WO2024047910A1 - Disease risk calculation system and method - Google Patents

Disease risk calculation system and method Download PDF

Info

Publication number
WO2024047910A1
WO2024047910A1 PCT/JP2023/006903 JP2023006903W WO2024047910A1 WO 2024047910 A1 WO2024047910 A1 WO 2024047910A1 JP 2023006903 W JP2023006903 W JP 2023006903W WO 2024047910 A1 WO2024047910 A1 WO 2024047910A1
Authority
WO
WIPO (PCT)
Prior art keywords
saliva
disease risk
physical property
salivary
metabolite
Prior art date
Application number
PCT/JP2023/006903
Other languages
French (fr)
Inventor
Masahiro Sugimoto
Kanae Saito
Original Assignee
Salivatech Co., Ltd.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Salivatech Co., Ltd. filed Critical Salivatech Co., Ltd.
Publication of WO2024047910A1 publication Critical patent/WO2024047910A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • 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
    • G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment

Definitions

  • the present invention accurately predicts the risks of the current disease of a saliva donor using a combination of physical and chemical properties, such as transparency of saliva and chemical components.
  • ALT alanine aminotransferase
  • AST aspartate aminotransferase
  • Metabolites organic molecules smaller than protein, have also been used as molecular biomarkers, e.g., blood creatinine concentrations evaluate kidney function. Indexes combining multiple markers have also been developed to calculate cancer risks, showing higher accuracy than using a single metabolite.
  • Salivary cortisol one of the metabolites, has been widely used as a stress biomarker.
  • Various salivary metabolites were identified as biomarkers of diseases in the oral cavity (such as periodontal disease), respiratory diseases, and systemic diseases (e.g., cancers) (see, for example, Non-Patent Literatures 1 to 3).
  • the inventors have previously compared metabolomic profiles (i.e., the concentration pattern of metabolites) of various biofluid samples and invented methodologies to discriminate patients with multiple cancers from non-cancer subjects using salivary metabolites (Patent Literatures 1 to 4).
  • Saliva is a non-invasively available biofluid that does not limit the place of sample collection and suite the frequent tests for cancer detection, especially for improving medical examination rates and realizing early detection of various cancers. Recent studies also have yielded biomarkers to diagnose a wide variety of diseases rather than cancers.
  • Hyvarinen E Savolainen M, Mikkonen JJW, Kullaa AM. Salivary Metabolomics for Diagnosis and Monitoring Diseases: Challenges and Possibilities. Metabolites. 2021 Aug 31;11(9):587. doi: 10.3390/metabo11090587. PMID: 34564402. Li CX, Zhang L, Yan YR, Ding YJ, Lin YN, Zhou JP, Li N, Li HP, Li SQ, Sun XW, Li QY. A narrative review of exploring potential salivary biomarkers in respiratory diseases: still on its way. J Thorac Dis. 2021 Jul;13(7):4541-4553. doi: 10.21037/jtd-21-202.
  • Patent Literature 1 proposed a method to use the alanine concentration for dividing salivary metabolite concentrations; however, using a single metabole alone is insufficient to eliminate the overall fluctuation.
  • Saliva specimens frequently show high cloudiness and high viscosity. The conventional method is difficult to eliminate these effects of the changes in salivary physical properties.
  • the present invention enables accurate disease risk calculation using saliva specimens.
  • the combination of multiple metabolites and physical properties, i.e., cloudiness and transparency, are used to eliminate fluctuation of the overall concentration of saliva.
  • the conventional method using only a single metabolite for the fluctuation elimination does not function when this metabolite is not detected.
  • the use of multiple metabolites solves this problem, which also drastically improves the risk estimating accuracies.
  • the first invention for achieving the above-described object is a disease risk calculation system that calculates a disease risk using saliva, and includes: a photographing device that photographs a specimen container containing the saliva and obtains image data; a concentration measurement device that measures salivary metabolite concentration data; and at least one computer that processes the image data and the metabolite concentration data, and the computer includes a physical property evaluation unit that evaluates a salivary physical property as to whether the saliva is transparent or cloudy based on the image data, a concentration correction unit that corrects the metabolite concentration data based on the evaluation of the salivary physical property, a concentration normalization unit that normalizes the corrected metabolite concentration data using multiple metabolite concentration data for normalization, and a disease risk calculation unit that calculates the disease risk based on the normalized metabolite concentration data.
  • the concentration correction unit corrects the metabolite concentration data by multiplying the metabolite concentration data with correction coefficients.
  • the specimen container includes a bottom part and a tubular part that extends from the bottom part
  • the image data includes the bottom part of the specimen container and part of the tubular part
  • the physical property evaluation unit extracts pixel values of the inside of the specimen container along with the vertical direction and evaluates the salivary physical property using the pixel values.
  • the image data includes a horizontal direction that is perpendicular to the vertical direction, and the physical property evaluation unit repeats, along the vertical direction, extraction of multiple pixel values in the horizontal direction and calculation of statistical values, and evaluates the salivary physical property using the statistical values.
  • the second invention is a disease risk calculation method for calculating a disease risk using saliva, and is the disease risk calculation method that includes: a step of, at a photographing device, photographing a specimen container containing the saliva and obtaining image data; a step of, at a computer, evaluating a salivary physical property as to whether the saliva is transparent or cloudy based on the image data; a step of, at a concentration measurement device, measuring salivary metabolite concentration data; a step of, at the computer, correcting the metabolite concentration data based on the evaluation of the salivary physical property; a step of, at the computer, normalizing the corrected metabolite concentration data using multiple metabolite concentration data for normalization; and a step of, at the computer, calculating the disease risk based on the normalized metabolite concentration data.
  • the present invention normalizes the salivary metabolite concentration by combining salivary physical properties and multiple salivary metabolite concentrations, which enables accurate disease risk estimation even in the use of saliva, whose overall concentration shows larger fluctuations than those of blood and urine.
  • Fig. 1 is an overview of a disease risk calculation system.
  • Fig. 2 is a flowchart of a disease risk calculation system.
  • Fig. 3 is a schematic view of a photographing device for a specimen container.
  • Fig. 4 is a flowchart for the physical property evaluation processing.
  • Fig. 5 shows examples of metabolite concentration data and correction coefficients.
  • Fig. 6 shows examples of image data and pixel value data.
  • Fig. 7 shows receiver operating characteristic (ROC) curves and the area under ROC curves (AUCs) to discriminate patients with cancer diseases from healthy controls.
  • Fig. 8 shows box plots of disease risks of prediction models.
  • Fig. 9 shows box plots of test data.
  • ROC receiver operating characteristic
  • Fig. 1 depicts an overview of a disease risk calculation system.
  • System 1 is a disease risk calculation system using saliva and comprises Units 2 to 6.
  • Unit 2 is a photographing device and photographs a specimen container containing the saliva and obtains image data.
  • Unit 3 is a concentration measurement device that measures salivary metabolite concentration data.
  • Units 4 to 6 are computers and process the image data and the metabolite concentration data.
  • Unit 4 is a first terminal.
  • Unit 5 is a second terminal.
  • Unit 6 is a server.
  • the computer consists of a central processing unit (CPU) to control the system, memories as primary storage, a hard disk drive (HDD) or flash memories as additional storage, displays to show the information, keyboards and mouses to input information, a touch panel display, and network connection with a local area network (LAN) or wireless module.
  • An Operating System (OS) application programs, and the information for data processing are stored in the HDD or the flash memory.
  • the CPU reads the OS and the application programs from main and additional storage and controls other equipment to execute the following procedures.
