[go: up one dir, main page]

CN119851925B - Septic shock diagnosis and treatment image information structured data processing system and method - Google Patents

Septic shock diagnosis and treatment image information structured data processing system and method Download PDF

Info

Publication number
CN119851925B
CN119851925B CN202510333815.XA CN202510333815A CN119851925B CN 119851925 B CN119851925 B CN 119851925B CN 202510333815 A CN202510333815 A CN 202510333815A CN 119851925 B CN119851925 B CN 119851925B
Authority
CN
China
Prior art keywords
data
diagnosis
treatment
module
structured
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202510333815.XA
Other languages
Chinese (zh)
Other versions
CN119851925A (en
Inventor
康焰
邱德松
缪海昕
唐辛群
唐超
黎方舟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sichuan Smart Medicine Technology Co ltd
Original Assignee
Sichuan Smart Medicine Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sichuan Smart Medicine Technology Co ltd filed Critical Sichuan Smart Medicine Technology Co ltd
Priority to CN202510333815.XA priority Critical patent/CN119851925B/en
Publication of CN119851925A publication Critical patent/CN119851925A/en
Application granted granted Critical
Publication of CN119851925B publication Critical patent/CN119851925B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/092—Reinforcement learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
    • G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
    • G06V10/451—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters with interaction between the filter responses, e.g. cortical complex cells
    • G06V10/454—Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/52—Scale-space analysis, e.g. wavelet analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/62—Extraction of image or video features relating to a temporal dimension, e.g. time-based feature extraction; Pattern tracking
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
    • G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/94—Hardware or software architectures specially adapted for image or video understanding
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/98—Detection or correction of errors, e.g. by rescanning the pattern or by human intervention; Evaluation of the quality of the acquired patterns
    • G06V10/993—Evaluation of the quality of the acquired pattern
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10081—Computed x-ray tomography [CT]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10132—Ultrasound image
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Multimedia (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Medical Informatics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Quality & Reliability (AREA)
  • Databases & Information Systems (AREA)
  • Public Health (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Pathology (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)

Abstract

The invention relates to the technical field of medical information processing, and discloses an infectious shock diagnosis and treatment image information structured data processing system and method. The image acquisition module acquires CT images, ultrasonic images and hemodynamic time sequence data, the multimode data fusion module generates multimode fusion data after processing, the dynamic feature extraction module generates dynamic feature vectors, and the structuring processing engine generates structuring diagnosis and treatment records. The abnormality detection module monitors the recording parameters, the diagnosis and treatment decision module generates diagnosis and treatment suggestions according to the abnormality signals, the early warning feedback module provides visual early warning, and the performance monitoring module evaluates the system performance. The method realizes multi-mode data fusion processing, generates a structured diagnosis and treatment record, provides accurate diagnosis and treatment decision support, has real-time monitoring and system optimization functions, and improves the level and efficiency of infectious shock diagnosis and treatment.

Description

System and method for processing infectious shock diagnosis and treatment image information structured data
Technical Field
The invention relates to the technical field of medical information processing, in particular to an infectious shock diagnosis and treatment image information structured data processing system and method.
Background
The infectious shock is taken as a serious clinical syndrome, the disease progress is rapid, the death rate is high, and timely and accurate diagnosis and treatment are important for improving the prognosis of patients. However, many difficulties in data processing and decision support are faced in the current diagnosis and treatment of septic shock.
In terms of medical image data processing, the diagnosis of an septic shock patient often depends on various medical image data, such as CT images, ultrasound images, and hemodynamic time series data. These data have different modalities and characteristics, which are difficult to integrate and analyze effectively in traditional processing modes. The CT image can clearly show the morphological structure of the organ but can not directly reflect the dynamic change of the hemodynamics, the ultrasonic image can observe the blood flow condition of partial tissues and organs in real time but has relatively limited information, and the hemodynamic time sequence data can show the change of the heart function and the circulation state along with time but lacks visual anatomical structure information. The spatial-temporal alignment and feature normalization of these multimodal data to fully and accurately analyze patient conditions is a current challenge.
From the point of view of clinical data management, most of the existing diagnosis and treatment records are in unstructured or semi-structured form, and contain a large amount of text description and scattered data. When viewing and analyzing these data, doctors need to spend a great deal of time and effort to screen and integrate the key information, which is inefficient and easy to miss important details. For example, when recording physiological parameters, hemodynamic indexes and image features of a patient, the data format lacks of unified specification, and the recording modes of different doctors are greatly different, which brings great inconvenience to subsequent data analysis and clinical decision. In addition, unstructured data is not beneficial to informatization management and big data analysis of medical data, potential value behind the data cannot be fully mined, and accurate medical treatment based on big data is difficult to achieve.
In the aspect of diagnosis and treatment decision support, the condition of the infective shock is complex and changeable, and the conditions of all patients are different, so that high requirements are put on the diagnosis and treatment decision capability of doctors. At present, doctors mainly make diagnosis and treatment schemes according to own experience and limited clinical guidelines, and personalized and accurate decision support tools are lacked. Due to lack of sufficient consideration for individual differences of patients, the traditional diagnosis and treatment scheme may not achieve the optimal treatment effect, and even delay the illness state. Moreover, how to quickly and accurately select the optimal solution when facing multiple possible diagnosis and treatment solutions, and coordinate resource conflict and time sequence conflict between different solutions is also a problem to be solved in the clinical diagnosis and treatment process.
In addition, in the whole diagnosis and treatment process, an effective real-time monitoring and early warning mechanism is lacked. The slight change of the patient's condition cannot be found in time, so that corresponding intervention measures cannot be taken in the early stage of the deterioration of the condition. Meanwhile, the performance evaluation of the diagnosis and treatment system is not perfect, the problems of the system in the aspects of data processing, decision support and the like cannot be found in time, the system is difficult to be optimized and improved in a targeted manner, and the improvement of the diagnosis and treatment level of the infective shock is limited. In view of the above, it is an urgent need to develop a system and method capable of effectively processing the image information of the septic shock diagnosis and treatment, realizing the data structure management, providing the accurate diagnosis and treatment decision support, and having the functions of real-time monitoring and system optimization.
Disclosure of Invention
The invention aims to provide a system and a method for processing infectious shock diagnosis and treatment image information structured data, which are used for solving the problems in the background technology.
The system comprises a processor, an image acquisition module, a multi-mode data fusion module, a dynamic characteristic extraction module, a structured processing engine, a diagnosis and treatment decision module, an abnormality detection module, an early warning feedback module and a performance monitoring module;
the system comprises an image acquisition module, a multi-mode data fusion module, a dynamic feature extraction module, a structuring processing engine, a structural diagnosis and treatment record, a data processing module and a data processing module, wherein the image acquisition module is used for acquiring medical image data of a patient, including CT (computed tomography) images, ultrasonic images and hemodynamic time sequence data;
The system comprises an abnormality detection module, a diagnosis and treatment decision module, an early warning feedback module, a performance monitoring module and a system performance log, wherein the abnormality detection module monitors physiological parameters, hemodynamic indexes and image characteristics in a structured diagnosis and treatment record in real time, generates an abnormal signal and sends the abnormal signal to a processor if the detected parameters deviate from a preset threshold range, the diagnosis and treatment decision module calls a preset clinical path algorithm according to the abnormal signal to generate an adaptive diagnosis and treatment suggestion, the early warning feedback module receives the abnormal signal and triggers a visual early warning interface, and the performance monitoring module records response time of the abnormal signal and correction times of diagnosis and treatment decisions to generate the system performance log.
Preferably, the specific operation process of the dynamic feature extraction module is as follows:
The method comprises the steps of carrying out time dependency modeling on hemodynamic time sequence data by adopting a long-short-term memory network to output time sequence feature vectors, carrying out multi-scale feature extraction on CT images and ultrasonic images through a three-dimensional convolutional neural network to output space feature vectors, carrying out tensor stitching on the time sequence feature vectors and the space feature vectors, and generating dynamic feature vectors after dimension reduction through a full-connection layer.
Preferably, the structuring processing engine comprises a template matching sub-module and a semantic parsing sub-module;
The template matching submodule maps physiological parameter classification in the dynamic feature vector to cardiac function indexes, microcirculation states and organ perfusion labels based on data fields defined by clinical guidelines, and the semantic analysis submodule carries out entity identification and relation extraction on the unstructured text report through a natural language processing technology to generate standardized semantic data;
The structured medical record is generated by combining standardized semantic data with mapped data fields.
Preferably, the specific operation process of the abnormality detection module is as follows:
respectively setting a dynamic threshold interval for the heart function index, the microcirculation state and the organ perfusion label, wherein the dynamic threshold interval is adaptively adjusted based on the historical data and the population statistics data of the patient;
And calculating the deviation degree of the current parameter value and the dynamic threshold interval in real time, and if the deviation degree of any parameter exceeds a preset tolerance coefficient, generating an abnormal signal and marking the abnormal type.
Preferably, the diagnosis and treatment decision module comprises a path optimization sub-module and a conflict resolution sub-module;
The conflict resolution sub-module coordinates resource conflict and time sequence conflict among multiple schemes through a fuzzy logic algorithm and outputs final diagnosis and treatment advice.
Preferably, the specific analysis process of the performance monitoring module is as follows:
recording response time from generation of abnormal signals to diagnosis and treatment advice output, and marking the abnormal signals as delay events if the response time exceeds a preset response threshold;
counting the times of clinical correction of the diagnosis and treatment advice, and marking the diagnosis and treatment advice as an inefficient decision event if the correction times exceed a preset correction threshold;
And carrying out weighted calculation on the occurrence frequency of the delay event and the low-efficiency decision event to generate a system performance score.
Preferably, the processor is in communication connection with a data tracing module, and the data tracing module is configured to:
Labeling each item of data in the structured diagnosis and treatment record with a source label comprising an image equipment model, acquisition time and operator information;
When the performance score of the system is lower than a preset qualification threshold value, triggering data tracing analysis and positioning data source abnormal nodes of a delay event or an inefficient decision event.
Preferably, the early warning feedback module comprises a multi-stage early warning sub-module;
the multi-stage early warning submodule divides early warning grades according to the severity of the abnormal signals, the primary early warning triggers interface popup prompt, the secondary early warning is synchronously sent to the mobile terminal, and the tertiary early warning is directly communicated with the first-aid system.
Preferably, the system further comprises a model iteration module for:
and periodically collecting system performance logs and clinical feedback data, updating a deep learning model and strengthening a learning strategy through an online learning algorithm, and optimizing the accuracy of dynamic feature extraction and diagnosis and treatment decision.
Preferably, the invention also comprises a method for processing the structural data of the infectious shock diagnosis and treatment image information, which comprises the following steps:
acquiring multi-mode medical image data through an image acquisition module;
Performing space-time alignment and feature normalization processing on the multi-mode data to generate fusion data;
Extracting space-time dynamic characteristics of the fusion data by using a deep learning model to generate a characteristic vector;
Mapping the feature vector to a clinical data template to generate a structured diagnosis and treatment record;
monitoring parameter deviation conditions of diagnosis and treatment records in real time, and triggering abnormal signals and adaptive diagnosis and treatment suggestions;
recording system response events and generating performance logs, and positioning abnormal nodes through data tracing;
Triggering a multi-stage feedback mechanism according to the early warning level, and iteratively optimizing algorithm performance through a model.
Compared with the prior art, the invention has the beneficial effects that:
According to the invention, various medical image data are acquired through the image acquisition module, and the multi-mode data fusion module is utilized to perform space-time alignment and feature normalization processing to generate multi-mode fusion data. The process realizes the advantage complementation of the data of different modes, so that the data can reflect the illness state of the patient more comprehensively. For example, the anatomical structure information of the CT image, the real-time blood flow information of the ultrasonic image and the dynamic change information of the hemodynamic time sequence data are combined, so that a rich and accurate data basis is provided for subsequent analysis. The dynamic feature extraction module further performs time sequence analysis and spatial feature extraction on the fusion data to generate dynamic feature vectors, and the deep data mining mode can capture the disease state change trend more accurately.
The structured processing engine maps the dynamic feature vectors to predefined clinical data templates, generating a structured medical record. The original complicated and unordered medical data becomes standard and ordered, and a doctor can conveniently and quickly review and analyze the data. The doctor does not need to spend a great deal of time searching key information in unstructured text records, and can directly acquire various physiological parameters, hemodynamic indexes, image features, diagnostic results and the like of a patient from the structured diagnosis and treatment record, so that the diagnosis and treatment efficiency is remarkably improved. Meanwhile, the structured data is more convenient for the informatization management and big data analysis of the medical data, provides powerful support for medical research and clinical decision, and is favorable for realizing accurate medical treatment based on big data. The abnormality detection module monitors various parameters in the structured diagnosis and treatment record in real time, and can judge whether the parameters are abnormal or not more accurately by adaptively adjusting the dynamic threshold interval based on the historical data of patients and the group statistical data. When the parameter is detected to deviate from the preset threshold range, an abnormal signal is immediately generated and the abnormal type is marked, and the function is helpful for doctors to discover the slight change of the patient in time. The early warning feedback module divides multi-stage early warning according to the severity of the abnormal signal, prompts the emergency system to be communicated from the interface popup window, ensures that doctors can take corresponding intervention measures at the first time, effectively avoids the worsening of the illness state and improves the success rate of the treatment of patients.
The diagnosis and treatment decision module calls a preset clinical path algorithm according to the abnormal signals and generates an adaptive diagnosis and treatment suggestion by combining individual differences of patients. The path optimization submodule adopts a reinforcement learning algorithm to prioritize the diagnosis and treatment schemes, fully considers the illness state characteristics and individual differences of different patients, and enables the diagnosis and treatment schemes to be more targeted. The conflict resolution sub-module coordinates resource conflict and time sequence conflict among multiple schemes through a fuzzy logic algorithm, and ensures that the finally output diagnosis and treatment suggestion is feasible. The personalized and accurate diagnosis and treatment decision support mode can help doctors to make more scientific and reasonable diagnosis and treatment schemes, improve treatment effects and improve prognosis of patients.
The performance monitoring module records response time of the abnormal signals and correction times of diagnosis and treatment decisions, and a system performance log is generated. Through analysis of the data, the performance of the system can be comprehensively evaluated, and problems existing in the running process of the system can be timely found. And when the performance score of the system is lower than a preset qualification threshold value, the data tracing module performs tracing analysis on the data, positions abnormal data source nodes of delay events or low-efficiency decision events, and provides a clear direction for system optimization. The model iteration module periodically collects system performance logs and clinical feedback data, updates a deep learning model and a reinforcement learning strategy through an online learning algorithm, continuously optimizes the accuracy of dynamic feature extraction and diagnosis and treatment decision, realizes continuous optimization and self-promotion of the system, and ensures that the system always maintains a high-efficiency and reliable running state.
Drawings
FIG. 1 is a schematic diagram of the system for processing the structural data of the septic shock diagnostic image information according to the present invention;
FIG. 2 is a flow chart of diagnosis and treatment decision conflict resolution and solution optimization;
FIG. 3 is a flow chart of iterative optimization of a system model;
FIG. 4 is a flow chart of a method of treating data for septic shock.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1-4, the present invention provides a technical solution, an infectious shock diagnosis and treatment image information structured data processing system, the system includes:
And the image acquisition module is responsible for acquiring medical image data of a patient, wherein the medical image data comprises CT images, ultrasonic images and hemodynamic time sequence data. These data are collected from imaging equipment and monitoring equipment in the hospital to provide the basis for subsequent analysis.
And the multi-mode data fusion module is used for carrying out space-time alignment and feature normalization processing on the acquired medical image data. The space-time alignment is used for ensuring the consistency of different modal data in time and space, and the feature normalization is used for unifying the value ranges of different features, so that the subsequent analysis and processing are convenient, and the multi-modal fusion data is finally generated.
And the dynamic characteristic extraction module is used for carrying out time sequence analysis and space characteristic extraction on the multi-mode fusion data through a preset deep learning model. The time sequence analysis is used for mining the change rule of data along with time, and the space feature extraction is focused on the space information in the image, so that the dynamic feature vector is finally generated.
And the structuring processing engine maps the dynamic feature vector into a predefined clinical data template to generate a structuring diagnosis and treatment record. The clinical data template is designed according to clinical guidelines and actual diagnosis and treatment requirements, can present complex medical data in a structured form, and is convenient for doctors to check and analyze.
And the abnormality detection module is used for monitoring physiological parameters, hemodynamic indexes and image characteristics in the structured diagnosis and treatment record in real time. Upon detecting that the parameter deviates from the preset threshold range, an abnormality signal is generated and sent to the processor for timely finding the patient's abnormality.
And the diagnosis and treatment decision module is used for calling a preset clinical path algorithm according to the abnormal signal to generate an adaptive diagnosis and treatment suggestion. The algorithms are formulated based on a large amount of clinical experience and research results, and can provide scientific and reasonable diagnosis and treatment references for doctors.
And the early warning feedback module is used for receiving the abnormal signal and triggering the visual early warning interface. Through visual visualization mode, doctor can know patient's abnormal conditions fast, takes corresponding measure in time.
And the performance monitoring module is used for recording response time of the abnormal signals and correction times of diagnosis and treatment decisions and generating a system performance log. By recording and analyzing the data, the performance of the system can be evaluated, and a basis is provided for optimizing the system.
The invention is further illustrated in the following in connection with examples 1 to 5:
Example 1:
The embodiment is used for elaborating how the dynamic feature extraction module extracts features from data of different modes and generates dynamic feature vectors, so as to provide accurate data support for subsequent structural processing and diagnosis and treatment decisions.
The dynamic feature extraction module operates as follows, and time-dependent modeling is performed on hemodynamic time series data by adopting a long-short-term memory network (LSTM). The hemodynamic time series data contains important information of the heart function, vascular resistance and the like of the patient along with time. The long-term and short-term memory network can effectively capture long-term dependency in data, and overcomes the problems of gradient disappearance and gradient explosion which are easy to occur when the traditional neural network processes time sequence data. Assuming hemodynamic time series data as a sequenceThe inputs to the LSTM network areThrough a series of gating mechanisms, e.g. forgetting gatesInput doorAnd an output doorFor cell stateUpdating and finally outputting the time sequence feature vector. The core formula is as follows:
Wherein, As a function of the Sigmoid,Representing the multiplication by element,AndRespectively a weight matrix and a bias vector.
And respectively carrying out multi-scale feature extraction on the CT image and the ultrasonic image through a three-dimensional convolutional neural network (3D-CNN). CT images and ultrasound images contain rich human organ structure and lesion information. The three-dimensional convolution neural network can carry out convolution operation on the image in a three-dimensional space, and extract spatial features of different scales. Taking a CT image as an example, the spatial feature vector is finally output through the processing of a multi-layer convolution layer and a pooling layer. Assume CT image asThrough convolution layerAnd a pooling layerOutputting a spatial feature vector。
Tensor splicing is carried out on the time sequence feature vector and the space feature vector, and the dynamic feature vector is generated after the dimension reduction of the full connection layer. In particular, a timing feature vector derived from hemodynamic timing dataAnd spatial feature vectors derived from CT images and ultrasound images、Tensor stitching is carried out to obtain a high-dimensional vector. Then the full connection layer is used for carrying out dimension reduction treatment to the dynamic feature vector, redundant information is removed, and a final dynamic feature vector is generated。
Example 2:
The embodiment has the function of detailing how the structural processing engine converts the dynamic feature vector into the structural diagnosis and treatment record, so that medical data is more convenient for doctors to check and analyze, and diagnosis and treatment efficiency is improved.
The structuring processing engine comprises a template matching sub-module and a semantic parsing sub-module.
The template matching submodule maps physiological parameter classification in the dynamic feature vector to cardiac function indexes, microcirculation states and organ perfusion labels based on data fields defined by clinical guidelines. Clinical guidelines are specifications and standards, which have been summarized through long-term practice and research in the medical field, defining a series of data fields related to the diagnosis and treatment of septic shock. The template matching submodule classifies physiological parameters in the dynamic feature vector according to the data fields. For example, parameters reflecting the systolic and diastolic function of the heart are mapped to cardiac function index labels, parameters related to the state of microcirculation blood flow are mapped to microcirculation state labels, and parameters reflecting the condition of organ blood perfusion are mapped to organ perfusion labels.
The semantic analysis sub-module performs entity recognition and relation extraction on the unstructured text report through a natural language processing technology to generate standardized semantic data. In medical records, there are a large number of unstructured text reports, such as a doctor's diagnostic description, examination result specifications, etc. The semantic parsing sub-module analyzes the text using natural language processing techniques. The method comprises the steps of firstly identifying entities, identifying key entities in texts, such as disease names, symptoms, examination items and the like, and then extracting relations, such as causal relations, accompanying relations and the like, between the entities. Through these operations, unstructured text is converted into standardized semantic data. For example, regarding the text "fever, hypotension occurs in the patient, and the heart ultrasound shows the left ventricular contractile function is reduced in consideration of the septic shock", the semantic analysis sub-module can identify entities such as "fever", "hypotension", "septic shock", "left ventricular contractile function is reduced", and the like, and extract the relationship between them, so as to generate standardized semantic data.
The structured medical record is generated by combining standardized semantic data with mapped data fields. And integrating the standardized semantic data generated by the semantic analysis submodule with the data field mapped by the template matching submodule to form a complete structured diagnosis and treatment record. In this way, doctors can quickly acquire the disease information of patients through the structured diagnosis record, including various physiological parameters, diagnosis results and relations among the physiological parameters and the diagnosis results.
Example 3:
The abnormal detection module operates as follows, and dynamic threshold intervals are respectively set for heart function indexes, microcirculation states and organ perfusion labels. These dynamic threshold intervals are not fixed, but rather are adaptively adjusted based on patient history data and population statistics. The patient's condition is dynamically changing, and there is also an individual difference between different patients, so the fixed threshold value cannot accurately reflect the patient's real condition. The dynamic threshold interval of each patient is determined by collecting historical data of the patient, including data of cardiac function indexes, microcirculation states, organ perfusion labels and the like at different time points, and combining population statistics data, such as the average value and standard deviation of various indexes of the patient with the same illness state. For example, for a patient's cardiac performance, a dynamic threshold interval is determined over time based on historical ejection fraction data and ejection fraction statistics for the same type of septic shock patient.
And calculating the deviation degree of the current parameter value and the dynamic threshold interval in real time, and if the deviation degree of any parameter exceeds a preset tolerance coefficient, generating an abnormal signal and marking the abnormal type. The deviation degree can be calculated in various ways, such as calculating the ratio of the difference value between the current parameter value and the dynamic threshold interval boundary value to the threshold interval width. Assume that the current parameter value isThe dynamic threshold interval isThe preset tolerance coefficient isThe deviation calculation formula is:
When the deviation is greater than When the current parameter is abnormal, the abnormality detection module generates an abnormality signal, and marks the abnormality type, such as ' abnormal heart function index ', ' abnormal microcirculation state ', abnormal organ perfusion ', and the like, according to the label category to which the parameter belongs.
Example 4:
The embodiment has the effects of detailing how the diagnosis and treatment decision module generates reasonable diagnosis and treatment schemes according to abnormal types and individual differences of patients, coordinating conflicts among multiple schemes, providing final diagnosis and treatment suggestions for doctors, and assisting the doctors in making scientific diagnosis and treatment decisions.
The diagnosis and treatment decision module comprises a path optimization sub-module and a conflict resolution sub-module.
The path optimization submodule adopts a reinforcement learning algorithm to generate diagnosis and treatment scheme priority ordering according to abnormal types and individual differences of patients. Reinforcement learning is a method of learning an optimal strategy by an agent interacting with an environment and rewarding. In the invention, the path optimization sub-module takes the abnormal type and the individual difference of the patient as input, explores different diagnosis and treatment schemes through a reinforcement learning algorithm, and obtains rewards for improving the illness state of the patient according to the schemes. For example, for patients with different types of abnormalities (such as heart dysfunction, microcirculation abnormality, etc.) and different physical conditions (age, basic diseases, etc.), different drug treatment schemes, liquid resuscitation schemes, etc. are tried, and corresponding rewards are given according to changes of patient condition indexes (such as blood pressure rise, organ perfusion improvement, etc.). Through continuous learning and optimization, diagnosis and treatment scheme priority orders aiming at different conditions are generated, and references are provided for doctors.
The conflict resolution sub-module coordinates resource conflict and time sequence conflict among multiple schemes through a fuzzy logic algorithm and outputs final diagnosis and treatment suggestions. In the actual diagnosis and treatment process, various diagnosis and treatment schemes are applicable at the same time, but resource conflict (such as insufficient medicine supply, limited medical equipment and the like) or time sequence conflict (such as contradiction between implementation sequences of different treatment measures) may occur. The fuzzy logic algorithm can process fuzzy and uncertain information, and conflict factors of different schemes are subjected to fuzzy processing, and reasoning and decision are performed according to expert experience and preset rules. For example, for conflict of the drug treatment scheme and the operation treatment scheme in terms of resources and time, various factors such as emergency degree of patient illness, resource availability and the like are comprehensively considered through a fuzzy logic algorithm, the final diagnosis and treatment proposal is determined, and feasibility and effectiveness of the diagnosis and treatment scheme are ensured.
Example 5:
The analysis process of the performance monitoring module is as follows, the response time from the generation of the abnormal signal to the output of the diagnosis and treatment advice is recorded, and if the response time exceeds a preset response threshold value, the abnormal signal is marked as a delay event. The preset response threshold is set according to clinical actual demands and experience, and in general, the condition of the patient suffering from the septic shock is critical, and diagnosis and treatment decisions need to be made quickly. If the system gives diagnosis and treatment advice for a long time after detecting the abnormal signal, the treatment effect of the patient may be affected. For example, the preset response threshold is set to be 5 minutes, when the abnormal signal is generated and the system outputs diagnosis and treatment advice only after 5 minutes, the event is marked as a delay event.
Counting the number of times the diagnosis and treatment suggestion is clinically corrected, and marking the diagnosis and treatment suggestion as an inefficient decision event if the correction number exceeds a preset correction threshold. The diagnosis and treatment advice is generated by the system according to the algorithm, but in actual clinical application, doctors can correct the diagnosis and treatment advice according to own experience and further observation of patients. If the number of corrections is excessive, there may be inaccurate or unreasonable places for the diagnosis and treatment advice generated by the system. The preset correction threshold can be adjusted according to the actual conditions and clinical data of different hospitals. For example, through statistical analysis, in a hospital, if the number of revisions of a diagnosis and treatment recommendation exceeds 10 times in one month, then the decision event is marked as an inefficient decision event.
And carrying out weighted calculation on the occurrence frequency of the delay event and the low-efficiency decision event to generate a system performance score. Different types of events have different degrees of impact on system performance, and therefore require weighted calculation of the frequency of occurrence of delay events and inefficient decision events. For example, the weight of the delay event is set to 0.6, the weight of the low-efficiency decision event is set to 0.4, and the calculation formula of the system performance score is as follows:
The processor is in communication connection with the data tracing module, and the data tracing module is used for labeling source labels for each item of data in the structured diagnosis and treatment record, and the source labels comprise image equipment models, acquisition time and operator information. When the performance score of the system is lower than a preset qualification threshold value, triggering data tracing analysis and positioning data source abnormal nodes of a delay event or an inefficient decision event. By labeling the data source label, the data acquisition process and related information can be traced. When the system performance is problematic, the data tracing module can help find the root cause of the problem. For example, these problems may be ameliorated if a delay event is found due to the too slow rate of data acquisition by a particular imaging device, or if an inefficient decision event is caused by operator entry of data errors.
The early warning feedback module comprises a multi-stage early warning sub-module. The multi-stage early warning submodule divides early warning grades according to the severity of the abnormal signal, the primary early warning triggers interface popup prompt, the secondary early warning is synchronously sent to the mobile terminal, and the tertiary early warning is directly communicated with the first-aid system. The severity of the abnormal signal can be evaluated according to the degree of deviation, the importance of the abnormal parameter and other factors. The first-level early warning is suitable for some relatively light abnormal conditions and prompts a doctor through a popup window on a system interface, the second-level early warning aims at serious abnormal conditions and sends early warning information to a mobile terminal of the doctor in addition to interface prompt, so that the doctor can know in time, and the third-level early warning aims at emergency conditions endangering the life of a patient and is directly communicated with an emergency system to start an emergency rescue program.
The system also comprises a model iteration module, wherein the model iteration module is used for periodically collecting system performance logs and clinical feedback data, updating a deep learning model and a reinforcement learning strategy through an online learning algorithm, and optimizing the accuracy of dynamic feature extraction and diagnosis and treatment decision. With the continuous accumulation of clinical data and the operation of the system, there is a need to continuously optimize models and algorithms to improve the performance of the system. The model iteration module updates the deep learning model and the reinforcement learning strategy by an online learning algorithm through collecting data in a system performance log, such as information of delay events, low-efficiency decision events and the like, and clinical feedback data, such as evaluation of diagnosis and treatment suggestions by doctors, treatment effects of patients and the like. For example, by adopting an online learning algorithm such as random gradient descent and the like, parameters of a deep learning model are adjusted, so that the model can extract dynamic characteristics more accurately, a reinforcement learning strategy is optimized, and accuracy and rationality of diagnosis and treatment decisions are improved.
The invention also comprises a method for processing the structural data of the infectious shock diagnosis and treatment image information, which comprises the following steps:
acquiring multi-mode medical image data through an image acquisition module;
Performing space-time alignment and feature normalization processing on the multi-mode data to generate fusion data;
Extracting space-time dynamic characteristics of the fusion data by using a deep learning model to generate a characteristic vector;
Mapping the feature vector to a clinical data template to generate a structured diagnosis and treatment record;
monitoring parameter deviation conditions of diagnosis and treatment records in real time, and triggering abnormal signals and adaptive diagnosis and treatment suggestions;
recording system response events and generating performance logs, and positioning abnormal nodes through data tracing;
Triggering a multi-stage feedback mechanism according to the early warning level, and iteratively optimizing algorithm performance through a model.
Reference is made to the above embodiments for implementation of the method, and the description is not repeated.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (7)

1.一种感染性休克诊疗图像信息结构化数据处理系统，其特征在于，包括处理器、图像采集模块、多模态数据融合模块、动态特征提取模块、结构化处理引擎、诊疗决策模块、异常检测模块、预警反馈模块及性能监控模块；1. A septic shock diagnosis and treatment image information structured data processing system, characterized by comprising a processor, an image acquisition module, a multimodal data fusion module, a dynamic feature extraction module, a structured processing engine, a diagnosis and treatment decision module, an abnormality detection module, a warning feedback module and a performance monitoring module; 所述图像采集模块用于获取患者的医学影像数据，包括CT图像、超声图像和血流动力学时序数据；所述多模态数据融合模块对医学影像数据进行时空对齐和特征归一化处理，生成多模态融合数据；所述动态特征提取模块通过预设的深度学习模型对多模态融合数据进行时间序列分析和空间特征提取，生成动态特征向量；所述结构化处理引擎将动态特征向量映射至预定义的临床数据模板中，生成结构化诊疗记录；The image acquisition module is used to obtain the patient's medical imaging data, including CT images, ultrasound images and hemodynamic time series data; the multimodal data fusion module performs time-space alignment and feature normalization processing on the medical imaging data to generate multimodal fusion data; the dynamic feature extraction module performs time series analysis and spatial feature extraction on the multimodal fusion data through a preset deep learning model to generate a dynamic feature vector; the structured processing engine maps the dynamic feature vector to a predefined clinical data template to generate a structured diagnosis and treatment record; 所述异常检测模块实时监测结构化诊疗记录中的生理参数、血流动力学指标及影像特征，若检测到参数偏离预设阈值范围，则生成异常信号并发送至处理器；所述诊疗决策模块根据异常信号调用预设的临床路径算法，生成适应性诊疗建议；所述预警反馈模块接收异常信号并触发可视化预警界面；所述性能监控模块记录异常信号的响应时间及诊疗决策的修正次数，生成系统性能日志；The abnormality detection module monitors the physiological parameters, hemodynamic indicators and imaging features in the structured diagnosis and treatment records in real time. If it is detected that the parameters deviate from the preset threshold range, an abnormality signal is generated and sent to the processor; the diagnosis and treatment decision module calls the preset clinical pathway algorithm according to the abnormal signal to generate adaptive diagnosis and treatment suggestions; the early warning feedback module receives the abnormal signal and triggers the visual early warning interface; the performance monitoring module records the response time of the abnormal signal and the number of corrections of the diagnosis and treatment decision, and generates a system performance log; 所述动态特征提取模块的具体运行过程如下：The specific operation process of the dynamic feature extraction module is as follows: 对血流动力学时序数据采用长短期记忆网络进行时间依赖性建模，输出时序特征向量；对CT图像和超声图像分别通过三维卷积神经网络进行多尺度特征提取，输出空间特征向量；将时序特征向量与空间特征向量进行张量拼接，并通过全连接层降维后生成动态特征向量；Long short-term memory network is used to model the time dependency of hemodynamic time series data, and the time series feature vector is output; multi-scale feature extraction is performed on CT images and ultrasound images through three-dimensional convolutional neural network, and spatial feature vector is output; the time series feature vector and the spatial feature vector are tensor spliced, and the dynamic feature vector is generated after dimensionality reduction through the fully connected layer; 所述结构化处理引擎包括模板匹配子模块和语义解析子模块；The structured processing engine includes a template matching submodule and a semantic parsing submodule; 所述模板匹配子模块基于临床指南定义的数据字段，将动态特征向量中的生理参数分类映射至心功能指标、微循环状态及器官灌注标签；所述语义解析子模块通过自然语言处理技术对非结构化文本报告进行实体识别和关系抽取，生成标准化语义数据；The template matching submodule maps the physiological parameters in the dynamic feature vector to cardiac function indicators, microcirculation status and organ perfusion labels based on the data fields defined in the clinical guidelines; the semantic parsing submodule performs entity recognition and relationship extraction on the unstructured text report through natural language processing technology to generate standardized semantic data; 所述结构化诊疗记录由标准化语义数据与映射后的数据字段合并生成；The structured medical record is generated by merging the standardized semantic data with the mapped data fields; 所述性能监控模块的具体分析过程如下：The specific analysis process of the performance monitoring module is as follows: 记录异常信号从生成至诊疗建议输出的响应时间，若响应时间超过预设响应阈值，则标记为延迟事件；Record the response time from the generation of abnormal signals to the output of diagnosis and treatment recommendations. If the response time exceeds the preset response threshold, it is marked as a delay event. 统计诊疗建议被临床修正的次数，若修正次数超过预设修正阈值，则标记为低效决策事件；Count the number of times the diagnosis and treatment recommendations are clinically revised. If the number of revisions exceeds the preset revision threshold, it will be marked as an inefficient decision-making event. 将延迟事件和低效决策事件的发生频率进行加权计算，生成系统性能评分。The frequencies of delay events and inefficient decision-making events are weighted and calculated to generate a system performance score. 2.根据权利要求1所述的感染性休克诊疗图像信息结构化数据处理系统，其特征在于，所述异常检测模块的具体运行过程如下：2. The septic shock diagnosis and treatment image information structured data processing system according to claim 1, characterized in that the specific operation process of the abnormality detection module is as follows: 对心功能指标、微循环状态及器官灌注标签分别设置动态阈值区间，所述动态阈值区间基于患者历史数据及群体统计数据进行自适应调整；Dynamic threshold intervals are set for cardiac function indicators, microcirculation status and organ perfusion labels respectively, and the dynamic threshold intervals are adaptively adjusted based on patient historical data and group statistical data; 实时计算当前参数值与动态阈值区间的偏离度，若任一参数的偏离度超过预设容差系数，则生成异常信号并标记异常类型。The deviation between the current parameter value and the dynamic threshold interval is calculated in real time. If the deviation of any parameter exceeds the preset tolerance coefficient, an abnormal signal is generated and the abnormal type is marked. 3.根据权利要求2所述的感染性休克诊疗图像信息结构化数据处理系统，其特征在于，所述诊疗决策模块包括路径优化子模块和冲突消解子模块；3. The septic shock diagnosis and treatment image information structured data processing system according to claim 2, characterized in that the diagnosis and treatment decision module includes a path optimization submodule and a conflict resolution submodule; 所述路径优化子模块采用强化学习算法，根据异常类型及患者个体差异生成诊疗方案优先级排序；所述冲突消解子模块通过模糊逻辑算法对多方案间的资源冲突和时序冲突进行协调，输出最终诊疗建议。The path optimization submodule adopts a reinforcement learning algorithm to generate a priority ranking of diagnosis and treatment plans according to the abnormality type and individual differences of patients; the conflict resolution submodule coordinates the resource conflicts and timing conflicts among multiple plans through a fuzzy logic algorithm and outputs the final diagnosis and treatment recommendations. 4.根据权利要求3所述的感染性休克诊疗图像信息结构化数据处理系统，其特征在于，所述处理器与数据溯源模块通信连接，所述数据溯源模块用于：4. The septic shock diagnosis and treatment image information structured data processing system according to claim 3, characterized in that the processor is communicatively connected with a data tracing module, and the data tracing module is used to: 对结构化诊疗记录中的每项数据标注来源标签，包括影像设备型号、采集时间及操作人员信息；Label each data item in the structured medical record with its source, including imaging equipment model, acquisition time, and operator information; 当系统性能评分低于预设合格阈值时，触发数据溯源分析，定位延迟事件或低效决策事件的数据来源异常节点。When the system performance score is lower than the preset qualified threshold, data traceability analysis is triggered to locate abnormal nodes of the data source of delayed events or inefficient decision-making events. 5.根据权利要求4所述的感染性休克诊疗图像信息结构化数据处理系统，其特征在于，所述预警反馈模块包括多级预警子模块；5. The septic shock diagnosis and treatment image information structured data processing system according to claim 4, characterized in that the early warning feedback module includes a multi-level early warning submodule; 所述多级预警子模块根据异常信号的严重程度划分预警等级，一级预警触发界面弹窗提示，二级预警同步发送至移动终端，三级预警直接连通急救系统。The multi-level warning submodule divides the warning level according to the severity of the abnormal signal. The first-level warning triggers a pop-up window prompt on the interface, the second-level warning is sent to the mobile terminal simultaneously, and the third-level warning is directly connected to the emergency system. 6.根据权利要求5所述的感染性休克诊疗图像信息结构化数据处理系统，其特征在于，所述系统还包括模型迭代模块，所述模型迭代模块用于：6. The septic shock diagnosis and treatment image information structured data processing system according to claim 5, characterized in that the system further comprises a model iteration module, wherein the model iteration module is used to: 定期采集系统性能日志及临床反馈数据，通过在线学习算法更新深度学习模型和强化学习策略，优化动态特征提取及诊疗决策的准确性。Regularly collect system performance logs and clinical feedback data, update deep learning models and reinforcement learning strategies through online learning algorithms, and optimize the accuracy of dynamic feature extraction and diagnosis and treatment decisions. 7.一种基于权利要求1~6任一所述的感染性休克诊疗图像信息结构化数据处理系统的感染性休克诊疗图像信息结构化数据处理方法，其特征在于，包括以下步骤：7. A method for processing structured data of septic shock diagnosis and treatment image information based on the system for processing structured data of septic shock diagnosis and treatment image information according to any one of claims 1 to 6, characterized in that it comprises the following steps: 通过图像采集模块获取多模态医学影像数据；Acquire multimodal medical imaging data through an image acquisition module; 对多模态数据进行时空对齐和特征归一化处理，生成融合数据；Perform spatiotemporal alignment and feature normalization on multimodal data to generate fused data; 利用深度学习模型提取融合数据的时空动态特征，生成特征向量；Use deep learning models to extract spatiotemporal dynamic features of fused data and generate feature vectors; 将特征向量映射至临床数据模板，生成结构化诊疗记录；Map feature vectors to clinical data templates to generate structured diagnosis and treatment records; 实时监测诊疗记录的参数偏离情况，触发异常信号及适应性诊疗建议；Real-time monitoring of parameter deviations in diagnosis and treatment records, triggering abnormal signals and adaptive diagnosis and treatment recommendations; 记录系统响应事件并生成性能日志，通过数据溯源定位异常节点；Record system response events and generate performance logs, and locate abnormal nodes through data tracing; 根据预警等级触发多级反馈机制，并通过模型迭代优化算法性能。A multi-level feedback mechanism is triggered according to the warning level, and the algorithm performance is optimized through model iteration.
CN202510333815.XA 2025-03-20 2025-03-20 Septic shock diagnosis and treatment image information structured data processing system and method Active CN119851925B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202510333815.XA CN119851925B (en) 2025-03-20 2025-03-20 Septic shock diagnosis and treatment image information structured data processing system and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202510333815.XA CN119851925B (en) 2025-03-20 2025-03-20 Septic shock diagnosis and treatment image information structured data processing system and method

Publications (2)

Publication Number Publication Date
CN119851925A CN119851925A (en) 2025-04-18
CN119851925B true CN119851925B (en) 2025-06-13

Family

ID=95358410

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202510333815.XA Active CN119851925B (en) 2025-03-20 2025-03-20 Septic shock diagnosis and treatment image information structured data processing system and method

Country Status (1)

Country Link
CN (1) CN119851925B (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119339873A (en) * 2024-12-16 2025-01-21 北京凯普顿医药科技开发有限公司 A method and device for analyzing nervous system images

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200097651A1 (en) * 2018-09-26 2020-03-26 General Electric Company Systems and methods to achieve robustness and security in medical devices
CN113611401A (en) * 2021-08-26 2021-11-05 中国医学科学院阜外医院 A system and method for perioperative blood management
CN113744873B (en) * 2021-11-08 2022-02-11 浙江大学 Heating to-be-checked auxiliary differential diagnosis system based on task decomposition strategy
CN118016280A (en) * 2024-03-11 2024-05-10 中国医科大学 Breast cancer clinical data analysis diagnosis and treatment platform based on artificial intelligence large language model
CN118629634A (en) * 2024-08-06 2024-09-10 合肥长月科技有限公司 An intelligent diagnosis system based on multimodal data fusion
CN119153117A (en) * 2024-11-02 2024-12-17 东莞市东南部中心医院(东莞市东南部中医医疗服务中心、广东医科大学附属东莞第一医院) Cerebral hemorrhage personalized treatment scheme optimization method and system based on big data analysis

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119339873A (en) * 2024-12-16 2025-01-21 北京凯普顿医药科技开发有限公司 A method and device for analyzing nervous system images

Also Published As

Publication number Publication date
CN119851925A (en) 2025-04-18

Similar Documents

Publication Publication Date Title
CN119377894B (en) Multi-mode data dynamic fusion and labeling method and system for medical decision
CN111292821B (en) Medical diagnosis and treatment system
US11232856B2 (en) Methods for automatic cohort selection in epidemiologic studies and clinical trials
US7630947B2 (en) Medical ontologies for computer assisted clinical decision support
US8670997B2 (en) Quality metric extraction and editing for medical data
CN113744873A (en) Heating to-be-checked auxiliary differential diagnosis system based on task decomposition strategy
CN118609743A (en) Medical data management method, system and storage medium based on artificial intelligence
CN113990502A (en) A prediction system of ICU heart failure based on heterogeneous graph neural network
CN116543917A (en) An Information Mining Method for Heterogeneous Time Series Data
US12362068B2 (en) Interpretation of intraoperative sensor data using concept graph neural networks
CN119129728A (en) A method and system for constructing a knowledge base for brain disease nursing
CN118762826A (en) Method for predicting coronary artery stenosis and FFR based on big data of coronary artery diagnosis and treatment
CN118016274B (en) Disease knowledge retrieval and diagnosis decision-making auxiliary system
JP2025022871A (en) Apparatus and method for determining patient survival profile using artificial intelligence enabled electrocardiogram (ECG) - Patent Application 20070123633
CN119851925B (en) Septic shock diagnosis and treatment image information structured data processing system and method
Sahu et al. Application of fuzzy logic to healthcare industry
Pan et al. EH-Benchmark: Ophthalmic hallucination benchmark and agent-driven top-down traceable reasoning workflow
CN119964744B (en) Self-inspection method and system for medical units based on data analysis
US12367953B1 (en) Apparatus and a method for the generation of a medical report
CN119153127B (en) Infectious disease data intelligent coding method, device, computer equipment and storage medium
EP4239642A1 (en) Method for generating protocol data of a radiological image data measurement
Gulati et al. Research Applications in Healthcare with Machine Learning Approaches
CN120727176A (en) Medical decision-making method, device and storage medium based on neural symbolic hybrid model
Putrawansyah et al. Implementation of the Bayesian Network Algorithm to Predict Chronic Diseases Using Electronic Medical Record Data at UPTD RSD Besemah, Pagar Alam City
CN120748762A (en) Artificial intelligence corpus construction method, device and equipment based on multi-modal data

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant