CN117405896A - Auxiliary diagnosis method, device, equipment, system and medium - Google Patents
Auxiliary diagnosis method, device, equipment, system and medium Download PDFInfo
- Publication number
- CN117405896A CN117405896A CN202210806511.7A CN202210806511A CN117405896A CN 117405896 A CN117405896 A CN 117405896A CN 202210806511 A CN202210806511 A CN 202210806511A CN 117405896 A CN117405896 A CN 117405896A
- Authority
- CN
- China
- Prior art keywords
- specific protein
- protein detection
- sample
- detected
- detection
- 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.)
- Pending
Links
- 238000003745 diagnosis Methods 0.000 title claims abstract description 101
- 238000000034 method Methods 0.000 title claims abstract description 19
- 238000002331 protein detection Methods 0.000 claims abstract description 258
- 238000001514 detection method Methods 0.000 claims abstract description 192
- 238000004458 analytical method Methods 0.000 claims abstract description 15
- 238000000605 extraction Methods 0.000 claims description 72
- 230000004927 fusion Effects 0.000 claims description 30
- 238000012549 training Methods 0.000 claims description 29
- 238000002405 diagnostic procedure Methods 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 13
- 230000000875 corresponding effect Effects 0.000 description 108
- 102000004169 proteins and genes Human genes 0.000 description 25
- 108090000623 proteins and genes Proteins 0.000 description 25
- 108010074051 C-Reactive Protein Proteins 0.000 description 18
- 102100032752 C-reactive protein Human genes 0.000 description 18
- 238000006243 chemical reaction Methods 0.000 description 14
- 210000004369 blood Anatomy 0.000 description 13
- 239000008280 blood Substances 0.000 description 13
- 230000002159 abnormal effect Effects 0.000 description 12
- 238000010586 diagram Methods 0.000 description 11
- 108700028909 Serum Amyloid A Proteins 0.000 description 10
- 102000054727 Serum Amyloid A Human genes 0.000 description 10
- 238000012360 testing method Methods 0.000 description 10
- 238000005065 mining Methods 0.000 description 9
- 239000004816 latex Substances 0.000 description 6
- 229920000126 latex Polymers 0.000 description 6
- 230000004044 response Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000004848 nephelometry Methods 0.000 description 5
- 238000005070 sampling Methods 0.000 description 5
- 230000000877 morphologic effect Effects 0.000 description 4
- 239000002245 particle Substances 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 206010020751 Hypersensitivity Diseases 0.000 description 2
- 239000012491 analyte Substances 0.000 description 2
- 239000000427 antigen Substances 0.000 description 2
- 102000036639 antigens Human genes 0.000 description 2
- 108091007433 antigens Proteins 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 210000001124 body fluid Anatomy 0.000 description 2
- 239000010839 body fluid Substances 0.000 description 2
- 239000003153 chemical reaction reagent Substances 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 208000015181 infectious disease Diseases 0.000 description 2
- 230000002757 inflammatory effect Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 239000004005 microsphere Substances 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 210000002966 serum Anatomy 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000004879 turbidimetry Methods 0.000 description 2
- 230000009385 viral infection Effects 0.000 description 2
- 208000035143 Bacterial infection Diseases 0.000 description 1
- 241001465754 Metazoa Species 0.000 description 1
- 206010029719 Nonspecific reaction Diseases 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000001154 acute effect Effects 0.000 description 1
- 238000000149 argon plasma sintering Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 208000022362 bacterial infectious disease Diseases 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 210000000601 blood cell Anatomy 0.000 description 1
- 230000023555 blood coagulation Effects 0.000 description 1
- 238000011088 calibration curve Methods 0.000 description 1
- 210000001175 cerebrospinal fluid Anatomy 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000013399 early diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 208000019622 heart disease Diseases 0.000 description 1
- 239000003219 hemolytic agent Substances 0.000 description 1
- 238000003018 immunoassay Methods 0.000 description 1
- 208000027866 inflammatory disease Diseases 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 210000004185 liver Anatomy 0.000 description 1
- 210000005229 liver cell Anatomy 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000004393 prognosis Methods 0.000 description 1
- 238000002731 protein assay Methods 0.000 description 1
- 230000009257 reactivity Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 210000001179 synovial fluid Anatomy 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 210000002700 urine Anatomy 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
- G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
- G01N33/68—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving proteins, peptides or amino acids
- G01N33/6893—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving proteins, peptides or amino acids related to diseases not provided for elsewhere
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/47—Scattering, i.e. diffuse reflection
- G01N21/49—Scattering, i.e. diffuse reflection within a body or fluid
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
- G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
- G01N33/53—Immunoassay; Biospecific binding assay; Materials therefor
- G01N33/543—Immunoassay; Biospecific binding assay; Materials therefor with an insoluble carrier for immobilising immunochemicals
- G01N33/54313—Immunoassay; Biospecific binding assay; Materials therefor with an insoluble carrier for immobilising immunochemicals the carrier being characterised by its particulate form
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
- G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
- G01N33/53—Immunoassay; Biospecific binding assay; Materials therefor
- G01N33/543—Immunoassay; Biospecific binding assay; Materials therefor with an insoluble carrier for immobilising immunochemicals
- G01N33/544—Immunoassay; Biospecific binding assay; Materials therefor with an insoluble carrier for immobilising immunochemicals the carrier being organic
-
- 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
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/40—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for data related to laboratory analysis, e.g. patient specimen analysis
-
- 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2333/00—Assays involving biological materials from specific organisms or of a specific nature
- G01N2333/435—Assays involving biological materials from specific organisms or of a specific nature from animals; from humans
- G01N2333/46—Assays involving biological materials from specific organisms or of a specific nature from animals; from humans from vertebrates
- G01N2333/47—Assays involving proteins of known structure or function as defined in the subgroups
- G01N2333/4701—Details
- G01N2333/4737—C-reactive protein
-
- 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
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Immunology (AREA)
- Biomedical Technology (AREA)
- Chemical & Material Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Hematology (AREA)
- Molecular Biology (AREA)
- Urology & Nephrology (AREA)
- Pathology (AREA)
- Physics & Mathematics (AREA)
- Biochemistry (AREA)
- General Physics & Mathematics (AREA)
- Analytical Chemistry (AREA)
- Medicinal Chemistry (AREA)
- Food Science & Technology (AREA)
- Microbiology (AREA)
- Cell Biology (AREA)
- Biotechnology (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Proteomics, Peptides & Aminoacids (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The embodiment of the application provides an auxiliary diagnosis method, device and equipment, a sample analysis system and medium, wherein the method comprises the steps of acquiring clinical information data corresponding to a sample to be detected; acquiring curve detection data of at least one specific protein detection item of the sample to be detected; calculating a specific protein detection value of the sample to be detected, which corresponds to the specific protein detection item, according to the curve detection data of the specific protein detection item; and combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item and the specific protein detection value through an auxiliary diagnosis model to obtain a first prediction category of the sample to be detected and associated characteristic information of the sample to be detected.
Description
Technical Field
The present disclosure relates to the field of medical assistance, and more particularly, to an assistance diagnosis method and apparatus, an assistance diagnosis device, a sample analysis system, and a computer-readable storage medium.
Background
In traditional immunoassays, specific protein assays are one of the important indicators for the evaluation of inflammatory diseases. For example, C-reactive protein (CRP) is an acute inflammatory timing phase response protein synthesized by the liver. The CRP concentration in normal human blood is very low, but can be rapidly increased in synthesis when the organism is suddenly stressed, tissue wound and various inflammatory stimuli, and can be secreted into the blood from liver cells, and high-level CRP can be detected 12-18 hours after infection occurs. CRP, which increases 12-14 days after infection, can drop to baseline levels, and thus, CRP can be one of the important markers for diagnosing bacterial infection, and is also a clinically important index for assessing heart disease incidence, recurrence rate, and mortality rate.
Common detection methods for specific proteins are various and mainly comprise an immunoturbidimetry method, a latex enhancement technology and the like. When detecting the specific protein concentration of a sample, an automatic detection device is generally adopted, a specific protein reaction curve of the sample is obtained first, and a curve characteristic is determined from the specific protein reaction curve and substituted into a calibration curve to obtain the specific protein concentration of the sample. However, when signal interference occurs in a certain period of time in the sample detection process, if the point of selecting the calculated curve characteristic is just located in the interference period of time, the calculation result is inaccurate, and a false result is caused.
Disclosure of Invention
In order to solve the technical problems, the application provides an auxiliary diagnosis method and device, auxiliary diagnosis equipment, a sample analysis system and a computer readable storage medium, wherein the efficiency can be improved, the dependence on personal experience is reduced, and the result accuracy can be improved.
In a first aspect of embodiments of the present application, there is provided an auxiliary diagnostic method, including:
a method of aiding diagnosis, comprising:
acquiring clinical information data corresponding to a sample to be detected;
acquiring curve detection data of at least one specific protein detection item of the sample to be detected;
Calculating a specific protein detection value of the sample to be detected, which corresponds to the specific protein detection item, according to the curve detection data of the specific protein detection item;
and combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item and the specific protein detection value through an auxiliary diagnosis model to obtain a first prediction category of the sample to be detected and associated characteristic information of the sample to be detected.
In a second aspect, there is provided an auxiliary diagnostic apparatus comprising:
the first acquisition module is used for acquiring clinical information data corresponding to a sample to be detected;
the second acquisition module is used for acquiring curve detection data of at least one specific protein detection item of the sample to be detected;
the detection module is used for calculating a specific protein detection value of the sample to be detected, which corresponds to the specific protein detection item, according to the curve detection data of the specific protein detection item;
the diagnosis module is used for obtaining a first prediction category of the sample to be detected and associated characteristic information thereof by combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item and the specific protein detection value through an auxiliary diagnosis model.
In a third aspect, there is further provided an auxiliary diagnostic apparatus, including a processor and a memory, where the memory stores a computer program executable by the processor, and when the computer program is executed by the processor, the auxiliary diagnostic method according to any embodiment of the present application is implemented.
In a fourth aspect, there is also provided a sample analysis system comprising an auxiliary diagnostic device as described in any of the embodiments of the present application and a specific protein detection analyzer connected to the auxiliary diagnostic device.
In a fifth aspect, there is also provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the auxiliary diagnostic method according to any of the embodiments of the present application.
In the auxiliary diagnosis method provided by the embodiment of the invention, the corresponding clinical information data of the sample to be detected and the curve detection data of at least one specific protein detection item are obtained, the corresponding specific protein detection value is calculated according to the curve detection data, the clinical information data, the specific protein detection value and the curve detection data corresponding to the sample to be detected are taken as input by the auxiliary diagnosis model, the auxiliary diagnosis model is used for analysis and prediction, the first prediction category of the sample to be detected is obtained, the associated characteristic information determined to be the first prediction category is correspondingly output, wherein the first prediction category can be the potential risk category determined to exist based on the specific protein detection result, the associated characteristic information can comprise the curve detection data protruding to display an abnormal region, so that the reliability of the auxiliary diagnosis result obtained at present can be conveniently and rapidly judged according to the associated characteristic information, the auxiliary diagnosis model can carry out more comprehensive characteristic extraction and identification on the curve detection data, the intelligent learning is used for obtaining the more comprehensive characteristic of the curve detection data, and the traditional inadequately providing the accurate diagnosis result based on the specific calculation of the characteristic calculation point of the characteristic detection data is avoided, and the inadequately providing the accurate diagnosis result is provided.
The auxiliary diagnostic apparatus, the device, the sample analysis system and the computer readable storage medium provided in the above embodiments have the same technical ideas as that of the corresponding auxiliary diagnostic method, so that they have the same beneficial technical effects, and are not described in detail herein.
Drawings
FIG. 1 is an alternative application scenario diagram of an auxiliary diagnostic method provided in one embodiment;
FIG. 2 is a schematic diagram of curve detection data obtained by using an immune nephelometry method in an embodiment;
FIG. 3 is a flow chart of an auxiliary diagnostic method in one embodiment;
FIG. 4 is a schematic diagram of detection curve data in an alternative embodiment;
FIG. 5 is a schematic diagram of an auxiliary diagnostic model in one embodiment;
FIG. 6 is a schematic diagram illustrating the operation of an auxiliary diagnostic model according to an embodiment;
FIG. 7 is a schematic diagram of an auxiliary diagnostic device according to an embodiment;
fig. 8 is a schematic structural diagram of an auxiliary diagnostic device in an embodiment.
Detailed Description
The technical scheme of the application is further elaborated below by referring to the drawings in the specification and the specific embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the description of the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the implementations of the present application. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.
In the following description, reference is made to the expression "some embodiments" which describe a subset of the possible embodiments, but it should be understood that "some embodiments" may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.
Referring to fig. 1, a schematic diagram of an optional application scenario of an auxiliary diagnostic method according to an embodiment of the present application is shown. The auxiliary diagnostic system includes a test information system (Laboratory Information System, LIS) 10, a specific protein detection analyzer 20, and an auxiliary diagnostic device 30. The test information system 10 is typically located at a clinical laboratory of a hospital and is operable to receive test data, enter stored patient test information, and assist the hospital in information management. The specific protein detection analyzer 20 is a device for performing intelligent detection analysis on a detection sample of a detection object to obtain curve detection data of a specific protein of the sample to be detected. Referring to fig. 2, a schematic diagram of a specific protein (CRP/SAA) detection analyzer for specific protein detection is provided as an alternative specific example, and the specific protein detection analyzer mainly adopts an immunoscattering nephelometry and a latex enhancement technique. The principle of the immune scattering turbidimetry is that an antigen-antibody reaction solution is irradiated by incident light with a certain wavelength along the horizontal direction, the refraction and diffraction of the incident light by antigen-antibody complex particles in the solution form scattered light, and the scattered light intensity is positively related to the amount of the antigen-antibody complex. The latex enhancement technique is to attach antibodies to tiny latex particles, which when antigen-antibody are combined to form antigen-antibody-latex particle complexes, increase the diameter of immune complexes, increase detection sensitivity, and reduce the influence of other non-specific reactions. The whole blood sample containing the antigen (antibody) to be detected is added into a hemolytic agent by the detection analyzer, the influence of blood cells on light scattering is removed, then a reagent is added, the specific combination of the antigen-antibody-microsphere immune complex is formed by the analyte in the sample and the latex microsphere coated with the antibody in the reagent, the turbidity of the reaction solution is caused, and along with the increase of the concentration of the analyte, the turbidity of the reaction is also increased, namely, the amount of the antigen to be detected is positively correlated with the turbidity of the reaction solution. The detection analyzer emits light with a certain wavelength to irradiate the solution along the horizontal direction, the light is refracted by the compound particles to form scattered light, the receiving end collects the light intensity of the scattered light in real time and converts the scattered light into an electric signal, the change (i.e. the reactivity) of the scattered light intensity is calculated, the change is in direct proportion to the content of the to-be-detected object, and then the concentration of the to-be-detected substance in the sample can be calculated, as shown in a graph of fig. 2, a specific protein reaction curve is shown, the horizontal axis is time/point, and the vertical axis is the scattered light intensity.
