Computer device, operation guidance method, operation guidance apparatus, storage medium, and program product

Abstract

The invention relates to a computer device, a surgical guidance method, a surgical guidance apparatus, a storage medium, and a program product. The computer equipment is configured to execute the following operations: acquiring an intraoperative image of a target object; if the agent determines that an operation guiding condition is met according to the intraoperative image, the agent displays operation guiding information; the surgical operation guide information is determined by the intelligent agent based on a surgical strategy, the surgical strategy is determined based on a recommended surgical strategy generated by the intelligent agent, and the recommended surgical strategy is generated by the intelligent agent according to a pre-constructed surgical knowledge base and preoperative state information of the target object; and after the operation is finished, the intelligent agent updates the operation knowledge base according to the operation record information of the target object.

Description

Technical Field

[0001] This application relates to the field of medical technology, and in particular to a computer device, surgical guidance method, surgical guidance apparatus, storage medium, and program product. Background Technology

[0002] Traditional surgical techniques rely primarily on the physician's accumulated clinical experience and medical knowledge. For example, the surgical plan is planned preoperatively based on the physician's personal experience and knowledge; during the intraoperative procedure, the physician needs to process multiple imaging information simultaneously, relying on real-time visual observation and emergency judgment to make immediate adjustments to key decisions during the operation.

[0003] While related technologies provide surgical guidance to physicians through medical image processing and navigation systems, this processing mainly provides isolated or fragmented guidance information based on preoperative planning or intraoperative image information. There is still considerable room for improvement in the safety and success rate of surgical guidance. Summary of the Invention

[0004] According to various embodiments of this application, a computer device, a surgical guidance method, a surgical guidance apparatus, a storage medium, and a program product are provided.

[0005] In a first aspect, this application provides a computer device configured to perform the following operations:

[0006] Acquire intraoperative images of the target subject;

[0007] If the intelligent agent determines that the operation guidance conditions are met based on the intraoperative images, the intelligent agent displays surgical operation guidance information; the surgical operation guidance information is determined by the intelligent agent based on the surgical strategy, the surgical strategy is determined based on the recommended surgical strategy generated by the intelligent agent, and the recommended surgical strategy is generated by the intelligent agent based on the pre-built surgical knowledge base and the preoperative status information of the target object;

[0008] After the surgery is completed, the intelligent agent updates the surgical knowledge base based on the surgical record information of the target object.

[0009] Secondly, this application also provides a surgical guidance method, comprising:

[0010] Acquire intraoperative images of the target subject;

[0011] If the intelligent agent determines that the operation guidance conditions are met based on the intraoperative images, the intelligent agent displays surgical operation guidance information; the surgical operation guidance information is determined by the intelligent agent based on the surgical strategy, the surgical strategy is determined based on the recommended surgical strategy generated by the intelligent agent, and the recommended surgical strategy is generated by the intelligent agent based on the pre-built surgical knowledge base and the preoperative status information of the target object;

[0012] After the surgery is completed, the intelligent agent updates the surgical knowledge base based on the surgical record information of the target object.

[0013] Thirdly, this application also provides a surgical guidance device, comprising:

[0014] The image acquisition module is used to acquire intraoperative images of the target object;

[0015] The intraoperative prompting module is used to display surgical operation guidance information by the intelligent agent if the intelligent agent determines that the operation guidance conditions are met based on the intraoperative images. The surgical operation guidance information is determined by the intelligent agent based on the surgical strategy, which is determined based on the recommended surgical strategy generated by the intelligent agent. The recommended surgical strategy is generated by the intelligent agent based on the pre-built surgical knowledge base and the preoperative status information of the target object.

[0016] The knowledge base update module is used to update the surgical knowledge base by the intelligent agent based on the surgical record information of the target object after the surgery.

[0017] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the surgical guidance method described in any embodiment of this application.

[0018] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the surgical guidance method described in any embodiment of this application.

[0019] Details of one or more embodiments of this application are set forth in the following drawings and description. Other features, objects, and advantages of this application will become apparent from the specification, drawings, and claims. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the disclosed drawings without creative effort. The additional details or examples used to describe the drawings should not be considered as a limitation on the scope of any of the disclosed invention, the currently described embodiments and / or examples, and the best mode of these inventions as currently understood.

[0021] Figure 1 This is a flowchart illustrating a surgical guidance method configured to be performed by a computer device according to one or more embodiments.

[0022] Figure 2 This is a flowchart illustrating the operations of building a surgical knowledge base, performed by a computer device configured according to one or more embodiments.

[0023] Figure 3 This is a schematic flowchart of a surgical guidance method configured to be performed by a computer device according to one or more other embodiments.

[0024] Figure 4 This is a structural block diagram of a surgical guidance device provided according to one or more embodiments.

[0025] Figure 5 This is an internal structural diagram of a computer device provided according to one or more embodiments.

[0026] Figure 6 This is an internal structural diagram of a computer device provided according to one or more other embodiments. Detailed Implementation

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

[0028] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0029] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various objects, but these objects are not limited by these terms. These terms are only used to distinguish the first object from the second object. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the solutions, or any combination of multiple solutions; in other words, the term "and / or" used in this application can be understood as at least one of multiple solutions.

[0030] In one embodiment, a computer device is provided. In some possible application scenarios, the computer device may include a terminal, or it may include a server, and it may also be applied to a system including a terminal and a server. The terminal and server in the system can interact through a network and implement the relevant steps configured to be executed. In this embodiment, as... Figure 1 As shown, the computer device can be configured to perform the following operations:

[0031] Operate S101 to acquire intraoperative images of the target object.

[0032] The target group refers to the person who will undergo the surgery.

[0033] Intraoperative imaging refers to image data acquired during surgery using medical imaging equipment to observe the internal state of the target object and / or the state of surgical instruments, such as X-ray images, ultrasound images, computed tomography (CT) images, magnetic resonance imaging (MRI), digital subtraction angiography (DSA), or one or more of these.

[0034] In practice, during surgery on the target patient, intraoperative images of the target patient can be acquired in a timely manner using medical imaging equipment. In some possible embodiments, after acquiring the intraoperative images, preprocessing can be performed on the images, such as noise reduction, enhancement, image registration, or one or more other operations, to improve image clarity and accuracy.

[0035] In operation S102, if the agent determines that the operation guidance conditions are met based on the intraoperative images, the agent will display the surgical operation guidance information. The surgical operation guidance information is determined by the agent based on the surgical strategy, which is determined based on the recommended surgical strategy generated by the agent. The recommended surgical strategy is generated by the agent based on the pre-built surgical knowledge base and the preoperative state information of the target object.

[0036] In this context, an intelligent agent can be understood as a computer system or software module capable of data processing, logical judgment, and instruction generation. The intelligent agent can make decisions based on preset rules and data. For example, an intelligent agent can be an artificial intelligence module with data analysis and strategy generation capabilities, which can be learned based on deep learning models and large amounts of training data. In this embodiment, the intelligent agent can be deployed in a computer device.

[0037] A surgical strategy can be understood as a specific plan, determined by the user, used to guide the implementation of surgery. For example, a surgical strategy may include one or more types of information such as surgical instruments, surgical procedures, surgical pathways, surgical risk assessment information, and risk management strategies. A recommended surgical strategy can be understood as a preliminary surgical plan generated by an agent based on a surgical knowledge base and the preoperative state information of the target object, providing a reference for the formulation of surgical strategies.

[0038] A surgical knowledge base can refer to a pre-built database for storing surgical-related data. In some possible embodiments, the surgical knowledge base can contain a variety of surgical-related information, such as historical surgical cases, surgical procedure guidelines, anatomical data, instrument usage guidelines, evidence-based medicine evidence, surgical complications, etc., which can provide sufficient data support for surgical decisions during, before and after surgery.

[0039] Preoperative status information refers to the status information of the target subject before surgery. This status information can be represented through one or more modalities such as text, numerical values, images, and videos. For example, preoperative status information may include one or more of the following: preoperative clinical diagnostic conclusions, preoperative imaging data, preoperative physiological indicators (such as clinical test results), basic information about the target subject (such as age and weight), and medical history.

[0040] In related technologies, during the preoperative planning stage, the setting of surgical plans often relies on the operator's personal experience and limited medical knowledge. However, personal experience is subjective and varies; different operators may formulate different surgical plans for the same case, and some surgical plans may lack evidence-based medical support, leading to inconsistent surgical outcomes. In this embodiment, a surgical knowledge base can be pre-constructed. This knowledge base can store a large number of surgical cases, standard surgical procedures, surgical risk management plans, surgical anatomical data, evidence-based medical evidence, and other multi-dimensional knowledge, providing high-quality, structured medical knowledge support for the decision-making of the intelligent agent and the operator, ensuring the scientific validity and authority of the decision-making basis. Furthermore, during the preoperative planning stage, the intelligent agent can generate recommended surgical strategies based on the surgical knowledge base and the preoperative status information of the target object. The operator can adjust or confirm the recommended surgical strategies generated by the intelligent agent to obtain the surgical strategy actually applied during the operation. By having the intelligent agent initially generate recommended surgical strategies based on the massive surgical knowledge information in the surgical knowledge base during the preoperative planning stage, the reliability, scientific validity, and authority of the final surgical strategy can be effectively improved, contributing to improved surgical safety and success rate.