  • the computer transfers the data via a network 9.
  • the photographing device (unit 2) attached to the terminal (unit 4) transfers image data by universal serial bus (USB) memory and/or the network (unit 9).
  • USB universal serial bus
  • the measurement device (unit 3) consists of, e.g., a mass spectrometry (MS) and computers.
  • the MS measures metabolites in preprocessed saliva.
  • the measurement results are transferred to the computer via the network (unit 9).
  • the computer of unit 3 calculates the metabolite concentrations using data analysis software.
  • the calculation method of metabolite concentration is available in Patent Literatures 1 to 4, etc.
  • the evaluations of salivary physical properties following the present invention should be conducted before the saliva processing.
  • the measurement device (unit 3) transfers the data to the server (unit 6).
  • the first terminal (unit 4) with an application program evaluates the salivary physical property as to whether the saliva is transparent or cloudy using the image data collected by the photographing device (unit 2).
  • the first terminal (unit 4) transfers the image and analyzed data to the server (unit 6).
  • the second terminal (unit 5) with an application program has three functions. Firstly, the second terminal (unit 5) corrects the metabolite concentration data based on the salivary physical property. Secondly, the second terminal (unit 5) further normalizes the corrected metabolite concentration data using multiple metabolite concentration data for normalization to eliminate the fluctuation of overall concentrations. Thirdly, the second terminal (unit 5) calculates the disease risk using the normalized metabolite concentration data. The second terminal (unit 5) transfers these data to the server (unit 6).
  • the server (unit 6) stores the metabolite concentration data, the image data, the physical property, and the disease risks.
  • the server also sends the data by responding to the corresponding requirements.
  • the disease risks are, e.g., printed out and sent to the saliva donors.
  • the number of used computers has no limitations, i.e., one or more computers are acceptable.
  • the analytical software and application can be installed into a or multiple computer(s). Instead of the application problems in terminals (units 4 and 5), the server (unit 6) or cloud service (not described in the figure) also can provide the same functions.
  • Fig. 2 illustrates a flowchart of the disease risk calculation system.
  • photographing device 2 obtains image data of a specimen container containing saliva.
  • Fig. 3 depicts a schematic view of a photographing instrument.
  • a saliva specimen container 10 includes a bottom part 11, a tubular part 12 that extends from the bottom part 11, and a cover part 13. Image data is obtained under the condition that specimen container 10 is in a photographing box 30 to protect environmental lighting.
  • the photographing box 30 includes a box part 31 whose back, upper, lower, left, and right are covered by light shielding plates and a supporting part 32 that is coupled to an interior of the box part 31 to place the specimen container 10.
  • the specimen container 10 is fixed to a specific position by the supporting part 32.
  • the supporting part 32 supports the specimen container 10, such that an extension direction of the tubular part 12 of specimen container 10 is a substantially vertical direction.
  • the photographing range R includes the bottom part 11 of the specimen container 10 and part of the tubular part 12, and preferably includes a liquid surface 21 of the saliva specimen 20.
  • step S2 the terminal 4 evaluates the salivary physical property using the image data obtained in step S1.
  • the terminal 4 evaluates the transparency and cloudiness of the saliva specimen contained in the specimen container 10 using image data. That is, a physical property evaluation unit of the termninal 4 evaluates the salivary physical property as to whether the saliva is transparent or cloudy.
  • Fig. 4 is a flowchart illustrating the flow of physical property evaluation processing.
  • the terminal 4 extracts the pixel values of the inside of the specimen container 10 in the image data and evaluates the physical property of the saliva 20 using these values.
  • the terminal 4 reads the image data when image data from the photographing device 2 is received (step S11).
  • the terminal 4 initializes the calculation reference position (step S12) and calculates a statistical value of pixel values in the surroundings of the calculation reference position (step S13).
  • a range of the surroundings of the calculation reference position is, e.g., a range of a square of eight pixels in a horizontal direction and eight pixels in a vertical direction at the calculation reference position.
  • the statistical value of the pixel values is, for example, an average value of the pixel values in the range of the square of 8 ⁇ 8 pixels, but not limited to this. Another statistical criterion, e.g., median value, is also acceptable.
  • step S14 the terminal 4 confirms whether or not the calculation reference position has reached an end position.
  • the terminal 4 shifts and sets the next calculation reference position along with the vertical direction (step S15) and repeats the flow from step S13.
  • the terminal 4 calculates a sum of all statistical values (step S16).
  • step S17 compares the sum of the statistical values of the pixel values and a predefined threshold.
  • the purpose of the processing in step S17 is described here.
  • the sum of the pixel values becomes drastically higher in a cloudy specimen than in a transparent specimen. Thus, this can be used as a criterion for distinguishing these specimens.
  • the sum of the statistical values of the pixel values determines the transparency or cloudiness of the saliva 20.
  • the terminal 4 regards the physical property of the saliva 20 as cloudiness. Thus, the metabolite concentration data needs to be corrected and transferred to the server 6 (step S18). On the other hand, in a case where the sum of the statistical values is less than the threshold (‘No’ in step S17), the terminal 4 decides that the physical property of the saliva 20 is evaluated as transparent. In this case, the metabolite concentration data does not need to be corrected and transferred to the server 6 (step S19).
  • the terminal 4 extracts the pixel values of the inside of the specimen container 10 along with the vertical direction of the image data and evaluates the physical property of the saliva 20 using the extracted pixel values (Fig. 4).
  • the pixel values along the vertical direction enable accurate evaluation the physical property of the saliva 20 even though only partial area of the saliva shows cloudiness.
  • the terminal 4 repeats, along with the vertical direction, extraction of multiple pixel values in the horizontal direction, and calculation of the statistical values.
  • the terminal 4 evaluates the physical property of the saliva 20 using the calculated statistical values.
  • the pixel values are computed along with the vertical and horizontal directions. This approach enables robust evaluations against outliers and abnormalities observed in a partial area. For example, the reflected light of a photograph is reflected by the specimen container 10 and becomes partially white, i.e., the outliers.
  • the measurement device 3 measures the metabolite concentrations in the saliva 20.
  • the measurement device 3 includes mass spectrometry and the computer.
  • the mass spectrometry measures preprocessed saliva specimens, and the computer calculates the concentration of each metabolite, and these metabolite concentration data are transferred to the server 6.
  • a concentration correction unit of the terminal 5 corrects the metabolite concentration data based on the evaluation of the salivary physical property.
  • Step 2 evaluates the physical property of the saliva 20.
  • the concentration correction unit corrects the metabolite concentration data using the predefined correction coefficients.
  • the concentration correction unit multiplies the metabolite concentration data with the correction coefficients.
  • a concentration normalization unit of the terminal 5 normalizes the corrected metabolite concentration data. When the correction is not required, the concentration normalization unit uses the metabolite concentration data without correction for the subsequent processes.
  • step S4 The purpose of step S4 is described here.
  • the saliva 20 collected from the same person at the same period revealed that the concentrations of the metabolites in cloudy saliva 20 were higher than those in transparent saliva 20.
  • the disease risk calculation processing in step S6 is affected by the abnormal metabolite concentrations caused by the cloudiness of the saliva 20 and may lead to the wrong risk calculation.
