A surgical plan evaluation method, device, system and storage medium

Abstract

The present invention discloses a surgical plan evaluation method, device, system and storage medium. The surgical plan evaluation method includes: obtaining medical record information, determining a surgical alternative plan data set based on the medical record information through a first knowledge graph; obtaining medical care information, determining a model input data set based on the medical care information, the surgical alternative plan data set and the medical record information, wherein one model input data of the model input data set contains a surgical alternative plan; taking each model input data or model input data set in the model input data set as input, classifying the surgical alternative plans contained in each model input data through a model, and determining the best surgical plan based on the classification result. In the surgical plan evaluation method proposed by the present invention, the surgical alternative plan is determined through a knowledge graph, and the model is used to determine whether the surgical alternative plan is the best surgical plan, and the reliability of the best surgical plan finally determined is high.

Description

Technical Field

[0001] Embodiments of the present invention relate to artificial intelligence technology, and more particularly to a surgical plan evaluation method, device, system and storage medium. Background Art

[0002] Surgery is a complex medical procedure. Since patients vary greatly, it is necessary to select a surgical plan that is suitable for the patient from a wide range of options.

[0003] At present, surgical plans are mostly evaluated manually. When determining some difficult surgical plans, it may require multiple departments and multiple doctors to jointly determine the surgical plan, which consumes a large amount of already scarce medical resources. At the same time, manual evaluation has great limitations. The reasons are: patients vary greatly. Even for the same surgical needs, different doctors or different patients, and different regions will have more or less different surgical plans. It is difficult to achieve comprehensive considerations during manual evaluation; manual evaluation often tends to choose experienced doctors, which is biased.

[0004] Based on the above problems, the market urgently needs an automated surgical plan evaluation method and system to replace doctors in performing feasibility analysis of surgeries. Summary of the invention

[0005] The present invention provides a surgical plan evaluation method, device, system and storage medium to achieve the purpose of automatically realizing surgical plan evaluation.

[0006] In a first aspect, an embodiment of the present invention provides a method for evaluating a surgical plan, comprising:

[0007] Acquire medical record information, and determine a surgical alternative plan dataset through a first knowledge graph based on the medical record information;

[0008] Acquire medical care information, and determine a model input data set based on the medical care information, the surgical alternative plan data set, and the medical record information, wherein one model input data of the model input data set includes a surgical alternative plan;

[0009] Each model input data in the model input data set or the model input data set is used as input, the surgical alternatives contained in each model input data are classified by the model, and the best surgical plan is determined based on the classification result.

[0010] Optionally, the model is a graph attention model;

[0011] The model input data set includes a specified number of nodes, each of which contains a specified number of features.

[0012] Optionally, obtaining medical record information, and determining a surgical alternative solution dataset based on the medical record information through a first knowledge graph includes:

[0013] Extracting information from the medical record information to obtain a first entity keyword and a first relationship keyword, and constructing a first graph data structure according to the first entity keyword and the first relationship keyword;

[0014] The first graph data structure is matched with the first knowledge graph, and the surgical alternative plan data set is determined according to the matching result.

[0015] Optionally, also include:

[0016] Extracting data from the medical care information, the surgical alternative plan data set, and the medical record information to obtain a second entity keyword and a second relationship keyword;

[0017] constructing a second graph data structure according to the second entity keyword and the second relationship keyword;

[0018] The second graph data structure is matched with the second knowledge graph, and an emergency data set is determined according to the matching result.

[0019] Optionally, also include:

[0020] The first knowledge graph is updated according to the optimal surgical plan and the model input data corresponding to the optimal surgical plan.

[0021] Optionally, also include:

[0022] The second knowledge graph is updated according to the optimal surgical plan, the model input data corresponding to the optimal surgical plan, and emergencies.

[0023] Optionally, the model input data is a word vector or a word matrix.

[0024] In a second aspect, an embodiment of the present invention further provides a surgical plan evaluation device, including a surgical plan evaluation unit, wherein the surgical plan evaluation unit is used to:

[0025] Acquire medical record information, and determine a surgical alternative plan dataset through a first knowledge graph based on the medical record information;

[0026] Acquire medical care information, and determine a model input data set based on the medical care information, the surgical alternative plan data set, and the medical record information;

[0027] Each model input data in the model input data set is taken as input respectively, each model input data is classified by the model, and the best surgical plan is determined based on the classification result.

[0028] In a third aspect, an embodiment of the present invention further provides a surgical plan evaluation system, including a terminal, wherein the terminal is configured with an executable program, and when the executable program is run, the surgical plan evaluation method recorded in the embodiment of the present invention is implemented.

[0029] In a fourth aspect, an embodiment of the present invention further provides a storage medium on which an executable program is stored, and when the executable program is run, the surgical plan evaluation method recorded in the embodiment of the present invention is implemented.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. In the surgical plan evaluation method proposed in the present invention, information extraction is performed on the medical record information, and triples are generated using the extracted information. The surgical alternatives are determined using the triples and the first knowledge graph. Based on the knowledge graph, feasible surgical alternatives can be determined to the greatest extent and as much as possible, which can avoid the problem of limitations in surgical alternatives due to incomplete considerations or experience limits when surgical alternatives are manually determined.

[0032] 2. In the surgical plan evaluation method proposed in the present invention, after the surgical alternatives are determined, model input data is generated based on the surgical alternatives, and the model is used to determine whether the surgical alternatives contained in the model input data are the best surgical plan. The reliability of the best surgical plan finally determined is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is a flow chart of a surgical plan evaluation method in an embodiment;

[0034] Figure 2 is another flow chart of a surgical plan evaluation method in an embodiment;

[0035] Figure 3 is a schematic structural diagram of a surgical plan evaluation device in an embodiment;

[0036] Figure 4 It is a schematic diagram of the structure of another surgical plan evaluation device in the embodiment. DETAILED DESCRIPTION

[0037] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0038] Embodiment 1

[0039] Figure 1is a flow chart of the surgical plan evaluation method in the embodiment, refer to Figure 1 , surgical plan evaluation methods include:

[0040] S101. Obtain medical record information, and determine a surgical alternative plan data set through a first knowledge graph based on the medical record information.

[0041] Illustratively, in this embodiment, the medical record information may include text information such as the patient's name, gender, medical history, diagnosis results, etc., and may also include image information such as angiography.

[0042] Exemplarily, different information included in the medical record information can be stored centrally in one electronic system, or distributedly in different electronic systems.

[0043] When obtaining medical record information, if the above-mentioned different information is stored in one electronic system, the medical record information is directly retrieved from the electronic system; if the above-mentioned different information is distributed and stored in different electronic systems, the required information is collected through all relevant electronic systems, and the collected information is integrated to form medical record information.

[0044] For example, a knowledge graph refers to a relational network formed by connecting different types of heterogeneous information. The essence of a knowledge graph is a graph-based data structure, which is mainly composed of nodes and edges. In a knowledge graph, each node represents an entity (or thing) that can exist independently, and each edge is a relationship between entities.

[0045] Exemplarily, in this embodiment, the storage form of the knowledge graph is not limited. The knowledge graph can be stored based on the RDF storage format, in the form of a graph database, in the form of a relational database, etc.

[0046] Exemplarily, in this embodiment, the nodes included in the first knowledge graph include: surgical alternatives, and some or all nodes that can be extracted from medical record information.

[0047] Exemplarily, in this embodiment, determining the surgical alternative solution dataset based on the medical record information through the first knowledge graph specifically includes:

[0048] Step 1: Extract entities from medical record information.

[0049] For example, taking medical record information as text information, when performing entity extraction on the medical record information, the entity can be determined by extracting keywords. For example, if the extracted word is included in the set vocabulary, the word is regarded as an entity;

[0050] When performing entity extraction on medical record information, the entity type can also be set in advance. After the words are extracted, the words are classified through the neural network model. If the classification result of the word belongs to the pre-set entity type, the word is regarded as an entity.

[0051] Step 2: Extract relationships from medical record information.

[0052] Exemplarily, the purpose of relation extraction is to determine the association between entities. In this step, the method used to extract relation from medical record information is the same as that of the prior art. For example, relation extraction can be implemented by using statistical machine learning methods, supervised relation extraction methods, etc.

[0053] Step 3: Use entities and relationships to construct triples, and determine the surgical alternative plan data set based on the matching relationship between the triples and the first knowledge graph.

[0054] Exemplarily, in this step, a triple refers to a data unit in a specified format. For example, if the specified format is entity-relationship-entity, the triple can be patient (Zhang San)-physiological index (blood pressure value)-index item (blood pressure), drug (drug name)-adverse reaction (symptom)-patient (Zhang San), medical history (symptom name)-drug use duration-drug, etc.

[0055] Exemplarily, in this embodiment, some entities obtained from a medical record information may not be included in the first knowledge graph;

[0056] For example, if the relationship network corresponding to the first knowledge graph includes patient-physiological indicator-indicator item, drug-adverse reaction-patient, medical history-duration of medication-drug, then the triples constructed through a medical record information can be drug-adverse reaction-patient, patient-genetic disease-family member.

[0057] Exemplarily, when matching the triples with the first knowledge graph, the data units in the first knowledge graph that have a higher degree of match with the triples are determined, and then at least two sub-networks containing the highly matched data units are determined, and the surgical alternatives contained in or corresponding to each sub-network are extracted to form a surgical alternatives data set.

[0058] Exemplarily, the degree of match between the triple and the data unit in the first knowledge graph can be determined based on whether the information corresponding to the entities is the same or similar, and whether the information corresponding to the relationships is the same or similar.

[0059] S102. Obtain medical information, and determine a model input data set based on the medical information, a surgical alternative plan data set, and medical record information.

[0060] Exemplarily, in this embodiment, the medical care information may include an alternative doctor information set and an alternative region information set, wherein the alternative doctor information set includes information on several alternative doctors, and the alternative region information set includes several alternative surgical regions.

[0061] When determining the model input data set, the candidate doctor information, the candidate surgical areas, and the candidate surgical plans are arranged and combined. One arrangement and combination method corresponds to one model input data, and all the model input data constitute the model input data set.

[0062] S103. Take each model input data in the model input data set as input, classify each model input data through the model, and determine the best surgical plan based on the classification result.

[0063] Exemplarily, in this embodiment, the model may be a neural network model.

[0064] In this embodiment, the configuration model is used to determine the classification category to which the model input data belongs, wherein the classification category may be recommended, general, not recommended, etc.

[0065] Illustratively, in this embodiment, the surgical alternatives included in the model input data of the recommendation category are taken as the best surgical plan, and there may be multiple best surgical plans corresponding to one model input data set.

[0066] In the surgical plan evaluation method proposed in this embodiment, information is extracted from the medical record information, triples are generated using the extracted information, and surgical alternatives are determined using the triples and the first knowledge graph. Based on the knowledge graph, feasible surgical alternatives can be determined to the greatest extent possible, thereby avoiding the problem of limitations in surgical alternatives due to incomplete considerations or experience limits when surgical alternatives are manually determined.

[0067] In the surgical plan evaluation method proposed in this embodiment, after the surgical alternatives are determined, model input data is generated based on the surgical alternatives, and the model is used to determine whether the surgical alternatives included in the model input data are the best surgical plan. The reliability of the best surgical plan finally determined is high.

[0068] Figure 2 is another flow chart of the surgical plan evaluation method in the embodiment, refer to Figure 2 As a feasible solution, the surgical plan evaluation method includes:

[0069] S201. Obtain medical record information, and determine a surgical alternative plan data set through a first knowledge graph based on the medical record information.

[0070] Exemplarily, in this solution, the medical record information may include the patient's basic information, physiological indicator information, angiography screening information, preoperative tracking information, and similar case information.

[0071] Exemplarily, based on the content recorded in step S101, the first knowledge graph is stored in the form of a graph database.

[0072] Accordingly, in this solution, obtaining medical record information and determining a surgical alternative plan data set based on the medical record information through the first knowledge graph includes:

[0073] Extracting information from the medical record information to obtain a first entity keyword and a first relationship keyword, and constructing a first graphic data structure according to the first entity keyword and the first relationship keyword;

[0074] The first graph data structure is matched with the first knowledge graph (graph data structure stored in the graph database), and a surgical alternative plan data set is determined according to the matching result.

[0075] Exemplarily, based on the graph database, the time required to determine surgical alternatives can be shortened and the execution efficiency of the surgical plan evaluation method can be improved.

[0076] Exemplarily, in this solution, the method of constructing the first graph data structure is the same as the method of constructing the knowledge graph, and its specific implementation process will not be repeated.

[0077] S202. Obtain medical information, and determine a model input data set based on the medical information, a surgical alternative plan data set, and medical record information.

[0078] Exemplarily, in this solution, the medical information may include doctor information and regional information.

[0079] Exemplarily, in this solution, the model uses the Graph Attention Networks (GAT).

[0080] Exemplarily, this solution does not involve improvements to the GAT model. The number of graph attention layers in the GAT model can be set according to demand. For example, 1 layer, 2 layers, or 3 layers of graph attention layers can be set.

[0081] For example, in the GAT model, the input of a graph attention layer is a node feature vector set (feature matrix), namely:

[0082]

[0083] In the above formula, N represents the number of nodes, and F represents the number of features corresponding to the nodes. is an F×1-order vector. The output of the graph attention layer is a new feature vector set of the node, namely:

[0084]

[0085]

[0086] a ij =softmax j (e ij )

[0087]

[0088] In the above formula, F′ represents the number of new features corresponding to the node, and F′ can be different from F; h′ i is a F′×1-order vector; a() represents a function. Before the GAT model is trained, a() is the set value. After the GAT model is trained, a() is the training value; softmax is an activation function. j (e ij ) represents the regularization of the neighboring nodes j of node i; σ() is a nonlinear activation function; j∈N i Represents the traversal of the adjacent nodes j of node i; W is a weight matrix of order F×F′. Before the GAT model is trained, W is the set value. After the GAT model is trained, W is the training value.

[0089] For example, in practical applications, F and F′ of each graph attention layer in the GAT model can be set freely, and F′ of the last graph attention layer can be set to the number of categories to be classified. At this time, h′ of the last graph attention layer is i The included features represent the probability that the node belongs to each category.

[0090] Based on the working principle of the GAT model, the model input data set can be determined as follows: select an information category in medical care information, medical record information, or surgical alternatives as a node, use the unselected information category as the feature of the node, and build a graphical network based on the nodes and features.

[0091] For example, surgical alternatives are selected as nodes, and basic information, physiological index information, angiography screening information, preoperative tracking information, similar case information, doctor information, and regional information are used as feature items of the nodes;

[0092] When constructing a graph network, if the surgical alternatives corresponding to two nodes are similar, the two nodes are connected by an edge so that the two nodes are neighbor nodes to each other;

[0093] When configuring the features of a node, if the feature item is empty, the value corresponding to the feature item is set to 0. Otherwise, the word vector technology can be used to map the words or phrases contained in each feature item to real numbers, thereby generating a feature vector.

[0094] S203. Taking the model input data set as input, classifying the surgical alternatives contained in the model input data set through the model, and determining the best surgical plan based on the classification results.

[0095] For example, based on the content recorded in step S202, the number of categories to be classified can be set to 3, and the classification categories can be recommended, general, and not recommended;

[0096] The input of the GAT model is a feature matrix composed of feature vectors corresponding to all nodes, and the output of the GAT model is the probability of each surgical alternative to each category.

[0097] S204. Determine the emergency situation data set through the second knowledge graph based on the medical care information, the surgical alternative plan data set, and the medical record information.

[0098] Exemplarily, in this solution, the nodes contained in the second knowledge graph include at least: surgical alternatives, emergencies, all nodes that can be extracted from medical record information, and all nodes that can be extracted from medical care information.

[0099] In this solution, the second knowledge graph is stored in the form of a graph database. Similar to the content recorded in step S201, the emergency situation data set determined by the second knowledge graph includes:

[0100] Extracting data from medical care information, surgical alternative plan data sets, and medical record information to obtain second entity keywords and second relationship keywords;

[0101] constructing a second graph data structure according to the second entity keyword and the second relationship keyword;

[0102] The second graph data structure is matched with the second knowledge graph, and the emergency data set is determined according to the matching result.

[0103] S205. Update the first knowledge graph according to the optimal surgical plan and the model input data corresponding to the optimal surgical plan.

[0104] Exemplarily, based on the content recorded in step S202, the model input data includes surgical alternatives, basic information, physiological index information, angiography screening information, preoperative tracking information, similar case information, doctor information, and regional information;

[0105] Among them, basic information, physiological index information, angiography screening information, preoperative tracking information, and similar case information correspond to entities and relationships extracted from medical record information pairs;

[0106] When updating the first knowledge graph, determine whether the extracted entities are included in the first knowledge graph. If there are entities not included in the first knowledge graph, find the surgical alternative plan node corresponding to the best surgical plan, and update the entities and relationships not included in the first knowledge graph to the subnetwork to which the surgical alternative plan node belongs.

[0107] S206. Update the second knowledge graph according to the best surgical plan, the model input data corresponding to the best surgical plan, and the surgical results.

[0108] Exemplarily, in this solution, the surgical result includes an emergency, and before updating the second knowledge graph, it is determined whether the emergency is included in the second knowledge graph;

[0109] If the emergency situation is not included in the second knowledge graph, then on the basis of the original second knowledge graph, the second knowledge graph is reconstructed in combination with the emergency situation corresponding to the best surgical plan, all nodes and relationships that can be extracted from the medical record information, and all entities and relationships that can be extracted from the medical care information, so that the new emergency situation is added to the second knowledge graph.

[0110] Exemplarily, in this solution, the method used to update the knowledge graph in step S205 and step S206 is the same as that in the prior art, and the specific process will not be repeated here.

[0111] exist Figure 1 On the basis of the beneficial effects of the scheme shown, in this scheme, the second knowledge graph is used to predict the emergencies that may occur during the operation, so that medical staff can predict the relevant work points during and after the operation to improve the success rate of the operation.

[0112] In addition, in this solution, after the surgical plan is evaluated by the GAT model, the first knowledge graph and the second knowledge graph are updated according to the evaluation results, and then the first knowledge graph and the second knowledge graph are gradually improved to further improve the accuracy of the surgical plan evaluation.

[0113] Embodiment 2

[0114] Figure 3 is a schematic diagram of the structure of the surgical plan evaluation device in the embodiment, refer to Figure 3 This embodiment proposes a surgical plan evaluation device, including a surgical plan evaluation unit, which includes an information acquisition module 100, a first knowledge graph module 200, and a surgical plan evaluation module 300.

[0115] The information acquisition module 100 is used to acquire medical record information and medical care information.

[0116] The first knowledge graph module 200 stores a first knowledge graph, and the first knowledge graph module 200 is used to determine a surgical alternative plan data set based on medical record information.

[0117] The surgical plan evaluation module 300 is configured with a model, and the surgical plan evaluation module 300 is used to determine the model input data set based on medical information, surgical alternative plan data set and medical record information, classify the surgical alternative plans in the model input data set, and determine the best surgical plan based on the classification results.

[0118] Figure 4 is a schematic diagram of another surgical plan evaluation device in the embodiment, referring to Figure 4 As a feasible implementation scheme, the surgical plan evaluation device includes an information acquisition module 100, a first knowledge graph module 200, a surgical plan evaluation module 300, and a second knowledge graph module 400.

[0119] The information acquisition module 100 is used to acquire medical record information and medical care information.

[0120] The first knowledge graph module 200 stores a first knowledge graph, and the first knowledge graph module 200 is used to determine a data set of alternative surgical plans based on medical record information; and update the first knowledge graph according to the best surgical plan and the model input data corresponding to the best surgical plan.

[0121] The surgical plan evaluation module 300 is configured with a GAT model, and is used to determine a model input data set based on medical information, surgical alternative plan data sets, and medical record information, classify the surgical alternative plans in the model input data set, and determine the best surgical plan based on the classification results.

[0122] The second knowledge graph module 400 stores a second knowledge graph, which is used to determine an emergency data set based on medical information, a surgical alternative plan data set, and medical record information; and update the second knowledge graph according to the best surgical plan, the model input data corresponding to the best surgical plan, and the surgical results.

[0123] In this embodiment, the working method and beneficial effects of the surgical plan evaluation device are the same as the corresponding contents recorded in the first embodiment.

[0124] Embodiment 3

[0125] This embodiment proposes a surgical plan evaluation system, including a terminal, and typical terminals include intelligent electronic devices, medical self-service terminals, etc. The terminal is configured with an executable program, and when the executable program is run, any surgical plan evaluation method recorded in the first embodiment is implemented.

[0126] Embodiment 4

[0127] This embodiment provides a storage medium on which an executable program is stored. When the executable program is run, any one of the surgical plan evaluation methods described in the first embodiment is implemented.

[0128] Exemplarily, the storage medium may adopt any combination of one or more computer-readable media, which may be a computer-readable signal medium or a computer-readable storage medium.

[0129] Computer-readable storage media are not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or device.

[0130] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0131] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0132] Computer program code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or terminal. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0133] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A surgical plan evaluation method, characterized in that: include: Acquire medical record information, and determine a surgical alternative plan dataset through a first knowledge graph based on the medical record information; Acquire medical care information, and determine a model input data set based on the medical care information, the surgical alternative plan data set, and the medical record information, wherein one model input data of the model input data set includes a surgical alternative plan; Taking each model input data in the model input data set or the model input data set as input, classifying the surgical alternatives contained in each model input data through the model, and determining the best surgical plan based on the classification result; The medical care information includes an alternative doctor information set and an alternative region information set, wherein the alternative doctor information set includes information of several alternative doctors, and the alternative region information set includes several alternative surgical regions; When determining the model input data set, the candidate doctor information, the candidate surgical area, and the candidate surgical plan are arranged and combined, one arrangement and combination method corresponds to one model input data, and all the model input data constitute the model input data set; Determining a surgical alternative solution dataset based on the medical record information through the first knowledge graph includes: Entities are extracted from the medical record information, relations are extracted from the medical record information, triples are constructed using entities and relations, and a surgical alternative plan data set is determined based on the matching relationship between the triples and the first knowledge graph.

2. The surgical plan evaluation method according to claim 1, characterized in that: The model is a graph attention model; The model input data set includes a specified number of nodes, each of which contains a specified number of features.

3. The surgical plan evaluation method according to claim 2, characterized in that: Acquiring medical record information, and determining a surgical alternative plan data set based on the medical record information through a first knowledge graph includes: Extracting information from the medical record information to obtain a first entity keyword and a first relationship keyword, and constructing a first graph data structure according to the first entity keyword and the first relationship keyword; The first graph data structure is matched with the first knowledge graph, and the surgical alternative plan data set is determined according to the matching result.

4. The surgical plan evaluation method according to claim 1, characterized in that: Also includes: Extracting data from the medical care information, the surgical alternative plan data set, and the medical record information to obtain a second entity keyword and a second relationship keyword; constructing a second graph data structure according to the second entity keyword and the second relationship keyword; The second graph data structure is matched with the second knowledge graph, and an emergency data set is determined according to the matching result.

5. The surgical plan evaluation method according to claim 1, characterized in that: Also includes: The first knowledge graph is updated according to the optimal surgical plan and the model input data corresponding to the optimal surgical plan.

6. The surgical plan evaluation method according to claim 4, characterized in that: Also includes: The second knowledge graph is updated according to the optimal surgical plan, the model input data corresponding to the optimal surgical plan, and emergencies.

7. The surgical plan evaluation method according to claim 1, characterized in that: The model input data is word vector or word matrix.

8. A surgical plan evaluation device, characterized in that: The method comprises a surgical plan evaluation unit, wherein the surgical plan evaluation unit is used for: Acquire medical record information, and determine a surgical alternative plan dataset through a first knowledge graph based on the medical record information; Acquire medical care information, and determine a model input data set based on the medical care information, the surgical alternative plan data set, and the medical record information; Taking each model input data in the model input data set as input respectively, classifying each model input data by the model, and determining the best surgical plan based on the classification result; The medical care information includes an alternative doctor information set and an alternative region information set, wherein the alternative doctor information set includes information of several alternative doctors, and the alternative region information set includes several alternative surgical regions; When determining the model input data set, the candidate doctor information, the candidate surgical area, and the candidate surgical plan are arranged and combined, one arrangement and combination method corresponds to one model input data, and all the model input data constitute the model input data set; Determining a surgical alternative solution dataset based on the medical record information through the first knowledge graph includes: Entities are extracted from the medical record information, relations are extracted from the medical record information, triples are constructed using entities and relations, and a surgical alternative plan data set is determined based on the matching relationship between the triples and the first knowledge graph.

9. A surgical plan evaluation system, characterized in that: It comprises a terminal, the terminal is configured with an executable program, and the executable program implements the surgical plan evaluation method according to any one of claims 1 to 7 when running.

10. A storage medium, characterized in that: An executable program is stored thereon, and when the executable program is run, the surgical plan evaluation method described in any one of claims 1 to 7 is implemented.

Images