WO2023153635A1 - Procédé et dispositif de classification de maladie inflammatoire utilisant une analyse spectroscopique raman basée sur l'apprentissage automatique - Google Patents

Procédé et dispositif de classification de maladie inflammatoire utilisant une analyse spectroscopique raman basée sur l'apprentissage automatique Download PDF

Info

Publication number
WO2023153635A1
WO2023153635A1 PCT/KR2022/021418 KR2022021418W WO2023153635A1 WO 2023153635 A1 WO2023153635 A1 WO 2023153635A1 KR 2022021418 W KR2022021418 W KR 2022021418W WO 2023153635 A1 WO2023153635 A1 WO 2023153635A1
Authority
WO
WIPO (PCT)
Prior art keywords
raman
machine learning
raman signal
signal
inflammatory disease
Prior art date
Application number
PCT/KR2022/021418
Other languages
English (en)
Korean (ko)
Inventor
김준기
이상화
주미연
Original Assignee
재단법인 아산사회복지재단
울산대학교 산학협력단
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 재단법인 아산사회복지재단, 울산대학교 산학협력단 filed Critical 재단법인 아산사회복지재단
Publication of WO2023153635A1 publication Critical patent/WO2023153635A1/fr

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
    • A61B5/0075—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence by spectroscopy, i.e. measuring spectra, e.g. Raman spectroscopy, infrared absorption spectroscopy
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/06—Devices, other than using radiation, for detecting or locating foreign bodies ; determining position of probes within or on the body of the patient
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/06—Devices, other than using radiation, for detecting or locating foreign bodies ; determining position of probes within or on the body of the patient
    • A61B5/061—Determining position of a probe within the body employing means separate from the probe, e.g. sensing internal probe position employing impedance electrodes on the surface of the body
    • A61B5/064—Determining position of a probe within the body employing means separate from the probe, e.g. sensing internal probe position employing impedance electrodes on the surface of the body using markers
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/20—Measuring for diagnostic purposes; Identification of persons for measuring urological functions restricted to the evaluation of the urinary system
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7271—Specific aspects of physiological measurement analysis
    • A61B5/7275—Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/74—Details of notification to user or communication with user or patient ; user input means
    • A61B5/742—Details of notification to user or communication with user or patient ; user input means using visual displays
    • A61B5/7445—Display arrangements, e.g. multiple display units
    • 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
    • G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
    • 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

Definitions

  • the present invention relates to a method and apparatus for classifying inflammatory diseases using machine learning-based Raman spectroscopy.
  • IC/BPS interstitial cystitis/bladder pain syndrome
  • Another urinary inflammatory disease such as urolithiasis, ureteral stricture, ureteral obstruction, and nephritis, may cause kidney damage and cause chronic or acute kidney disease, so quantified evaluation criteria for early diagnosis are further required.
  • metabolomics has been studied for early diagnosis of diseases, and metabolomics is an important study that reveals mechanisms of metabolome by relating metabolome to various physiological and pathological conditions.
  • metabolomics shows great potential as a biomarker for early diagnosis and monitoring of cancer, improved analysis methods for complex biological products are needed for clinical application as a quantified diagnosis method.
  • GC-MS gas chromatography mass spectrometry
  • mass spectrometry such as liquid chromatography, Raman spectroscopy, and surface enhanced Raman spectroscopy (Surface Enhanced Raman Spectroscopy).
  • the early diagnosis methods according to the prior art such as Raman spectroscopy, have a problem in that it is difficult to clearly classify and quantify a specific disease based on the test results for a specific sample or individual factors and diagnose it. Specifically, there is a problem in that accuracy is low in clearly quantifying and diagnosing a specific disease by category by checking only the spectrum distribution or detection signals detected in various ranges for a specific sample or specimen.
  • An object to be solved by the present invention is to provide a method and apparatus for classifying inflammatory diseases using machine learning-based Raman spectroscopic analysis.
  • the problem to be solved by the present invention is an inflammatory disease classification method capable of detecting chemical biomarkers for urinary inflammatory diseases and diseases by surface-enhanced Raman spectroscopy, and classifying inflammatory disease types and early diagnosis of patients. and to provide an apparatus.
  • An inflammatory disease classification method performed by an inflammatory disease classification apparatus for solving the above problems includes preparing and pre-treating specimens; Obtaining a Raman signal and a spectrum of the Raman signal through Raman spectroscopy; analyzing the Raman signal and a spectrum distribution of the Raman signal; Quantifying diagnostic data through a preset machine learning program; and extracting a diagnostic result by comparing the Raman signal and the spectrum distribution with quantified diagnostic data.
  • the Raman signal and the spectrum of the Raman signal may be obtained by performing a surface-enhanced Raman scattering test.
  • the acquiring of the Raman signal and the spectrum of the Raman signal may further include correcting the Raman spectroscopy signal by performing biomarker detection correction in the surface-enhanced Raman scattering test process.
  • nano-sized biomarkers on a surface-enhanced Raman scattering test kit are applied as Raman measurement targets, and intensity changes of amplified peaks of the nano-sized molecules are measured to measure the Raman spectroscopic signal.
  • the spectral signal can be calibrated.
  • the type of metabolite contained in each sample, inflammation level, and inflammation type are analyzed for each sample, and the type of metabolite, inflammation level, Depending on the type of inflammation, the progress of urinary inflammatory diseases such as nephritis, cystitis, cancer, and acute kidney disease can be analyzed.
  • the quantifying the diagnostic data through the preset machine learning program may include inputting the Raman signal and the Raman spectrum detection result into the machine learning program and the database as input values; additionally inputting additional learning factors including subject detection human information as input values of the machine learning program; selecting at least one or more machine learning models and machine learning programs; allowing the selected machine learning program to learn by reflecting the input values; and quantifying the diagnostic data according to an execution result of the machine learning program.
  • the step of additionally inputting additional learning factors including the sample detection human information as input values of the machine learning program includes age, gender, detection human information including diseases, findings, symptoms, sample detection amount, and detection site. Information can be additionally input as the input values.
  • the extracting of the diagnosis result may include comparing the Raman signal and the Raman spectrum detection result with the diagnosis data quantified through the machine learning program; and setting a diagnosis result according to the comparison result and displaying it on a display screen.
  • an apparatus for classifying inflammatory diseases for solving the above problems includes a Raman test module for obtaining a Raman signal and a spectrum of the Raman signal by performing a Raman spectroscopic test on specimens; a data learning module that receives learning factors including the Raman signal and a spectral distribution of the Raman signal and performs preset machine learning to quantify diagnostic data; and an early diagnosis module for extracting a diagnosis result by comparing the Raman signal and the spectrum distribution of the Raman signal with the quantified diagnosis data.
  • the Raman inspection module may include a sample inspection pre-processing unit including a surface-enhanced Raman scattering inspection kit; and a Raman signal detector for acquiring the Raman signal by performing a surface-enhanced Raman scattering test on the sample pretreated on the surface-enhanced Raman scattering test kit.
  • the Raman test module applies the nano-sized biomarkers on the surface-enhanced Raman scattering test kit as Raman measurement targets, and measures the intensity change of the amplified peak of the nano-sized molecule to obtain the Raman spectroscopy signal.
  • a spectral signal corrector for correcting and quantifying may be further included.
  • the early diagnosis module analyzes the type of metabolite, inflammation level, and inflammation type contained in each sample for each sample, and according to the type of metabolite, inflammation level, and inflammation type, nephritis, cystitis, cancer, etc. of urinary system inflammatory disease, and progression to acute kidney disease can be analyzed.
  • the data learning module may include: a learning factor input unit inputting the Raman signal and a spectrum detection result of the Raman signal to a machine learning program and a database as input values; a machine learning program input unit for storing machine learning programs for each of a plurality of machine learning models and selecting and compiling at least one or more machine learning models and machine learning programs; a machine learning processing unit executing the selected machine learning program so that the selected machine learning program reflects the input values and is machine trained; and a disease diagnosis model generating unit quantifying the diagnosis data according to an execution result of the machine learning program.
  • the learning factor input unit may additionally input detection human information including age, gender, disease, findings, and symptoms, a sample detection amount, and detection site information as the input values.
  • the early diagnosis module may include: a data processing unit that reads the Raman signal and the Raman spectrum detection result and classifies them according to subject personal information including age, gender, diseases, findings, and symptoms; a disease diagnosis model extractor that reads the quantified diagnosis data from the data learning module; a data comparison detection unit that compares and analyzes the Raman signal, a spectrum detection result of the Raman signal, and the quantified diagnostic data; and a diagnosis result extraction unit configured to set and display a diagnosis result according to a comparison result with the diagnosis data.
  • the inflammatory disease classification method and apparatus can quickly classify and quantify the patient's inflammatory disease type through machine learning-based Raman spectroscopic analysis results and machine learning, and perform early diagnosis using the quantified results. can Therefore, it is possible to prevent damage to the kidney by diagnosing inflammation of the ureter near the kidney at an early stage.
  • the inflammatory disease classification method and apparatus can more quickly and accurately classify and diagnose difficult-to-diagnose urinary diseases such as cystitis and nephritis, enabling appropriate and rapid treatment.
  • FIG. 1 is a flowchart for sequentially explaining a method for classifying inflammatory diseases according to an embodiment of the present invention.
  • 2 is a graph showing Raman signal detection results obtained through a surface-enhanced Raman scattering method.
  • Figure 3 is a diagram showing the random forest (Random forest) results of the Raman spectrum obtained through the surface-enhanced Raman scattering method.
  • 4A to 4C are diagrams showing results of sequentially performing principal component analysis and linear discriminant analysis of the Raman spectrum obtained through the surface-enhanced Raman scattering method.
  • 5A to 5C are diagrams showing results of sequentially performing non-negative matrix factorization-linear discriminant analysis of the Raman spectrum obtained through the surface-enhanced Raman scattering method.
  • FIG. 6 is a flowchart illustrating the method of analyzing spectral data and extracting a disease diagnosis model shown in FIG. 1 in more detail.
  • FIG. 7 is a block diagram showing a configuration of an inflammatory disease classification apparatus according to an embodiment of the present invention in detail.
  • the identification code is used for convenience of description, and the identification code does not explain the order of each step, and each step may be performed in a different order from the specified order unless a specific order is clearly described in context. there is.
  • FIG. 1 is a flowchart for sequentially explaining a method for classifying inflammatory diseases according to an embodiment of the present invention.
  • the inflammatory disease classification method includes preparing and preprocessing a specimen (ST1), obtaining a Raman signal through a Raman test (ST2), correcting a Raman spectroscopy signal (ST3), Raman signal and The step of processing Raman spectroscopy data by analyzing the spectrum of the Raman signal (ST4), the step of quantifying the diagnostic results (diagnostic data) analyzed by performing machine learning (ST5), and the step of extracting the diagnostic results (ST6).
  • the specimens detected from the inflammatory site or the specimens detected from the test subject are placed in the test kit for disease testing without a separate separation process.
  • urine detected from inflammation inducing sites such as the ureter or bladder can be applied as a sample.
  • inflammation levels and related diseases can be detected through Raman spectroscopy and analysis of urine samples.
  • urine of 8- to 10-week-old laboratory rats can be used as a sample in order to quantify the urinary diagnostic data through examination and analysis through the specimen.
  • urine of 23 laboratory rats aged 8 to 10 weeks could be used as the specimens applied to show the Raman test results in graphs and drawings. More specifically, 5 ⁇ l of 1.5 ml of urine from 4 animals with interstitial cystitis symptoms, 5 animals with mild ureteral obstruction symptoms, 5 animals with severe ureteral obstruction symptoms, 5 animals with control group, and 4 animals with normal group could be applied as samples.
  • a Raman signal may be obtained by performing a surface-enhanced Raman scattering test.
  • the surface-enhanced Raman scattering method is an analysis method that can overcome the detection sensitivity limit of Raman spectroscopy and provides molecular-specific information about biological and chemical samples.
  • a target material is quantified by measuring a change in intensity of amplified characteristic peaks of nano-sized molecules.
  • a Raman signal may be obtained by measuring a portion where the sample is widely distributed in the substrate using a Raman spectroscopic method.
  • background correction of the Raman signal may be performed.
  • the biomarker detection and correction may be performed to correct the Raman spectroscopy signal (ST3).
  • an enhancement factor (EF) used as a measure of the surface-enhanced Raman scattering scale is usually 10 4 to 10 8 , and sometimes reaches 10 14 , which enables single-molecule level detection.
  • EF enhancement factor
  • the intensity of the signal is dramatically increased by about 10 6 to about 10 8 compared to the general Raman spectroscopy method, so that even a small amount of inflammatory markers can be detected.
  • the Raman spectroscopy signal correction step (ST3) according to an embodiment of the present invention, nano-sized biomarkers on the surface-enhanced Raman scattering test kit are tested.
  • the Raman signal and spectral distribution can be calibrated and quantified by measuring a change in intensity of a characteristic peak of the amplified biomarker Raman signal.
  • the type of metabolite contained in the sample, the level of inflammation, the type of inflammation, etc. are confirmed for each sample collected.
  • the degree of progression of diseases eg, nephritis, cystitis, urinary inflammatory diseases such as cancer, and acute kidney disease
  • the type of metabolite, inflammation level, and type of inflammation can be confirmed.
  • Figure 2 is a graph showing Raman signal detection results obtained through a surface-enhanced Raman scattering method. Specifically, Figure 2 is a graph showing the surface-enhanced Raman signals of pyelonephritis and interstitial cystitis specimens.
  • inflammation levels due to severe ureteral obstruction can be confirmed through a sample (SO_K) taken between the kidney and the ureter, and collected from the bladder. Even one sample (SO_B) can confirm the level of inflammation caused by severe ureteral obstruction.
  • another sample (MO_K) taken between the kidney and ureter can confirm the level of inflammation due to mild ureteral obstruction, and another sample (MO_B) taken from the bladder can also show the level of inflammation due to mild ureteral obstruction. You can check.
  • the level of inflammation according to interstitial cystitis can be confirmed through a sample (IC/BPS) taken from the bladder.
  • the inflammatory signal of pyelonephritis is prominent in the range of about 650 cm -1 to 750 cm -1
  • the inflammatory signal of interstitial cystitis is dominant in the range of about 910 cm -1 to 1060 cm -1 .
  • Figure 3 is a diagram showing the random forest (Random forest) results of the Raman spectrum obtained through the surface-enhanced Raman scattering method.
  • a Raman spectroscopy equipment eg, 532nm laser, 785nm laser Raman Device
  • a preset wavelength range may be sampled and a spectrum distribution of the sampled wavelength range may be analyzed.
  • related diseases such as interstitial cystitis, mild ureteral obstruction, severe ureteral obstruction, and control group can be identified and analyzed.
  • the spectrum distribution of the Raman signal is analyzed through a pre-set program on a computer, the content and increase in the type of protein, lipid, RNA, DNA, etc. can be analyzed according to the unique Raman spectrum distribution and coverage of organic and inorganic molecules. there is.
  • a sample taken between the kidney and the ureter (SO_K), a sample taken from the bladder (SO_B), another sample taken between the kidney and the ureter (MO_K), and another sample taken from the bladder ( MO_B), and a disease classification result can be confirmed by calculating a disease prediction value through a random forest ensemble algorithm result of a Raman spectrum for each specific sample (IC/BPS) collected from the bladder.
  • 4A to 4C are diagrams showing results of sequentially performing principal component analysis and linear discriminant analysis of the Raman spectrum obtained through the surface-enhanced Raman scattering method.
  • a Raman spectroscopy equipment eg, 532nm laser, 785nm laser Raman Device
  • 500cm -1 ⁇ 3000cm - A preset wavelength range between 1 and 1 may be sampled, and the spectrum distribution of the sampled wavelength range may be analyzed.
  • the wavelength range of 500 cm ⁇ 1 to 3000 cm ⁇ 1 related diseases such as interstitial cystitis, mild ureteral obstruction, severe ureteral obstruction, and control group can be identified and analyzed.
  • the spectrum distribution of the Raman signal is analyzed through a pre-set program on a computer, the content and increase in the type of protein, lipid, RNA, DNA, etc. can be analyzed according to the unique Raman spectrum distribution and coverage of organic and inorganic molecules. there is.
  • a sample collected between the kidney and the ureter (SO_K), a sample collected from the bladder (SO_B), another sample collected between the kidney and the ureter (MO_K), and the bladder through principal component analysis (PCA) Linear discriminant analysis after obtaining 5 in FIG. 4A, 30 in FIG. 4B, and 50 in FIG.
  • PCA principal component analysis
  • 5A to 5C are diagrams showing results of sequentially performing non-negative matrix factorization-linear discriminant analysis of the Raman spectrum obtained through the surface-enhanced Raman scattering method.
  • the step of analyzing the Raman signal and the spectrum distribution of the Raman signal by using Raman spectroscopy equipment (eg, 532nm laser, 785nm laser Raman Device) 500cm -1 ⁇ 3000cm -
  • Raman spectroscopy equipment eg, 532nm laser, 785nm laser Raman Device
  • 500cm -1 ⁇ 3000cm - A preset wavelength range between 1 and 1 may be sampled, and the spectrum distribution of the sampled wavelength range may be analyzed.
  • related diseases such as interstitial cystitis, mild ureteral obstruction, severe ureteral obstruction, and control group can be identified and analyzed.
  • the spectrum distribution of the Raman signal is analyzed through a pre-set program on a computer, the content and increase in the type of protein, lipid, RNA, DNA, etc. can be analyzed according to the unique Raman spectrum distribution and coverage of organic and inorganic molecules. there is.
  • a sample collected between the kidney and the ureter through non-negative matrix decomposition (NMF) (SO_K), a sample collected from the bladder (SO_B), and another sample collected between the kidney and the ureter (MO_K) , another sample (MO_B) taken from the bladder, and the non-negative matrix decomposition component (NMFC) of the Raman spectrum for each specific sample (IC/BPS) taken from the bladder are 5 in FIG. 5, 30 in FIG. 5 b, and 30 in FIG. 5 c.
  • the disease classification result can be confirmed by deriving the result value through a linear discriminant analysis (LDA) model.
  • LDA linear discriminant analysis
  • FIG. 6 is a flowchart illustrating the method of analyzing spectral data and extracting a disease diagnosis model shown in FIG. 1 in more detail.
  • step ST5 of quantifying diagnostic results (diagnostic data) analyzed through a preset machine learning program the Raman signal and Raman spectrum detection results are entered into the machine learning program and database as input values.
  • step (SS1) step of additionally inputting additional learning factors including sample detection human information as input values of a machine learning program (SS2), step of selecting and setting at least one machine learning model and machine learning program (SS3) , making the set machine learning program reflect the input values for learning processing (SS4), and quantifying diagnostic data according to the execution result of the machine learning program (SS5).
  • the Raman signal detection result and the Raman spectrum detection result are entered as input values into an input program and a database (SS1).
  • SS1 an input program and a database
  • at least one machine learning program among preset machine learning programs you can additionally input the Raman signal detection result and Raman spectrum detection result, including sample-related information, as learning factors.
  • Machine learning programs include PCA (Principal Component Analysis)-LDA (Linear Discriminant Analysis) model, NMF (Non-Negative Matrix Factorization)-LDA (Linear Discriminant Analysis) model, RFML (Random Forest Machine Learning) model, deep learning (e.g. For example, Convolutional Neural Networks (CNN) and the like may be concurrently applied.
  • PCA Principal Component Analysis
  • LDA Linear Discriminant Analysis
  • NMF Non-Negative Matrix Factorization
  • RFML Random Forest Machine Learning
  • deep learning e.g. For example, Convolutional Neural Networks (CNN) and the like may be concurrently applied.
  • CNN Convolutional Neural Networks
  • the type of metabolite contained in the sample can be upgraded.
  • the level of inflammation can be upgraded.
  • the progress of urinary inflammatory diseases such as nephritis, cystitis, and cancer, and the progress of acute kidney disease, respectively, are diagnostic data. can be quantified and databased. (SS5)
  • the step of extracting the diagnosis result is the step of comparing and analyzing the Raman signal detection result and the Raman spectrum detection result and the diagnostic data quantified through a machine learning program (SS6), and the comparison result and setting a diagnostic result according to the method and displaying it on a display screen such as a monitor (SS7).
  • the Raman signals extracted by the surface-enhanced Raman scattering test for each sample and the spectrum distribution results of the Raman signals are compared and analyzed with diagnostic data quantified through a machine learning program.
  • the diagnosis result is finally extracted according to the comparison result of the Raman signal and the spectrum distribution result of the Raman signal and the quantified diagnosis data, and the finally extracted diagnosis result can be displayed on a separate monitor or display screen. there is.
  • FIG. 7 is a block diagram showing a configuration of an inflammatory disease classification apparatus according to an embodiment of the present invention in detail.
  • the inflammatory disease classification apparatus shown in FIG. 7 includes a Raman examination module 100, an early diagnosis module 200, and a data learning module 300.
  • the Raman inspection module 100 acquires a Raman signal by performing a surface-enhanced Raman scattering inspection on the pretreated sample.
  • the Raman inspection module 100 includes a sample inspection and preprocessing unit 101 and a Raman signal detection unit 103 .
  • the sample inspection and pre-processing unit 101 of the Raman test module 100 includes a test kit in which a sample detected from an inflammation-inducing site or a sample of a test subject is placed and pre-processed so as to spread for a predetermined period of time for disease inspection.
  • the Raman signal detector 103 obtains a Raman signal by performing a surface-enhanced Raman scattering test on a sample pretreated on a surface-enhanced Raman scattering test kit. Then, the obtained Raman signal is stored in the database.
  • the Raman signal detector 103 measures the change in intensity of the amplified characteristic peak of the Raman nano-sized molecule by performing a sample inspection using a preset surface-enhanced Raman scattering method, and measures the Raman signal and Raman signal for the target material. A spectrum of the signal can be acquired and quantified.
  • the Raman test module 100 filters molecules larger than the micro size on the surface-enhanced Raman scattering test kit, and nano-sized biomarkers. It may further include a spectroscopy signal correction unit 105 for correcting and quantifying the Raman spectroscopy signal by measuring a change in intensity of the amplified peak.
  • the intensity of the signal is dramatically increased by about 10 6 to about 10 8 compared to the general Raman spectroscopy method, so that even a small amount of inflammatory markers can be detected.
  • the spectral signal corrector 105 of the Raman test module 100 quantifies the target material by measuring intensity changes of amplified characteristic peaks of nano-sized biomarkers on the surface-enhanced Raman scattering test kit.
  • the data learning module 300 quantifies diagnostic data by receiving learning factors including a Raman signal and a spectral distribution of the Raman signal and performing preset machine learning.
  • the data learning module 300 may include a learning factor input unit 301, a machine learning processor 303, a machine learning program input unit 305, and a disease diagnosis model generator 307.
  • the learning factor input unit 301 of the data learning module 300 allows the Raman signal detection result and the Raman spectrum detection result to be input into a preset program and database as input values.
  • the learning factor input unit 301 may further input additional learning factors including detection human information when inputting the Raman signal detection result and the Raman spectrum detection result.
  • the sample detection amount, detection site information, and detection personal information eg, age, gender, disease, findings, symptoms, etc.
  • detection personal information eg, age, gender, disease, findings, symptoms, etc.
  • the machine learning program input unit 305 stores machine learning programs for each of a plurality of machine learning models, and selects and compiles at least one machine learning model and machine learning program.
  • the machine learning programs include PCA (Principal Component Analysis)-LDA (Linear Discriminant Analysis) model, NMF (Non-Negative Matrix Factorization)-LDA (Linear Discriminant Analysis) model, RFML (Random Forest Machine Learning) model, deep learning (eg, Convolutional Neural Networks (CNN)) may be concurrently applied.
  • the machine learning processing unit 303 executes the input and selected machine learning program from the machine learning program input unit 305 to determine the type of metabolite contained in each specimen, the type and type of inflammation, the level of inflammation, and the level of inflammation. Upgrade disease progression information.
  • the disease diagnosis model generation unit 307 quantifies diagnosis data according to a result of executing a machine learning program. At this time, the disease diagnosis model generation unit 307 provides progress information of urinary inflammatory diseases such as nephritis, cystitis, cancer, and acute kidney disease according to sample-related classification information such as age, gender, diseases possessed, findings, and symptoms, respectively. It is quantified as diagnostic data and stored and shared in a database.
  • urinary inflammatory diseases such as nephritis, cystitis, cancer, and acute kidney disease
  • sample-related classification information such as age, gender, diseases possessed, findings, and symptoms, respectively. It is quantified as diagnostic data and stored and shared in a database.
  • the early diagnosis module 200 compares and analyzes the Raman signal and the spectral distribution of the Raman signal and the quantified diagnostic data to extract a diagnostic result for an inflammatory disease.
  • the early diagnosis module 200 may include a data processing unit 202, a disease diagnosis model extraction unit 204, a data comparison detection unit 206, and a diagnosis result extraction unit 208.
  • the data processing unit 202 of the early diagnosis module 200 reads the Raman signal detection results and the Raman spectrum detection results and classifies them according to specimen-related classification information such as age, gender, diseases possessed, findings, and symptoms.
  • the disease diagnosis model extractor 204 reads the quantified diagnosis data from the data learning module 300 in real time.
  • the comparison detector 206 compares and analyzes the Raman signal detection result and the Raman spectrum detection result and the quantified diagnostic data.
  • the comparison detection unit 206 sequentially compares and analyzes the Raman signals extracted through the surface-enhanced Raman scattering test for each sample and the spectrum distribution result of the Raman signals with diagnostic data quantified through a machine learning program.
  • the diagnosis result extraction unit 208 sets and displays diagnosis results according to comparison results.
  • the diagnosis result extraction unit 208 may finally extract a diagnosis result according to a comparison result between the Raman signal and the spectrum distribution result of the Raman signal and the quantified diagnosis data, and display the result on a screen.
  • the early diagnosis module 200 and the data learning module 300 include a memory (not shown) storing data for an algorithm or a program reproducing the algorithm for controlling the operation of components in the processor of each module, and a memory ( Alternatively, it may include a processor (not shown) that performs the above-described operation using data stored in a database.
  • the memory and the processor may be implemented as separate chips.
  • the memory and the processor may be implemented as a single chip.
  • Each component refers to software and/or hardware components such as Field Programmable Gate Array (FPGA) and Application Specific Integrated Circuit (ASIC).
  • FPGA Field Programmable Gate Array
  • ASIC Application Specific Integrated Circuit
  • the disclosed embodiments may be implemented in the form of a recording medium storing commands executable by the early diagnosis module 200 and the data learning module 300 . Instructions may be stored in the form of program codes, and when executed by a processor, create program modules to perform operations of the disclosed embodiments.
  • the recording medium may be implemented as an early diagnosis module 200, a data learning module 300, and a computer-readable recording medium.
  • the early diagnosis module 200 and the data learning module 300 or computer-readable recording media include all types of recording media in which instructions readable by a computer are stored. For example, there may be read only memory (ROM), random access memory (RAM), magnetic tape, magnetic disk, flash memory, optical data storage device, and the like.
  • ROM read only memory
  • RAM random access memory
  • magnetic tape magnetic tape
  • magnetic disk magnetic disk
  • flash memory optical data storage device

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Public Health (AREA)
  • Medical Informatics (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Pathology (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • Biophysics (AREA)
  • Veterinary Medicine (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Physiology (AREA)
  • Primary Health Care (AREA)
  • Epidemiology (AREA)
  • Artificial Intelligence (AREA)
  • Psychiatry (AREA)
  • Signal Processing (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Human Computer Interaction (AREA)
  • Urology & Nephrology (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Mathematical Physics (AREA)
  • Fuzzy Systems (AREA)
  • Evolutionary Computation (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)

Abstract

L'invention concerne un procédé et un dispositif de classification de maladie inflammatoire, utilisant une analyse spectroscopique Raman basée sur l'apprentissage automatique. Le procédé de classification de maladie inflammatoire, selon la présente invention, comprend les étapes consistant à : préparer et prétraiter des échantillons ; acquérir un signal Raman et un spectre du signal Raman par une analyse spectroscopique Raman ; analyser le signal Raman et la distribution spectrale du signal Raman ; quantifier des données de diagnostic au moyen d'un programme d'apprentissage automatique prédéfini ; et extraire un résultat de diagnostic par comparaison du signal Raman et de la distribution spectrale avec les données de diagnostic quantifiées.
PCT/KR2022/021418 2022-02-08 2022-12-27 Procédé et dispositif de classification de maladie inflammatoire utilisant une analyse spectroscopique raman basée sur l'apprentissage automatique WO2023153635A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR1020220016224A KR20230119901A (ko) 2022-02-08 2022-02-08 머신 러닝 기반 라만 분광 분석을 이용한 염증 질환 분류 방법 및 장치
KR10-2022-0016224 2022-02-08

Publications (1)

Publication Number Publication Date
WO2023153635A1 true WO2023153635A1 (fr) 2023-08-17

Family

ID=87564558

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2022/021418 WO2023153635A1 (fr) 2022-02-08 2022-12-27 Procédé et dispositif de classification de maladie inflammatoire utilisant une analyse spectroscopique raman basée sur l'apprentissage automatique

Country Status (2)

Country Link
KR (1) KR20230119901A (fr)
WO (1) WO2023153635A1 (fr)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20050078924A (ko) * 2004-02-03 2005-08-08 재단법인서울대학교산학협력재단 라만 분광법을 이용한 소변 성분 분석 시스템 및 그 방법
KR20170130195A (ko) * 2016-05-18 2017-11-28 서울대학교산학협력단 표면증강라만산란을 이용한 폐암 조기 진단 나노센서 및 그 방법
KR20180053957A (ko) * 2016-11-14 2018-05-24 테라셈 주식회사 의료 진단 시스템, 서버 및 방법
KR20190016387A (ko) * 2017-08-08 2019-02-18 사회복지법인 삼성생명공익재단 플라즈모닉 효과를 이용한 질병의 진단 방법 및 질병 진단용 슬라이드
JP2021526628A (ja) * 2018-04-05 2021-10-07 イーエニエーエスセー テック − インスティチュート デ エンゲンハリア デ システマス エ コンピュータドレス テクノロジア エ シエンシアＩｎｅｓｃ Ｔｅｃ − Ｉｎｓｔｉｔｕｔｏ Ｄｅ Ｅｎｇｅｎｈａｒｉａ Ｄｅ Ｓｉｓｔｅｍａｓ Ｅ Ｃｏｍｐｕｔａｄｏｒｅｓ， Ｔｅｃｎｏｌｏｇｉａ Ｅ Ｃｉｅｎｃｉａ 試料からの成分の定量化値を予測する分光測光方法及び装置

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102428314B1 (ko) 2020-02-26 2022-08-03 사회복지법인 삼성생명공익재단 소변의 라만 신호를 이용한 암 진단 방법

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20050078924A (ko) * 2004-02-03 2005-08-08 재단법인서울대학교산학협력재단 라만 분광법을 이용한 소변 성분 분석 시스템 및 그 방법
KR20170130195A (ko) * 2016-05-18 2017-11-28 서울대학교산학협력단 표면증강라만산란을 이용한 폐암 조기 진단 나노센서 및 그 방법
KR20180053957A (ko) * 2016-11-14 2018-05-24 테라셈 주식회사 의료 진단 시스템, 서버 및 방법
KR20190016387A (ko) * 2017-08-08 2019-02-18 사회복지법인 삼성생명공익재단 플라즈모닉 효과를 이용한 질병의 진단 방법 및 질병 진단용 슬라이드
JP2021526628A (ja) * 2018-04-05 2021-10-07 イーエニエーエスセー テック − インスティチュート デ エンゲンハリア デ システマス エ コンピュータドレス テクノロジア エ シエンシアＩｎｅｓｃ Ｔｅｃ − Ｉｎｓｔｉｔｕｔｏ Ｄｅ Ｅｎｇｅｎｈａｒｉａ Ｄｅ Ｓｉｓｔｅｍａｓ Ｅ Ｃｏｍｐｕｔａｄｏｒｅｓ， Ｔｅｃｎｏｌｏｇｉａ Ｅ Ｃｉｅｎｃｉａ 試料からの成分の定量化値を予測する分光測光方法及び装置

Also Published As

Publication number Publication date
KR20230119901A (ko) 2023-08-16

Similar Documents

Publication Publication Date Title
Zhang et al. Evaluation of a novel, integrated approach using functionalized magnetic beads, bench-top MALDI-TOF-MS with prestructured sample supports, and pattern recognition software for profiling potential biomarkers in human plasma
Quinones et al. Metabolomics tools for identifying biomarkers for neuropsychiatric diseases
US8175816B2 (en) System and method for analyzing metabolomic data
CN108603859B (zh) 尿中代谢物在制备癌的评价方法所使用的试剂盒中的用途
US20050101023A1 (en) Methods for diagnosing urinary tract and prostatic disorders
Guleken et al. Identification of polycystic ovary syndrome from blood serum using hormone levels via Raman spectroscopy and multivariate analysis
JP2019105456A (ja) 尿中代謝物におけるバイオマーカー探索法
Sotirakis et al. Identification of motor progression in Parkinson’s disease using wearable sensors and machine learning
WO2023153635A1 (fr) Procédé et dispositif de classification de maladie inflammatoire utilisant une analyse spectroscopique raman basée sur l'apprentissage automatique
JP2007285922A (ja) 近赤外光を用いた臨床血液検査方法
Risoluti et al. Development of a “single-click” analytical platform for the detection of cannabinoids in hemp seed oil
Adigal et al. Protein profile analysis of tear fluid with hyphenated HPLC-UV LED-induced fluorescence detection for the diagnosis of dry eye syndrome
CN1965232A (zh) 质谱分析装置、质谱分析方法和质谱分析程序
US20220108099A1 (en) Maturity classification of stained reticulocytes using optical microscopy
WO2007066589A1 (fr) Procédé et appareil pour examiner et diagnostiquer une maladie liée au mode de vie utilisant une spectroscopie de proche infrarouge
RU2723139C2 (ru) Способ дифференциальной диагностики форм хронического тонзиллита с использованием раман-флюоресцентной спектроскопии
WO2015026171A1 (fr) Composition pour diagnostiquer le cancer des ovaires, contenant de la lysophosphatidylcholine et de l'acide homocystéique, et procédé pour diagnostiquer le cancer des ovaires utilisant celle-ci
JP2010117290A (ja) 液体試料成分の分析装置および分析方法
JP2007513399A (ja) 生化学画像の生成及びその使用方法
WO2006123611A1 (fr) Procede et dispositif de diagnostic du syndrome de fatigue chronique (sfc) utilisant un spectre de proche infrarouge
Naylor et al. Use of the Accutrend Plus point-of-care monitor for blood triglyceride measurement in horses
Zuleika Cross-Sectional Study as Research Design in Medicine
US20230194532A1 (en) Quantum optics profiles for screening, diagnosis, and prognosis of diseases
CN116500179B (zh) 一种血浆中神经酰胺相关因子的原始得分计算及转换方法和系统
EP4212864A1 (fr) Méthode de test du cancer utilisant une liste de métabolites

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22926231

Country of ref document: EP

Kind code of ref document: A1