  • the metabolite concentration data obtained by the measurement device 3 is normalized by multiplying with correction coefficients ranging from 0 to 1 when the saliva shows cloudiness, which leads to accurate risk disease predictions.
  • step S5 the concentration normalization unit of the terminal 5 normalizes the metabolite concentration data corrected as needed.
  • the metabolite concentrations of the saliva 20 show an individual difference in saliva cloudiness. An overall concentration of the saliva 20 is normalized, which would be a more significant difference between the disease and control groups.
  • the concentration normalization unit normalizes the metabolite concentration using multiple metabolite concentration data for normalization. In the embodiment of the present invention, the concentration normalization unit normalizes the metabolite concentration by dividing the multiple metabolite concentration data for normalization.
  • the terminal 5 multiplies each of the metabolite concentration data for disease risk calculation and the metabolite concentration data for normalization with the correction coefficient. This procedure realizes accurate disease risk calculations.
  • Fig. 5 shows examples of metabolite concentrations and the correction coefficients.
  • Fig. 5(a) shows metabolite concentrations
  • Fig. 5(b) shows the correction coefficients.
  • the correction coefficients are ranged from 0 to 1.
  • the concentration correction unit multiplies the concentration of the sum of normalization metabolites (M 1 ) and each metabolite for risk calculation (M 2 , M 3 , ..., and M n ) by the correlation coefficients, i.e., X 1 ⁇ C 1 , X 2 ⁇ C 2 , X 3 ⁇ C 3 , ... and, X n ⁇ C n (X n indicates the concentration of M n and C n means the corresponding correlation coefficients).
  • the concentration normalization unit divides these data by the corrected sum of normalization metabolites, i.e., X 2 ⁇ C 2 /(X 1 ⁇ C 1 ), X 3 ⁇ C 3 /(X 1 ⁇ C 1 ), ... and X n ⁇ C n /(X 1 ⁇ C 1 ). Otherwise, when step 2 does not require the correction, the metabolite concentrations for risk calculation are divided by the sum of the normalization metabolites, i.e., X 2 /X 1 , X 3 /X 1 , ... and X n /X 1 .
  • a disease risk calculation unit of the terminal 5 calculates the disease risk using multivariate analysis and machine learning methods using corrected and normalized metabolite concentrations. For example, the disease risk calculation unit calculates a risk value using the machine learning-based prediction model and considers the analytical result as a positive case when the risk value exceeds a threshold.
  • the computer includes the physical property evaluation unit to evaluate the physical property of the saliva 20 using the image data.
  • This computer also includes the concentration correction unit to correct the metabolite concentration data depending on the physical property of the saliva 20, and includes the concentration normalization unit to normalize the processed concentration with multiple metabolites for normalization.
  • This computer also includes the disease risk calculation unit to calculate the disease risk using the processed metabolite concentrations. Consequently, the accurate disease risk calculation is realized even in the saliva showing a specific physical property for which conventional methods miscalculate the risk.
  • This example shows the calculated results of cancer risks by the above-described system 1.
  • the results are compared with the analytical results using a conventional method that does not normalize and correlate the salivary metabolite concentrations using salivary physical properties.
  • Fig. 6 shows examples of image data and pixel value data.
  • image data 41 and 42 include the vertical direction along with the tubular part 12 and the horizontal direction. Both image data 41 and 42 include 400 pixels in the horizontal direction and 300 pixels in the vertical direction. They also include the bottom part 11 of the specimen container 10 and part of the tubular part 12 from the bottom part 11 to a position above the liquid surface 21 of the saliva 20.
  • Image data 41 and 42 in Figs. 6(a) and 6(b) shows transparent and cloudy saliva 20, respectively.
  • the pixel value data 51 is associated with image data 41
  • the pixel value data 52 is associated with image data 42.
  • the reference position for calculating pixel values was set as a square range of eight pixels in the horizontal direction and eight in the vertical direction. Furthermore, an averaged value in the square range of 8 ⁇ 8 pixels was used as a statistical value of the pixel values.
  • An end position of the calculation reference position was a position of pixel 200 in the lateral direction and a position of pixel 250 in the longitudinal direction.
  • the sum of the pixel values from pixel 150 to pixel 250 equals a waveform area from pixel 150 to pixel 250 in the graphs shown in Figs. 6(b) and 6(d).
  • This waveform area was drastically larger in the cloudy specimen (Fig. 6(d)) than in the transparent specimen (Fig. 6(b)). Therefore it was used as a criterion to evaluate the cloudiness of the specimens. Whether a specimen was transparent or cloudy was decided based on a sum of statistical values of these pixel values.
  • a preprocessing method for measuring metabolite substances of saliva includes collecting 20 ⁇ L of a saliva sample, mixing the saliva sample with 1% of ammonium hydroxide and 2.5 ⁇ M of each stable isotope in a methanol solution, and centrifuging the mixture at 4°C and at 12,000 rpm for 10 minutes. 40 ⁇ L of the supernatant was collected, mixed with 60 ⁇ L of ultrapure water, and obtained as a sample.
  • Table 1 shows the correction coefficient of a sum of metabolites for normalization for stomach cancer risk calculation and correction coefficients of stomach cancer markers, variables for a prediction model. The sum of creatinine and lysine concentrations was used for the normalization.
  • the stomach cancer markers included five metabolites, including N1,N8-diacetylspermidine, N1-acetylspermidine, N8-acetylspermidine, spermidine, and spermine.
  • the data of 41 patients with stomach cancer and 73 controls were used for the cancer risk calculation processing.
  • a machine learning method-based prediction model using bagging-ADTree was used, and the sensitivity of 80% was used as a threshold to detect positive cases.
  • an area under the receiver operating characteristic curve (AUC) was used to evaluate the prediction model.
  • the AUC is the area under the receiver operating characteristic (ROC) curve.
  • the AUC value ranges from 0 to 1, and a higher value indicates a higher discrimination accuracy.
  • the AUC value of the comparative example was 0.975 (Fig. 7(a)), and that of the example was 0.996 (Fig. 7(b)).
  • the AUC value using 10-fold cross-validation was 0.894 in the comparative example and 0.898 in the example (Fig. 7(c)). Both comparisons revealed higher AUC values in the example than in the comparative example.
  • Figs. 8(a) and 8(b) show box plots of predicted values in the comparative example and the example, respectively.
  • Figs. 9(a) and 9 (b) show the box plots of calculated risk values of test data in the comparison example and the example.
  • the calculated risks of cloudy specimens were significantly higher than those of transparent specimens among control data (Kruskal-Wallis test). Such a significant difference was not observed in the example.
  • Table 2 shows the comparison of calculated risks in the example and the comparative example regarding stomach cancer data.
  • Fig. 9(c) shows the bar graphs of the comparison of accuracy between the example and the comparative example.
  • accuracy substantially improved in the example compared to the comparative example.
  • the sensitivity improved from 50% to 67%, and the specificity improved from 61% to 97%.
  • Table 3 shows the correction coefficients of a sum of metabolites for calculating colorectal cancer risks and correction coefficients of normalization used for a prediction model. The sum of creatinine and arginine concentrations was used as normalization.
  • Table 4 shows the calculated risks of colorectal cancers in the example and the comparative example.
  • Table 5 shows a correction coefficient and correction coefficients of normalization for lung cancer used for the prediction model. The sum of creatinine and adenosine was used as normalization.
  • Table 6 shows the calculated risks of lung cancers in the example and the comparative example.
  • Tables 4 and 6 show the accuracy rates (sensitivity and specificity) of both colorectal cancer and lung cancer risk calculations substantially improved in the example compared to the comparative example; the sensitivity and specificity improved from 76% to 82% and from 69% to 72%, respectively, for colorectal cancer and from 80% to 87% and from 47% to 56% for the lung cancers.
  • the disease risk calculation system 1 can reduce false positives and negatives without collecting further saliva that would show transparency.
  • the system accurately calculates the cancer risks even in the saliva shows cloudiness.

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Epidemiology (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Biomedical Technology (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Investigating Or Analysing Biological Materials (AREA)

Abstract

The present invention accurately predicts the risks of the current disease of a saliva donor. A disease risk calculation system 1 is to calculate a disease risk using saliva. This system includes a photographing device 2 to obtain image data of a specimen container containing the saliva, a device 3 to measure the salivary metabolite concentrations, and one or multiple computers for processing the image data and the metabolite concentrations. The computer includes a unit to evaluate the salivary physical property based on the image data using the photographing device 2, a unit to correct the metabolite concentration data based on the salivary physical property, a unit to normalize the corrected metabolite concentrations and a unit to calculate the disease risk based on the normalized metabolite concentrations.

Description

DISEASE RISK CALCULATION SYSTEM AND METHOD
The present invention accurately predicts the risks of the current disease of a saliva donor using a combination of physical and chemical properties, such as transparency of saliva and chemical components.
Various diagnostic methodologies using biofluids, such as blood or urine, have been developed to estimate a health status. Especially liquid biopsies that use the endogenous information of cells or molecular markers to assess disease risks have been intensively developed. Examples of common blood biomarkers include alanine aminotransferase (ALT) and aspartate aminotransferase (AST) for evaluating liver health status. Another example s are various tumor markers for diagnosing cancers. Metabolites, organic molecules smaller than protein, have also been used as molecular biomarkers, e.g., blood creatinine concentrations evaluate kidney function. Indexes combining multiple markers have also been developed to calculate cancer risks, showing higher accuracy than using a single metabolite.
Salivary cortisol, one of the metabolites, has been widely used as a stress biomarker. Various salivary metabolites were identified as biomarkers of diseases in the oral cavity (such as periodontal disease), respiratory diseases, and systemic diseases (e.g., cancers) (see, for example, Non-Patent Literatures 1 to 3).
The inventors have previously compared metabolomic profiles (i.e., the concentration pattern of metabolites) of various biofluid samples and invented methodologies to discriminate patients with multiple cancers from non-cancer subjects using salivary metabolites (Patent Literatures 1 to 4). Saliva is a non-invasively available biofluid that does not limit the place of sample collection and suite the frequent tests for cancer detection, especially for improving medical examination rates and realizing early detection of various cancers. Recent studies also have yielded biomarkers to diagnose a wide variety of diseases rather than cancers.
JP 6443937 B2 JP 6659808 B2 JP 6659809 B2 JP 6851096 B2
Hyvarinen E, Savolainen M, Mikkonen JJW, Kullaa AM. Salivary Metabolomics for Diagnosis and Monitoring Diseases: Challenges and Possibilities. Metabolites. 2021 Aug 31;11(9):587. doi: 10.3390/metabo11090587. PMID: 34564402. Li CX, Zhang L, Yan YR, Ding YJ, Lin YN, Zhou JP, Li N, Li HP, Li SQ, Sun XW, Li QY. A narrative review of exploring potential salivary biomarkers in respiratory diseases: still on its way. J Thorac Dis. 2021 Jul;13(7):4541-4553. doi: 10.21037/jtd-21-202. PMID: 34422380. Panneerselvam K, Ishikawa S, Krishnan R, Sugimoto M. Salivary Metabolomics for Oral Cancer Detection: A Narrative Review. Metabolites. 2022 May 12;12(5):436. doi: 10.3390/metabo12050436. PMID: 35629940.
The use of saliva biomarkers presents a technical problem. Fluctuation of overall molecular concentration in a saliva specimen is frequently observed. This fluctuation should be eliminated before the use of biomarker concentrations. The same problem using urine specimens has already been overcome; the urinary metabolite concentrations are divided by creatinine concentration; however, such a method for saliva specimens has not been established. Patent Literature 1 proposed a method to use the alanine concentration for dividing salivary metabolite concentrations; however, using a single metabole alone is insufficient to eliminate the overall fluctuation. Saliva specimens frequently show high cloudiness and high viscosity. The conventional method is difficult to eliminate these effects of the changes in salivary physical properties.
Thus, the present invention enables accurate disease risk calculation using saliva specimens. The combination of multiple metabolites and physical properties, i.e., cloudiness and transparency, are used to eliminate fluctuation of the overall concentration of saliva. The conventional method using only a single metabolite for the fluctuation elimination does not function when this metabolite is not detected. The use of multiple metabolites solves this problem, which also drastically improves the risk estimating accuracies.
The first invention for achieving the above-described object is a disease risk calculation system that calculates a disease risk using saliva, and includes: a photographing device that photographs a specimen container containing the saliva and obtains image data; a concentration measurement device that measures salivary metabolite concentration data; and at least one computer that processes the image data and the metabolite concentration data, and the computer includes a physical property evaluation unit that evaluates a salivary physical property as to whether the saliva is transparent or cloudy based on the image data, a concentration correction unit that corrects the metabolite concentration data based on the evaluation of the salivary physical property, a concentration normalization unit that normalizes the corrected metabolite concentration data using multiple metabolite concentration data for normalization, and a disease risk calculation unit that calculates the disease risk based on the normalized metabolite concentration data.
Furthermore, preferably, in a case where the physical property evaluation unit evaluates that the saliva is cloudy, the concentration correction unit corrects the metabolite concentration data by multiplying the metabolite concentration data with correction coefficients.
Furthermore, preferably, the specimen container includes a bottom part and a tubular part that extends from the bottom part, the image data includes the bottom part of the specimen container and part of the tubular part, and a vertical direction along with the tubular part, and the physical property evaluation unit extracts pixel values of the inside of the specimen container along with the vertical direction and evaluates the salivary physical property using the pixel values.
Furthermore, preferably, the image data includes a horizontal direction that is perpendicular to the vertical direction, and the physical property evaluation unit repeats, along the vertical direction, extraction of multiple pixel values in the horizontal direction and calculation of statistical values, and evaluates the salivary physical property using the statistical values.
The second invention is a disease risk calculation method for calculating a disease risk using saliva, and is the disease risk calculation method that includes: a step of, at a photographing device, photographing a specimen container containing the saliva and obtaining image data; a step of, at a computer, evaluating a salivary physical property as to whether the saliva is transparent or cloudy based on the image data; a step of, at a concentration measurement device, measuring salivary metabolite concentration data; a step of, at the computer, correcting the metabolite concentration data based on the evaluation of the salivary physical property; a step of, at the computer, normalizing the corrected metabolite concentration data using multiple metabolite concentration data for normalization; and a step of, at the computer, calculating the disease risk based on the normalized metabolite concentration data.
The present invention normalizes the salivary metabolite concentration by combining salivary physical properties and multiple salivary metabolite concentrations, which enables accurate disease risk estimation even in the use of saliva, whose overall concentration shows larger fluctuations than those of blood and urine.
Fig. 1 is an overview of a disease risk calculation system. Fig. 2 is a flowchart of a disease risk calculation system. Fig. 3 is a schematic view of a photographing device for a specimen container. Fig. 4 is a flowchart for the physical property evaluation processing. Fig. 5 shows examples of metabolite concentration data and correction coefficients. Fig. 6 shows examples of image data and pixel value data. Fig. 7 shows receiver operating characteristic (ROC) curves and the area under ROC curves (AUCs) to discriminate patients with cancer diseases from healthy controls. Fig. 8 shows box plots of disease risks of prediction models. Fig. 9 shows box plots of test data.
The detail of the present invention is described in the figures. Fig. 1 depicts an overview of a disease risk calculation system. System 1 is a disease risk calculation system using saliva and comprises Units 2 to 6. Unit 2 is a photographing device and photographs a specimen container containing the saliva and obtains image data. Unit 3 is a concentration measurement device that measures salivary metabolite concentration data. Units 4 to 6 are computers and process the image data and the metabolite concentration data. Unit 4 is a first terminal. Unit 5 is a second terminal. Unit 6 is a server.
The computer consists of a central processing unit (CPU) to control the system, memories as primary storage, a hard disk drive (HDD) or flash memories as additional storage, displays to show the information, keyboards and mouses to input information, a touch panel display, and network connection with a local area network (LAN) or wireless module. An Operating System (OS), application programs, and the information for data processing are stored in the HDD or the flash memory. The CPU reads the OS and the application programs from main and additional storage and controls other equipment to execute the following procedures. The computer transfers the data via a network 9.
The photographing device (unit 2) attached to the terminal (unit 4) transfers image data by universal serial bus (USB) memory and/or the network (unit 9).
The measurement device (unit 3) consists of, e.g., a mass spectrometry (MS) and computers. The MS measures metabolites in preprocessed saliva. The measurement results are transferred to the computer via the network (unit 9). The computer of unit 3 calculates the metabolite concentrations using data analysis software. The calculation method of metabolite concentration is available in Patent Literatures 1 to 4, etc. The evaluations of salivary physical properties following the present invention should be conducted before the saliva processing. The measurement device (unit 3) transfers the data to the server (unit 6).
The first terminal (unit 4) with an application program evaluates the salivary physical property as to whether the saliva is transparent or cloudy using the image data collected by the photographing device (unit 2). The first terminal (unit 4) transfers the image and analyzed data to the server (unit 6).
The second terminal (unit 5) with an application program has three functions. Firstly, the second terminal (unit 5) corrects the metabolite concentration data based on the salivary physical property. Secondly, the second terminal (unit 5) further normalizes the corrected metabolite concentration data using multiple metabolite concentration data for normalization to eliminate the fluctuation of overall concentrations. Thirdly, the second terminal (unit 5) calculates the disease risk using the normalized metabolite concentration data. The second terminal (unit 5) transfers these data to the server (unit 6).
The server (unit 6) stores the metabolite concentration data, the image data, the physical property, and the disease risks. The server also sends the data by responding to the corresponding requirements. The disease risks are, e.g., printed out and sent to the saliva donors.
The number of used computers has no limitations, i.e., one or more computers are acceptable. The analytical software and application can be installed into a or multiple computer(s). Instead of the application problems in terminals (units 4 and 5), the server (unit 6) or cloud service (not described in the figure) also can provide the same functions.
Fig. 2 illustrates a flowchart of the disease risk calculation system. In step S1, photographing device 2 obtains image data of a specimen container containing saliva.
Fig. 3 depicts a schematic view of a photographing instrument. A saliva specimen container 10 includes a bottom part 11, a tubular part 12 that extends from the bottom part 11, and a cover part 13. Image data is obtained under the condition that specimen container 10 is in a photographing box 30 to protect environmental lighting. The photographing box 30 includes a box part 31 whose back, upper, lower, left, and right are covered by light shielding plates and a supporting part 32 that is coupled to an interior of the box part 31 to place the specimen container 10. The specimen container 10 is fixed to a specific position by the supporting part 32. The supporting part 32 supports the specimen container 10, such that an extension direction of the tubular part 12 of specimen container 10 is a substantially vertical direction.
An installation position, a photographing direction, a focus, and the like of the photographing device 2 (not illustrated in Fig. 3) are adjusted to enable the photographing device 2 to photograph the specimen container 10 within a predetermined photographing range R. In the embodiment of the present invention, the photographing range R includes the bottom part 11 of the specimen container 10 and part of the tubular part 12, and preferably includes a liquid surface 21 of the saliva specimen 20.
The explanation of Fig. 2 is described again. In step S2, the terminal 4 evaluates the salivary physical property using the image data obtained in step S1. In the embodiment of the present invention, the terminal 4 evaluates the transparency and cloudiness of the saliva specimen contained in the specimen container 10 using image data. That is, a physical property evaluation unit of the termninal 4 evaluates the salivary physical property as to whether the saliva is transparent or cloudy.
Fig. 4 is a flowchart illustrating the flow of physical property evaluation processing. The terminal 4 extracts the pixel values of the inside of the specimen container 10 in the image data and evaluates the physical property of the saliva 20 using these values. As illustrated in Fig. 4, the terminal 4 reads the image data when image data from the photographing device 2 is received (step S11).
The terminal 4 initializes the calculation reference position (step S12) and calculates a statistical value of pixel values in the surroundings of the calculation reference position (step S13). A range of the surroundings of the calculation reference position is, e.g., a range of a square of eight pixels in a horizontal direction and eight pixels in a vertical direction at the calculation reference position. Furthermore, the statistical value of the pixel values is, for example, an average value of the pixel values in the range of the square of 8×8 pixels, but not limited to this. Another statistical criterion, e.g., median value, is also acceptable.
Next, the terminal 4 confirms whether or not the calculation reference position has reached an end position (step S14). When the calculation reference position has not reached the end position (‘No’ in step S14), the terminal 4 shifts and sets the next calculation reference position along with the vertical direction (step S15) and repeats the flow from step S13. On the other hand, in a case where the calculation reference position has reached the end position (‘Yes’ in step S14), the terminal 4 calculates a sum of all statistical values (step S16).
Next, the terminal 4 compares the sum of the statistical values of the pixel values and a predefined threshold (step S17). The purpose of the processing in step S17 is described here. The sum of the pixel values becomes drastically higher in a cloudy specimen than in a transparent specimen. Thus, this can be used as a criterion for distinguishing these specimens. Hence, in the embodiment of the present invention, the sum of the statistical values of the pixel values determines the transparency or cloudiness of the saliva 20.
In the case where the sum of the statistical values is the threshold or more (‘Yes’ in step S17), the terminal 4 regards the physical property of the saliva 20 as cloudiness. Thus, the metabolite concentration data needs to be corrected and transferred to the server 6 (step S18). On the other hand, in a case where the sum of the statistical values is less than the threshold (‘No’ in step S17), the terminal 4 decides that the physical property of the saliva 20 is evaluated as transparent. In this case, the metabolite concentration data does not need to be corrected and transferred to the server 6 (step S19).
As described above, the terminal 4 extracts the pixel values of the inside of the specimen container 10 along with the vertical direction of the image data and evaluates the physical property of the saliva 20 using the extracted pixel values (Fig. 4). The pixel values along the vertical direction enable accurate evaluation the physical property of the saliva 20 even though only partial area of the saliva shows cloudiness.
Furthermore, the terminal 4 repeats, along with the vertical direction, extraction of multiple pixel values in the horizontal direction, and calculation of the statistical values. The terminal 4 evaluates the physical property of the saliva 20 using the calculated statistical values. The pixel values are computed along with the vertical and horizontal directions. This approach enables robust evaluations against outliers and abnormalities observed in a partial area. For example, the reflected light of a photograph is reflected by the specimen container 10 and becomes partially white, i.e., the outliers.
Fig. 2 is explained again. In step S3, the measurement device 3 measures the metabolite concentrations in the saliva 20. As described above, the measurement device 3 includes mass spectrometry and the computer. The mass spectrometry measures preprocessed saliva specimens, and the computer calculates the concentration of each metabolite, and these metabolite concentration data are transferred to the server 6.
In step S4, a concentration correction unit of the terminal 5 corrects the metabolite concentration data based on the evaluation of the salivary physical property. Step 2 evaluates the physical property of the saliva 20. When the correction is required, the concentration correction unit corrects the metabolite concentration data using the predefined correction coefficients. The concentration correction unit multiplies the metabolite concentration data with the correction coefficients. Subsequently, a concentration normalization unit of the terminal 5 normalizes the corrected metabolite concentration data. When the correction is not required, the concentration normalization unit uses the metabolite concentration data without correction for the subsequent processes.
The purpose of step S4 is described here. The saliva 20 collected from the same person at the same period revealed that the concentrations of the metabolites in cloudy saliva 20 were higher than those in transparent saliva 20. The disease risk calculation processing in step S6 is affected by the abnormal metabolite concentrations caused by the cloudiness of the saliva 20 and may lead to the wrong risk calculation. Hence, in the embodiment of the present invention, the metabolite concentration data obtained by the measurement device 3 is normalized by multiplying with correction coefficients ranging from 0 to 1 when the saliva shows cloudiness, which leads to accurate risk disease predictions.
In step S5, the concentration normalization unit of the terminal 5 normalizes the metabolite concentration data corrected as needed. The metabolite concentrations of the saliva 20 show an individual difference in saliva cloudiness. An overall concentration of the saliva 20 is normalized, which would be a more significant difference between the disease and control groups. As detailed, the concentration normalization unit normalizes the metabolite concentration using multiple metabolite concentration data for normalization. In the embodiment of the present invention, the concentration normalization unit normalizes the metabolite concentration by dividing the multiple metabolite concentration data for normalization.
In the case where the physical property evaluation unit evaluates that the saliva is cloudy, the terminal 5 multiplies each of the metabolite concentration data for disease risk calculation and the metabolite concentration data for normalization with the correction coefficient. This procedure realizes accurate disease risk calculations.
Fig. 5 shows examples of metabolite concentrations and the correction coefficients. Fig. 5(a) shows metabolite concentrations, and Fig. 5(b) shows the correction coefficients. The correction coefficients are ranged from 0 to 1. When step 2 requires the correction of metabolite concentrations based on the physical property of the saliva 20, the concentration correction unit multiplies the concentration of the sum of normalization metabolites (M1) and each metabolite for risk calculation (M2, M3, …, and Mn) by the correlation coefficients, i.e., X1×C1, X2×C2, X3×C3, … and, Xn×Cn (Xn indicates the concentration of Mn and Cn means the corresponding correlation coefficients). Subsequently, the concentration normalization unit divides these data by the corrected sum of normalization metabolites, i.e., X2×C2/(X1×C1), X3×C3/(X1×C1), … and Xn×Cn/(X1×C1). Otherwise, when step 2 does not require the correction, the metabolite concentrations for risk calculation are divided by the sum of the normalization metabolites, i.e., X2/X1, X3/X1, … and Xn/X1.
Fig. 2 is explained again. In step S6, a disease risk calculation unit of the terminal 5 calculates the disease risk using multivariate analysis and machine learning methods using corrected and normalized metabolite concentrations. For example, the disease risk calculation unit calculates a risk value using the machine learning-based prediction model and considers the analytical result as a positive case when the risk value exceeds a threshold.
As described, the disease risk calculation system 1 includes the photographing device 2 to obtain image data of the specimen container 10 containing the saliva 20, the concentration measurement device 3 to quantify metabolite concentrations in the saliva 20, at least one computer (= the terminals 4 and 5 and the server 6) to process the image and the metabolite concentration data. The computer includes the physical property evaluation unit to evaluate the physical property of the saliva 20 using the image data. This computer also includes the concentration correction unit to correct the metabolite concentration data depending on the physical property of the saliva 20, and includes the concentration normalization unit to normalize the processed concentration with multiple metabolites for normalization. This computer also includes the disease risk calculation unit to calculate the disease risk using the processed metabolite concentrations. Consequently, the accurate disease risk calculation is realized even in the saliva showing a specific physical property for which conventional methods miscalculate the risk.

Example
This example shows the calculated results of cancer risks by the above-described system 1. The results are compared with the analytical results using a conventional method that does not normalize and correlate the salivary metabolite concentrations using salivary physical properties.
Fig. 6 shows examples of image data and pixel value data. As shown in Figs. 6(a) and 6(c), image data 41 and 42 include the vertical direction along with the tubular part 12 and the horizontal direction. Both image data 41 and 42 include 400 pixels in the horizontal direction and 300 pixels in the vertical direction. They also include the bottom part 11 of the specimen container 10 and part of the tubular part 12 from the bottom part 11 to a position above the liquid surface 21 of the saliva 20. Image data 41 and 42 in Figs. 6(a) and 6(b) shows transparent and cloudy saliva 20, respectively. Pixel value data 51 in Fig. 6(b) and pixel value data 52 in Fig. 6(d) each are a graph whose horizontal axis indicates a pixel position in the longitudinal direction and whose vertical axis indicates a pixel value of a grayscale. The pixel value data 51 is associated with image data 41, and the pixel value data 52 is associated with image data 42.
As shown in the image data 41 in Fig. 6(a) and 42 in Fig. 6(c), a position of pixel 200 in the horizontal direction (= a center position in the horizontal direction) and a pixel 150 in the vertical direction (= a center position in the longitudinal direction) was used as the initial calculation reference position. The reference position for calculating pixel values was set as a square range of eight pixels in the horizontal direction and eight in the vertical direction. Furthermore, an averaged value in the square range of 8×8 pixels was used as a statistical value of the pixel values. An end position of the calculation reference position was a position of pixel 200 in the lateral direction and a position of pixel 250 in the longitudinal direction. The sum of the pixel values from pixel 150 to pixel 250 equals a waveform area from pixel 150 to pixel 250 in the graphs shown in Figs. 6(b) and 6(d). This waveform area was drastically larger in the cloudy specimen (Fig. 6(d)) than in the transparent specimen (Fig. 6(b)). Therefore it was used as a criterion to evaluate the cloudiness of the specimens. Whether a specimen was transparent or cloudy was decided based on a sum of statistical values of these pixel values.
A preprocessing method for measuring metabolite substances of saliva includes collecting 20 μL of a saliva sample, mixing the saliva sample with 1% of ammonium hydroxide and 2.5 μM of each stable isotope in a methanol solution, and centrifuging the mixture at 4℃ and at 12,000 rpm for 10 minutes. 40 μL of the supernatant was collected, mixed with 60 μL of ultrapure water, and obtained as a sample.
The concentration of each metabolite of the saliva specimen was measured using a liquid chromatography-mass spectrometer (Triple Quad LC/MS). A concentration ratio of cloudy and normal saliva specimens was used as a correction coefficient. The measurement method is described below.
1) Cationic metabolite measurement mode
<LC>
System: Agilent Technologies 1290 Infinity
Column: ACQUITY BEH C18 (inner diameter: 2.1 mm × 50 mm, 1.7 mm)
Solvent:
Pump A; Water containing 0.1% Formic acid and 1.5 mM HFBA
Pump B; MeOH containing 1.5 mM HFBA
Gradient:
Time Pump A
1.00 min 99.00%
2.00 min 90.00%
3.50 min 60.00%
4.00 min 5.00%
5.00 min 5.00%
Measurement time: 5 minutes
Flow rate: 0.4 ml/min
Column temperature: 40℃
<MS>
System : Agilent Technologies 6460 (QQQ)
Nitrogen gas temperature: 350℃
Nitrogen gas flow rate: 13 L/min
Nebulizer: 55 psig
VCap: 3500
Measurement mode: Positive
Twelve subjects who provided saliva showing high cloudiness were required to provide it again after light rinsing their mouths for 5 min or more. Table 1 shows the correction coefficient of a sum of metabolites for normalization for stomach cancer risk calculation and correction coefficients of stomach cancer markers, variables for a prediction model. The sum of creatinine and lysine concentrations was used for the normalization. The stomach cancer markers included five metabolites, including N1,N8-diacetylspermidine, N1-acetylspermidine, N8-acetylspermidine, spermidine, and spermine.
Figure JPOXMLDOC01-appb-T000001
The data of 41 patients with stomach cancer and 73 controls were used for the cancer risk calculation processing.
A machine learning method-based prediction model using bagging-ADTree was used, and the sensitivity of 80% was used as a threshold to detect positive cases. In addition, an area under the receiver operating characteristic curve (AUC) was used to evaluate the prediction model. The AUC is the area under the receiver operating characteristic (ROC) curve. The AUC value ranges from 0 to 1, and a higher value indicates a higher discrimination accuracy.
The AUC value of the comparative example was 0.975 (Fig. 7(a)), and that of the example was 0.996 (Fig. 7(b)). The AUC value using 10-fold cross-validation was 0.894 in the comparative example and 0.898 in the example (Fig. 7(c)). Both comparisons revealed higher AUC values in the example than in the comparative example.
Figs. 8(a) and 8(b) show box plots of predicted values in the comparative example and the example, respectively. The comparison between Figs. 8(a) and 8(b) revealed that the calculated risk values of a cloudy specimen (n = 46) of controls were lower in the example than in the comparative example, indicating the reduced false positives in the example.
Another comparison of patients with stomach cancer (n=6) and controls (n=33) was used as test data. The prediction model using the Bagging-ADTree prediction model (Figs. 7(a) and 7(b)) was evaluated using the test data. The calculated results of the example and the comparative example were compared.
Figs. 9(a) and 9 (b) show the box plots of calculated risk values of test data in the comparison example and the example. The comparison between Figs. 9(a) and 9(b) revealed that the calculated risk values of a cloudy specimen (n = 25) of controls were lower in the example than in the comparative example, indicating the reduced false positives in the example. In the comparative example, the calculated risks of cloudy specimens were significantly higher than those of transparent specimens among control data (Kruskal-Wallis test). Such a significant difference was not observed in the example.
Table 2 shows the comparison of calculated risks in the example and the comparative example regarding stomach cancer data. Fig. 9(c) shows the bar graphs of the comparison of accuracy between the example and the comparative example. As shown in Table 2 and Fig. 9(c), accuracy (sensitivity and specificity) substantially improved in the example compared to the comparative example. The sensitivity improved from 50% to 67%, and the specificity improved from 61% to 97%.
Figure JPOXMLDOC01-appb-T000002
The comparisons using the data of patients with colorectal cancer and lung cancer were also conducted. Cancer biomarkers and normalization metabolites depend on the cancer type.
Table 3 shows the correction coefficients of a sum of metabolites for calculating colorectal cancer risks and correction coefficients of normalization used for a prediction model. The sum of creatinine and arginine concentrations was used as normalization. Table 4 shows the calculated risks of colorectal cancers in the example and the comparative example. Table 5 shows a correction coefficient and correction coefficients of normalization for lung cancer used for the prediction model. The sum of creatinine and adenosine was used as normalization. Table 6 shows the calculated risks of lung cancers in the example and the comparative example.
Figure JPOXMLDOC01-appb-T000003
Figure JPOXMLDOC01-appb-T000004
Figure JPOXMLDOC01-appb-T000005
Figure JPOXMLDOC01-appb-T000006
Tables 4 and 6 show the accuracy rates (sensitivity and specificity) of both colorectal cancer and lung cancer risk calculations substantially improved in the example compared to the comparative example; the sensitivity and specificity improved from 76% to 82% and from 69% to 72%, respectively, for colorectal cancer and from 80% to 87% and from 47% to 56% for the lung cancers.
According to the embodiment of the present invention, the disease risk calculation system 1 can reduce false positives and negatives without collecting further saliva that would show transparency. The system accurately calculates the cancer risks even in the saliva shows cloudiness.
Although the preferred embodiment of the present invention showed examples to calculate cancer risks, the invention is not limited to the example. Anyone with common knowledge and skills may modify the examples within the scope of the technical idea of the present invention. Therefore, these modified examples also belong to the technical range of the present invention in nature.
1 ... disease risk calculation system
2 ... photographing device
3 ... concentration measurement device
4 ... first terminal
5 ... second terminal
6 ... server
9 ... network
10 ... specimen container
11 ... bottom part
12 ... tubular part
20 ... saliva
21 ... liquid surface
30 ... photographing box
31 ... box part
32 ... support part
41 and 42 ... image data
51 and 52 ... pixel value data

Claims (5)

  1. A disease risk calculation system that calculates a disease risk using saliva, the disease risk calculation system comprising:
    a photographing device that photographs a specimen container containing the saliva and obtains image data;
    a concentration measurement device that measures salivary metabolite concentration data; and
    at least one computer that processes the image data and the metabolite concentration data, wherein
    the computer includes
    a physical property evaluation unit that evaluates a salivary physical property as to whether the saliva is transparent or cloudy based on the image data,
    a concentration correction unit that corrects the metabolite concentration data based on the evaluation of the salivary physical property,
    a concentration normalization unit that normalizes the corrected metabolite concentration data using multiple metabolite concentration data for normalization, and
    a disease risk calculation unit that calculates the disease risk based on the normalized metabolite concentration data.
  2. The disease risk calculation system according to claim 1, wherein, in a case where the physical property evaluation unit evaluates that the saliva is cloudy, the concentration correction unit corrects the metabolite concentration data by multiplying the metabolite concentration data with correction coefficients.
  3. The disease risk calculation system according to claim 1 or 2, wherein,
    the specimen container includes a bottom part and a tubular part that extends from the bottom part,
    the image data includes the bottom part of the specimen container and part of the tubular part, and a vertical direction along with the tubular part, and
    the physical property evaluation unit extracts pixel values of the inside of the specimen container along with the vertical direction and evaluates the salivary physical property using the pixel values.
  4. The disease risk calculation system according to claim 3, wherein,
    the image data includes a horizontal direction that is perpendicular to the vertical direction, and
    the physical property evaluation unit repeats, along the vertical direction, extraction of multiple pixel values in the horizontal direction and calculation of statistical values, and evaluates the salivary physical property using the statistical values.
  5. A disease risk calculation method for calculating a disease risk using saliva, the disease risk calculation method comprising:
    a step of, at a photographing device, photographing a specimen container containing the saliva and obtaining image data;
    a step of, at a computer, evaluating a salivary physical property as to whether the saliva is transparent or cloudy based on the image data;
    a step of, at a concentration measurement device, measuring salivary metabolite concentration data;
    a step of, at the computer, correcting the metabolite concentration data based on the evaluation of the salivary physical property;
    a step of, at the computer, normalizing the corrected metabolite concentration data using multiple metabolite concentration data for normalization; and
    a step of, at the computer, calculating the disease risk based on the normalized metabolite concentration data.
PCT/JP2023/006903 2022-09-02 2023-02-27 Disease risk calculation system and method WO2024047910A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2022-139602 2022-09-02
JP2022139602A JP7233665B1 (en) 2022-09-02 2022-09-02 Disease risk determination system and disease risk determination method

Publications (1)

Publication Number Publication Date
WO2024047910A1 true WO2024047910A1 (en) 2024-03-07

Family

ID=85460132

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2023/006903 WO2024047910A1 (en) 2022-09-02 2023-02-27 Disease risk calculation system and method

Country Status (2)

Country Link
JP (1) JP7233665B1 (en)
WO (1) WO2024047910A1 (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080183101A1 (en) * 2006-08-17 2008-07-31 Jonathan Richard Stonehouse Salivary analysis
US20210118559A1 (en) * 2019-10-22 2021-04-22 Tempus Labs, Inc. Artificial intelligence assisted precision medicine enhancements to standardized laboratory diagnostic testing

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2300095A1 (en) * 1975-02-06 1976-09-03 Henkin Robert TOTAL SALIVARY PROTEIN FRACTIONATION PROCESS
JP2000009727A (en) * 1998-06-24 2000-01-14 Daikin Ind Ltd Method for converting saliva component into blood component, device therefor an diagnostic equipment
JP4626700B2 (en) * 2008-10-07 2011-02-09 ソニー株式会社 Biological rhythm information acquisition method
JP6443937B2 (en) * 2013-10-28 2018-12-26 株式会社サリバテック Salivary biomarker for pancreatic cancer and method for distinguishing pancreatic cancer patient from healthy person using the same
JPWO2016136991A1 (en) * 2015-02-27 2017-12-07 味の素株式会社 Method for determining malnutrition using marker protein in saliva and diagnostic agent used therefor
JP2016206074A (en) * 2015-04-24 2016-12-08 味の素株式会社 Method of measuring serum albumin level, method of diagnosing hypoalbuminemia, and reagents used therefor
JP6517095B2 (en) * 2015-07-01 2019-05-22 浜松ホトニクス株式会社 Saliva measurement method
CN112294258A (en) * 2020-06-03 2021-02-02 康健益智科技(北京)有限公司 Salivary cortisol stress hormone collection and evaluation system

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080183101A1 (en) * 2006-08-17 2008-07-31 Jonathan Richard Stonehouse Salivary analysis
US20210118559A1 (en) * 2019-10-22 2021-04-22 Tempus Labs, Inc. Artificial intelligence assisted precision medicine enhancements to standardized laboratory diagnostic testing

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
HYVARINEN ESAVOLAINEN MMIKKONEN JJWKULLAA AM, SALIVARY METABOLOMICS FOR DIAGNOSIS AND MONITORING DISEASES: CHALLENGES AND POSSIBILITIES. METABOLITES, vol. 11, no. 9, 31 August 2021 (2021-08-31), pages 587
LI CXZHANG LYAN YRDING YJLIN YNZHOU JPLI NLI HPLI SQSUN XW: "A narrative review of exploring potential salivary biomarkers in respiratory diseases: still on its way", J THORAC DIS, vol. 13, no. 7, July 2021 (2021-07-01), pages 4541 - 4553
MURATA TAKESHI ET AL: "Salivary metabolomics with alternative decision tree-based machine learning methods for breast cancer discrimination", BREAST CANCER RESEARCH AND TREATMENT, SPRINGER US, NEW YORK, vol. 177, no. 3, 8 July 2019 (2019-07-08), pages 591 - 601, XP036886189, ISSN: 0167-6806, [retrieved on 20190708], DOI: 10.1007/S10549-019-05330-9 *
PANNEERSELVAM KISHIKAWA SKRISHNAN RSUGIMOTO M: "Salivary Metabolomics for Oral Cancer Detection: A Narrative Review", METABOLITES, vol. 12, no. 5, 12 May 2022 (2022-05-12), pages 436

Also Published As

Publication number Publication date
JP7233665B1 (en) 2023-03-07
JP2024035256A (en) 2024-03-14

Similar Documents

Publication Publication Date Title
US11802839B2 (en) Quantifying color changes in chemical test pads induced by different concentrations of analytes under different lighting conditions
Kreil et al. DNA microarray normalization methods can remove bias from differential protein expression analysis of 2D difference gel electrophoresis results
JP4071186B2 (en) Method and system for identifying an object of interest in a biological specimen
TWI621851B (en) Biosensor, terminal and method for measuring biometric information
EP2849648B1 (en) Fish eye lens imaging apparatus and imaging method
US8743138B2 (en) Method and system for processing mass analysis data
US10983127B2 (en) Device for diagnosing colorectal cancer and method for providing colorectal cancer diagnosis information
Iwano et al. Breast cancer diagnosis based on lipid profiling by probe electrospray ionization mass spectrometry
WO2018166533A1 (en) Method for improving usability and accuracy for physiological measurement
EP3759465A1 (en) Methods and systems for calibrating and using a camera for detecting an analyte in a sample
Doorn et al. Ion chromatography for the precise analysis of chloride and sodium in sweat for the diagnosis of cystic fibrosis
WO2019221676A1 (en) Analysis of urine test strips with mobile camera analysys and providing recommendation by customising data
WO2024047910A1 (en) Disease risk calculation system and method
WO2017115775A1 (en) Analysis device, analysis method, and analysis system
US20190371004A1 (en) System and method for determining skin color, method for generating icc profile, and image capturing device
US11282683B2 (en) Imaging mass spectrometric data analyzer
Zhang et al. Maldi–TOF–MS for rapid screening and typing of β-globin variant and β-thalassemia through direct measurements of intact globin chains
Pais et al. An automated workflow for MALDI-ToF mass spectra pattern identification on large data sets: An application to detect aneuploidies from pregnancy urine
Surowiec et al. Quantification of run order effect on chromatography-mass spectrometry profiling data
WO2022217544A1 (en) Method and system for sample quality control
Wang et al. Interfering effect of bilirubin on the determination of alkaline phosphatase
Valenzuela et al. Urine test strip analysis using image processing for mobile application
WO2022054183A1 (en) Cancer test method using metabolite list
CN109493326B (en) Mobile recognition analysis system for medical detection field
CN114778629A (en) Be used for clinical detection device that draws blood of medical treatment endocrinology branch of academic or vocational study

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: 23710474

Country of ref document: EP

Kind code of ref document: A1