The sample to be detected may refer to a blood sample, such as whole blood, serum, etc. The auxiliary diagnostic device 30 is communicatively connected to the test information system 10 and the specific protein detection analyzer 20, and is configured to obtain clinical information data corresponding to a sample to be detected from the test information system 10 and curve detection data corresponding to a specific protein detection item from the specific protein detection analyzer 20, and perform more comprehensive and deep mining analysis and recognition on the curve detection data of the specific protein detection item of the sample to be detected by combining the clinical information data of the object to be detected and the curve detection data corresponding to the specific protein detection item with an auxiliary diagnostic model constructed based on a historical diagnosis example and an expert consensus library, so as to obtain auxiliary diagnostic information capable of corresponding prediction results of a plurality of preset prediction categories. Wherein the prediction categories may include all categories in which whether there is a problem/potential risk may be determined based on a detection result of a specific protein detection item, the first prediction category may be determining whether a sample to be detected is one of a plurality of target prediction categories, for example, the target prediction category may include n categories, and the first prediction category is determining that the sample to be detected is one of the n categories; the first prediction category may also include prediction results corresponding to a plurality of target prediction categories, for example, the target prediction category includes n categories, the first prediction category is an n-dimensional vector, and the corresponding prediction result of the currently determined category is represented by a value on each bit in the n-dimensional vector.
Wherein, the auxiliary diagnostic device 30 may be various intelligent devices with storage and calculation capabilities, which can be physically separated from the inspection information system or the specific protein detection analyzer, such as a smart phone, a personal computer, etc. loaded with a computer program for implementing the auxiliary diagnostic method of the embodiment of the present application; it is also possible that the test information system or the specific protein detection analyzer is integrated as one body, such as the test information system or the specific protein detection analyzer loaded with a computer program for realizing the auxiliary diagnostic method of the embodiment of the present application.
Referring to fig. 3, an auxiliary diagnostic method provided in an embodiment of the present application may be applied to the auxiliary diagnostic apparatus shown in fig. 1, where the auxiliary diagnostic method includes:
s101: and acquiring clinical information data corresponding to the sample to be detected.
The sample to be tested may be a blood sample of a human body, and the person to be tested is the person providing the blood sample. The sample to be detected may be other body fluid sample such as urine, or blood sample of an animal. The age, sex, type of test sample, past medical history, etc. of the person to whom the different samples to be tested belong are different, and the biological information features contained in the blood sample are different.
S103, acquiring curve detection data of at least one specific protein detection item of the sample to be detected.
The sample to be detected is a blood sample, and can be whole blood or serum. Specific protein detection items may include: CRP (C-reactive protein)/Hs-CRP (hypersensitive C-reactive protein), SAA (serum amyloid A). The obtaining of the curve detection data of at least one specific protein detection item of the sample to be detected may be obtaining one or more of the curve detection data of the CRP detection item, the curve detection data of the Hs-CRP detection item, and the curve detection data of the SAA detection item.
Alternatively, the specific protein detection items may be as follows: CRP, SAA, CRP +SAA. The detection curve data corresponds to the specific protein detection items respectively, and optionally, the detection curve data can be obtained from detection results corresponding to the specific protein detection items after the specific protein detection items are executed by each specific protein detection analyzer respectively.
In some embodiments, the detection of the specific protein detection item is performed on the sample to be detected by a specific protein detection analyzer, wherein the specific protein detection analyzer may comprise a plurality of detection modes, and different detection modes may respectively comprise different numbers and types of specific protein detection items. For one sample to be detected, a specific protein detection analyzer can be set in a detection mode containing required specific protein detection items, and the sample to be detected is detected by the specific protein detection analyzer to obtain detection results corresponding to the specific protein detection items. Optionally, the specific protein detection analyzer may also support configuration of a detection mode, and the user selects a specific protein detection item to be executed for a current sample to be detected from a plurality of specific protein detection items that the specific protein detection analyzer can support to detect, configures the specific protein detection analyzer to form a detection mode of the specific protein detection analyzer, and detects the sample to be detected in the configured detection mode by the specific protein detection analyzer to obtain a detection result corresponding to each specific protein detection item.
The obtaining the detection curve data corresponding to at least one specific protein detection item of the sample to be detected may refer to that the auxiliary detection device obtains the detection curve data corresponding to a plurality of specific protein detection items from the detection result output after the specific protein detection analyzer performs the detection of the specific protein detection item on the sample to be detected.
S105: and calculating a specific protein detection value of the sample to be detected, which corresponds to the specific protein detection item, according to the curve detection data of the specific protein detection item.
Calculating a specific protein detection value corresponding to a specific protein detection item according to curve detection data of the specific protein detection item may refer to selecting a preset curve characteristic point from the curve detection data, and calculating the specific protein detection value corresponding to the specific protein detection item of the sample to be detected through the value of the selected curve characteristic point. In this embodiment, the specific protein detection value refers to a specific protein concentration value corresponding to the specific protein detection item.
The calculating the specific protein detection value of the sample to be detected, which corresponds to the specific protein detection item, may be determined by a conventional calculation method, for example, taking the curve detection data as a specific protein reaction curve of the sample to be detected, which is obtained by performing specific protein detection on the sample to be detected based on a nephelometry, for example, the specific protein reaction curve is used to represent a change of a voltage obtained by the nephelometry in a total sampling time, and the calculating the specific protein detection value may include the following steps: obtaining the minimum value of the first derivative of the specific protein response curve; determining whether a specific protein response curve of a sample to be detected has fluctuation according to the relation between the minimum value of the first derivative and a preset threshold value, if the specific protein response curve does not have fluctuation, selecting an extraction time period within the sampling total time, if the specific protein response curve of the sample to be detected has fluctuation, acquiring a second derivative of the specific protein response curve of the sample to be detected, determining sampling effective time within the sampling total time according to the second derivative, and selecting the extraction time period within the sampling effective time; extracting curve characteristics corresponding to a specific protein reaction curve of the sample to be detected in the extraction time period; extracting curve characteristics corresponding to specific protein reaction curves of the known concentration samples in the extraction time period according to specific protein reaction curves of a plurality of known concentration samples obtained in advance based on a scattering turbidimetry method; establishing a scaling function relation between curve characteristics of specific protein reaction curves of all known concentration samples and the corresponding specific protein concentrations; and inputting the curve characteristics of the sample to be tested into the scaling function relation to obtain the specific protein concentration of the sample to be tested.
S107, combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item and the specific protein detection value through an auxiliary diagnosis model to obtain a first prediction category of the sample to be detected and associated characteristic information of the first prediction category.
The auxiliary diagnosis model takes clinical information data corresponding to a sample to be detected, curve detection data of a specific protein detection item and a specific protein detection value as inputs, performs feature extraction on the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item and the specific protein detection value, obtains a feature vector set corresponding to the sample to be detected, performs classification prediction based on the feature vector set to determine a first prediction category of the sample to be detected, and determines associated feature information representing the first prediction category of the sample to be detected from the feature vector set. The auxiliary diagnosis model can be a deep learning model obtained by training sample data formed by a large number of historical diagnosis examples, expert knowledge and the like, and the deep learning model intelligently learns rules for extracting features from input formed by the training sample data and classifying and predicting based on the extracted features through training of the training sample data, so that the auxiliary diagnosis model takes clinical information data corresponding to a sample to be detected, curve detection data of a specific protein detection item and a specific protein detection value as input, extracts the features of the input, predicts and determines a first prediction category of the sample to be detected and associated feature information of the sample to be detected based on the intelligently learned rules; optionally, the auxiliary diagnostic model may also be a knowledge graph database formed based on historical diagnosis examples, expert knowledge, and the like, and the knowledge graph database is constructed by taking clinical information data corresponding to a detection sample, curve detection data of a specific protein detection item, and a specific protein detection value as indexes, and mapping relations between prediction categories and associated characteristic information index values, so that the auxiliary diagnostic model predicts and determines the first prediction category and associated characteristic information of the sample to be detected based on the mapping relations by taking the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item, and the specific protein detection value as inputs.
In the above embodiment, the clinical information data corresponding to the sample to be detected and the curve detection data of at least one specific protein detection item are obtained, the corresponding specific protein detection value is calculated according to the curve detection data, the clinical information data, the specific protein detection value and the curve detection data corresponding to the sample to be detected are taken as input by an auxiliary diagnosis model, the first prediction category of the sample to be detected is obtained through analysis and prediction by the auxiliary diagnosis model, and associated characteristic information determined as the first prediction category is correspondingly output, wherein the first prediction category can be a potential risk category determined to exist based on the specific protein detection result, the associated characteristic information can include curve detection data protruding to display an abnormal region, so that the reliability of the currently obtained prediction category as an auxiliary diagnosis result is conveniently and rapidly judged according to the associated characteristic information of the first prediction category, the auxiliary diagnosis model can carry out more comprehensive characteristic extraction and identification on the curve detection data, the intelligent learning can be used for excavating more curve characteristics, and accordingly, the situation that the conventional risk is not accurately supported by calculating specific protein detection points based on the specific characteristic selected from the detection data can be avoided, and the risk is not accurately provided.
In some embodiments, the auxiliary diagnostic method further comprises:
outputting a secondary diagnostic report comprising the first predictive category and associated characteristic information including an anomaly identification result for the curve detection data for the particular protein detection item.
Wherein the auxiliary diagnostic device may generate an auxiliary diagnostic report based on the first prediction category and its associated feature information based on a pre-customized report template. The first prediction category may be one of a plurality of preset target prediction categories, for example, the target prediction category includes a prediction category a, a prediction category B, and a prediction category C, for the prediction category a, the corresponding prediction result is 1, the risk represented as the target prediction category a is greater than a threshold, for the prediction category B, C, the corresponding prediction result is 0, the risk represented as the target prediction category B, C is less than the threshold, the associated feature information for determining that the first prediction category is the prediction category a and the corresponding prediction result of the prediction category a is 1 in the auxiliary diagnostic report includes: clinical information x1 and specific protein detection value characteristics a1 are contained in clinical information data corresponding to a sample to be detected, abnormal characteristics a2 are contained in detection curve data of the sample to be detected, and the bulge display of the area containing the abnormal characteristics a2 is contained in the detection curve data. The abnormal recognition result of the curve detection data correspondingly refers to a convex display result of a region containing the abnormal feature a2 in the detection curve data, the depth mining and the visual display of the abnormal feature a2 contained in the detection curve data are carried out through an auxiliary diagnosis model, the accuracy of the auxiliary diagnosis result is greatly improved, the risk type judged based on the specific protein detection result can be clearly and directly obtained through reading the auxiliary diagnosis report, the basis with corresponding risk is determined, the reliability of the rapid judgment result according to the basis is facilitated, or the adjustment according to the basis is facilitated, and other practical conditions are combined.
In an alternative specific example, referring to fig. 4, a schematic diagram of detection curve data of specific protein detection obtained by an immunothermal nephelometry method is shown, the specific protein detection value feature a1 may be obtained by performing first-order derivation on a selected feature point in a curve, and the detection curve data of a sample to be detected may include background section curve features, light intensity variation features, reaction section curve area features and the like. The number of target prediction categories to be predicted by the auxiliary diagnosis model is n, the output of the auxiliary diagnosis model is a prediction category determined based on an n-dimensional vector, the n-dimensional vector corresponds to the prediction results corresponding to the n target prediction categories one by one, the auxiliary diagnosis model carries out classification prediction to obtain an n-dimensional vector with the corresponding value of each target prediction category, and the vector carries out normalization processing (the corresponding value of each target prediction category in the vector, namely, the sum of the numerical values of each bit in the vector is 1) through an output layer (softmax) of the auxiliary diagnosis model for prediction classification. A threshold may be set in the output layer, and if the probability is greater than the threshold, the corresponding target prediction category is determined to be at risk. And if the corresponding bit in the vector is greater than the threshold value, the corresponding prediction result of the target prediction category is positive, the corresponding position 1 of the vector is output, and if the corresponding bit in the vector is greater than the threshold value, the corresponding prediction result of the target prediction category is negative, and the corresponding position 0 of the vector is output. Taking n as 5 and a threshold value as 0.45, taking the target prediction category to be predicted as an example including category 1, category 2, category 3, category 4 and category 5, after a certain sample to be detected is normalized by an output layer of the prediction classification of the auxiliary diagnosis model, the output vector is 0.4, 0.0, 0.5, 0.0 and 0.1, wherein the numerical value in the 3 rd bit is greater than 0.45, so that the auxiliary diagnosis model obtains a prediction result of each target prediction category as 0 0 1 0 0, and sequentially and correspondingly judging the first prediction category as category 3.
For example, according to the CRP and SAA detection results, a plurality of target prediction categories are set as shown in the following table one:
the auxiliary diagnosis information obtained by the auxiliary diagnosis model is beneficial to early diagnosis and prognosis judgment of light virus infection, severe virus infection and the like according to the detection results of specific protein detection items of the sample to be detected.
In the above embodiment, the auxiliary diagnostic device outputs the auxiliary diagnostic information and the auxiliary diagnostic report representing the diagnostic basis of the auxiliary diagnostic information, where the diagnostic basis includes the associated feature information corresponding to the prediction category, so that the readability of the auxiliary diagnostic result can be improved, and the user can conveniently and rapidly identify the reliability of the current auxiliary diagnostic result through the feature information associated with the prediction category displayed in the auxiliary diagnostic report, or can conveniently adjust according to the basis in combination with other actual situations.
In some embodiments, the auxiliary diagnostic method further comprises:
determining a second prediction category of the sample to be detected based on a preset category prediction rule according to the specific protein detection value and clinical information data of the object to be detected;
outputting an auxiliary diagnostic report comprising the first predictive category and associated characteristic information thereof, including an anomaly identification result of the curve detection data for the particular protein detection item, and the second predictive category.
The auxiliary diagnosis device can predict the category of the sample to be detected according to the specific protein detection value, and obtain a second prediction category which is relatively independent to the result of the first prediction category output by the auxiliary diagnosis model. The category prediction rule of the category is determined based on the specific protein detection value, and the corresponding rule refers to the rule that the specific protein detection value is compared with the corresponding standard value and the category is determined according to the comparison result. The first prediction category and the second prediction category may be the same or different, if the first prediction category and the second prediction category are the same, the probability that the category of the current sample to be detected belongs to the current prediction category is high can be indicated, and the prediction category result output by the auxiliary diagnosis model can be used for verifying the prediction result obtained by the specific protein detection value; if the first prediction category and the second prediction category are different, the associated feature information of the first prediction category in the auxiliary diagnosis report can be used for quickly identifying the determination basis of the first prediction category, for example, the associated feature information comprises an abnormal recognition result of curve detection data of a specific protein detection item, an abnormal region contained in the curve detection data is determined by mining feature information of a deeper layer in the curve detection data, and the abnormal region is projected and displayed in the curve detection data to serve as a key attention point so as to judge the reliability of the first prediction category, so that the comparison of the first prediction category and the second prediction category is realized, and the comparison result is combined with the actual situation to adjust.
In the above embodiment, the auxiliary diagnosis method includes determining whether the sample to be detected has the second prediction category of which potential risks according to the specific protein detection value conventionally, and determining whether the sample to be detected has the first prediction category of which potential risks according to the auxiliary diagnosis model mining the technical features of the deeper layers in the curve detection data of the specific protein detection item, so that the auxiliary diagnosis model is used for realizing the feature mining of the deeper layers in the curve detection data of the specific protein detection item, thereby obtaining more complete and comprehensive feature information to obtain the specific protein detection result, the auxiliary diagnosis model can verify the prediction category obtained based on the specific protein detection value, and the first prediction category and the second prediction category can be compared to improve the accuracy and the reference value of the auxiliary diagnosis result.
In some embodiments, the specific protein detection items comprise a first specific protein detection item and a second specific protein detection item; the obtaining, by an auxiliary diagnostic model, the first prediction category of the sample to be detected and associated feature information thereof by combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item, and the specific protein detection value, includes:
The auxiliary diagnosis model takes curve detection data corresponding to the first specific protein detection item and the second specific protein detection item as two paths of input, takes clinical information data corresponding to the sample to be detected as one path of input and takes the specific protein detection value as one path of input, and performs feature extraction in parallel;
combining the extracted features according to the inputs of each path to obtain a fused feature set;
and carrying out classification prediction according to the fusion characteristic set to obtain a first prediction category of the sample to be detected and associated characteristic information thereof.
The auxiliary diagnosis model takes curve detection data corresponding to the first specific protein detection item and the second specific protein detection item as input to extract characteristics, including deeper characteristics which cannot be obtained from the curve form of the corresponding detection curve data. The detection curve data respectively correspond to the specific protein detection items, the curve forms of the detection curve data correspondingly obtained are often different for different specific protein detection items, and for each specific protein detection item, different types and different numbers of curve features for finally determining the prediction type of the sample to be detected can be extracted from the detection curve data corresponding to each specific protein detection item. Referring to fig. 5, the auxiliary diagnostic model may include feature extraction networks respectively corresponding to a plurality of inputs, a feature fusion layer connected to the feature extraction networks, and a classification prediction layer connected to the feature fusion layer. In an alternative specific example, the feature extraction network includes a first feature extraction network that takes the first specific protein detection item as an input, and a second feature extraction network that takes curve detection data corresponding to the second specific protein detection item as an input. The feature fusion layer comprises two paths of inputs respectively corresponding to the first feature extraction network and the second feature extraction network, and two paths of inputs corresponding to clinical information data corresponding to a sample to be detected and a specific protein detection value determined based on curve morphological feature calculation. The feature fusion layer is used for carrying out feature fusion on the features extracted by the first feature extraction network and the second feature extraction network and the clinical information data and the specific protein detection value to obtain a fusion feature set. The classification prediction layer determines a first prediction category of the sample to be detected and associated feature information thereof based on an intelligently learned rule for determining the prediction category according to the fusion feature set output by the feature fusion layer, wherein the associated feature information comprises curve detection data for performing projection display on a region containing abnormal features.
The first feature extraction network and the second feature extraction network may be algorithm models for extracting features included in the image based on an image recognition algorithm, for example, the first feature extraction network and the second feature extraction network may respectively use classification models for recognizing the image, and the first feature extraction network and the second feature extraction network extract corresponding detection curve data of the first specific protein detection item and sample features included in the detection curve data of the second specific protein detection item, so that sample features included in the detection curve data, which cannot be calculated based on a conventional algorithm, may be obtained by learning and mining through the first feature extraction network and the second feature extraction network.
Optionally, the combining the features extracted according to each path of input to obtain a fused feature set includes:
combining the first feature, the second feature, the third feature and the fourth feature according to a first feature which is extracted by taking the first specific protein detection item as input, a second feature which is extracted by taking the second specific protein detection item as input, a third feature which is extracted by taking the clinical information data corresponding to the sample to be detected as input and a fourth feature which is extracted by taking the specific protein detection value as input to obtain a fusion feature set;
When one of the first specific protein detection item and the second specific protein detection item is missing, the feature extraction result of the first specific protein detection item or the second specific protein detection item corresponding to the missing is a feature value 0.
For each sample to be detected, the fusion feature set obtained through the auxiliary diagnosis model comprises a first feature extracted through first feature extraction network mining, a second feature extracted through second feature extraction network mining, a third feature extracted and determined from clinical information data and a fourth feature which can be calculated and determined based on the curve form of detection curve data, so that all features of detection curve data of two types of specific protein items are more fully considered to obtain auxiliary diagnosis information, on one hand, diagnosis basis of the obtained auxiliary diagnosis information can be determined and consistent, and the situation that different analysis results appear on the same detection result due to personal experience difference of doctors reading the detection results of the specific protein detection items is avoided; on the other hand, all the characteristics of the detection curve data can be fully considered to obtain auxiliary diagnosis information, so that the accuracy of the detection result of the specific protein can be improved.
In an alternative specific example, referring to fig. 6, the first specific protein detection item may be CRP (C-reactive protein)/Hs-CRP (hypersensitive C-reactive protein), the second specific protein detection item may be SAA (serum amyloid a), the first feature extraction network and the second feature extraction network may use an MLP (Multi-layer perceptron neural networks, multi-layer sensing) neural network model, or a 1D-CNN (one-dimensional convolutional neural network model), and the first feature extraction network and the second feature extraction network extract detection curve data of the first specific protein detection item and the second specific protein detection item, respectively, so as to obtain 8-dimensional feature information F1 and F2, respectively. In the actual application process, only one of the first specific protein detection item and the second specific protein detection item can be detected, and at this time, the auxiliary diagnosis model sets 0 to the result of feature extraction of the input corresponding to the other item, so that the auxiliary diagnosis model can be compatible with different actual application scenes, and auxiliary diagnosis information can be obtained by performing category prediction on detection data obtained by executing one or more specific protein detection items on a sample to be detected. In addition, for the curve detection data of the first specific protein detection item and the second specific protein detection item, detection values, such as CRP concentration and SAA concentration, obtained by calculation by using a conventional algorithm are used to obtain 2-dimensional characteristic information F3; the age, sex, test sample type and past medical history of the object to which the sample to be detected belongs can be extracted from the clinical information data corresponding to the sample to be detected to obtain 8-dimensional characteristic information F4. The feature fusion layer combines the extracted feature information F1, F2, F3 and F4 to obtain an 18-dimensional fusion feature set.
The auxiliary diagnosis model can be a knowledge graph model constructed based on known judgment rules, such as judgment rules determined according to historical diagnosis examples, expert consensus libraries and the like; or the auxiliary diagnosis AI model is obtained after training the neural network model through the historical diagnosis examples and the expert consensus library. The target prediction category can be a potential risk type which needs to be predicted based on a specific protein detection result, the potential risk type comprises a plurality of potential risk types, each potential risk type is correspondingly formed into a target prediction category, the auxiliary diagnosis information comprises prediction results which correspond to all preset potential risk types respectively, and one or more of the target prediction categories possibly correspond to the prediction results which are at risk, so that the auxiliary diagnosis information can be conveniently and purposefully screened and judged, and support is provided for improving the diagnosis accuracy.
According to the auxiliary diagnosis method provided by the embodiment, according to the detection curve data obtained by a plurality of specific protein detection projects executed by a sample to be detected, sample characteristics which are determined based on traditional calculation according to curve shapes corresponding to the detection curve data and deeper sample characteristics which cannot be obtained from the curve shapes by traditional calculation in the corresponding detection curve data and are extracted by a characteristic extraction network can be combined with clinical information data corresponding to the sample to be detected to form a characteristic set, so that auxiliary diagnosis information containing a prediction type result is obtained, and therefore, all the characteristics of the detection curve data of the specific protein detection projects are fully considered in the obtaining of the prediction type result, and the accuracy of analyzing and utilizing the specific protein detection result can be improved.
In some embodiments, the sample to be detected is a blood sample, and may also be an extravascular body fluid sample, such as cerebrospinal fluid, synovial fluid, and the like. The obtaining curve detection data of at least one specific protein detection item of the sample to be detected comprises:
and acquiring detection curve data corresponding to a plurality of specific protein detection items respectively, wherein the detection curve data correspond to a plurality of specific protein detection items.
In this embodiment, the auxiliary diagnostic method uses detection curve data in corresponding detection results of a specific protein detection item as analysis data by acquiring detection results of a plurality of specific protein detection items of a sample to be detected by a specific protein detection analyzer. In practical applications, the auxiliary diagnostic device may be connected to a specific protein detection analyzer that performs a specific protein detection project on the blood sample, and may cooperate with the specific protein detection analyzer to realize auxiliary diagnosis.
In the above embodiment, by using the blood coagulation detection result output by the known specific protein detection analyzer as analysis data, more complete characteristics are extracted from the detection curve data output by the specific protein detection analyzer to perform standard and consistent analysis, so that the dependence of the identification of the detection result of the specific protein detection analyzer on the personal experience of a checking doctor is reduced, and support is provided for improving the diagnosis accuracy.
In some embodiments, the specific protein detection item comprises a first specific protein detection item or a second specific protein detection item;
the obtaining, by an auxiliary diagnostic model, the first prediction category of the sample to be detected and associated feature information thereof by combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item, and the specific protein detection value, includes:
the auxiliary diagnosis model takes the curve detection data of a first specific protein detection item or a second specific protein detection item as one input, takes the clinical information data corresponding to the sample to be detected as one input, and takes the specific protein detection value as one input, and performs feature extraction in parallel;
combining the extracted features according to the inputs of each path to obtain a fused feature set;
and carrying out classification prediction according to the fusion characteristic set to obtain a first prediction category of the sample to be detected and associated characteristic information thereof.
The auxiliary diagnosis model takes the curve detection data corresponding to the first specific protein detection item or the curve detection data corresponding to the second specific protein detection item as input to perform feature extraction, so that deeper features which cannot be obtained from the curve form of the corresponding detection curve data are obtained. The auxiliary diagnosis model comprises a feature extraction network corresponding to curve detection data corresponding to a specific protein detection project, a feature fusion layer connected with the feature extraction network and a classification prediction layer connected with the feature fusion layer. In an alternative specific example, the feature extraction network takes as input detection curve data corresponding to the first specific protein detection item or the second specific protein detection item. The feature fusion layer comprises one input corresponding to a feature extraction network, one input corresponding to clinical information data corresponding to a sample to be detected, and one input corresponding to a specific protein detection value calculated and determined based on curve morphological features. The feature fusion layer is used for carrying out feature fusion on the features extracted by the feature extraction network and the clinical information data and the specific protein detection value to obtain a fusion feature set. The classification prediction layer determines a first prediction category of the sample to be detected and associated feature information thereof based on an intelligently learned rule for determining the prediction category according to the fusion feature set output by the feature fusion layer, wherein the associated feature information comprises curve detection data for performing projection display on a region containing abnormal features.
In the above embodiment, the auxiliary diagnostic model includes a feature extraction network corresponding to a detection result obtained by executing a specific protein detection item, so that sample features included in the detection curve data and not obtained based on a traditional algorithm calculation may be obtained by learning and mining through the feature extraction network, and may be suitable for a use scenario in which only a specific protein detection item is executed, and the corresponding auxiliary diagnostic model may only set one path of feature extraction network to perform feature extraction on the detection curve data, so that the architecture of the auxiliary diagnostic model is simpler, and the training convergence speed of the auxiliary diagnostic model may be improved.
In some embodiments, the auxiliary diagnostic method further comprises:
constructing an initial classification network model comprising a plurality of paths of parallel feature extraction modules, and respectively forming training samples according to each historical diagnosis example and the diagnosis results thereof as labels; the input of the classification network model comprises corresponding clinical information data of a detection object contained in the training sample, specific protein detection values obtained by calculation of detection curve data of each specific protein detection item of the detection object and the detection curve data;
And training the classification network model through the training sample, and obtaining an auxiliary diagnosis model after training.
The classification network model includes two parallel feature extraction networks, the detection curve data corresponding to two specific protein detection items are respectively used as input, the clinical information data corresponding to the sample to be detected and the specific protein detection value determined by transmission calculation based on the morphological feature of the detection curve data are simultaneously used as input of the classification network model, and the output of the two parallel feature extraction networks, the clinical information data corresponding to the sample to be detected and the specific protein detection value determined by transmission calculation are jointly connected with an output layer for realizing classification prediction, and please refer to fig. 5, which is a schematic diagram of the architecture of the classification network model.
For each historical diagnosis example, executing detection curve data output by each specific protein detection item aiming at a specific protein detection analyzer, taking the detection curve data of two detection items as the input of a first feature extraction network and a second feature extraction network of a classification network model respectively, carrying out feature extraction on the corresponding detection curve data by the first feature extraction network to obtain a first feature, carrying out feature extraction on the corresponding detection curve data by the second feature extraction network to obtain a second feature, extracting features from clinical information data corresponding to a detection sample to obtain a third feature, calculating and determining a detection value as a fourth feature according to the morphological features of the detection curve data of the two detection items, accessing the output of the first feature extraction network and the second feature extraction network and the third feature and the fourth feature together into a feature fusion layer, accessing the output layer for classification prediction through the feature fusion layer, and obtaining deeper features corresponding to the detection curve data through the combined action of the second feature extraction network and the first feature extraction network, forming the clinical information data corresponding to the detection curve data corresponding to the detection sample, the detection curve data of each specific protein detection item of the detection sample, and determining the detection curve data according to the specific protein detection value as a diagnosis example, and taking the detection curve data as a diagnosis result of the diagnosis sample as a training label. And obtaining a large number of training samples through a large number of verified historical diagnosis examples, and performing iterative training on the classification network model to obtain a final trained auxiliary diagnosis model.
In this embodiment, the first feature extraction network and the second feature extraction network for feature extraction of the detection curve data in the classification network model may be a feature fusion layer connecting the feature extraction layer and the auxiliary diagnostic model after training by using an independent classification network, or may be a feature extraction module integrated in the auxiliary diagnostic model, and may be an auxiliary diagnostic model after training based on iterative training of the same round by integrating in the auxiliary diagnostic model.
In the above embodiment, a classification network model including two paths of parallel feature extraction networks is constructed, and the trained auxiliary diagnosis model can be obtained based on iterative training of the same round, so that the feature extraction module for extracting the feature of the detection curve data is integrated in the auxiliary diagnosis model, which is beneficial to improving the accuracy of the auxiliary diagnosis result.
Referring to fig. 7, in some embodiments, the present application further provides an auxiliary diagnostic device. The auxiliary diagnostic device includes: a first obtaining module 51, configured to obtain clinical information data corresponding to a sample to be detected; a second obtaining module 52, configured to obtain curve detection data of at least one specific protein detection item of the sample to be detected; a detection module 53, configured to calculate a specific protein detection value of the sample to be detected, corresponding to the specific protein detection item, according to the curve detection data of the specific protein detection item; the diagnosis module 54 is configured to obtain, through an auxiliary diagnosis model, a first prediction category of the sample to be detected and associated feature information thereof by combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item, and the specific protein detection value.
Optionally, the auxiliary diagnostic device further comprises a reporting module for outputting an auxiliary diagnostic report comprising the first prediction category and associated characteristic information thereof, the associated characteristic information comprising an abnormality recognition result of the curve detection data of the specific protein detection item.
Optionally, the diagnostic module 54 is further configured to determine, based on a preset class prediction rule, a second prediction class of the sample to be detected according to the specific protein detection value and clinical information data of the object to be detected; the report module is further configured to output an auxiliary diagnostic report including the first prediction category and associated feature information thereof, and the second prediction category, where the associated feature information includes an anomaly identification result of the curve detection data of the specific protein detection item.
Optionally, the specific protein detection items include a first specific protein detection item and a second specific protein detection item; the auxiliary diagnosis model takes curve detection data corresponding to the first specific protein detection item and the second specific protein detection item as two paths of input, takes clinical information data corresponding to the sample to be detected as one path of input and takes the specific protein detection value as one path of input, and performs feature extraction in parallel; combining the extracted features according to the inputs of each path to obtain a fused feature set; and carrying out classification prediction according to the fusion characteristic set to obtain a first prediction category of the sample to be detected and associated characteristic information thereof.
Optionally, the auxiliary diagnostic model is further configured to combine the first feature, the second feature, the third feature and the fourth feature according to a first feature extracted by taking the first specific protein detection item as input, a second feature extracted by taking the second specific protein detection item as input, a third feature extracted by taking the clinical information data corresponding to the sample to be detected as input, and a fourth feature extracted by taking the specific protein detection value as input, so as to obtain a fusion feature set; when one of the first specific protein detection item and the second specific protein detection item is missing, the feature extraction result of the first specific protein detection item or the second specific protein detection item corresponding to the missing is a feature value 0.
Optionally, the system further comprises a training module for constructing an initial classification network model comprising a plurality of paths of parallel feature extraction modules, and respectively forming training samples according to each historical diagnosis example and the diagnosis results thereof as labels; the input of the classification network model comprises corresponding clinical information data of a detection object contained in the training sample, specific protein detection values obtained by calculation of detection curve data of each specific protein detection item of the detection object and the detection curve data; and training the classification network model through the training sample, and obtaining an auxiliary diagnosis model after training.
It should be noted that, the module structure provided in the embodiment of the present application does not constitute a limitation of the auxiliary diagnostic device, and each module may be implemented in whole or in part by software, hardware, and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules. In other embodiments, more or fewer modules than shown may be included in the auxiliary diagnostic device.
In another aspect of the embodiments of the present application, referring to fig. 8, there is further provided an auxiliary diagnostic apparatus, where the auxiliary diagnostic apparatus includes a processor 201 and a memory 202, and the memory 202 stores a computer program executable by the processor 201, and the computer program is executed by the processor 201 to perform the steps of the auxiliary diagnostic method according to any embodiment of the present application.
In another aspect of the embodiments, a sample analysis system is provided, which includes an auxiliary diagnostic device and a specific protein detection analyzer connected to the auxiliary diagnostic device. The auxiliary diagnostic apparatus is the auxiliary diagnostic apparatus described in the foregoing embodiment.
In another aspect of the embodiments, there is also provided a computer program product, the instructions in which, when executed by a processor of an electronic device, cause the electronic device to perform the steps according to the auxiliary diagnostic method described in any of the embodiments of the present application.
In addition, in another aspect of the embodiments of the present application, there is also provided a computer-readable storage medium storing computer program instructions; the computer program instructions, when executed by a processor, implement the steps of the auxiliary diagnostic method provided in accordance with any of the embodiments of the present application.
The processor may be a CPU (central processing unit ) or ASIC (application specific integrated circuit, application Specific Integrated Circuit) or one or more integrated circuits configured to implement embodiments of the present invention. The one or more processors included in the moving object detection apparatus may be the same type of processor, such as one or more CPUs; but may also be different types of processors such as one or more CPUs and one or more ASICs.
The Memory may include a high-speed RAM (random access Memory ) and may further include an NVM (Non-Volatile Memory), such as at least one magnetic disk Memory.
The foregoing is merely specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think about changes or substitutions within the technical scope of the present application, and the changes and substitutions are intended to be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (11)
1. An auxiliary diagnostic method, comprising:
acquiring clinical information data corresponding to a sample to be detected;
acquiring curve detection data of at least one specific protein detection item of the sample to be detected;
calculating a specific protein detection value of the sample to be detected, which corresponds to the specific protein detection item, according to the curve detection data of the specific protein detection item;
and combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item and the specific protein detection value through an auxiliary diagnosis model to obtain a first prediction category of the sample to be detected and associated characteristic information of the sample to be detected.
2. The auxiliary diagnostic method as set forth in claim 1, further comprising:
Outputting a secondary diagnostic report comprising the first predictive category and associated characteristic information including an anomaly identification result for the curve detection data for the particular protein detection item.
3. The auxiliary diagnostic method as set forth in claim 1, further comprising:
determining a second prediction category of the sample to be detected based on a preset category prediction rule according to the specific protein detection value and clinical information data of the object to be detected;
outputting an auxiliary diagnostic report comprising the first predictive category and associated characteristic information thereof, including an anomaly identification result of the curve detection data for the particular protein detection item, and the second predictive category.
4. The diagnostic aid of claim 1, wherein the specific protein detection items comprise a first specific protein detection item and a second specific protein detection item;
the obtaining, by an auxiliary diagnostic model, the first prediction category of the sample to be detected and associated feature information thereof by combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item, and the specific protein detection value, includes:
The auxiliary diagnosis model takes curve detection data corresponding to the first specific protein detection item and the second specific protein detection item as two paths of input, takes clinical information data corresponding to the sample to be detected as one path of input and takes the specific protein detection value as one path of input, and performs feature extraction in parallel;
combining the extracted features according to the inputs of each path to obtain a fused feature set;
and carrying out classification prediction according to the fusion characteristic set to obtain a first prediction category of the sample to be detected and associated characteristic information thereof.
5. The aided diagnosis method of claim 4 wherein combining features extracted from each input to obtain a fused feature set comprises:
combining the first feature, the second feature, the third feature and the fourth feature according to a first feature which is extracted by taking the first specific protein detection item as input, a second feature which is extracted by taking the second specific protein detection item as input, a third feature which is extracted by taking the clinical information data corresponding to the sample to be detected as input and a fourth feature which is extracted by taking the specific protein detection value as input to obtain a fusion feature set;
When one of the first specific protein detection item and the second specific protein detection item is missing, the feature extraction result of the first specific protein detection item or the second specific protein detection item corresponding to the missing is a feature value 0.
6. The diagnostic aid of claim 1, wherein the specific protein detection items comprise a first specific protein detection item or a second specific protein detection item;
the obtaining, by an auxiliary diagnostic model, the first prediction category of the sample to be detected and associated feature information thereof by combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item, and the specific protein detection value, includes:
the auxiliary diagnosis model takes the curve detection data of a first specific protein detection item or a second specific protein detection item as one input, takes the clinical information data corresponding to the sample to be detected as one input, and takes the specific protein detection value as one input, and performs feature extraction in parallel;
combining the extracted features according to the inputs of each path to obtain a fused feature set;
And carrying out classification prediction according to the fusion characteristic set to obtain a first prediction category of the sample to be detected and associated characteristic information thereof.
7. The auxiliary diagnostic method according to any one of claims 1 to 6, further comprising:
constructing an initial classification network model comprising a plurality of paths of parallel feature extraction modules, and respectively forming training samples according to each historical diagnosis example and the diagnosis results thereof as labels; the input of the classification network model comprises corresponding clinical information data of a detection object contained in the training sample, specific protein detection values obtained by calculation of detection curve data of each specific protein detection item of the detection object and the detection curve data;
and training the classification network model through the training sample, and obtaining an auxiliary diagnosis model after training.
8. An auxiliary diagnostic device, comprising:
the first acquisition module is used for acquiring clinical information data corresponding to a sample to be detected;
the second acquisition module is used for acquiring curve detection data of at least one specific protein detection item of the sample to be detected;
the detection module is used for calculating a specific protein detection value of the sample to be detected, which corresponds to the specific protein detection item, according to the curve detection data of the specific protein detection item;
The diagnosis module is used for obtaining a first prediction category of the sample to be detected and associated characteristic information thereof by combining the clinical information data corresponding to the sample to be detected, the curve detection data of the specific protein detection item and the specific protein detection value through an auxiliary diagnosis model.
9. An auxiliary diagnostic device comprising a processor and a memory, wherein the memory stores a computer program executable by the processor, which when executed by the processor, implements the steps of the auxiliary diagnostic method as claimed in any one of claims 1 to 7.
10. A sample analysis system comprising the auxiliary diagnostic device of claim 9 and a specific protein detection analyzer coupled to the auxiliary diagnostic device.
11. A computer-readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, implements the steps of the auxiliary diagnostic method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210806511.7A CN117405896A (en) | 2022-07-08 | 2022-07-08 | Auxiliary diagnosis method, device, equipment, system and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210806511.7A CN117405896A (en) | 2022-07-08 | 2022-07-08 | Auxiliary diagnosis method, device, equipment, system and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117405896A true CN117405896A (en) | 2024-01-16 |
Family
ID=89498480
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210806511.7A Pending CN117405896A (en) | 2022-07-08 | 2022-07-08 | Auxiliary diagnosis method, device, equipment, system and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117405896A (en) |
-
2022
- 2022-07-08 CN CN202210806511.7A patent/CN117405896A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11769596B2 (en) | Plasma based protein profiling for early stage lung cancer diagnosis | |
| JP7286863B2 (en) | Automated validation of medical data | |
| CN101981446A (en) | Method and system for analysis of flow cytometry data using support vector machines | |
| AU2005236067A1 (en) | Pattern recognition method for diagnosis of systemic autoimmune diseases | |
| CN115099331A (en) | Auxiliary diagnosis system for malignant pleural effusion based on interpretable machine learning algorithm | |
| Wei et al. | Multi-scale sequential feature selection for disease classification using Raman spectroscopy data | |
| Bunch et al. | Artificial intelligence applications in clinical chemistry | |
| Chadaga et al. | COVID-19 diagnosis using clinical markers and multiple explainable artificial intelligence approaches: A case study from Ecuador | |
| Nashwan et al. | Using artificial intelligence to improve body iron quantification: A scoping review | |
| CN117253625A (en) | Construction device of lung cancer screening model, lung cancer screening device, equipment and medium | |
| Wang et al. | Survival risk prediction model for ESCC based on relief feature selection and CNN | |
| WO2023186051A1 (en) | Auxiliary diagnosis method and apparatus, and construction apparatus, analysis apparatus and related product | |
| US10235495B2 (en) | Method for analysis and interpretation of flow cytometry data | |
| CN117405896A (en) | Auxiliary diagnosis method, device, equipment, system and medium | |
| CN116030957A (en) | Knowledge graph construction method, knowledge graph construction device, auxiliary diagnosis equipment and storage medium | |
| Raju et al. | Chronic kidney disease prediction using ensemble machine learning | |
| Zeb et al. | Towards the Selection of the Best Machine Learning Techniques and Methods for Urinalysis | |
| Naing et al. | The Evaluation of Acute Myeloid Leukaemia (AML) Blood Cell Detection Models Using Different YOLO Approaches | |
| CN117409951A (en) | Auxiliary diagnosis method, device, equipment, product, system and medium | |
| EP4195219A1 (en) | Means and methods for the binary classification of ms1 maps and the recognition of discriminative features in proteomes | |
| CN117438067A (en) | Auxiliary diagnosis method, device, equipment, product, system and medium | |
| KHALEL et al. | IMPROVE DIABETES DIAGNOSIS BY INTEGRATING MULTIPLE MACHINE LEARNING ALGORITHMS | |
| Mohammed et al. | Two-Stage Classification of Breast Tumor Biomarkers for Iraqi Women | |
| CN117524456A (en) | Auxiliary diagnosis method and system | |
| CN116030958A (en) | Auxiliary diagnosis method, device, equipment and system and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |