Acute infection rapid diagnosis method, device and equipment

Abstract

The invention relates to the technical field of medical data, in particular to an acute infection rapid diagnosis method, device and equipment, and the method comprises the steps: collecting an electronic medical record, blood routine, inflammation indexes, a microbiological detection result and imaging data of a target patient; respectively extracting corresponding target basic features based on the electronic medical record, the blood routine, the inflammation index, the microbiological detection result and the imaging data; based on the basic features, a convolutional neural network and a long-short-term memory network, target depth features are extracted; obtaining an acute infection prediction model, wherein the acute infection prediction model is obtained through model training; based on the target depth feature and the acute infection prediction model, the acute infection type of the target patient is determined, the accuracy and efficiency of diagnosis are further improved, scientific antibiotic use guidance is provided for doctors, waste of medical resources is reduced, and prognosis of the patient is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data, and in particular, to a method, device, and equipment for rapid diagnosis of acute infections. Background Art

[0002] With the rapid development of medical informatization, the application of electronic medical record systems in medical institutions has become increasingly widespread. Electronic medical records contain important health information of patients, including medical history, diagnosis results, treatment plans, medication conditions, and laboratory test results, etc. These information provide important decision-making support for doctors and help improve the efficiency and accuracy of medical work. However, existing electronic medical record data has problems such as data inconsistency, incompleteness, and incorrect information, which affect the accuracy of auxiliary diagnosis information and may mislead doctors to make incorrect diagnosis results.

[0003] In the medical field, acute infection is one of the important reasons for patients to seek medical treatment. However, existing diagnostic methods often rely on manual judgment, which not only consumes a large amount of manpower but also causes doctors to ignore existing thrombus conditions. In addition, in most cases, doctors need to conduct multiple examinations and referrals for patients to obtain an accurate diagnosis, which not only increases the burden on medical resources but also delays the condition. Especially in areas with relatively backward medical resources, the differential diagnosis of potential causes of fever of unknown origin remains an urgent problem to be solved.

[0004] Therefore, how to extract key information from electronic medical records to assist in the diagnosis of acute infections is an urgent technical problem to be solved at present. Summary of the Invention

[0005] In view of the above problems, the present invention provides a method, device, and equipment for rapid diagnosis of acute infections that overcome the above problems or at least partially solve the above problems.

[0006] In a first aspect, the present invention provides a method for rapid diagnosis of acute infections, including: Collecting the electronic medical record, blood routine, inflammatory index, microbial test result, and imaging data of a target patient; Based on the electronic medical record, blood routine, inflammatory index, microbial test result, and imaging data, respectively extracting corresponding target basic features; Based on the target basic features, convolutional neural network, and long short-term memory network, extracting target deep features; Obtaining an acute infection prediction model, which is obtained through model training; Based on the target deep features and the acute infection prediction model, determining the type of acute infection of the target patient.

[0007] Preferably, based on the electronic medical record, blood routine, inflammation indicators, microbial test results, and imaging data, corresponding target basic features are extracted respectively, including: Based on the electronic medical record, medical history features, disease features, and diagnosis and treatment process features are extracted; Based on the blood routine and inflammation indicators, quantitative index features of the body's immune response are extracted; Based on the microbial test results, features of the presence or absence of pathogens, pathogen type features, and drug resistance features are extracted; Based on the imaging data, infection site features, macroscopic pathological change features, and microscopic pathological change features are extracted.

[0008] Preferably, based on the target basic features, convolutional neural network, and long short-term memory network, target deep features are extracted, including: Use the convolutional neural network to extract local features in the target basic features; Use the long short-term memory network to capture long-range dependence features in the target basic features; Based on the local features and the long-range dependence features, target deep features are obtained.

[0009] Preferably, based on the target basic features, convolutional neural network, and long short-term memory network, target deep features are extracted, including: Based on the medical history features, disease features, and diagnosis and treatment process features, use the convolutional neural network and long short-term memory network to extract the correlation between symptom features and the logical relationship between disease course progressions; Based on the quantitative index features of the body's immune response, use the convolutional neural network and long short-term memory network to extract the correlation between the quantitative indexes of each immune response; Based on the features of the presence or absence of pathogens, pathogen type features, and drug resistance features, use the convolutional neural network and long short-term memory network to extract etiological diagnosis index features; Based on the infection site features, macroscopic pathological change features, and microscopic pathological change features, use the convolutional neural network and long short-term memory network to locate the infection focus and identify the nature of the infection.

[0010] Preferably, an acute infection prediction model is obtained, including: Collect the historical electronic medical records, historical blood routine, historical inflammation indicators, historical microbial test results, and historical imaging data of historical patients; Based on the historical electronic medical records, historical blood routine, historical inflammation indicators, historical microbial test results, and historical imaging data, corresponding historical basic features are extracted respectively; Based on the historical basic features, convolutional neural network, and long short-term memory network, historical deep features are extracted; Train a prediction model based on the historical depth features to obtain an acute infection prediction model, which is used to predict the infection type, predict the types of pathogen subtypes, and the drug resistance judgment results.

[0011] Preferably, based on the target depth features and the acute infection prediction model, determine the acute infection type of the target patient, including: Input the target depth features into the acute infection prediction model to output the preliminary pathogen candidate range of the target patient; Extract the pathogen feature sequences from the etiological diagnosis index features; Use the basic local alignment algorithm to align the pathogen feature sequences with the full gene sequences of known pathogens in the pathogen gene database to determine the similarity between the pathogen feature sequences and the full gene sequences of known pathogens; Based on the similarity, determine the acute infection type of the target patient.

[0012] Preferably, based on the similarity, determine the acute infection type of the target patient, including: Based on the similarity, the quantitative index features of each immune response, the infection focus and the nature of the infection, judge whether there are contradictions to obtain a judgment result; Based on the judgment result, determine the accuracy of the acute infection type of the target patient.

[0013] Preferably, after determining the acute infection type of the target patient based on the target depth features and the acute infection prediction model, it further includes: Based on the drug resistance characteristics of the target patient, provide guidance for the antibiotic use of the target patient.

[0014] In a second aspect, the present invention also provides an acute infection rapid diagnosis device, including: An acquisition module, configured to acquire the electronic medical record, blood routine, inflammatory indicators, microbial test results, and imaging data of the target patient; A first extraction module, configured to respectively extract the corresponding target basic features based on the electronic medical record, blood routine, inflammatory indicators, microbial test results, and imaging data; A second extraction module, configured to extract target depth features based on the target basic features, convolutional neural network, and long short-term memory network; An acquisition module, configured to acquire an acute infection prediction model, which is obtained through model training; A determination module, configured to determine the acute infection type of the target patient based on the target depth features and the acute infection prediction model.

[0015] In a third aspect, the present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the method described in the first aspect is implemented.

[0016] In a fourth aspect, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the method described in the first aspect is implemented.

[0017] One or more technical solutions in the embodiments of the present invention have at least the following technical effects or advantages: The present invention provides a method for rapid diagnosis of acute infections, including collecting the electronic medical records, blood routine, inflammatory indicators, microbial test results, and imaging data of a target patient; respectively extracting corresponding target basic features based on the electronic medical records, blood routine, inflammatory indicators, microbial test results, and imaging data; extracting target deep features based on the target basic features, convolutional neural network, and long short-term memory network; obtaining an acute infection prediction model, which is obtained through model training; and determining the type of acute infection of the target patient based on the target deep features and the acute infection prediction model, thereby improving the accuracy and efficiency of diagnosis, providing scientific guidance for doctors on the use of antibiotics, reducing the waste of medical resources, and improving the prognosis of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings: Figure 1 shows a schematic flow chart of the steps of the method for rapid diagnosis of acute infections in the embodiments of the present invention; Figure 2 shows a schematic structural diagram of the device for rapid diagnosis of acute infections in the embodiments of the present invention; Figure 3 shows a schematic structural diagram of the computer device for implementing the method for rapid diagnosis of acute infections in the embodiments of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] The exemplary embodiments of the present invention will be described in more detail below with reference to the drawings. Although the exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided so that the present invention can be more thoroughly understood and the scope of the present invention can be fully communicated to those skilled in the art.

[0020] Example 1

[0021] An embodiment of the present invention provides a rapid diagnosis method for acute infections, as Figure 1 shown, including: S101, collecting the electronic medical record, blood routine, inflammatory indicators, microbial test results, and imaging data of the target patient; S102, respectively extracting corresponding target basic features based on the electronic medical record, blood routine, inflammatory indicators, microbial test results, and imaging data; S103, extracting target deep features based on the target basic features, convolutional neural network, and long short-term memory network; S104, obtaining an acute infection prediction model, which is obtained through model training; S105, determining the type of acute infection of the target patient based on the target deep features and the acute infection prediction model.

[0022] In a specific implementation manner, multi-modal data of the target patient is collected, including: electronic medical record, blood routine, inflammatory indicators, microbial test results, and imaging data.

[0023] Among them, the electronic medical record is unstructured text data, specifically including patient basic information, such as: age, gender, past medical history, allergy history, exposure history (travel history, contact history); the time line of medical treatment: time stamp information such as symptom onset time, medication time, examination time, etc.

[0024] Clinical manifestation characteristics: symptom description, sign record, diagnosis-related text. Among them, symptom description, such as fever, cough, expectoration (sputum color / character), dyspnea, fatigue, headache, etc.; sign record, such as heart rate, blood pressure, respiratory rate, body temperature, lung rales, tonsil enlargement, etc.; diagnosis-related documents, such as preliminary diagnosis conclusion, differential diagnosis record, key judgments in doctor's remarks.

[0025] The blood routine is structured data, including: cell count indicators, red blood cell-related indicators, and platelet count.

[0026] Among them, the cell count indicators include: total white blood cell count (WBC) and classification: absolute value / percentage of neutrophils (NEU), absolute value / percentage of lymphocytes (LYM), absolute value / percentage of monocytes (MON), absolute value / percentage of eosinophils (EOS).

[0027] Red blood cell-related indicators: hemoglobin (HGB), hematocrit (HCT) (assisting in judging whether anemia is accompanied by infection).

[0028] Platelet count (PLT) is used to indicate the severity of infection or coagulation function.

[0029] Inflammatory indicators are structured data. Classic inflammatory markers include: C-reactive protein (CRP), which is significantly elevated during bacterial infections and mildly elevated or normal during viral infections; procalcitonin (PCT), which is a specific indicator of sepsis and is usually normal during viral infections; erythrocyte sedimentation rate (ESR), a non-specific inflammatory indicator that can be elevated in both infections and autoimmune diseases.

[0030] It also includes: cytokines and chemokines, specifically including: interleukin-6 (IL-6), a sensitive indicator of early infection, and an elevation indicates a systemic inflammatory response; tumor necrosis factor-α (TNF-α), which reflects the excessive inflammatory state caused by infection.

[0031] Microbiological test results are semi-structured data or sequence data. Traditional microbiological test characteristics: culture results have pathogen growth status (such as "aerobic / anaerobic bacteria growth"), colony morphology (size, color, edge characteristics); smear microscopy; Gram staining results (Gram-positive / negative bacteria), bacterial morphology (cocci, bacilli, spirochetes), fungal spore / hyphal morphology; serum test: antibody type (IgM / IgG) and titer (such as "positive for Mycoplasma pneumoniae IgM antibody").

[0032] Imaging data are specifically image data.

[0033] Next, feature extraction is performed on each type of data respectively to extract the corresponding target basic features. Specifically, in S102, based on the electronic medical record, blood routine, inflammatory indicators, microbiological test results, and imaging data, the corresponding target basic features are extracted respectively.

[0034] Before extracting the target basic features, preprocessing is first performed, which specifically includes data cleaning and standardization processing, removing outliers and missing values, and using different filling strategies for categorical variables and numerical variables to handle missing values. For categorical variables, the mode is used for filling, and for numerical variables, if their distribution conforms to a normal distribution, the mean filling method is adopted, and if their distribution does not conform to a normal distribution, the median filling method is used.

[0035] Next, feature extraction is performed on each type of data respectively, including: based on the electronic medical record, extracting medical history features, disease characteristics, and treatment process features; Based on the blood routine and inflammatory indicators, extracting quantitative index features of the body's immune response; Based on the microbiological test results, extracting features of the presence or absence of pathogens, pathogen type features, and drug resistance features; Based on the imaging data, extracting infection site features, macroscopic pathological change features, and microscopic pathological change features.

[0036] Specifically, natural language processing techniques can be used to parse unstructured text data, and deep learning models can be combined to extract target basic features.

[0037] For example, for electronic medical records, the pathogen name is non - Streptococcus pneumoniae, symptom terms such as sepsis, and the examination item is blood culture are extracted.

[0038] For blood routine tests, some derivative features are extracted: Neutrophil - to - lymphocyte ratio (NLR); it usually increases during bacterial infections and may decrease during viral infections.

[0039] Abnormal cell ratio: such as blast cells (indicating blood system infection or leukemia complicated with infection).

[0040] For microbial test results, nucleic acid detection features can be extracted: target gene fragments for PCR detection (such as bacterial 16srRNA gene, viral ORFlab gene), fluorescence quantitative Ct value (reflecting pathogen load); genomic sequence features: pathogen characteristic sequences extracted by metagenomic sequencing (such as viral spike protein gene, bacterial drug - resistant gene fragment), which are used for subsequent comparison with the pathogen genome database.

[0041] For imaging data, features such as lung imaging (CT / X - ray) can be extracted: lesion location, density, size, edge features, presence or absence of pleural effusion, etc.

[0042] For images of other parts, such as hepatosplenomegaly in abdominal ultrasound and hydronephrosis in urinary system CT (indicating the infection site).

[0043] Next, for the target basic features of various types of data, target deep features are extracted. That is, S103 is executed to extract target deep features based on the basic features, convolutional neural network, and long - short - term memory network.

[0044] Specifically, a convolutional neural network is used to extract local features in the target basic features; a long - short - term memory network is used to capture long - distance dependence features in the target basic features; based on the local features and long - distance dependence features, target deep features are obtained.

[0045] By associating the local features with the long - distance dependence features, target deep features are thus obtained.

[0046] Next, the specific target deep features are described in detail: Based on the medical history features, disease symptoms features, and diagnosis and treatment process features, the association relationships between symptom features and the logical relationships between disease course progressions are extracted through a convolutional neural network and a long - short - term memory network; Based on the quantitative index characteristics of the body's immune response, the correlation relationships between the quantitative indexes of each immune response are extracted through convolutional neural networks and long short-term memory networks; Based on the characteristics of the presence or absence of pathogens, pathogen type characteristics, and drug resistance characteristics, the etiological diagnosis index characteristics are extracted through convolutional neural networks and long short-term memory networks; Based on the characteristics of the infection site, macroscopic pathological changes, and microscopic pathological changes, the infection focus is located and the nature of the infection is differentiated through convolutional neural networks and long short-term memory networks.

[0047] Specifically, the extraction process of the corresponding target deep features is described for each type of data: Symptom descriptions in the electronic medical record (such as "chills, high fever, purulent sputum" → Streptococcus pneumoniae infection), exposure history (such as "raw freshwater fish and shrimp → Clonorchis sinensis infection"), and drug reaction (such as "ineffective penicillin treatment → drug-resistant bacteria"). This process can be transformed into an "symptom feature → etiology" association rule extraction process that can be learned by the model after being extracted through natural language processing (NLP).

[0048] In blood routine and inflammatory indicators, for example, an increase in neutrophils → bacterial infection, and a decrease in lymphocytes → viral infection. These two are classic infection type differentiation indicators; procalcitonin (PCT) > 0.5 ng / ml → indicates bacterial sepsis, and it is usually < 0.1 ng / ml during viral infection. By using convolutional neural networks and long short-term memory networks to construct a corresponding classification model, the initial determination and identification of the infection type are achieved.

[0049] For microbial detection features, the culture results, Gram staining, and nucleic acid sequences directly indicate the presence of pathogens (such as Gram-negative bacilli → suggesting Enterobacteriaceae infection, and 16s rRNA gene sequence matching bacterial classification), which are "gold standard" level evidence for etiological diagnosis. This process also uses convolutional neural networks and long short-term memory networks to construct a corresponding recognition model to obtain the specific subdivided types of the corresponding infections.

[0050] For imaging data: the lesion morphology (such as ground glass opacity in the lungs → viral pneumonia, lobar consolidation → bacterial pneumonia) is directly related to the pathological mechanism of pathogen infection, and CT / MRI features have been proven to assist in the determination of the infection type.

[0051] The above-extracted target depth features are all based on existing medical consensus, thus ensuring that the learning process in the above process of extracting target depth features is a clinically recognized diagnostic logic. For example, "procalcitonin (PCT)" is included as a core indicator for bacterial infection, and indicators with low specificity such as "erythrocyte sedimentation rate (SER)" are excluded to reduce noise. In the electronic medical record, "antibiotic use history" is mainly extracted to evaluate the risk of drug-resistant bacterial infection, thus meeting the clinical requirement of "precision medication".

[0052] Next, execute S104 to obtain an acute infection prediction model, which is obtained through model training.

[0053] Specifically, collect the historical electronic medical records, historical blood routine, historical inflammatory indicators, historical microbial test results, and historical imaging data of historical patients; Based on the historical electronic medical records, historical blood routine, historical inflammatory indicators, historical microbial test results, and historical imaging data, extract the corresponding historical basic features respectively; Based on the historical basic features, convolutional neural network, and long short-term memory network, extract historical depth features; Based on the historical depth features, train the prediction model to obtain an acute infection prediction model, and the acute infection prediction model is used to predict the infection type, predict the types of pathogen subtypes, and the drug resistance judgment result.

[0054] Through the multi-task learning training of the prediction model, simultaneously learn multiple related tasks such as "infection type classification", "pathogen subtype classification", and "drug resistance judgment" to obtain the multi-level requirements of clinical diagnosis.

[0055] Then, execute S105 to determine the acute infection type of the target patient based on the target depth features and the acute infection prediction model.

[0056] Specifically, input the target depth features into the acute infection prediction model to output the preliminary pathogen candidate range of the target patient; Extract the pathogen feature sequence from the etiological diagnosis index features; Use the basic local alignment search tool method to align the pathogen feature sequence with the complete gene sequences of known pathogens in the pathogen gene database to determine the similarity between the pathogen feature sequence and the complete gene sequences of known pathogens; Based on the similarity, determine the acute infection type of the target patient.

[0057] This type of acute infection has multiple categories. Classified by pathogen type, it includes bacterial infections, viral infections, fungal infections, and other microbial infections; classified by the site of infection, it includes: respiratory tract infections, digestive tract infections, urogenital tract infections, skin and soft tissue infections, blood infections, central nervous system infections, and infections in other sites; special types of acute infections: sepsis, mixed infections, opportunistic infections, and hospital-acquired infections.

[0058] In a specific implementation, the target depth features of the target patient are input into the acute infection prediction model, and thereby, a preliminary pathogen candidate range of the target patient is output.

[0059] For example, the output preliminary pathogen candidate range: a high possibility of Gram-negative bacterial infection and RNA virus, or a specific list of candidate pathogens, such as Escherichia coli, Klebsiella pneumoniae, and influenza virus.

[0060] Next, pathogen feature sequences are extracted from the etiological diagnosis index features. For example, pathogen feature sequences are extracted from the judgment of PCR-amplified nucleic acids and metagenomic sequence data, such as 16s rRNA gene fragments and viral spike protein gene sequences.

[0061] Within the preliminary pathogen candidate range, the extracted pathogen feature sequences are aligned with the complete gene sequences of known pathogens in the pathogen gene database. In this alignment process, the basic local alignment search tool (BLAST) algorithm is specifically used. Among them, blastp is used for protein sequence alignment, and blastn is used for nucleic acid sequence alignment. The extracted feature sequences are aligned with the complete genome sequences of known pathogens in the database. Among them, this pathogen genome database stores a large number of verified pathogen complete genome sequences, typing markers (serotype, genotype), virulence factor gene sequences, etc., and supports rapid retrieval according to the taxonomic hierarchy (kingdom - phylum - class - order - family - genus - species).

[0062] Specifically, the BLAST algorithm evaluates the similarity between the sequence to be detected and the sequences in the database by calculating indicators such as the alignment length, identity, and E-value of the sequences. For example, if the identity between the nucleic acid sequence to be detected and the 16s rRNA gene sequence of Escherichia coli in the database is ≥97% and the E-value < 1e-10, it is determined as an Escherichia coli infection.

[0063] Next, based on the similarity, the acute infection type of the target patient is determined, specifically including: Based on the similarity, the quantitative index characteristics of each immune response, the infection site, and the nature of the infection, it is judged whether there are contradictions to obtain a judgment result; Based on the judgment result, the accuracy of the acute infection type of the target patient is determined.

[0064] Among them, after determining the type of acute infection based on similarity, it can be used as the final infection result. To verify the accuracy of this infection result, some of the data are subjected to a contradiction judgment.

[0065] For example, if the sequence alignment indicates an influenza virus, but the CRP in the inflammatory indicators is normal and the IL-6 is elevated, and its manifestation is consistent with the characteristics of viral infection, then the confidence level of the diagnosis is enhanced; if the sequence alignment result is a bacterial infection, but the imaging result is a viral infection, at this time, a secondary verification needs to be triggered. For example, it is recommended to supplement nucleic acid testing and so on.

[0066] Finally, after determining the type of acute infection of the target patient, it further includes: Providing guidance for the use of antibiotics for the target patient based on the drug resistance characteristics of the target patient.

[0067] When determining the drug resistance characteristics of the target patient, specifically, it is to judge whether the patient carries the β-lactamase gene.

[0068] Next, provide guidance for the use of antibiotics, such as recommending oseltamivir for influenza virus and avoiding using cephalosporins for ESBL-producing bacteria.

[0069] The following are the classic scenarios for diagnosis: Fever of unknown origin: Rapidly identify the type of infection (bacteria / viruses / fungi) and pathogens (such as differentiating influenza virus from Streptococcus pneumoniae).

[0070] Severe infection: Through blood routine (changes in white blood cells), inflammatory indicators (PCT / CRP), imaging (lung lesions), and microbial detection (nucleic acid sequences), rapidly identify the pathogens of sepsis and guide the use of antibiotics.

[0071] Infection in immunocompromised patients: Integrate the history of underlying diseases (such as diabetes, HIV), medication history (immunosuppressants), and other data to screen for opportunistic infections (such as Cryptococcus, CMV).

[0072] Mixed infection: Through sequence alignment of microbial detection and fusion of other data evidence (such as imaging showing multiple pathological changes), identify co-infections (such as bacteria + fungi).

[0073] From single pathogen infections (such as Streptococcus pneumoniae) to complex mixed infections involving multiple organs. Through the technical solution of the present invention, various data, including electronic medical records, blood routine, inflammatory indicators, microbial detection, and imaging, are integrated to achieve rapid and accurate identification of various types of infections, especially having significant advantages in difficult or severe infections.

[0074] One or more technical solutions in the embodiments of the present invention have at least the following technical effects or advantages: The present invention provides a rapid diagnosis method for acute infections, including collecting the electronic medical record, blood routine, inflammatory indicators, microbial test results, and imaging data of the target patient; respectively extracting the corresponding target basic features based on the electronic medical record, blood routine, inflammatory indicators, microbial test results, and imaging data; extracting target deep features based on the target basic features, convolutional neural network, and long short-term memory network; obtaining an acute infection prediction model, where the acute infection prediction model is obtained through model training; determining the type of acute infection of the target patient based on the target deep features and the acute infection prediction model, thereby improving the accuracy and efficiency of diagnosis, and also providing scientific guidance for doctors on the use of antibiotics, reducing the waste of medical resources, and improving the prognosis of patients.

[0075] Embodiment Two

[0076] Based on the same inventive concept, the embodiments of the present invention also provide a rapid diagnosis device for acute infections, as Figure 2 shown, including: A collection module 201, configured to collect the electronic medical record, blood routine, inflammatory indicators, microbial test results, and imaging data of the target patient; A first extraction module 202, configured to respectively extract the corresponding target basic features based on the electronic medical record, blood routine, inflammatory indicators, microbial test results, and imaging data; A second extraction module 203, configured to extract target deep features based on the target basic features, convolutional neural network, and long short-term memory network; An acquisition module 204, configured to acquire an acute infection prediction model, where the acute infection prediction model is obtained through model training; A determination module 205, configured to determine the type of acute infection of the target patient based on the target deep features and the acute infection prediction model.

[0077] In an alternative embodiment, the first extraction module 202 is configured to: Extract the medical history features, disease features, and diagnosis and treatment process features based on the electronic medical record; Extract the quantitative index features of the body's immune response based on the blood routine and inflammatory indicators; Extract the features of the presence or absence of pathogens, pathogen type features, and drug resistance features based on the microbial test results; Extract the infection site features, macroscopic pathological change features, and microscopic pathological change features based on the imaging data.

[0078] In an alternative embodiment, the second extraction module 203 is configured to: Use a convolutional neural network to extract local features in the target basic features; Use a long short-term memory network to capture long-range dependence features in the target basic features; Based on the local features and the long-range dependence features, obtain the target depth features.

[0079] In an optional implementation manner, the second extraction module 203 is used for: Based on the medical history features, disease symptoms features, and diagnosis and treatment process features, use a convolutional neural network and a long short-term memory network to extract the association relationships between symptom features and the logical relationships between disease progression; Based on the quantitative index features of the body's immune response, use a convolutional neural network and a long short-term memory network to extract the association relationships between the quantitative indexes of each immune response; Based on the features of the presence or absence of pathogens, pathogen type features, and drug resistance features, use a convolutional neural network and a long short-term memory network to extract etiological diagnosis index features; Based on the infection site features, macroscopic pathological change features, and microscopic pathological change features, use a convolutional neural network and a long short-term memory network to locate the infection focus and identify the nature of the infection.

[0080] In an optional implementation manner, the acquisition module 204 is used for: Collect the historical electronic medical records, historical blood routine, historical inflammatory indexes, historical microorganism test results, and historical imaging data of historical patients; Based on the historical electronic medical records, historical blood routine, historical inflammatory indexes, historical microorganism test results, and historical imaging data, respectively extract the corresponding historical basic features; Based on the historical basic features, a convolutional neural network, and a long short-term memory network, extract historical depth features; Train the prediction model based on the historical depth features to obtain an acute infection prediction model, and the acute infection prediction model is used to predict the infection type, predict the types of pathogen subtypes, and the drug resistance judgment result.

[0081] In an optional implementation manner, the determination module 205 is used for: Input the target depth features into the acute infection prediction model, and output the preliminary pathogen candidate range of the target patient; Extract the pathogen feature sequence from the etiological diagnosis index features; Use a basic local alignment-like algorithm to compare the pathogen feature sequence with the whole gene sequences of known pathogens in the pathogen gene database to determine the similarity between the pathogen feature sequence and the whole gene sequences of known pathogens; Based on the similarity, determine the acute infection type of the target patient.

[0082] In an alternative embodiment, the determining module 205 is further configured to: Based on the similarity, the quantitative index characteristics of each immune response, the infection focus and the nature of the infection, determine whether there is a contradiction, and obtain a determination result; Based on the determination result, determine the accuracy of the acute infection type of the target patient.

[0083] In an alternative embodiment, it further includes: a guiding module, configured to: Based on the drug resistance characteristics of the target patient, provide guidance for the use of antibiotics for the target patient.

[0084] Embodiment III

[0085] Based on the same inventive concept, an embodiment of the present invention provides a computer device, as Figure 3 shown, including a memory 304, a processor 302, and a computer program stored on the memory 304 and executable on the processor 302. When the processor 302 executes the program, the steps of the above-mentioned rapid diagnosis method for acute infection are implemented.

[0086] Wherein, in Figure 3 , a bus architecture (represented by bus 300), bus 300 may include any number of interconnected buses and bridges. Bus 300 links together various circuits including one or more processors represented by processor 302 and a memory represented by memory 304. Bus 300 may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and thus will not be further described herein. Bus interface 306 provides an interface between bus 300 and receiver 301 and transmitter 303. Receiver 301 and transmitter 303 may be the same element, i.e., a transceiver, providing a unit for communicating with various other devices over a transmission medium. Processor 302 is responsible for managing bus 300 and general processing, while memory 304 may be used to store data used by processor 302 when performing operations.

[0087] Embodiment IV

[0088] Based on the same inventive concept, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the above-mentioned rapid diagnosis method for acute infection are implemented.

[0089] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general-purpose systems may also be used in conjunction with the teachings based hereon. The structure required to construct such systems will be apparent from the above description. In addition, the present invention is not directed to any particular programming language. It should be appreciated that the present invention as described herein may be implemented in various programming languages, and the description of specific languages above is for the purpose of disclosing the best mode of the present invention.

[0090] In the specification provided herein, numerous specific details are set forth. However, it can be understood that embodiments of the present invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0091] Similarly, it should be understood that in order to streamline the present invention and assist in understanding one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present invention, the various features of the present invention are sometimes grouped together in a single embodiment, figure, or description thereof. However, the disclosed method should not be construed as reflecting an intention that the claimed invention requires more features than are expressly recited in each embodiment. Rather, as reflected in each embodiment, the inventive aspects lie in less than all of the features of the previously disclosed single embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into the detailed description, where each claim itself is considered a separate embodiment of the present invention.

[0092] Those skilled in the art will appreciate that the modules in the devices in the embodiments can be adaptively changed and disposed in one or more devices different from the embodiments. The modules or units or components in the embodiments can be combined into one module or unit or component, and in addition, they can be divided into multiple sub-modules or sub-units or sub-components. Except that at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all the features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all the processes or units of any method or device so disclosed. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) can be replaced by an alternative feature that provides the same, equivalent, or similar purpose.

[0093] In addition, those skilled in the art can understand that, although some embodiments herein include certain features included in other embodiments rather than other features, the combination of features of different embodiments means that it is within the scope of the present invention and forms different embodiments. For example, in the detailed description, any one of the claimed embodiments can be used in any combination.

[0094] Each component embodiment of the present invention can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor (DSP) can be used in practice to implement some or all of the functions of some or all of the components of the rapid diagnosis device for acute infections and computer equipment according to the embodiments of the present invention. The present invention can also be implemented as a device or device program (for example, a computer program and a computer program product) for executing part or all of the methods described herein. Such a program for implementing the present invention can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0095] It should be noted that the above embodiments illustrate the present invention rather than limit the present invention, and those skilled in the art can design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in the claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In the unit claims listing several devices, several of these devices can be embodied by the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words can be interpreted as names.

Claims

1. A rapid diagnosis method for acute infections, characterized in that, Including: Collecting the electronic medical records, blood routine, inflammatory indicators, microbial test results, and imaging data of the target patient; Based on the electronic medical records, blood routine, inflammatory indicators, microbial test results, and imaging data, respectively extracting the corresponding target basic features; Based on the target basic features, convolutional neural network, and long short-term memory network, extracting the target deep features; Obtaining an acute infection prediction model, which is obtained through model training; Based on the target deep features and the acute infection prediction model, determining the acute infection type of the target patient.

2. The method according to claim 1, wherein Based on the electronic medical records, blood routine, inflammatory indicators, microbial test results, and imaging data, respectively extracting the corresponding target basic features, including: Based on the electronic medical records, extracting the medical history features, disease characteristics, and diagnosis and treatment process features; Based on the blood routine and inflammatory indicators, extracting the quantitative index features of the body's immune response; Based on the microbial test results, extracting the features of the presence or absence of pathogens, pathogen type features, and drug resistance features; Based on the imaging data, extracting the infection site features, macroscopic pathological change features, and microscopic pathological change features.

3. The method according to claim 1, characterized in that, Based on the target basic features, convolutional neural network, and long short-term memory network, extracting the target deep features, including: Using a convolutional neural network to extract local features from the target basic features; Using a long short-term memory network to capture long-distance dependence features in the target basic features; Based on the local features and the long-distance dependence features, obtaining the target deep features.

4. The method according to claim 2, wherein Based on the target basic features, convolutional neural network, and long short-term memory network, extracting the target deep features, including: Based on the medical history features, disease characteristics, and diagnosis and treatment process features, through a convolutional neural network and a long short-term memory network, extracting the correlation between symptom features and the logical relationship between disease course progress; Based on the quantitative index features of the body's immune response, through a convolutional neural network and a long short-term memory network, extracting the correlation between the quantitative indexes of each immune response; Based on the features of the presence or absence of pathogens, pathogen type features, and drug resistance features, through a convolutional neural network and a long short-term memory network, extracting the etiological diagnosis index features; Based on the infection site features, macroscopic pathological change features, and microscopic pathological change features, through a convolutional neural network and a long short-term memory network, locating the infection focus and differentiating the nature of the infection.

5. The method according to claim 1, characterized in that, Obtaining an acute infection prediction model, including: Collecting the historical electronic medical records, historical blood routine, historical inflammatory indicators, historical microbial test results, and historical imaging data of historical patients; Based on the historical electronic medical records, historical blood routine, historical inflammatory indicators, historical microbial test results, and historical imaging data, respectively extracting the corresponding historical basic features; Based on the historical basic features, convolutional neural network, and long short-term memory network, extracting the historical deep features; Based on the historical deep features, training the prediction model to obtain an acute infection prediction model, which is used to predict the infection type, predict the types of pathogen subtypes, and the drug resistance judgment results.

6. The method according to claim 4, wherein Based on the target depth features and the acute infection prediction model, determine the acute infection type of the target patient, including: Input the target depth features into the acute infection prediction model to output the preliminary pathogen candidate range of the target patient; Extract the pathogen feature sequence from the etiological diagnosis index features; Use a basic local alignment-like algorithm to align the pathogen feature sequence with the complete gene sequences of known pathogens in the pathogen gene database to determine the similarity between the pathogen feature sequence and the complete gene sequences of known pathogens; Based on the similarity, determine the acute infection type of the target patient.

7. The method according to claim 6, wherein Based on the similarity, determine the acute infection type of the target patient, including: Based on the similarity, the quantitative index features of each immune response, the infection focus and the nature of the infection, judge whether there are contradictions to obtain a judgment result; Based on the judgment result, determine the accuracy of the acute infection type of the target patient.

8. The method according to claim 2, wherein After determining the acute infection type of the target patient based on the target depth features and the acute infection prediction model, it further includes: Based on the drug resistance characteristics of the target patient, provide guidance for the use of antibiotics for the target patient.

9. A rapid diagnosis device for acute infections, characterized in that, Including: A collection module for collecting the electronic medical record, blood routine, inflammatory index, microbial test results and imaging data of the target patient; A first extraction module for respectively extracting the corresponding target basic features based on the electronic medical record, blood routine, inflammatory index, microbial test results and imaging data; A second extraction module for extracting target depth features based on the target basic features, convolutional neural network and long short-term memory network; An acquisition module for acquiring an acute infection prediction model, which is obtained through model training; A determination module for determining the acute infection type of the target patient based on the target depth features and the acute infection prediction model.

10. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method described in any one of claims 1 to 8.

Images