[0041] During the intraoperative phase, while related technologies can provide navigation information, they heavily rely on the operator's real-time visual observation. For example, the operator needs to process multiple images simultaneously, comparing the tissue structures appearing in the field of vision with the pre-planned surgical path to determine whether the surgical instruments have reached their target positions. This approach depends on the operator's individual skill and on-the-spot judgment, and its limitations are particularly pronounced when dealing with complex tissue structures or high-risk surgeries.

[0042] In this embodiment, the intelligent agent can determine whether the operation guidance conditions are met based on intraoperative images. For example, the intelligent agent can acquire one or more key pieces of information from the intraoperative images, such as the anatomical features of the surgical site, the current position of surgical instruments, and the progress nodes of the surgical operation. Then, it can determine whether the operation guidance conditions are met based on the acquired key information. When the operation guidance conditions are determined to be met, the intelligent agent can display surgical operation guidance information based on a surgical strategy determined for the target object. This surgical operation guidance information guides the user in subsequent surgical operations for the target object. The surgical operation guidance information can be specific guidance information used to guide the user to perform surgical operations according to a pre-determined surgical strategy. For example, it can guide the user to adjust the angle and depth of the surgical instruments, or prompt the user to perform the next specific surgical action.

[0043] In some embodiments, surgical operation guidance information may be displayed when the operation guidance conditions are determined to be met, and not displayed when the operation guidance conditions are determined not to be met, thereby avoiding interference from the surgical operation guidance information to the operator when the operator's attention is highly focused.

[0044] In some optional embodiments, the user operating the surgical instruments can wear an augmented reality (AR) device, such as AR glasses. When the intelligent agent determines that the current conditions for operation guidance are met based on the intraoperative images, it can trigger the AR navigation system to overlay and display the surgical path and specified tissue structures on the augmented reality device, thereby realizing the display of surgical operation guidance information. For example, the AR navigation system can register the surgical path and key tissue structures to be displayed with the intraoperative images displayed in the augmented reality device, determine the position of the surgical path and key tissue structures in the intraoperative images based on the registration results, and overlay them onto the intraoperative images displayed in the augmented reality device based on the position.

[0045] Operation S103: After the surgery is completed, the intelligent agent updates the surgical knowledge base based on the surgical record information of the target object.

[0046] The surgical record information can refer to various information related to the target patient's surgery, such as information related to one or more stages, including preoperative, intraoperative, and postoperative procedures. In some optional embodiments, the surgical record information can be obtained based on the target patient's status information, the surgical strategy confirmed by the user, the surgical recommendation strategy generated by the agent, and relevant data recorded during and after the surgery.

[0047] In some embodiments, during the postoperative stage, such as after surgery or during postoperative monitoring, the agent can acquire surgical record information of the target object, and then the agent can update the pre-built surgical knowledge base based on the acquired surgical record information.

[0048] For example, if the surgical record information describes a new and effective treatment method for a specific surgical complication, and this treatment method is not included in the existing surgical knowledge base, the agent can add the treatment method to the corresponding knowledge module of the surgical knowledge base. If the surgical record information shows that a certain traditional surgical procedure has room for optimization in the surgery for the target subject, and the optimized procedure achieves better surgical results, the agent can update or supplement the description of the corresponding surgical procedure in the surgical knowledge base. This allows the surgical knowledge base to continuously accumulate experience data from actual surgeries, continuously improve the accuracy and applicability of subsequent generated recommended surgical strategies, and provide more reliable knowledge support for surgical guidance for more target subjects.

[0049] In this embodiment, an intelligent agent can integrate preoperative planning, intraoperative navigation, and postoperative knowledge management into a closed-loop system, sharing data across all stages to achieve intelligent support throughout the entire process from preoperative to postoperative. Specifically, on the one hand, the intelligent agent generates recommended surgical strategies based on the surgical knowledge base and the preoperative status information of the target patient. This leverages the vast knowledge in the surgical knowledge base to enhance the reliability, accuracy, and authority of preoperative surgical planning. On the other hand, when the intelligent agent determines that the conditions for operative guidance are met based on intraoperative images, it displays corresponding surgical operation guidance information. This allows for timely and accurate guidance of surgical operations based on reliable preoperative surgical planning, reducing the surgeon's workload and improving surgical safety. Furthermore, after the surgery, the intelligent agent updates the surgical knowledge base based on surgical record information, enabling the surgical knowledge base to continuously accumulate experience data from actual surgeries, thereby continuously improving the accuracy and applicability of subsequent recommended surgical strategies generated by the intelligent agent based on the surgical knowledge base. Therefore, this embodiment can effectively improve the overall success rate and safety of the surgery.

[0050] In one embodiment, the computer device may also be configured to perform at least one of the following operations:

[0051] If the agent determines that the position of the surgical instruments in the intraoperative image meets the preset position conditions, then the operation guidance conditions are met; if the agent determines that the tissue structure in the intraoperative image is a complex tissue structure, then the operation guidance conditions are met.

[0052] In practice, after acquiring intraoperative images, these images can be analyzed to obtain image analysis results. These results can be understood as information obtained after analyzing and processing the intraoperative images. In one embodiment, the image analysis results may include the location of surgical instruments in the intraoperative images, or they may include the tissue structures within the images. For example, object detection algorithms can be used to analyze the intraoperative images, identifying one or more types of information from the surgical instruments and tissue structures in the images to form the image analysis results.

[0053] After obtaining the image analysis results, the intelligent agent can determine whether the position of the surgical instruments in the intraoperative images meets the preset position conditions. For example, the position of the surgical instruments can reflect the spatial coordinates or positional relationship of the surgical instruments relative to other objects (such as the tissue structure or surgical area of ​​the target object) in the current surgical operation scenario; the preset position conditions can be pre-set and used to determine whether the surgical instruments are in a position that triggers the display of guidance information.

[0054] If the intelligent agent determines that the position of the surgical instrument meets the preset position conditions, it can then display surgical operation guidance information accordingly. For example, if the position of the surgical instrument reaches the preset range of the target tissue structure to be operated on, it can determine that the target position has been reached.

[0055] After obtaining the image analysis results, the agent can also determine whether the tissue structures appearing in the intraoperative images are complex. Complex tissue structures can be those identified by the agent based on the surgical knowledge base, indicating complex surgical procedures (e.g., long operation time, high difficulty) or complex structures themselves. Alternatively, complex tissue structures can be pre-selected and marked by the user. For example, the agent can identify the aortic arch as a complex tissue structure. Furthermore, when the presence of complex tissue structures in the intraoperative images is confirmed, the agent can determine that the conditions for surgical guidance are met and display corresponding surgical guidance information.

[0056] In this embodiment, when the intelligent agent determines that the position of the surgical instruments in the intraoperative image meets the preset position conditions, and when it determines that the tissue structure in the intraoperative image is a complex tissue structure, it determines that the operation guidance conditions are met. This can provide the operator with objective operation references in a timely manner, reduce reliance on personal experience, help the user maintain operational accuracy in complex surgical scenarios, and help avoid operational errors caused by fatigue or limited field of vision.

[0057] In one embodiment, such as Figure 2 As shown, the surgical knowledge base can be obtained in the following ways:

[0058] Operate S201 to acquire knowledge in the field of multimodal medicine.

[0059] Multimodal medical domain knowledge can be understood as a collection of medical-related knowledge existing in one or more forms, such as text, images, videos, audio, structured data, and unstructured data. In some possible embodiments, textual medical domain knowledge may include medical literature, surgical guidelines, medical record reports, clinical research texts, expert consensus information, etc.; image-based medical domain knowledge may include surgical anatomical diagrams, pathological slide diagrams, and various medical images acquired through medical imaging equipment; videos may include surgical operation videos; and structured data may include surgical instrument parameters and pre- and post-operative indicators of the surgical subject.

[0060] In some possible implementations, multimodal medical knowledge can be acquired in various ways. For example, structured and unstructured medical knowledge can be extracted from authoritative medical databases. Alternatively, multimodal large language model (MLLM) technology can be used to perform semantic analysis on imaging reports and surgical records to extract key diagnostic information and surgical procedure descriptions.

[0061] In some possible embodiments, taking the construction of an interventional surgery knowledge base as an example, the knowledge system framework of the interventional surgery knowledge base can include dimensions such as disease classification, surgical procedures, instrument parameters, complication management, and evidence-based evidence. In other words, multimodal medical knowledge can be collected for the above multiple dimensions. Through the setting of the above dimensions, the key knowledge categories required in the clinical application and research of interventional surgery can be fully covered, providing a foundation for the subsequent construction and application of the knowledge base.

[0062] Operation S202 extracts various entities related to surgery from multimodal medical domain knowledge, and constructs a surgical knowledge graph based on the entity association relationships between these entities.

[0063] In this context, "surgery-related entities" refers to entities involved in the surgical scenario. For example, various surgery-related entities may include, but are not limited to, one or more of the following: surgery name entity (e.g., "vascular interventional surgery"), surgical instrument entity, anatomical location entity, surgical procedure entity, complication entity, and surgeon entity. Entity relationships can be understood as logical or business connections between different entities, such as "use," "cause," and "prevention" relationships. Specifically, some possible entity relationships might include "laparoscopic cholecystectomy - use - laparoscopic grasping forceps." The surgical knowledge graph can be a structured model that uses entities as nodes and entity relationships as edges, employing a graph structure to store surgery-related knowledge, and can intuitively present the logical connections between surgical knowledge.

[0064] In some embodiments, after obtaining multimodal medical domain knowledge, key entities related to surgery and entity relationships between entities can be extracted from the multimodal medical domain knowledge using MLLM technology.

[0065] In some exemplary embodiments, when constructing a surgical knowledge graph, to improve the logic and relevance of the knowledge system, a hierarchical and networked knowledge modeling method can be adopted to design a multi-layered knowledge graph containing "disease-surgery-instrument-risk-evidence" from top to bottom. The hierarchical design ensures the orderly division of knowledge across different levels, allowing each level of knowledge to be relatively independent while also providing mutual support. The networked design strengthens the relationships between knowledge nodes, facilitating subsequent knowledge retrieval, analysis, and application. Through the combination of hierarchy and networking, a structured presentation of surgical domain knowledge (such as interventional surgery knowledge) can be achieved.

[0066] Operate S203 to construct a surgical knowledge base based on the surgical knowledge graph and multimodal medical domain knowledge.

[0067] After obtaining the surgical knowledge graph, a surgical knowledge base can be constructed based on the surgical knowledge graph and multimodal medical domain knowledge. The surgical knowledge graph can intuitively display the relationships between various knowledge nodes, making it easy for users to quickly understand the logic between knowledge; multimodal medical domain knowledge can provide a reserve of various surgical knowledge for the surgical knowledge base, providing a foundation for knowledge query and retrieval for subsequent clinical decision support after surgery.

[0068] In this embodiment, by collecting multimodal medical knowledge, covering various forms such as text, images, and videos, it is possible to more comprehensively cover surgery-related knowledge, ensuring the completeness and reliability of the knowledge dimensions of the surgical knowledge base. Furthermore, by extracting various entities related to surgery from multimodal medical knowledge and constructing a surgical knowledge graph based on entity relationships, it is possible to form a structured knowledge network from scattered surgery-related entities, improving the efficiency of knowledge understanding and application. Subsequently, based on the surgical knowledge graph and multimodal medical knowledge, a surgical knowledge base is constructed, enabling the surgical knowledge base to provide reliable support for the accurate output of surgical decision-making information while providing knowledge storage functions.

[0069] In one embodiment, the computer device is also configured to perform the following operations:

[0070] The process involves: obtaining an agent scheduling request; triggering the agent to acquire prompts related to surgical decisions; the agent scheduling request carrying prompt query information; responding to the agent scheduling request, determining the target entity associated with the prompt query information based on the entity matching results of the prompt query information and the surgical knowledge graph; obtaining the first feature representation corresponding to the target entity, and obtaining the target multimodal medical domain knowledge related to the prompt query information based on the matching results of the first feature representation and the second feature representations corresponding to the multimodal medical domain knowledge; and generating surgical decision prompts based on the target multimodal medical domain knowledge.

[0071] The agent scheduling request can be used to trigger an agent in a computer device to obtain information related to surgical decisions. In some possible embodiments, the agent scheduling request can include instructions actively sent by the user or information automatically generated by the computer device based on preset events or information.

[0072] The agent scheduling request may carry prompting information, which can be used to ask questions to obtain surgical decision-making prompts. This information can clearly define the specific need for obtaining prompts. For example, it may include one or more pieces of information such as the scenario, purpose, and function of obtaining prompts. For example, the prompting information may be one or more of the following related to a specific surgical type (such as interventional surgery): surgical strategy inquiry, operation procedure optimization inquiry, postoperative complication prevention inquiry, etc. In some embodiments, the prompting information may be user-defined input or a preset inquiry template. For example, when the user actively triggers the generation of the agent scheduling request, the user can set the prompting information. When the computer device automatically triggers the generation of the agent scheduling request, the prompting information can be generated according to the preset inquiry template, preset events, or information.

[0073] Feature representation can transform the semantic and attribute information of an entity into a recognizable vector form, which can be used to characterize the main features of the entity. For ease of distinction, in this embodiment, the feature representation of the target entity is referred to as the first feature representation, and the feature representations corresponding to the various multimodal medical domain knowledge are referred to as the second feature representations.

[0074] In some embodiments, in response to a received agent scheduling request, the agent can use the prompt query information as a matching basis, combine it with a pre-built surgical knowledge graph to perform entity matching, and then determine the target entity that is associated with the prompt query information. When performing entity matching between the prompt query information and the surgical knowledge graph, in some possible embodiments, keyword matching can be used to obtain a first matched entity. Then, the first entity and a second entity that is associated with the first entity can be determined as the target entity.

[0075] In some embodiments, a feature representation extraction model can be pre-trained to obtain the second feature representations corresponding to each of the multimodal medical domain knowledge. This enables multimodal knowledge representation, integrating multi-source information such as text, images, and surgical parameters into a knowledge representation within the same feature space. Furthermore, after identifying the target entity, the similarity between the first feature representation corresponding to the target entity and the second feature representations corresponding to each of the multimodal medical domain knowledge can be calculated. Based on the multimodal medical domain knowledge whose similarity satisfies a preset similarity condition, a matching result is obtained, thereby acquiring the target multimodal medical domain knowledge related to the prompt query information, achieving semantic-based accurate knowledge retrieval. Then, an agent can generate surgical decision prompts based on the target multimodal medical domain knowledge.

[0076] In some optional embodiments, the agent may embed a multimodal large model. When generating surgical decision prompts, the agent can integrate and analyze the target multimodal medical domain knowledge through the multimodal large model, and combine it with the specific needs of the prompt inquiry information to form targeted and practical prompt content.

[0077] In this embodiment, on the one hand, the agent can be flexibly triggered to obtain surgical decision prompts in the required scenario through agent scheduling requests. On the other hand, through knowledge matching based on knowledge graphs and feature representations, it is ensured that the selected multimodal knowledge is highly correlated with the target entity, reducing interference from irrelevant information and improving the reliability of decision prompts. Furthermore, the agent can simultaneously process multimodal medical domain knowledge such as images, text, and surgical parameters, and perform knowledge integration and decision prompt output, realizing seamless transformation from medical knowledge to surgical decisions, which can improve the accuracy and timeliness of decision-making.

[0078] In one embodiment, obtaining an agent scheduling request may include the following steps:

[0079] During the surgery, an agent scheduling request is generated based on the intraoperative status information of the target object, including a first prompt inquiry message. The first prompt inquiry message is used to trigger the agent to generate an intraoperative risk warning based on the intraoperative status information and surgical knowledge related to the type of surgery in the surgical knowledge base.

[0080] Intraoperative status information may include the status information of the target object collected in real time during the operation. For example, intraoperative status information may include one or more of the following: imaging data (such as DSA imaging data), intraoperative laboratory indicators, and physiological monitoring data (such as blood pressure, heart rate, liver and kidney function monitoring data, etc.).

[0081] In some embodiments, intraoperative status information of the target object can be collected during surgery using monitoring devices such as sensors and medical imaging equipment. For example, intraoperative status information of the target object can be collected in real time using an electrocardiogram monitor, a pulse oximeter, and a bleeding monitoring device for the surgical area connected to the target object.

[0082] In some embodiments, the agent scheduling request can be constructed based on known information such as "intraoperative status information" and "intraoperative risk identification based on intraoperative status information". For example, after obtaining the intraoperative status information, an agent scheduling request can be generated based on the intraoperative status information at preset time intervals. Alternatively, the agent scheduling request can be triggered when the intraoperative status information meets preset conditions (e.g., the indicator data exceeds a threshold).

[0083] In some embodiments, based on the collected intraoperative status information, an agent scheduling request containing a first prompt inquiry can be generated and provided to the agent. After receiving the first prompt inquiry, the agent can call the surgical knowledge base, extract surgical knowledge related to the type of surgery (e.g., interventional surgery, laparoscopic surgery, etc.) from the surgical knowledge base, and then analyze and process it in conjunction with the acquired intraoperative status information to finally generate an intraoperative risk warning for the current surgical procedure. The intraoperative risk warning may include possible operational risks during the operation (e.g., improper guidewire movement angle or depth leading to vascular perforation), abnormal physiological indicators, such as vascular stenosis, hemodynamic abnormalities, etc., to assist the physician in timely avoidance of risks.

[0084] In this embodiment, the risk assessment bias caused by relying solely on physician experience is avoided by relying on the dual basis of intraoperative status information and surgical knowledge base. The intelligent agent is triggered by the first prompt inquiry information to generate intraoperative risk prompts in combination with surgical type-related knowledge. This can accurately identify potential risks under specific surgical procedures and specific individuals, effectively improving the accuracy and efficiency of intraoperative risk identification, while also helping to prevent risks in advance and improve surgical safety.

[0085] In one embodiment, obtaining an agent scheduling request may include the following steps:

[0086] During the surgery, an agent scheduling request is generated based on the intraoperative status information of the target subject, including a second prompt inquiry message. The second prompt inquiry message is used to trigger the agent to generate a first postoperative complication risk warning for the target subject based on the intraoperative status information and evidence-based medical knowledge in the surgical knowledge base.

[0087] Among them, the postoperative complication risk warning can be tailored to the individual circumstances of the target subject, specifying the types of specific complications that the target subject may experience after surgery, as well as the risk level or probability of occurrence. For ease of distinction, the postoperative complication risk warning generated during the intraoperative stage is referred to as the first postoperative complication risk warning.

[0088] In some embodiments, intraoperative status information of the target object during surgery can be obtained. This intraoperative status information can be referred to the foregoing embodiments and will not be repeated here. Then, an agent scheduling request containing second prompt inquiry information is generated based on the intraoperative status information and provided to the agent.

