diagnostic records with a specificity greater than a preset specificity threshold are selected to obtain at least one specific diagnostic record; For each specific diagnostic record, records among all medical diagnostic records whose similarity to the specific diagnostic record is greater than a preset similarity threshold are identified as the relevant diagnostic records corresponding to the specific diagnostic record.

7. The diagnostic and surgical matching verification method based on graph coding according to claim 6, characterized in that, The graph network-based coding algorithm encodes the diagnostic change features and surgical step features into a corresponding node relationship graph, including: Each of the surgical steps described is defined as a surgical node; The surgical key points corresponding to each surgical step are determined as the node data corresponding to the surgical node. For each specific diagnostic record, the specific diagnostic record is determined as a record node, and the record data corresponding to the specific diagnostic record and all the related diagnostic records is determined as the node data corresponding to the record node; For any surgical node and any recording node, the node data corresponding to the surgical node and the recording node are respectively input into the trained feature association model to output the node connection features between the surgical node and the recording node; The network, which includes all the surgical nodes and the recording nodes, as well as the corresponding node connection features, is output as a node relationship graph.

8. The diagnostic and surgical matching verification method based on graph coding according to claim 7, characterized in that, The graph neural network includes an inter-node matching prediction model and an overall matching calculation model; the inter-node matching prediction model is used to calculate and output the node matching degree between any surgical node and any recording node; the overall matching calculation model is used to perform the following steps: For each node matching degree, determine the record data corresponding to all the relevant diagnostic records in the node data corresponding to the record node with that matching degree; The diagnostic overlap is obtained by calculating the average of the data similarity between each pair of the record data corresponding to all the relevant diagnostic records; Calculate the average of the matching scores of all other nodes corresponding to the record node corresponding to the matching score of the node, and obtain the record matching score; The surgical matching degree is obtained by calculating the average of the matching degrees of all other nodes corresponding to the surgical node corresponding to the matching degree of the node. Calculate the product of the diagnostic overlap, the record matching degree, and the surgical matching degree to obtain the calculation weight corresponding to the matching degree of the node; Based on the calculated weights, the matching degrees of all the nodes are weighted and averaged to obtain the matching degree parameters of the output diagnosis and surgical plan.

9. A diagnostic and surgical matching verification system based on graph coding, characterized in that, The system includes: The acquisition module is used to acquire diagnostic data and surgical plan data of the target object; The extraction module is used to extract diagnostic change features and surgical step features from the diagnostic data and surgical plan data based on a feature extraction algorithm. The encoding module is used to encode the diagnostic change features and surgical step features into a corresponding node relationship graph based on a graph network encoding algorithm; The matching module is used to input the node relationship graph into the trained graph neural network to obtain the matching degree parameters of the output diagnosis and surgical plan.

10. A diagnostic and surgical matching verification system based on graph coding, characterized in that, The system includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the graph coding-based diagnostic and surgical matching verification method as described in any one of claims 1-8.

Images