[0089] In some embodiments, after receiving the second prompt query, the intelligent agent can retrieve evidence-based medical knowledge stored in the surgical knowledge base. This evidence-based medical knowledge may include clinically validated medical conclusions, such as postoperative complication patterns and risk assessment models summarized from extensive clinical research data. The evidence-based medical knowledge can then be combined with intraoperative status information for data analysis and risk assessment, thereby generating a postoperative complication risk warning for the target individual. For ease of distinction, this postoperative complication risk warning is also referred to as the first postoperative complication risk warning. For example, the first postoperative complication risk warning may include information such as the specific types of complications that the target individual may experience after surgery and their probability of occurrence.

[0090] In this embodiment, the intelligent agent is triggered by the second prompt inquiry information to generate a first postoperative complication risk warning for the target object based on the real-time intraoperative status information of the target object and the evidence-based medicine knowledge in the surgical knowledge base. This can predict the risk of postoperative complications in advance during the operation and provide early warning of potential complications, which helps to formulate preventive measures in advance. At the same time, the postoperative complication risk warning is based on a large amount of evidence and cases in the surgical knowledge base, rather than subjective judgment, which improves the scientificity and accuracy of complication risk prediction.

[0091] In one embodiment, obtaining an agent scheduling request may include the following steps:

[0092] During the surgery, based on the intraoperative status information of the target object, an agent scheduling request including a third prompt inquiry is generated. The third prompt inquiry is used to trigger the agent to determine the current surgical scenario based on the intraoperative status information and the type of surgery, determine the target surgical equipment parameters that match the current surgical scenario based on the surgical knowledge base, and generate parameter configuration prompts based on the target surgical equipment parameters.

[0093] The surgery type can be understood as the macro-level classification information of the surgery, which is the category to which the currently performed surgery belongs, such as laparoscopic surgery, interventional surgery, knee replacement surgery, etc. The target surgical equipment parameters can refer to the specific parameter values ​​that match the current surgical scenario and can be directly used to adjust the equipment parameters.

[0094] In some embodiments, during the operation, the intraoperative status information of the target object can be continuously collected by the monitoring device. When an abnormality is detected in the intraoperative status information (such as exceeding the normal range or meeting the preset trigger conditions), the intraoperative status information can be integrated with the recorded surgical type information to generate an intelligent agent scheduling request containing third prompt inquiry information and provide it to the intelligent agent.

[0095] In some embodiments, intraoperative status information may include the current stage of the surgical operation (such as the surgical separation stage, anastomosis stage, etc.), the intraoperative position information of the target object, and the exposure of anatomical structures in the surgical area. Then, based on the intraoperative status information and the type of surgery, the current surgical scenario is determined. For example, in orthopedic joint replacement surgery, the intraoperative status information determines that the current scenario is a prosthesis implantation scenario. Subsequently, an agent scheduling request containing a third prompt query is generated, which triggers the agent to perform subsequent processing.

[0096] In some embodiments, after receiving the third prompt query information, the intelligent agent can search the knowledge base for target surgical equipment parameters, such as pressure and speed, that match the currently determined surgical scenario. For example, the target surgical equipment parameters may include the operating parameters of surgical instruments and the setting parameters of auxiliary equipment (such as the magnification of the surgical microscope or the injection speed of the intraoperative drug injection device).

[0097] After determining the parameters of the target surgical equipment, the intelligent agent can generate parameter configuration prompts to guide the physician in setting the correct parameters for the corresponding surgical equipment, ensuring the stable operation of the surgical equipment and the precise implementation of the surgical procedure. For example, if the clarity of the medical images acquired during surgery (such as DSA images) does not meet the preset clarity requirements, the agent can determine the injection rate of the contrast agent based on intraoperative status information and matching surgical cases in the surgical knowledge base, and generate parameter configuration prompts based on this injection rate.

[0098] In this embodiment, the intelligent agent is triggered by a third prompt inquiry to determine the surgical scenario, and the target surgical equipment parameters matching the current surgical scenario are determined based on the surgical knowledge base to generate parameter configuration prompts. Based on the real-time data of the target object and the surgical knowledge base, dynamic adjustment of surgical plan and parameter optimization can be provided, and precise decision support can be provided according to the individual differences of the target object, thereby improving the safety and effect of surgery and effectively avoiding surgical risks caused by improper equipment parameters.

[0099] In one embodiment, the computer device is also configured to perform the following operations:

[0100] Acquire preoperative images of the target patient; in response to surgical planning instructions, the agent acquires three-dimensional reconstruction results of the surgical site within a preset range based on the preoperative images, and obtains recommended surgical strategies for the target patient based on the three-dimensional reconstruction results and the surgical knowledge base.

[0101] Preoperative images can be image data acquired through medical imaging equipment, used to present the state of tissue structures within a preset range of the surgical site of the target object. Surgical planning instructions can refer to operation commands triggered by the user to initiate the surgical planning process. In some examples, surgical planning instructions may include information such as the type of surgery and the surgical site area. The preset range of the surgical site can be understood as the area that needs to be reconstructed in three dimensions, determined according to the type of surgery and the surgical site. For example, the preset range of the surgical site may include the target tissue structure involved in the surgical operation and other tissue structures within a certain range around the target tissue structure.

[0102] In some embodiments, preoperative images of the target object can be acquired using medical imaging equipment. For example, a preoperative image can be obtained by scanning a preset area, including the surgical site, using medical imaging equipment. Then, in response to a surgical planning instruction for surgical planning, the intelligent agent can be triggered to perform corresponding surgical strategy generation processing.

[0103] In some embodiments, after receiving surgical planning instructions, the agent can convert preoperative images into a three-dimensional model using a three-dimensional reconstruction algorithm to obtain a three-dimensional reconstruction result. This result can be used to clarify the spatial structure of the surgical site. For example, during three-dimensional reconstruction, the agent can process pixel or voxel information in the preoperative images using algorithms such as image segmentation, feature extraction, and three-dimensional modeling to construct a three-dimensional model that intuitively reflects the spatial relationship and morphological features of the anatomical structures within a preset range of the surgical site. This three-dimensional reconstruction result clearly presents the three-dimensional structure of the surgical site, which helps to formulate a more precise surgical strategy.

[0104] Furthermore, based on the 3D reconstruction results and one or more pieces of information contained in the surgical knowledge base, such as surgical operation guidelines, typical surgical plans for different surgical types, association rules between anatomical structures and surgical strategies, postoperative complication prevention measures, and past successful surgical cases, a recommended surgical strategy for the target object can be obtained. In some exemplary embodiments, during the generation of the recommended surgical strategy, the agent can combine the individual anatomical features of the target object reflected in the 3D reconstruction results with general surgical knowledge and clinical experience in the surgical knowledge base through data matching, logical reasoning, machine learning analysis, etc., thereby generating a recommended surgical strategy adapted to the specific situation of the target object. This recommended surgical strategy may include one or more pieces of information such as surgical path, surgical operation step suggestions, key anatomical structure avoidance prompts, and surgical instrument options.

[0105] In this embodiment, the intelligent agent can perform 3D reconstruction based on preoperative images to accurately restore the spatial structure of the surgical site, providing precise spatial reference for surgical strategy formulation and reducing surgical risks caused by anatomical structure misjudgment during the intelligent agent's preoperative planning process. On the other hand, the intelligent agent combines the 3D reconstruction results reflecting the individual anatomical characteristics of the target subject with a surgical knowledge base containing a large amount of professional knowledge and cases to generate recommended surgical strategies. This overcomes the limitations of traditional technologies that rely on the limited experience and medical knowledge of individual physicians for surgical planning. While achieving personalized surgical strategy planning, it obtains scientifically reliable recommended surgical strategies that are more in line with the target subject and follow medical principles, thus helping to improve surgical outcomes.

[0106] In one embodiment, the agent obtains a recommended surgical strategy for the target object based on the 3D reconstruction results and a surgical knowledge base, which may include the following operations:

[0107] The system acquires descriptions of the abnormal conditions and preoperative status information of the target object. Based on these information, the agent retrieves similar cases from the surgical knowledge base that meet preset similarity criteria. Then, based on the surgical strategies of these similar cases and the results of 3D reconstruction, the system generates a recommended surgical strategy for the target object.

[0108] Among them, abnormal condition description information can refer to information used to characterize the abnormal physical condition of the target object. The information modality of abnormal condition description information can include one or more, such as text, data, images, videos, etc. The content indicated by abnormal condition description information can include one or more types of information such as lesion type, lesion location, lesion degree and symptom manifestation.

[0109] The preset similarity condition refers to the threshold used by the agent to determine whether cases are similar. In one example, a similarity score can be used. The preset similarity threshold can be adjusted according to the type of surgery. Similar cases can refer to historical cases in the surgical knowledge base whose similarity scores with the target object's abnormal condition description information and preoperative status information reach the preset threshold.

[0110] In some possible embodiments, the intelligent agent can establish data interaction with the electronic medical record system of the target object, quickly read the description of abnormal conditions and preoperative examination data recorded in the electronic medical record of the target object, thereby obtaining the description information of abnormal conditions and the preoperative status information of the target object; of course, the intelligent agent can also receive information input by the user through the system interaction interface. For example, if some information is missing in the electronic medical record, the user can supplement the missing information through the system interaction interface.

[0111] After obtaining the description of the abnormal condition and the preoperative state information of the target object, the agent can perform case matching in the surgical knowledge base based on the description of the abnormal condition and the preoperative state information to obtain similar cases whose similarity meets preset similarity conditions. In some exemplary embodiments, the agent can convert the description of the abnormal condition and the preoperative state information of the target object into structured data. For example, it can obtain the third feature representation of the description of the abnormal condition and the preoperative state information through a feature representation extraction model, and achieve case matching in the surgical knowledge base by calculating the similarity between the third feature representation and the feature representations of multiple cases in the surgical knowledge base.

[0112] Furthermore, the agent can generate recommended surgical strategies based on surgical strategies for similar cases and 3D reconstruction results. For example, the agent can extract surgical strategies for similar cases and then adjust and optimize the surgical strategies for similar cases by combining them with the anatomical features of the target object in the 3D reconstruction results, thus forming a recommended scheme adapted to the target object.

[0113] In this embodiment, by acquiring the abnormal conditions and preoperative status information of the target object, reliable data support is provided for the intelligent agent to perform case matching in the knowledge base. By screening similar cases through preset similarity conditions, the reliability of the surgical strategy used as a reference is ensured, which helps to reduce the interference of subjective experience. Furthermore, by generating a recommended surgical strategy for the target object based on the surgical strategies of similar cases and the 3D reconstruction results, a shift from experience-based to individualized surgical planning is realized, which effectively reduces intraoperative risks, improves the adaptability and safety of surgical plans, and provides scientific support for clinical surgical decisions.

[0114] In one embodiment, the 3D reconstruction result characterizes the 3D spatial relationship between the surgical site and adjacent tissue structures. For example, the 3D reconstruction result can determine the relative position, distance, angle, size, and other geometric information of multiple tissue structures (including the surgical site and adjacent tissue structures) in the target object. Based on the surgical strategies of similar cases and the 3D reconstruction results, a recommended surgical strategy for the target object can be generated, which may include:

[0115] Based on the surgical strategies of similar cases, reference surgical instruments and reference surgical paths are determined; based on the three-dimensional spatial relationships represented by the three-dimensional reconstruction results, at least one of the reference surgical instruments and reference surgical paths is adjusted, and the target surgical instruments and target surgical paths are determined based on the adjustment results; based on the target surgical instruments and target surgical paths, a recommended surgical strategy for the target patient is generated.

[0116] In some embodiments, after obtaining surgical strategies for similar cases, the surgical instruments and surgical paths used in the similar cases can be determined by analyzing the surgical strategies of the similar cases, thereby obtaining reference surgical instruments and reference surgical paths. For example, reference surgical instruments may include, but are not limited to, surgical blades, puncture needles, endoscopes, and other instrument types proven effective in similar case surgeries. Reference surgical paths may include surgical incision locations, the channels through which instruments enter the target object, and their movement trajectories within the body, among other surgical operation information.

[0117] In some embodiments, after determining the reference surgical instruments and reference surgical path, at least one of the reference surgical instruments and reference surgical path can be adaptively adjusted based on the three-dimensional spatial relationship between the surgical site and adjacent tissue structures represented by the three-dimensional reconstruction results. For example, if the three-dimensional reconstruction results show that the location of a blood vessel near the surgical site of the target patient differs significantly from that in similar cases, potentially leading to a risk of the reference surgical path touching the blood vessel, the trajectory of the reference surgical path can be adjusted based on the three-dimensional spatial relationship to avoid the blood vessel area. Similarly, if the three-dimensional reconstruction results show that the anatomical dimensions of the surgical site of the target patient differ from those in similar cases, making it impossible for the specifications of the reference surgical instruments to be fully compatible, the model, size, and other parameters of the reference surgical instruments can be adjusted to ensure that the adjusted surgical instruments meet the surgical operation needs of the target patient.

[0118] After adjusting at least one of the reference surgical instruments and reference surgical pathways, target surgical instruments and target surgical pathways that match the specific circumstances of the target patient can be determined based on the adjustment results. Finally, based on the determined target surgical instruments and target surgical pathways, a recommended surgical strategy for the target patient can be generated. This recommended surgical strategy can effectively combine clinical experience from similar cases with the individual anatomical characteristics of the target patient, providing a scientific and reasonable guidance for clinical surgical procedures.

[0119] In this embodiment, the three-dimensional spatial relationship between the surgical site and adjacent tissue structures can be accurately characterized based on the three-dimensional reconstruction results, providing an objective and detailed anatomical basis for personalized adjustment of surgical strategies. This avoids the compatibility problems that may result from relying solely on similar cases to determine the surgical plan. Furthermore, determining reference instruments and surgical paths based on the surgical strategies of similar cases allows for flexible adjustments based on the established paths of similar cases, reducing the workload of surgical plan planning while improving the reliability of the plan. By guiding adjustments through three-dimensional spatial relationships, both the reliability and personalization of the plan are taken into account.

[0120] In one embodiment, the computer device is also configured to perform the following operations:

[0121] The intelligent agent generates a second postoperative complication risk warning for the target subject based on the target subject's preoperative status information and evidence-based medicine knowledge in the surgical knowledge base. It also obtains the rationale for the recommended surgical strategy based on the target subject's preoperative status information, 3D reconstruction results, and the surgical knowledge base. Furthermore, it identifies at least one piece of literature information associated with the second postoperative complication risk warning and the surgical strategy. Finally, based on the recommended surgical strategy, the rationale for the surgical strategy, the second postoperative complication risk warning, and the literature information, it generates a surgical planning report for the target subject.

[0122] Postoperative complication risk warnings can be tailored to the individual circumstances of the target patient, specifying the types of specific complications that may occur after surgery, as well as the risk level or probability of occurrence. For ease of distinction, postoperative complication risk warnings generated during the preoperative planning stage are referred to as second postoperative complication risk warnings.

[0123] The reasons for the recommendation can be used to explain the basis and rationale for selecting the recommended surgical strategy. For example, it can combine individual data of the target subject with medical knowledge to clarify the advantages of the recommended surgical strategy in terms of safety, effectiveness, and applicability. Literature information can include information used to identify the source or attributes of the literature. For example, literature information can include key information such as the title, author, publication journal, publication year, and main conclusions of the literature.

[0124] In some embodiments, the intelligent agent can acquire the preoperative status information of the target object, and then generate a second postoperative complication risk warning for the target object based on the preoperative status information and evidence-based medicine knowledge in the surgical knowledge base. The evidence-based medicine knowledge in the surgical knowledge base is formed based on a large amount of clinical research data, authoritative medical guidelines, and literature summaries, possessing scientific validity and reliability, and can provide an effective basis for assessing the risk of postoperative complications.

[0125] In some possible embodiments, during the preoperative planning stage, a fourth prompt inquiry can be generated based on the preoperative status information of the target object, and the intelligent agent can be triggered to generate a corresponding second postoperative complication risk warning based on the fourth prompt inquiry. The specific process can refer to the process of generating the first postoperative complication risk warning in the aforementioned embodiments, which will not be elaborated here.

[0126] Furthermore, the intelligent agent can combine the preoperative state information of the target object, the 3D reconstruction results, and the surgical knowledge base to further obtain the reasons for recommending the surgical strategy. In some embodiments, the process of obtaining the reasons for recommendation can be an analytical process in which the intelligent agent judges the individual suitability of the target object based on the preoperative state information, analyzes the feasibility of the surgical operation based on the 3D reconstruction results, and verifies the rationality of the surgical strategy based on clinical experience and evidence-based knowledge in the surgical knowledge base. This process can verify the pertinence and applicability of the recommended surgical strategy.

[0127] After obtaining the risk warning for the second postoperative complication and the rationale for the recommended surgical strategy, further literature information related to at least one of these risk warnings and the surgical strategy can be identified. For example, if the risk warning for the second postoperative complication includes the statement "high risk of postoperative infection," then literature information related to "postoperative infection risk factor analysis" can be obtained; if the recommended surgical strategy is "minimally invasive surgery," then literature information related to "clinical efficacy of minimally invasive surgery" and "standard operating procedures for minimally invasive surgery" can be obtained. By linking these literature information, additional theoretical support can be provided for the rationality of the risk warning for the second postoperative complication and the scientific validity of the recommended surgical strategy.

[0128] Furthermore, the intelligent agent can integrate and generate a surgical planning report tailored to the target individual based on the acquired recommended surgical strategies, the rationale behind those strategies, secondary postoperative complication risk warnings, and relevant literature information. In some examples, the surgical planning report can cover key information required during the surgical procedure, providing a comprehensive and reliable reference for developing the actual surgical plan and assessing surgical risks, thus helping to improve the accuracy of surgical decisions and the safety of the surgical procedure.

[0129] In this embodiment, on the one hand, the second postoperative complication risk warning generated based on evidence-based medicine avoids the generality of generic risk disclosure and helps physicians develop preventive measures in advance. On the other hand, the recommended surgical strategy and rationale, combined with the 3D reconstruction results, ensure the anatomical suitability and clinical effectiveness of the surgical plan. Simultaneously, the associated literature information provides authoritative medical evidence for the surgical plan and risk warning, enhancing the credibility of the surgical plan. By integrating the above information to generate a surgical planning report, physicians can quickly grasp the individual circumstances of the target patient and the appropriate surgical plan, reducing the risk of postoperative complications and improving the success rate of the surgery.

[0130] In one embodiment, the computer device is further configured to perform at least one of the following operations:

[0131] The intelligent agent generates a third postoperative complication risk warning for the target object based on the target object's postoperative status information, pathological data, and evidence-based medicine knowledge in the surgical knowledge base. Based on the third postoperative complication risk warning, the agent obtains the postoperative monitoring results of the target object. The intelligent agent also generates a corresponding postoperative follow-up strategy for the target object based on the target object's postoperative status information, pathological data, and evidence-based medicine knowledge in the surgical knowledge base.

[0132] In related technologies, postoperative evaluation of surgical outcomes and monitoring of complications for target subjects often rely on manual recording and analysis, lacking a systematic approach and resulting in high rates of postoperative complications and low efficiency in postoperative management. In this embodiment, postoperative status information and pathological data of the target subject can be obtained.

[0133] Postoperative status information refers to the target subject's condition after surgery. This status information can be represented through one or more modalities, such as text, numerical values, images, and videos. For example, postoperative status information may include one or more of the following: postoperative clinical diagnosis, postoperative imaging data, postoperative physiological indicators, the target subject's baseline condition, and medical history. Pathological data may include postoperative histopathological examination results and other medical data related to the disease itself. Postoperative status information and pathological data can be obtained automatically through medical data acquisition devices, manually entered into the system, or retrieved from the target subject's electronic medical record system. The specific method of acquisition can be selected according to the needs of the actual application scenario and is not limited here.

[0134] In some embodiments, after obtaining the postoperative status information and pathological data of the target object, the intelligent agent can generate a postoperative complication risk warning for the target object based on the postoperative status information, pathological data, and evidence-based medicine knowledge in the surgical knowledge base. For ease of distinction, the postoperative complication risk warning generated in the postoperative stage is referred to as the third postoperative status information and pathological data. In some embodiments, when generating the third postoperative complication risk warning, the intelligent agent can match and analyze the postoperative status information and pathological data corresponding to the target object with the evidence-based medicine knowledge in the surgical knowledge base. The intelligent agent can determine the risk of postoperative infection complications of the target object based on the matched evidence-based medicine knowledge and generate the corresponding third postoperative complication risk warning. For example, the risk warning may include information such as the possible types of complications, risk levels, and risk assessment criteria.

[0135] After generating the third postoperative complication risk warning, postoperative monitoring results for the target subject can be obtained based on this warning. Specifically, the third postoperative complication risk warning can be directly used as the postoperative monitoring result. By triggering the agent to retrieve the third postoperative complication risk warning at preset time intervals, potential postoperative complication risks for the target subject can be identified promptly and prevented in a timely manner. Alternatively, based on the high-risk complication types identified in the third postoperative complication risk warning, the monitoring frequency or dimensions of corresponding monitoring items can be increased. For example, if the risk warning indicates a high risk of postoperative bleeding, the monitoring frequency of information such as blood pressure and heart rate can be increased, and the data obtained from each monitoring session can be recorded as the postoperative monitoring result to achieve timely understanding of the occurrence of complications.

[0136] In this embodiment, the intelligent agent integrates the postoperative status information, pathological data, and evidence-based medicine knowledge in the surgical knowledge base of the target object to generate a third postoperative complication risk warning. This enables targeted postoperative monitoring of the target object. On the one hand, it can realize real-time monitoring of potential postoperative abnormalities, effectively reduce the incidence of postoperative complications, and improve the efficiency of postoperative management and monitoring. On the other hand, it can also improve the targeting and accuracy of monitoring items, avoid the subjectivity of relying solely on experience judgment, and make the third postoperative complication risk warning more in line with the individual situation, thereby improving the accuracy and reliability of postoperative complication prediction.

[0137] Furthermore, after obtaining the postoperative status information and pathological data of the target subject, the intelligent agent can generate a postoperative follow-up strategy corresponding to the target subject based on the postoperative status information, pathological data, and evidence-based medicine knowledge in the surgical knowledge base. In some embodiments, during the postoperative stage, an intelligent agent scheduling request including a fifth prompt inquiry can be generated based on the postoperative status information and pathological data of the target subject, triggering the intelligent agent to generate a postoperative follow-up strategy.

[0138] In some embodiments, the intelligent agent can comprehensively analyze the possibility of postoperative changes in the target subject's condition and the timeliness of required medical intervention based on evidence-based medical knowledge related to the type of surgery, pathological stage, and postoperative recovery status of the surgery performed on the target subject in the surgical knowledge base, thereby determining the postoperative follow-up strategy.

[0139] In this embodiment, the intelligent agent generates a postoperative follow-up strategy for the target object based on the target object's postoperative status information, pathological data, and evidence-based medicine knowledge in the surgical knowledge base. This enables the intelligent agent to configure a personalized follow-up plan for the target object after surgery, ensuring the effectiveness of subsequent follow-up.

[0140] In one embodiment, a surgical guidance method is also provided, which may include the following operations:

[0141] Operation S1 obtains the image analysis results corresponding to the intraoperative images of the target object; the image analysis results include the instrument positions of surgical instruments in the intraoperative images.

[0142] In operation S2, if the agent determines that the instrument position meets the preset position conditions, the agent will trigger the display of surgical operation guidance information based on the surgical strategy of the target object to guide the surgery for the target object. The surgical strategy is determined based on the recommended surgical strategy generated by the agent, which is generated by the agent according to the pre-built surgical knowledge base and the preoperative state information of the target object.

[0143] Operation S3 involves updating the surgical knowledge base by the intelligent agent based on the surgical record information of the target object after the surgery.

[0144] The solution provided by this surgical guidance method is similar to the solution described in the above embodiments. Therefore, the specific limitations of the surgical guidance method can be found in the above description of the limitations on the operations configured to be performed by the computer device, and will not be repeated here.

[0145] To enable those skilled in the art to better understand this application, the following uses interventional surgery as an example to illustrate this application, but it should be understood that the embodiments of this application are not limited thereto.

[0146] In order to improve surgical safety and operational efficiency, some medical image processing systems and navigation systems have been provided, but the following shortcomings still exist: (1) Surgical decisions still rely to a certain extent on the subjective experience and limited knowledge of physicians, making it difficult to provide objective basis for surgical decisions; (2) Guidance information generated based on image fusion technology depends on preoperative registration, and surgical guidance information is difficult to flexibly adapt to the dynamic changes brought about by intraoperative tissue structure (such as blood vessels) deformation; (3) The preoperative, intraoperative and postoperative systems are independent of each other, resulting in information silos and decision-making fragmentation; (4) Medical knowledge is updated rapidly, but in related technologies, the system or physician lacks the ability to continuously absorb and apply massive amounts of new medical evidence and new medical knowledge in a timely manner, making surgical decision support relatively lagging; (5) Surgical plans and surgical guidance information are relatively simple and fixed, making it difficult to adapt to different target groups.

[0147] This embodiment provides an evidence-based intelligent workflow system for interventional surgery. This system can realize intelligent assistance throughout the entire process, including preoperative CT image 3D reconstruction and surgical planning, intraoperative DSA image real-time analysis and navigation, and postoperative efficacy prediction and suggestion generation. It provides strong support for improving the accuracy and safety of interventional surgery, reducing surgical risks, and improving treatment outcomes.

[0148] The interventional surgery intelligent agent provided in this embodiment (i.e., the "intelligent agent" mentioned in one or more of the foregoing embodiments, or simply the intelligent agent) may include an evidence-based medicine question-and-answer system, a preoperative 3D reconstruction and analysis system, a surgical plan recommendation and risk prediction system, an intraoperative real-time image fusion navigation and decision-making system, and a postoperative efficacy evaluation and long-term management system. The interventional surgery intelligent agent can adopt a layered system architecture design. In one optional embodiment, the system architecture specifically includes a knowledge base layer, an intelligent processing layer, and an application layer. These layers cooperate with each other to jointly ensure the stable operation and functional implementation of the interventional surgery intelligent agent.

[0149] The knowledge base layer can serve as the data and knowledge foundation for interventional intelligent agents. In some examples, it includes interventional knowledge graphs and multimodal large models. Interventional knowledge graphs can structure and associate various types of knowledge in the field of interventional diagnosis and treatment, while multimodal large models can process various types of medical data such as text, images, and videos, providing rich knowledge and data support for subsequent intelligent processing.

[0150] The intelligent processing layer can possess functions such as evidence retrieval, intelligent question answering, image analysis, and risk prediction. Specifically, the evidence retrieval function can obtain evidence-based medical evidence from the knowledge base layer according to user needs; the intelligent question answering function can accurately respond to user-submitted interventional diagnosis and treatment-related questions (such as prompts and inquiries) based on existing knowledge and data; the image analysis function can professionally analyze various types of image data generated during interventional diagnosis and treatment to extract key information; and the risk prediction function can scientifically predict surgical risks by combining the patient's individual condition and treatment data, providing a reference for medical decision-making.

[0151] As the link between interventional intelligent agents and medical personnel, the application layer can provide application services such as preoperative planning, intraoperative navigation, and postoperative management for the entire process of interventional diagnosis and treatment in practical applications.

[0152] like Figure 3 As shown, this embodiment may include the following steps:

[0153] S301, Building a knowledge base for interventional surgery.

[0154] In this step, the following operations can be performed:

[0155] Acquire knowledge from the multimodal medical domain; extract various entities related to surgery from the multimodal medical domain knowledge, and construct a surgical knowledge graph based on the entity relationships between these entities; construct a surgical knowledge base based on the surgical knowledge graph and the multimodal medical domain knowledge.

[0156] This step allows us to collect knowledge from various multimodal medical fields related to interventional surgery, and to build a knowledge graph and knowledge base that covers multiple types of interventional surgery and evidence-based medicine.

[0157] S302, Preoperative stage intelligent agent-generated surgical planning report.

[0158] In this step, the following operations can be performed:

[0159] First, in response to surgical planning instructions, the agent can acquire 3D reconstruction results of the surgical site within a preset range based on the preoperative images of the target object. Based on the description information of the target object's abnormalities and preoperative state information, it can retrieve similar cases in the surgical knowledge base that meet the preset similarity conditions. Based on the surgical strategies of similar cases, it can determine reference surgical instruments and reference surgical paths. Based on the 3D spatial relationships represented by the 3D reconstruction results, it can adjust at least one of the reference surgical instruments and reference surgical paths. Based on the adjustment results, it can determine the target surgical instruments and target surgical paths. Based on the target surgical instruments and target surgical paths, it can generate a recommended surgical strategy for the target object.

[0160] Subsequently, the agent can generate a second postoperative complication risk warning for the target subject based on the target subject's preoperative status information and evidence-based medicine knowledge in the surgical knowledge base. The agent can also obtain the rationale for the recommended surgical strategy based on the target subject's preoperative status information, 3D reconstruction results, and the surgical knowledge base. Then, it can determine at least one piece of literature information associated with the second postoperative complication risk warning and the surgical strategy. Finally, based on the recommended surgical strategy, the rationale for the surgical strategy, the second postoperative complication risk warning, and the literature information, a surgical planning report for the target subject can be generated.

[0161] For example, image data of the target patient can be input into the intelligent agent. The agent can then perform automatic image analysis and 3D reconstruction based on a large visual model (such as a multimodal large model). Based on the 3D reconstruction results, it automatically matches similar cases in the interventional surgery knowledge base. Then, combined with the individual characteristics of the target patient, it adjusts the surgical strategies for similar cases and recommends personalized surgical plans (such as stent selection and surgical pathways). The risk prediction system within the agent automatically associates patient characteristics with evidence-based medicine in the knowledge base, assesses the risk of surgical complications (such as vascular injury and in-stent thrombosis), and generates a surgical planning report with evidence-based medicine support, including the reasons for the recommendation, risk assessment, and references.

[0162] S303, the intraoperative intelligent agent provides real-time decision support.

[0163] In this step, the following operations can be performed:

[0164] The system acquires intraoperative images of the target patient. If the agent determines that the conditions for surgical guidance are met based on the intraoperative images, it displays surgical guidance information. For example, if the agent determines that the position of the surgical instruments in the intraoperative images meets preset positional conditions, or if the tissue structure in the intraoperative images is a complex tissue structure, the agent can display surgical guidance information.

[0165] During the surgery, based on the intraoperative status information of the target object, an agent scheduling request including a first prompt inquiry information, an agent scheduling request including a second prompt inquiry information, and an agent scheduling request including a third prompt inquiry information are generated.

[0166] In some embodiments, during the intraoperative phase, multimodal image fusion and real-time display can be achieved. The fused image types may include, but are not limited to, DSA images, CT images, and ultrasound images. Based on the real-time display of multimodal image fusion, an AR navigation system can be combined to overlay the surgical path and key structures. When the instrument positions of surgical instruments meet preset conditions, the AR navigation system can overlay the pre-planned surgical path and key anatomical structures (such as important blood vessels and nerve tissues) within the surgical area onto the surgeon's field of vision as virtual images. This allows the surgeon to simultaneously obtain surgical path guidance and key structure location information while observing the real surgical scene, reducing the risk of surgical operation deviations due to limited field of vision and further improving the accuracy of surgical operations. Simultaneously, during the intraoperative phase, the intelligent agent can automatically and in real-time identify the instrument positions used during the surgery and dynamically optimize the movement path of the instruments based on the identified instrument position information and surgical path planning requirements. Through real-time instrument tracking and path optimization, it can be ensured that surgical instruments always move along the optimal path, avoiding unnecessary instrument deviations during surgery, reducing accidental damage to surrounding normal tissues, and ensuring the safety and efficiency of the surgical operation.

[0167] During surgery, the intelligent agent can provide real-time warnings of potential risks based on data collected in the surgical area. For example, when vascular stenosis or hemodynamic abnormalities are detected in the surgical area based on the target patient's intraoperative status information (such as real-time medical images), a warning signal can be issued in a timely manner to remind the physician to pay attention and take appropriate measures. On the other hand, it can also combine the physiological data monitored in real-time during the operation of the target patient with evidence-based medicine evidence (i.e., evidence-based medicine instructions) stored in the preset knowledge base to predict and warn of potential complications that may occur during the operation (such as postoperative infection, bleeding risk, etc.), so as to make preparations in advance and reduce the probability of complications.

[0168] During the intraoperative phase, relevant treatment parameters can be dynamically adjusted. This adjustment process relies on recommended surgical practices stored in the interventional surgery knowledge base. It involves real-time analysis of surgical progress, changes in the target patient's physiological state, and feedback from the surgical area. This allows for dynamic optimization of key treatment parameters, such as pressure parameters, instrument movement speed, and drug injection rate, ensuring that each parameter remains optimally suited to the current surgical stage and the individual patient's condition. This further enhances the surgical outcome and guarantees the safety and reliability of the procedure.

[0169] S304, the postoperative intelligent agent monitors potential complications in real time, configures personalized follow-up strategies, and updates the knowledge base based on the experience of this surgery.

[0170] In this step, the following operations can be performed:

[0171] The intelligent agent generates a third postoperative complication risk warning for the target subject based on the target subject's postoperative status information, pathological data, and evidence-based medicine knowledge in the surgical knowledge base. Based on the third postoperative complication risk warning, the agent obtains the postoperative monitoring results of the target subject. The intelligent agent also generates a corresponding postoperative follow-up strategy for the target subject based on the target subject's postoperative status information, pathological data, and evidence-based medicine knowledge in the surgical knowledge base.

[0172] In the postoperative stage, the intelligent agent's effectiveness evaluation system can automatically analyze and process the surgical outcome data. Specifically, the system can collect various key parameters recorded during the surgery (such as surgical operation time, completion rate of target lesion treatment, intraoperative blood loss, etc.) and real-time postoperative physiological indicators of the target patient. Through preset analysis algorithms, the system comprehensively calculates the above data and generates an evaluation report containing analytical information such as surgical outcome indicators and the completion status of surgical goals, providing data support for users to quickly grasp the surgical results.

[0173] The intelligent agent's early complication identification system can monitor various vital signs, laboratory results, and examination results of the target subject in real time based on pre-stored complication characteristic information and evidence-based medicine in the interventional surgery knowledge base. For example, the complication characteristic information in the interventional surgery knowledge base may include typical clinical manifestations and abnormal threshold ranges of common postoperative complications (such as infection, bleeding, and organ dysfunction) of different types of interventional surgeries. The evidence-based medicine may include clinically validated risk factors for complication occurrence and early warning indicators. When the intelligent agent detects an abnormality in the target subject's status information that matches the aforementioned complication characteristics or evidence-based medicine, it can promptly issue an early warning so that medical personnel can intervene as early as possible.

[0174] In the follow-up and prognosis management phase, the intelligent agent's long-term follow-up and prognosis prediction system can develop personalized follow-up plans for target subjects based on their individual characteristics and evidence-based medicine data from the interventional surgery knowledge base. For example, individual characteristics of the target subject may include patient age, type of underlying disease, surgical procedure, and initial postoperative recovery status. Evidence-based medicine data from the interventional surgery knowledge base may include postoperative recovery patterns of patients with different individual characteristics, prognostic influencing factors, and the impact of different follow-up frequencies on prognostic outcomes. Based on this information, the intelligent agent's long-term follow-up and prognosis prediction system can determine the follow-up time intervals and content, and dynamically adjust the follow-up plan according to the target subject's recovery progress during subsequent follow-ups.

[0175] After the relevant procedures in the postoperative stage of this surgery are completed, the intelligent agent can filter and organize various data generated during the surgery, such as surgical operation details, complication management process, and feedback on the implementation effect of personalized follow-up plan, and incorporate the latest experience into the interventional surgery knowledge base according to the preset knowledge base update rules. This enriches the content dimensions of the knowledge base and improves the accuracy and adaptability of system decision-making in subsequent surgical assistance, postoperative management and other aspects.

[0176] In this embodiment, the interventional surgery intelligent agent workflow system based on evidence-based medicine constructs a domain knowledge graph covering the entire interventional surgery process. During the construction process, it integrates multi-dimensional knowledge such as evidence-based medicine evidence, surgical operation standards, and complication management. Through the integration of the above multi-dimensional knowledge, it can provide high-quality and structured medical knowledge support for the entire interventional surgery intelligent agent workflow system, thereby ensuring the scientific nature and authority of the information on which the system bases its decision-making in each stage.

[0177] Furthermore, this embodiment also achieves deep integration of the multimodal large model and the interventional surgery knowledge base. By developing a multimodal large model specifically for the interventional surgery field, the multimodal large model can simultaneously process multimodal data such as images, text, and surgical parameters. Through synchronous processing of multimodal data, seamless transformation from medical knowledge to surgical decisions can be achieved, thereby helping to improve the accuracy and timeliness of the system in the surgical-related decision-making process and ensuring the efficient advancement of each stage of the surgery.

[0178] In addition, this embodiment has the function of intelligent intervention throughout the entire process. It integrates preoperative planning, intraoperative navigation and postoperative management into a closed-loop system. In this closed-loop system, data at each stage can be shared in real time, realizing intelligent support for the entire process from preoperative to postoperative, avoiding delays or deviations in surgical decisions caused by data fragmentation at each stage, thereby helping to improve the overall success rate of the surgery.

[0179] Furthermore, this embodiment also provides real-time decision support and personalized customization services. Specifically, for example, the intelligent agent can provide dynamic suggestions for adjusting surgical plans and optimizing surgical parameters based on the patient's individual data and the aforementioned interventional surgery knowledge base. It can also provide accurate and detailed decision support according to the individual differences of the target subject. Through this dynamic adjustment and personalized support, the safety and treatment effect of the surgery are further improved, meeting the surgical treatment needs of different target subjects.

[0180] Based on the same inventive concept, this application also provides a surgical guidance device for implementing the surgical guidance method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in the surgical guidance device embodiments provided below can be found in the limitations on the computer equipment and surgical guidance method in one or more of the above embodiments, and will not be repeated here.

[0181] In one exemplary embodiment, such as Figure 4 As shown, a surgical guidance device is provided, comprising:

[0182] Image acquisition module 401 is used to acquire intraoperative images of the target object;

[0183] The intraoperative prompting module 402 is used to display surgical operation guidance information by the intelligent agent if the intelligent agent determines that the operation guidance conditions are met based on the intraoperative images; the surgical operation guidance information is determined by the intelligent agent based on the surgical strategy, the surgical strategy is determined based on the recommended surgical strategy generated by the intelligent agent, and the recommended surgical strategy is generated by the intelligent agent based on the pre-built surgical knowledge base and the preoperative status information of the target object;

[0184] The knowledge base update module 403 is used to update the surgical knowledge base by the intelligent agent based on the surgical record information of the target object after the surgery.

[0185] The modules in the aforementioned surgical guidance device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0186] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores surgical data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When executed by the processor, the computer program implements a surgical guidance method.

[0187] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 6 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a surgical guidance method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0188] Those skilled in the art will understand that Figure 5 and Figure 6The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0189] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0190] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0191] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0192] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0193] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0194] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0195] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A computer device configured to perform the following operations: Acquire intraoperative images of the target subject; If the intelligent agent determines that the operation guidance conditions are met based on the intraoperative images, the intelligent agent displays surgical operation guidance information; the surgical operation guidance information is determined by the intelligent agent based on the surgical strategy, the surgical strategy is determined based on the recommended surgical strategy generated by the intelligent agent, and the recommended surgical strategy is generated by the intelligent agent based on the pre-built surgical knowledge base and the preoperative status information of the target object; After the surgery, the intelligent agent updates the surgical knowledge base based on the surgical record information of the target object.

2. The computer device according to claim 1, further configured to perform at least one of the following operations: If the agent determines that the position of the surgical instruments in the intraoperative image meets the preset position conditions, then it determines that the operation guidance conditions are met; If the agent determines that the tissue structure of the intraoperative image is a complex tissue structure, then it determines that the operation guidance conditions are met.

3. The computer device according to claim 1, wherein the surgical knowledge base is obtained in the following manner: Acquire knowledge in the field of multimodal medicine; Extract various entities related to surgery from the multimodal medical domain knowledge, and construct a surgical knowledge graph based on the entity association relationships between these entities; The surgical knowledge base is constructed based on the surgical knowledge graph and the multimodal medical domain knowledge.

4. The computer device according to claim 3, further configured to perform the following operations: Obtain an agent scheduling request; the agent scheduling request is used to trigger the agent to obtain prompts related to surgical decisions, and the agent scheduling request carries prompt query information; In response to the agent scheduling request, the agent determines the target entity associated with the prompt inquiry information based on the entity matching result of the prompt inquiry information and the surgical knowledge graph; Obtain the first feature representation corresponding to the target entity, and based on the matching results of the first feature representation and the second feature representation corresponding to the multimodal medical domain knowledge, obtain the target multimodal medical domain knowledge related to the prompt query information; The intelligent agent generates surgical decision suggestions based on the target's multimodal medical domain knowledge.

5. The computer device according to claim 4, wherein obtaining the agent scheduling request includes: During the surgery, based on the intraoperative status information of the target object, an agent scheduling request including a first prompt inquiry is generated; The first prompting query information is used to trigger the intelligent agent to generate an intraoperative risk warning based on the intraoperative status information and surgical knowledge related to the surgical type in the surgical knowledge base.

6. The computer device according to claim 4, wherein obtaining the agent scheduling request includes: During the surgery, based on the intraoperative status information of the target object, an agent scheduling request including a second prompt inquiry is generated; The second prompting information is used to trigger the intelligent agent to generate a first postoperative complication risk warning for the target object based on the intraoperative status information and evidence-based medicine knowledge in the surgical knowledge base.

7. The computer device according to claim 4, wherein obtaining the agent scheduling request includes: During the surgery, based on the intraoperative status information of the target object, an intelligent agent scheduling request including third prompt inquiry information is generated; The third prompting information is used to trigger the intelligent agent to determine the current surgical scenario based on the intraoperative status information and the surgical type, determine the target surgical equipment parameters that match the current surgical scenario based on the surgical knowledge base, and generate parameter configuration prompts based on the target surgical equipment parameters.

8. The computer device of claim 1, further configured to perform the following operations: Acquire preoperative images of the target object; In response to surgical planning instructions, the intelligent agent obtains a three-dimensional reconstruction result within a preset range of the surgical site based on the preoperative images, and obtains a recommended surgical strategy for the target object based on the three-dimensional reconstruction result and the surgical knowledge base.

9. The computer device according to claim 8, wherein the step of obtaining a recommended surgical strategy for the target object by the intelligent agent based on the three-dimensional reconstruction result and the surgical knowledge base includes: Obtain the abnormal condition description information and the preoperative status information of the target object; The intelligent agent obtains similar cases that meet preset similarity conditions from the surgical knowledge base based on the abnormal situation description information and the preoperative state information, and generates a recommended surgical strategy for the target object based on the surgical strategies of the similar cases and the three-dimensional reconstruction results.

10. The computer device according to claim 9, wherein the three-dimensional reconstruction result characterizes the three-dimensional spatial relationship between the surgical site and adjacent tissue structures; The step of generating a recommended surgical strategy for the target patient based on the surgical strategies of similar cases and the three-dimensional reconstruction results includes: Based on the surgical strategies of the similar cases, determine the reference surgical instruments and reference surgical pathways; Based on the three-dimensional spatial relationships characterized by the three-dimensional reconstruction results, at least one of the reference surgical instruments and the reference surgical path is adjusted, and the target surgical instrument and the target surgical path are determined based on the adjustment results. Based on the target surgical instruments and the target surgical path, a recommended surgical strategy is generated for the target patient.

11. The computer device of claim 8, further configured to perform the following operations: The intelligent agent generates a second postoperative complication risk warning for the target object based on the preoperative status information of the target object and the evidence-based medicine knowledge in the surgical knowledge base. The intelligent agent also obtains the recommendation reasons for the recommended surgical strategy based on the preoperative status information of the target object, the three-dimensional reconstruction results, and the surgical knowledge base. Identify literature information associated with at least one of the second postoperative complication risk warning and the surgical strategy; Based on the recommended surgical strategy, the rationale for recommending the surgical strategy, the second postoperative complication risk warning, and the literature information, a surgical planning report for the target subject is generated.

12. The computer device according to claims 1 to 11, wherein the computer device is further configured to perform at least one of the following operations: The intelligent agent generates a third postoperative complication risk warning for the target object based on the target object's postoperative status information, the target object's pathological data, and evidence-based medicine knowledge in the surgical knowledge base. Based on the third postoperative complication risk warning, the agent obtains the target object's postoperative monitoring results. The intelligent agent generates a postoperative follow-up strategy for the target object based on the target object's postoperative status information, the target object's pathological data, and evidence-based medicine knowledge in the surgical knowledge base.

13. A surgical guidance method, the method comprising: Acquire intraoperative images of the target subject; If the intelligent agent determines that the operation guidance conditions are met based on the intraoperative images, then the intelligent agent displays the surgical operation guidance information; The surgical operation guidance information is determined by the intelligent agent based on the surgical strategy, which is determined based on the recommended surgical strategy generated by the intelligent agent. The recommended surgical strategy is generated by the intelligent agent based on a pre-built surgical knowledge base and the preoperative status information of the target object. After the surgery, the intelligent agent updates the surgical knowledge base based on the surgical record information of the target object.

14. A surgical guidance device, the device comprising: The image acquisition module is used to acquire intraoperative images of the target object; The intraoperative prompt module is used to display surgical operation guidance information by the intelligent agent if the intelligent agent determines that the operation guidance conditions are met based on the intraoperative images. The surgical operation guidance information is determined by the intelligent agent based on the surgical strategy, which is determined based on the recommended surgical strategy generated by the intelligent agent. The recommended surgical strategy is generated by the intelligent agent based on a pre-built surgical knowledge base and the preoperative status information of the target object. The knowledge base update module is used to update the surgical knowledge base by the intelligent agent based on the surgical record information of the target object after the surgery.

15. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method of claim 13.

16. A computer program product comprising a computer program that, when executed by a processor, implements the steps of the method of claim 13.

Images