CN116773440A

CN116773440A - Analyte analysis device, analyte analysis method, and program

Info

Publication number: CN116773440A
Application number: CN202211691840.8A
Authority: CN
Inventors: 木村考伸; 铃木健一郎; 中西利志
Original assignee: Sysmex Corp
Current assignee: Sysmex Corp
Priority date: 2022-03-17
Filing date: 2022-12-28
Publication date: 2023-09-19
Also published as: CN116773441A; JP2023137001A

Abstract

Provided are an object analysis device, an object analysis method, and a program, which can reduce the load on a computer that analyzes data obtained by measuring an object using an artificial intelligence algorithm. An analyte analysis device (4000) for analyzing an analyte in an analyte is provided with: a measurement unit (400) including an optical detection unit for acquiring an optical signal from a subject; and an analysis unit (300) that analyzes the 1 st and 2 nd data corresponding to the optical signal. An analysis unit (300) performs a 1 st analysis operation based on an artificial intelligence algorithm on the 1 st data, and performs a 2 nd analysis operation of processing a representative value corresponding to a characteristic of the analyte in the 2 nd data.

Description

Analyte analysis device, analyte analysis method, and program

Technical Field

The present invention relates to a sample analyzer for analyzing a sample, a sample analysis method, and a program.

Background

Patent document 1 describes the following method: the signals obtained by measuring the cells using a flow cytometer (flow cytometer) are analyzed by an artificial intelligence algorithm, and the cells are classified according to categories.

Prior art literature

Patent literature

Patent document 1: international publication No. 2018/203568

Disclosure of Invention

Technical problem to be solved by the invention

In the case of processing data using an artificial intelligence algorithm, the more the data capacity increases, the more the load of a computer processing the data increases. For example, in order to improve the classification accuracy, when the amount of information for classifying components (for example, cells and formed components) in a subject is increased, the data capacity per subject increases in the subject such as blood, urine, or the like containing a plurality of components due to an increase in the amount of information obtained from 1 component. When the number of objects to be inspected increases, the data capacity also increases. Patent document 1 does not disclose a technique capable of reducing the computer load when data processing is performed using an artificial intelligence algorithm.

In view of the technical problems, an object of the present invention is to provide an object analysis device, an object analysis method, and a program, which can reduce the load on a computer that analyzes data obtained by measuring an object by an artificial intelligence algorithm.

Technical solution for solving technical problems

The present invention relates to an analyte analyzer 4000 for analyzing an analyte in an analyte. The object analysis device (4000) of the present invention comprises: a measurement unit (400) including optical detection units (410, 470) for acquiring optical signals (80 a, 80b, 80 c) from an object to be examined; and analysis means (300, 600) for analyzing the 1 st and 2 nd data (82 a, 82b, 82 c) corresponding to the optical signals (80 a, 80b, 80 c). The analysis units (300, 600) execute the 1 st analysis operation based on the artificial intelligence algorithm (60) on the 1 st data (82 a, 82b, 82 c), and execute the 2 nd analysis operation for processing the representative value corresponding to the characteristic of the analyte in the 2 nd data (82 a, 82b, 82 c).

According to the object analysis device of the present invention, the analysis processing of the data corresponding to the optical signal acquired from the object is shared by the 1 st analysis operation by the artificial intelligence algorithm and the 2 nd analysis operation for processing the representative value corresponding to the characteristic of the analyte, and the load of the analysis unit, which is a computer for processing the data, can be reduced as compared with the case where the data corresponding to the optical signal is analyzed by the artificial intelligence algorithm alone.

The present invention relates to a method for analyzing an analyte in a sample. The method for analyzing an analyte of the present invention comprises: the method comprises a step (S1, S11, S121, S131, S141, S302) of acquiring optical signals (80 a, 80b, 80 c) from an object to be examined, and an analysis step (S2, S3, S14, S16, S71, S74, S81, S84, S91, S95, S122, S124, S132, S134, S142, S143, S201, S202) of analyzing the 1 st and 2 nd data (82 a, 82b, 82 c) corresponding to the optical signals (80 a, 80b, 80 c). In the analysis steps (S2, S3, S14, S16, S71, S74, S81, S84, S91, S95, S122, S124, S132, S134, S142, S143, S201, S202), the 1 st analysis operation by the artificial intelligence algorithm (60) is performed on the 1 st data (82 a, 82b, 82 c), and the 2 nd analysis operation of processing the representative value corresponding to the characteristic of the analyte in the 2 nd data (82 a, 82b, 82 c) is performed.

According to the object analysis method of the present invention, the analysis processing of the data corresponding to the optical signal acquired from the object is shared by the 1 st analysis operation by the artificial intelligence algorithm and the 2 nd analysis operation for processing the representative value corresponding to the characteristic of the analyte, and the load of the computer for processing the data can be reduced as compared with the case where the data corresponding to the optical signal is analyzed by the artificial intelligence algorithm alone.

The program of the present invention relates to a program for causing a computer (300, 600, 3001, 3002, 6001, 6002) to execute a process of analyzing an analyte in a subject. The program of the present invention includes: and processing for analyzing the 1 st and 2 nd data (82 a, 82b, 82 c) corresponding to the optical signals (80 a, 80b, 80 c) acquired from the subject. The process performs a 1 st analysis operation based on the artificial intelligence algorithm (60) on the 1 st data (82 a, 82b, 82 c), and performs a 2 nd analysis operation for processing a representative value corresponding to the characteristic of the analyte in the 2 nd data (82 a, 82b, 82 c).

According to the program of the present invention, the analysis processing of the data corresponding to the optical signal acquired from the subject is shared by the 1 st analysis operation based on the artificial intelligence algorithm and the 2 nd analysis operation for processing the representative value corresponding to the characteristic of the analyte, and the load on the computer for processing the data can be reduced as compared with the case where the data corresponding to the optical signal is analyzed by using only the artificial intelligence algorithm.

The present invention relates to an analyte analyzer 4000 for analyzing an analyte in an analyte. The object analysis device (4000) of the present invention comprises: a measurement unit (400) including optical detection units (410, 470) for acquiring optical signals (80 a, 80b, 80 c) from an object to be examined; and analysis means (300, 600) for analyzing the data (82 a, 82b, 82 c) corresponding to the optical signals (80 a, 80b, 80 c). The analysis unit (300, 600) analyzes the data (82 a, 82b, 82 c) using a 1 st analysis based on the artificial intelligence algorithm (60) or a 2 nd analysis processing a representative value corresponding to a characteristic of the analyte in the data (82 a, 82b, 82 c) according to an analysis mode of the data (82 a, 82b, 82 c).

According to the analyte analyzing device of the present invention, by performing the analysis processing of the data corresponding to the optical signal acquired from the analyte by using the 1 st analysis based on the artificial intelligence algorithm or the 2 nd analysis that processes the representative value corresponding to the characteristic of the analyte, the load of the analysis unit, which is a computer that processes the data, can be reduced as compared with the case where the data corresponding to the optical signal is analyzed by using only the artificial intelligence algorithm.

The present invention relates to a method for analyzing an analyte in a sample. The object analysis method of the present invention includes steps (S1, S11, S121, S131, S141, S302) of acquiring optical signals (80 a, 80b, 80 c) from an object to be examined, and analysis steps (S2, S3, S14, S16, S71, S74, S81, S84, S91, S95, S122, S124, S132, S134, S142, S143, S201, S202) of analyzing data (82 a, 82b, 82 c) corresponding to the optical signals (80 a, 80b, 80 c). In the analysis steps (S2, S3, S14, S16, S71, S74, S81, S84, S91, S95, S122, S124, S132, S134, S142, S143, S201, S202), the data (82 a, 82b, 82 c) are analyzed by using the 1 st analysis by the artificial intelligence algorithm (60) or the 2 nd analysis of the data (82 a, 82b, 82 c) which is processed by the representative value corresponding to the characteristic of the analyte, according to the analysis mode of the data (82 a, 82b, 82 c).

According to the object analysis method of the present invention, the analysis processing of the data corresponding to the optical signal acquired from the object is performed by using the 1 st analysis by the artificial intelligence algorithm or the 2 nd analysis of processing the representative value corresponding to the characteristic of the analyte, and the load of the computer processing the data can be reduced as compared with the case of analyzing the data corresponding to the optical signal by using only the artificial intelligence algorithm.

The program of the present invention relates to a program for causing a computer (300, 600, 3001, 3002, 6001, 6002) to execute a process of analyzing an analyte in a subject. The program of the present invention includes a process of analyzing data (82 a, 82b, 82 c) corresponding to optical signals (80 a, 80b, 80 c) acquired from an object. The process analyzes the data (82 a, 82b, 82 c) using a 1 st analysis based on the artificial intelligence algorithm (60) or a 2 nd analysis of the data (82 a, 82b, 82 c) that processes representative values corresponding to the characteristics of the analyte, according to an analysis mode of the data (82 a, 82b, 82 c).

According to the program of the present invention, by performing analysis processing of data corresponding to an optical signal acquired from an object by using the 1 st analysis based on an artificial intelligence algorithm or the 2 nd analysis that processes a representative value corresponding to a characteristic of the analyte, the load on a computer that processes the data can be reduced as compared with a case where data corresponding to an optical signal is analyzed using only an artificial intelligence algorithm.

Effects of the invention

According to the present invention, the load on a computer that analyzes data obtained by measuring an object using an artificial intelligence algorithm can be reduced.

Drawings

Fig. 1 is a diagram schematically showing a configuration example of an analyte analyzer according to embodiment 1.

Fig. 2 is a diagram showing an outline of analysis in the case where the optical detection unit according to embodiment 1 is a detection unit by flow cytometry (flow cytometry).

Fig. 3 is a diagram schematically showing waveform data and representative values in embodiment 1.

Fig. 4 is a diagram showing an outline of analysis in the case where the optical detection unit according to embodiment 1 is a detection unit that detects transmitted light or scattered light from a measurement sample.

Fig. 5 is a flowchart showing an example of the method for analyzing a sample according to embodiment 1.

Fig. 6 is a flowchart showing an example of setting analysis work based on rules set in the analysis unit according to embodiment 2.

Fig. 7 is a flowchart showing an example of analysis performed on the basis of measurement items in embodiment 3.

Fig. 8 is an explanatory diagram schematically showing a screen for setting AI analysis or calculation process analysis for each measurement item according to embodiment 3.

Fig. 9 is a flowchart showing an example of analysis performed according to a measurement instruction in embodiment 3.

Fig. 10 is an exemplary view schematically showing a screen for setting an analysis mode for a measurement instruction according to embodiment 3.

Fig. 11 is a flowchart showing an example of performing analysis according to the analysis mode of the apparatus of embodiment 3.

Fig. 12 is an explanatory diagram schematically showing a screen for setting an analysis mode of the analysis unit according to embodiment 3.

Fig. 13 is a flowchart showing an example of analysis performed according to the type of measurement instruction in embodiment 3.

Fig. 14 is an explanatory diagram schematically showing a screen for setting an analysis mode for each type of measurement instruction according to embodiment 3.

Fig. 15 is a flowchart showing an example of analysis performed according to the type of measurement item and measurement instruction in embodiment 3.

Fig. 16 is an explanatory diagram schematically showing a screen for setting AI analysis or calculation processing analysis for each measurement item and each category of measurement instruction according to embodiment 3.

Fig. 17 is a flowchart showing an example of determining whether or not AI analysis is necessary based on a flag obtained by calculation processing analysis in embodiment 3.

Fig. 18 is an explanatory diagram schematically showing a screen for setting AI analysis for each flag of the analysis result in embodiment 3.

Fig. 19 is a flowchart showing an example of performing AI analysis on a specific analyte classified by calculation processing analysis according to embodiment 3.

Fig. 20 is an explanatory diagram schematically showing a screen for setting whether or not AI analysis is performed for each type of analyte according to embodiment 3.

Fig. 21 is a diagram illustrating a classification method using calculation processing analysis and AI analysis performed in the processing shown in fig. 19 according to embodiment 3.

Fig. 22 is a flowchart showing an example of performing AI analysis in the case where specific classification is performed by calculation processing analysis in embodiment 3.

Fig. 23 is an explanatory diagram schematically showing a screen for setting whether or not AI analysis is performed on a specific class of analytes in embodiment 3.

Fig. 24 is a block diagram showing the structure of a measurement unit according to embodiment 4.

Fig. 25 is a diagram schematically showing the configuration of an optical system of the FCM detection unit according to embodiment 4.

Fig. 26 is a block diagram showing the structure of an analysis unit according to embodiment 4.

Fig. 27 is a block diagram showing the configuration of a measurement unit in the case where the sample analyzer according to embodiment 4 performs counting and sorting of blood cells in a blood sample.

Fig. 28 is a block diagram showing the structures of the specimen suction unit and the sample preparation unit in the measurement unit of fig. 27 according to embodiment 4.

Fig. 29 is a block diagram showing another configuration of the sample preparation section shown in fig. 28 according to embodiment 4.

Fig. 30 is a flowchart showing an example of analysis performed according to the measurement channel in embodiment 4.

Fig. 31 is an explanatory diagram schematically showing a screen for setting AI analysis or calculation processing analysis for each measurement channel in embodiment 4.

Fig. 32 is a schematic diagram for explaining waveform data used in the analysis method according to embodiment 4.

Fig. 33 is a schematic diagram showing an example of a training data generation method used for training an AI algorithm for determining the type of analyte in a subject according to embodiment 4.

Fig. 34 is a graph showing the label values corresponding to the cell types in embodiment 4.

Fig. 35 is a diagram schematically showing a method of analyzing waveform data of an analyte in a subject using the AI algorithm according to embodiment 4.

Fig. 36 is a flowchart showing an example of performing AI analysis on waveform data acquired in a WDF channel according to embodiment 4.

Fig. 37 is a flowchart showing an example of classifying nucleated erythrocytes and basophils by AI analysis and classifying others by computational processing analysis based on waveform data acquired in the WDF channel in embodiment 4.

Fig. 38 is a flowchart showing an example of performing AI analysis on neutrophils/basophils determined by analysis of calculation processing on a WDF channel according to embodiment 4.

Fig. 39 is a block diagram schematically showing the structure of a measurement unit according to embodiment 5.

Fig. 40 is a side view schematically showing measurement by a detection block according to embodiment 5.

Fig. 41 is a flowchart showing an analysis example of embodiment 5.

Fig. 42 is a block diagram showing the structure of the analyte analyzer of embodiment 6.

Fig. 43 is a block diagram showing the structure of an analysis unit according to embodiment 6.

Fig. 44 is a block diagram showing another configuration of the analyte analyzing device according to embodiment 6.

Fig. 45 is a diagram showing a configuration example of a parallel processing processor according to embodiment 6.

Fig. 46 is a diagram schematically showing an example of mounting the parallel processing processor according to embodiment 6.

Fig. 47 is a diagram schematically showing an example of mounting a parallel processing processor according to embodiment 6.

Fig. 48 is a diagram schematically showing an example of mounting the parallel processing processor according to embodiment 6.

Fig. 49 is a diagram showing another mounting example of the parallel processing processor of embodiment 6.

Fig. 50 is a diagram showing a configuration example of a parallel processing processor that executes arithmetic processing according to embodiment 6.

Fig. 51 is a diagram showing an outline of matrix operations performed by the parallel processing processor of embodiment 6.

Fig. 52 is a conceptual diagram showing that a plurality of arithmetic processes of embodiment 6 are executed in parallel by a parallel processing processor.

Fig. 53 is a diagram schematically showing an outline of the calculation processing related to the convolutional layer in embodiment 6.

Fig. 54 is a flowchart showing the analysis operation of the analysis unit and the measurement unit according to embodiment 6.

Fig. 55 is a flowchart showing details of AI analysis in step S201 of fig. 54 of embodiment 6.

Fig. 56 is a flowchart showing details of step S2011 in fig. 55 according to embodiment 6.

Fig. 57 is a block diagram showing another configuration of the measurement unit according to embodiment 6.

Fig. 58 is a block diagram showing another configuration of the analysis unit of embodiment 6.

Fig. 59 is a block diagram showing another configuration of the measurement unit according to embodiment 6.

Fig. 60 is a block diagram showing another configuration of the analysis unit of embodiment 6.

Fig. 61 is a block diagram showing another configuration of the analyte analyzer of embodiment 6.

Fig. 62 is a block diagram showing another configuration of the measurement unit according to embodiment 6.

Fig. 63 is a block diagram showing another configuration of the analysis unit of embodiment 6.

Fig. 64 is a diagram showing a configuration example of a parallel processing processor that executes arithmetic processing according to embodiment 6.

Fig. 65 is a block diagram showing the structure of a computer according to embodiment 6.

Fig. 66 is a block diagram showing another configuration of the measurement unit according to embodiment 6.

Fig. 67 is a block diagram showing another configuration of the analysis unit of embodiment 6.

Fig. 68 is a block diagram showing another configuration of the measurement unit according to embodiment 6.

Fig. 69 is a block diagram showing another configuration of the analysis unit of embodiment 6.

Fig. 70 is a block diagram showing another configuration of the measurement unit according to embodiment 6.

Fig. 71 is a block diagram showing another configuration of the analysis unit of embodiment 6.

Fig. 72 is a block diagram showing another configuration of the measurement unit according to embodiment 6.

Fig. 73 is a block diagram showing another configuration of the analysis unit of embodiment 6.

Fig. 74 is a diagram schematically showing the configuration of a waveform data analysis system according to embodiment 7.

Fig. 75 is a block diagram showing the structure of the deep learning device according to embodiment 7.

Fig. 76 is a functional block diagram of the deep learning device according to embodiment 7.

Fig. 77 is a flowchart showing the processing performed by the deep learning device according to embodiment 7.

Fig. 78 is a schematic diagram illustrating the structure of the neural network, a schematic diagram illustrating operations at each node, and a schematic diagram illustrating operations between nodes according to embodiment 7.

Reference numerals

60: AI algorithm (artificial intelligence algorithm); 80a, 80b, 80c: an optical signal; 82a, 82b, 82c: waveform data (data 1, data 2); 300. 600: an analysis unit; 400: a measurement unit; 410: FCM detection unit (optical detection unit); 440: a sample preparation unit; 470: a detection unit (optical detection unit); 471: a light source unit (light source); 475: a light receiving section (photodetector); 3001. 6001: a processor (host processor); 3002. 6002: a parallel processing processor; 4000: a sample analyzer; 4111: a light source; 4113: a flow cell; 4116. 4121, 4122: light receiving element (light detector)

Detailed Description

The outline and embodiments of the present invention are described in detail below with reference to the drawings. In the following description and drawings, the same reference numerals denote the same or similar components, and a description thereof will be omitted for convenience.

Embodiment 1

The present embodiment discloses a sample analyzer, a sample analysis method, and a program as follows: both analysis based on an artificial intelligence algorithm (AI (Artificial Intelligence, artificial intelligence) algorithm) and analysis without using the AI algorithm can be performed on data acquired by measuring an object.

In AI algorithm-based analysis, data is analyzed, for example, by a large number of matrix arithmetic processes. For convenience, the analysis based on the AI algorithm will be referred to as "AI analysis" hereinafter. In AI analysis, for example, a convolution operation based on an AI algorithm is performed.

In the analysis not using the AI algorithm, data is analyzed by, for example, calculation processing for a representative value corresponding to a feature of an analyte. Hereinafter, for convenience, an analysis method of analyzing data by calculation processing for a representative value corresponding to a feature of an analyte without using an AI algorithm is referred to as "calculation processing analysis" or "non-AI analysis". The amount of data of the representative value processed in the calculation processing analysis is smaller than the data input to the AI algorithm in the AI analysis. In the calculation processing analysis, the amount of data and the amount of calculation processing of the processing target are small compared to the AI analysis, and therefore the load on the computer performing the analysis is small compared to the AI analysis. This can shorten the TAT (Turn Around Time) of analysis of the measurement results.

According to the object analysis device, the object analysis method, and the program of embodiment 1, the load of the computer performing the analysis can be reduced by performing the analysis of the data acquired by measuring the object in a shared manner by the AI analysis and the calculation processing analysis.

Fig. 1 is a diagram schematically showing a configuration example of an analyte analyzer 4000 according to embodiment 1. In fig. 1, the upper diagram shows a configuration example of the sample analyzer 4000 of the embodiment, and the lower diagram shows a modification of the configuration of embodiment 1.

As shown in the upper diagram of fig. 1, the analyte analyzer 4000 according to embodiment 1 includes, for example, a measurement unit 400 and an analysis unit 300. As shown in the lower diagram of fig. 1, the analyte analyzer 4000 may include a measurement unit 400 and an analysis unit 300 that are integrally formed. Examples of the analyte analyzer 4000 include a blood cell analyzer, a urine analyzer, a blood coagulation analyzer, an immunoassay analyzer, a biochemical analyzer, and a gene analyzer. The analyte to be analyzed by the analyte analyzer 4000 is, for example, a cell, a physical component, a protein, a gene, or the like.

The measurement unit 400 measures the object to be examined and acquires data related to the object to be examined. The analysis unit 300 analyzes the data acquired by the measurement unit 400. The analysis unit 300 may have a function of setting measurement conditions of the measurement sample in the measurement unit 400 and performing control of measurement. The analysis unit 300 is configured as a device (e.g., a computer) separate from the measurement unit 400 and is connected to the measurement unit 400. The analysis unit 300 and the measurement unit 400 are connected by a wire or wireless.

The measurement unit 400 includes an optical detection unit for measuring a measurement sample prepared from a subject.

The optical detection unit is, for example, a detection unit based on flow cytometry, and is used for measuring a blood test object or a urine test object. The optical detection unit irradiates the measurement sample flowing through the flow cell with light to obtain an optical signal. For example, the optical detection unit irradiates light on a measurement sample including an analyte (e.g., a cell or a formed component) flowing in the flow cell, so that forward scattered light, side scattered light, and fluorescence are generated from the analyte. The photodetector provided in the optical detection unit receives the generated light and outputs an optical signal corresponding to the intensity of the received light. The optical signal is an analog signal of a waveform corresponding to time variation of forward scattered light, side scattered light, and fluorescence. An a/D conversion unit provided in the optical detection unit performs digital conversion on the optical signal to obtain digital data of waveforms (hereinafter referred to as "waveform data") corresponding to the respective analytes. The waveform data in this case is used for, for example, classification of leukocyte types in a blood sample, classification of the number of erythrocytes and leukocytes in a blood sample, classification of the physical components in a urine sample, and the like.

The optical detection unit may be configured to irradiate light to the measurement sample accommodated in the container and detect light transmitted from the measurement sample or scattered light scattered from the measurement sample by the photodetector. In this case, the optical detection unit irradiates light to the measurement sample containing the analyte which is left standing in the state of being accommodated in the container. A photodetector provided in the optical detection unit receives the transmitted light transmitted through the measurement sample or the scattered light generated by the measurement sample for a predetermined period of time, and outputs an optical signal corresponding to the intensity of the received light. The optical signal in this case is an analog signal of a waveform corresponding to a temporal change in transmitted light or scattered light with solidification of the measurement sample. An a/D conversion unit provided in the optical detection unit performs digital conversion on the optical signal to obtain digital data of a waveform corresponding to the temporal change of the transmitted light or the scattered light (hereinafter referred to as "solidification waveform data"). The coagulation waveform data in this case is used for analysis of the coagulation ability of the blood sample, for example.

Next, an example in which the analysis unit 300 performs analysis using the data acquired by the measurement unit 400 will be described.

Fig. 2 is a diagram showing an outline of analysis in the case where the optical detection unit is a detection unit based on flow cytometry.

In fig. 2, the left diagram shows an outline of the calculation processing analysis, and the right diagram shows an outline of the AI analysis. FSC, SSC, and FL in FIG. 2 show optical signals corresponding to forward scattered light intensity, side scattered light intensity, and fluorescence, respectively, acquired by the optical detection unit of measurement unit 400.

As shown in the upper graph of fig. 3, in the digital data obtained by digitally converting the optical signal, the measurement unit 400 determines a region having a value greater than a predetermined threshold value as a region corresponding to the analyte in the subject. The region in which the value of the digital data of the digitally converted optical signal is greater than the threshold value is a region corresponding to each analyte in the subject. Each graph of fig. 3 schematically shows a region corresponding to 1 analyte in the subject (for example, a region of "waveform data" in the graph of the upper part of fig. 3) determined in the digital data. Further, the determination of the region whose value is greater than the predetermined threshold value may also be made for the optical signal.

The measurement unit 400 acquires, as waveform data, areas corresponding to the respective analytes in the object from the digital data in which the optical signals are digitally converted. Waveform data is acquired corresponding to a plurality of analytes in the subject. The analysis unit 300 calculates a representative value corresponding to a characteristic of the analyte in the waveform data by calculation processing. As shown in each graph of fig. 3, the analysis unit 300 calculates, for example, the amounts of peak, width, area, and the like of waveform data as representative values. The peak value is the maximum value of the waveform data, the width is the width of the waveform data in the time axis direction, and the area is the area surrounded by the waveform data.

In the calculation processing analysis, a representative value corresponding to the characteristic of the analyte is predetermined. For example, in the case of classifying and counting blood cells as analytes, a representative value predetermined in an algorithm for calculation processing analysis is a peak value. The analysis unit 300 acquires a predetermined representative value from the waveform data by a predetermined calculation, and processes the representative value acquired for analyzing the analyte. The analysis unit 300 obtains predetermined representative values for each of the plurality of waveform data obtained by the measurement unit 400. That is, representative values (e.g., peak values) of the same class are acquired from the plurality of waveform data, respectively, by predetermined calculation performed by the analysis unit 300. The measurement unit 400 may acquire a predetermined representative value, and the acquired representative value and waveform data may be transmitted to the analysis unit 300.

On the other hand, in AI analysis, since the AI algorithm extracts characteristics of waveform data, the representative value is not predetermined. Since the characteristics of the waveform data extracted by the AI algorithm (i.e., the characteristics corresponding to the analyte) can be changed according to the learning content of the AI algorithm, the representative value does not need to be decided in advance in the AI analysis. Since the AI algorithm can extract various features of waveform data according to learning contents, not only the representative value but also the waveform data itself are input to the AI algorithm. Since waveform data itself is input to the AI algorithm, the computer load for data operation becomes higher in the AI analysis and TAT (turn around time) required for the operation becomes longer as compared with the calculation processing analysis.

As shown in the left diagram of fig. 2, in the calculation processing analysis, the analysis unit 300 acquires a representative value from waveform data acquired in correspondence with an analyte, and generates, for example, a scatter diagram SC based on the acquired representative value. In the scatter diagram SC illustrated in fig. 2, the SSCP on the horizontal axis represents the peak value of the waveform data based on the side scattered light, and the FLP on the vertical axis represents the peak value of the waveform data based on the fluorescence. A plurality of analytes are plotted in a scatter plot SC. The analysis unit 300 performs classification and analysis of analytes in the subject based on the scatter diagram SC.

As shown in the right-hand diagram of fig. 2, in the AI analysis, the analysis unit 300 inputs waveform data corresponding to the analyte to the AI algorithm 60, and performs classification and analysis of the analyte in the subject. The AI algorithm 60 is a learned AI algorithm, and the AI algorithm 60 is generated by inputting the AI algorithm before training and learning the waveform data as described above. The amount of data of the representative value acquired in the calculation processing analysis is smaller than the waveform data input to the AI algorithm in the AI analysis.

The types of analytes classified by the calculation processing analysis and the AI analysis are, for example, the types of blood cells in a blood sample, the types of physical components in a urine sample, and the like. For example, the analysis unit 300 performs AI analysis on measurement items for classifying the types of white blood cells in the blood test object, and performs calculation process analysis on other measurement items.

Fig. 4 is a diagram showing an outline of analysis in the case where the optical detection unit is a detection unit that detects transmitted light or scattered light from a measurement sample.

The measurement unit 400 acquires digital data obtained by digitally converting the optical signal as solidification waveform data. In 1 measurement, 1 coagulation waveform data was obtained from 1 measurement sample.

Fig. 4 is a graph showing an example of solidification waveform data of transmitted light detected based on irradiation of light to a measurement sample. The horizontal axis represents elapsed time, and the vertical axis represents absorbance. The absorbance is a value indicating how much light irradiated to the measurement sample is absorbed by the measurement sample. The state where the absorbance is 0% indicates a state where almost all of the light irradiated to the measurement sample reaches the photodetector, and the state where the absorbance is 100% indicates a state where almost all of the light irradiated to the measurement sample does not reach the photodetector.

Furthermore, it is also possible to transmit light intensity instead of absorbance. In this case, when the ratio (transmitted light intensity) set to the vertical axis increases with the upward movement, the solidification waveform data has a shape that decreases with the passage of time, as in fig. 4.

The coagulation waveform data includes at least data corresponding to an optical signal acquired from a timing T2 indicating the start of coagulation of the object to a timing T3 indicating the end of coagulation of the object. The coagulation waveform data may include data corresponding to an optical signal acquired from the start timing T1 of photometry by the measurement unit 400 to the end timing T4 of photometry.

In the calculation processing analysis, the analysis unit 300 calculates a representative value corresponding to the characteristic of the analyte in the solidification waveform data by calculation processing, and performs analysis based on the calculated representative value. In the calculation processing analysis, the analysis unit 300 determines, as a representative value, the solidification waveform data when the intensity of the detected light satisfies a predetermined condition. For example, the analysis unit 300 acquires, as a representative value, a time (T-T2) required until the absorbance of the coagulation waveform data decreases to a predetermined value (for example, 50%), and supplies the acquired representative value as a result indicating a time until the blood specimen is coagulated.

In the AI analysis, the analysis unit 300 analyzes the solidification waveform data based on the AI algorithm 60 (refer to fig. 2). The analysis unit 300 obtains whether or not there is an abnormality related to measurement, for example, based on the feature amount extracted from the solidification waveform data by the AI algorithm 60. The analysis unit 300 determines the possibility of occurrence of a non-specific reaction based on the presence or absence of an abnormality related to measurement.

For example, the analysis unit 300 analyzes whether or not there is an abnormality in the blood sample due to an interfering substance. Specifically, the analysis unit 300 analyzes the presence or absence of abnormality using coagulation waveform data related to PT (prothrombin time), which is an item for measuring coagulation ability related to prothrombin as a coagulation factor.

Further, the analysis unit 300 may input coagulation waveform data to the AI algorithm 60 to acquire the time until the blood test object is coagulated in the AI analysis. In addition, the analysis unit 300 may input the coagulation waveform data to the AI algorithm 60 in AI analysis to acquire the cause of extension when the coagulation time is extended.

In step S1, the measurement unit 400 acquires an optical signal by the optical detection unit, and acquires waveform data from the acquired optical signal.

In step S2, the analysis unit 300 performs AI analysis on the waveform data (data 1) as the AI analysis target among the waveform data acquired by the measurement unit 400. For example, the analysis unit 300 determines the waveform data corresponding to the measurement item as the AI analysis target as the 1 st data, and performs AI analysis on the determined 1 st data.

In step S3, the analysis unit 300 performs calculation process analysis on waveform data (data 2) as a calculation process analysis object among the waveform data acquired by the measurement unit 400. For example, the analysis unit 300 determines the waveform data corresponding to the measurement item as the analysis target of the calculation process as the 2 nd data, and performs the calculation process analysis on the determined 2 nd data.

In the above steps S2 and S3, a case where the measurement for classifying the leukocyte types in the blood sample is the object of the AI analysis will be described as an example. The measurement unit 400 prepares a blood sample with a reagent corresponding to measurement of leukocyte classification, and measures the prepared measurement sample with an optical detection unit by flow cytometry, for example. For example, since the measurement related to the leukocyte classification is the object of AI analysis, the analysis unit 300 determines the waveform data of the measurement sample based on the leukocyte classification as the 1 st data. The analysis unit 300 analyzes the 1 st data by using the AI algorithm 60 to classify the white blood cells. On the other hand, the analysis unit 300 determines, for example, waveform data of a measurement sample based on a measurement sample other than the white blood cell classification as the 2 nd data. The analysis unit 300 identifies a representative value corresponding to the characteristic of the analyte from the 2 nd data, performs calculation processing analysis for processing the identified representative value, and classifies blood cells other than leukocytes.

In step S4, the analysis unit 300 supplies the analysis results acquired through steps S2, S3. In step S4, for example, the analysis unit 300 displays the analysis result on the display unit, transmits the analysis result to another computer, and the like.

In step S1, the measurement unit 400 may acquire an optical signal from 1 measurement sample by using the optical detection unit and acquire waveform data from the acquired optical signal. In this case, the 1 st data and the 2 nd data may be a plurality of data, respectively, and a part of the data may be the same data as each other.

In step S1, optical signals may be acquired from a plurality of measurement samples containing an object acquired from the same subject by the optical detection unit, and waveform data may be acquired from the acquired optical signals. In this case, in step S2, the analysis unit 300 performs AI analysis on waveform data (data 1) acquired from one measurement sample, and in step S3, the analysis unit 300 performs calculation process analysis on waveform data (data 2) acquired from another measurement sample. The plurality of measurement samples containing the sample collected from the same subject may be prepared using reagents of the same kind as each other or may be prepared using reagents of different kinds from each other.

In step S1, optical signals may be acquired from a plurality of measurement samples containing samples acquired from subjects different from each other by an optical detection unit, and waveform data may be acquired from the acquired optical signals. In this case, in step S2, the analysis unit 300 performs AI analysis on waveform data (data 1) acquired from one measurement sample, and in step S3, the analysis unit 300 performs calculation process analysis on waveform data (data 2) acquired from another measurement sample. The plurality of measurement samples containing the samples collected from the subjects different from each other may be prepared using the same kind of reagent as each other or may be prepared using different kinds of reagents as each other.

In addition, in embodiment 1 described above, as the analysis of the calculation processing, in step S3, the analysis unit 300 specifies the representative value corresponding to the characteristic of the analyte from the 2 nd data, and processes the specified representative value, but is not limited thereto. For example, in step S1, the measurement unit 400 may acquire a representative value from the waveform data, output the waveform data and the representative value to the analysis unit 300, and perform analysis as calculation processing, and in step S3, the analysis unit 300 may process the representative value acquired from the measurement unit 400.

Embodiment 2

In embodiment 2, AI analysis and calculation processing analysis are selected based on the rule set in the analysis unit 300.

For example, a rule for selecting an analysis job is set by the user via the analysis unit 300. For example, the user can set rules corresponding to the operation guidelines of the laboratory in the analysis unit 300. This makes it possible to appropriately change the contribution of the AI analysis and the calculation processing analysis according to the operation policy of the laboratory.

Since the rule for analysis work can be set, the sharing of the AI analysis and the calculation processing analysis can be flexibly changed while the load of the analysis unit 300 is reduced. For example, when the accuracy of the AI analysis is improved by performing the additional learning by the AI algorithm 60, a rule may be set so that the data to be AI-analyzed increases. In addition, for example, when priority is given to shortening the TAT (turn around time) of the analysis of the measurement result, a rule may be set so that the data to be analyzed by the calculation process increases.

Fig. 6 is a flowchart showing an example of setting an analysis job based on a rule set in the analysis unit 300.

In step S11, the measurement unit 400 acquires an optical signal by the optical detection unit, and acquires waveform data from the acquired optical signal. In step S12, the analysis unit 300 refers to the rule for selecting the analysis operation, and determines waveform data to be analyzed by the AI analysis and the calculation processing, respectively, for the waveform data acquired in step S11 based on the referenced rule.

In step S13, the analysis unit 300 determines whether or not the waveform data determined in step S12 includes waveform data to be AI-analyzed. When the waveform data to be AI-analyzed is included (yes in S12), in step S14, the analysis unit 300 performs AI analysis on the waveform data to be AI-analyzed determined in step S12.

Next, in step S15, the analysis unit 300 determines whether or not there is waveform data as a calculation processing analysis target in addition to the waveform data analyzed by the AI. When there is waveform data to be subjected to the calculation process analysis (yes in S15), in step S16, the analysis unit 300 performs the calculation process analysis on the waveform data to be subjected to the calculation process analysis determined in step S12.

In step S12, the measurement unit 400 may acquire both the waveform data to be the AI analysis target and the waveform data to be the calculation processing analysis target. For example, when measurement related to leukocyte classification and measurement related to reticulocyte (reticulocyte) are performed according to a measurement instruction, the measurement unit 400 acquires waveform data for leukocyte classification and waveform data for reticulocyte measurement. When the white blood cells are classified as the object of the AI analysis and the reticulocyte measurement is the object of the calculation process analysis, the analysis unit 300 determines that the waveform data for the white blood cell classification as the object of the AI analysis is included (S13: yes), and performs the AI analysis on the waveform data. Further, the analysis unit 300 determines that the waveform data for reticulocyte classification, which is the object of the calculation process analysis, is also included (S15: yes), and performs the calculation process analysis on the waveform data.

On the other hand, in the case where the waveform data to be the AI analysis target is not included in the waveform data acquired by the measurement unit 400 (S13: no), in step S16, the analysis unit 300 performs the calculation process analysis on the waveform data to be the calculation process analysis target determined in step S12. If the AI analysis is performed and waveform data to be analyzed by the calculation process is not included (S15: no), the calculation process analysis is not performed, and the process proceeds to step S17.

In step S17, the analysis unit 300 provides an analysis result.

The analysis unit 300 may determine whether or not waveform data to be analyzed by the calculation process is included in step S13, and determine whether or not waveform data to be analyzed by the AI is included in step S15. In this case, when it is determined in step S13 that the waveform data as the analysis target of the calculation process is included, the analysis unit 300 performs the calculation process analysis in step S14. Further, when it is determined in step S15 that the waveform data to be the AI-analysis target is included, the analysis unit 300 performs AI-analysis in step S16.

Embodiment 3

In embodiment 3, various examples of sharing AI analysis and calculation processing analysis will be described.

For example, the contribution of AI analysis and calculation processing analysis is determined by a software program for the analysis unit 300 to perform analysis of waveform data. The software program of the analysis unit 300 determines waveform data as an AI analysis object and waveform data as a calculation processing analysis object, respectively, and performs analysis. For example, a software program is designed based on the requirements related to the test (e.g., TAT is increased, analysis accuracy is improved).

Fig. 7 is a flowchart showing an example of performing analysis based on a measurement item.

In fig. 7, steps S21, S22, S23 are added instead of steps S12, S13, S15, respectively, as compared with fig. 6. The following describes the change point with respect to fig. 6.

In step S21, the analysis unit 300 refers to a rule including which of the AI analysis and the calculation process analysis is performed based on the measurement item, and determines waveform data of the measurement item to be subjected to the AI analysis and waveform data of the measurement item to be subjected to the calculation process analysis, respectively, based on the referenced rule, with respect to the waveform data acquired in step S11.

Fig. 8 is an explanatory diagram schematically showing a screen for setting AI analysis or calculation processing analysis for each measurement item. The measurement items illustrated in fig. 8 are measurement items related to a blood cell analysis device.

The screen of fig. 8 is displayed on a display unit provided in the analysis unit 300, for example. In the screen of fig. 8, a check box for setting AI analysis and a check box for setting calculation processing analysis are provided for each measurement item. The check box for AI analysis of 1 measurement item and the check box for calculation processing analysis are configured such that only one of them can be selected. The user operates a check box to select which of the AI analysis and the calculation process analysis is to be performed for each measurement item, and operates a setting button. Thereby storing the rule in the storage section of the analysis unit 300.

Further, although the user sets one of the AI analysis and the calculation process analysis for each measurement item via the screen shown in fig. 8, a screen may be configured so that both the AI analysis and the calculation process analysis can be set. This makes it possible to compare the result of the AI analysis with the result of the calculation processing analysis. The selection of the analysis for each measurement item may be set in advance at the time of shipment of the device, or may be set so that only the administrator can change the setting.

Returning to fig. 7, in step S22, the analysis unit 300 determines whether or not the waveform data of the measurement item as the AI analysis target is included in the waveform data determined in step S21. When the waveform data of the measurement item to be AI-analyzed is included (yes in S22), in step S14, the analysis unit 300 performs AI analysis on the waveform data of the measurement item to be AI-analyzed determined in step S21, with respect to the measurement item.

For example, the measurement unit 400 prepares a leukocyte measurement sample by mixing a reagent and a sample corresponding to the classification based on a measurement instruction for classifying leukocytes (for example, 5 classes of neutrophils, lymphocytes, monocytes, eosinophils, and basophils). The measurement unit 400 acquires an optical signal corresponding to the leukocyte measurement sample by the optical detection unit. The measurement unit 400 acquires waveform data corresponding to the acquired optical signal. In the case where measurement items (for example, counts and duty ratios of neutrophils, lymphocytes, monocytes, eosinophils, and basophils, respectively) related to classification of leukocytes are the subjects of AI analysis, the analysis unit 300 performs AI analysis on waveform data obtained by the measurement unit 400 by measuring a leukocyte measurement sample.

Next, in step S23, the analysis unit 300 determines whether or not the waveform data of the measurement item that is the object of the calculation processing analysis is present among the waveform data determined in step S12. When there is waveform data to be subjected to the calculation process analysis (yes in S23), in step S16, the analysis unit 300 performs calculation process analysis on the waveform data to be subjected to the calculation process analysis determined in step S21, in association with the measurement item.

For example, the measurement unit 400 mixes a reagent corresponding to the classification with a sample according to a measurement command for classifying reticulocytes, and prepares a reticulocyte measurement sample. The measurement unit 400 obtains an optical signal corresponding to the reticulocyte measurement sample by the optical detection unit. The measurement unit 400 acquires waveform data corresponding to the acquired optical signal. In the case where measurement items related to classification of reticulocytes (for example, the count and the duty ratio of reticulocytes) are the object of calculation process analysis, the analysis unit 300 performs calculation process analysis on waveform data obtained by the measurement unit 400 by measuring the reticulocyte measurement sample.

The analyte analyzer 4000 is not limited to the blood cell analyzer, and may be a urine analyzer or a blood coagulation measuring device. For example, in the case where the analyte analyzing device 4000 is a urine analyzing device, the analyzing unit 300 performs AI analysis for a part of the measurement items and performs calculation process analysis for the rest of the measurement items. In addition, in the case where the sample analyzer 4000 is a blood coagulation measuring device, the analysis unit 300 performs calculation processing analysis on all measurement items and performs AI analysis on some measurement items in addition to calculation processing analysis to determine the possibility of occurrence of a non-specific reaction.

Fig. 9 is a flowchart showing an example of performing analysis according to a measurement instruction.

As compared with fig. 6, steps S31 and S32 are added to fig. 9 instead of steps S12 and S13, respectively, and step S15 is deleted. The following describes the change point with respect to fig. 6.

In step S31, the analysis unit 300 determines which of the waveform data to be AI-analyzed and the waveform data to be calculation-processed-analyzed is the waveform data acquired in step S11, based on the measurement instruction. The analysis mode for the measurement command is one of an AI analysis mode and a calculation processing analysis mode, and is stored in the storage unit of the analysis unit 300 in association with the measurement command.

Fig. 10 is an exemplary view schematically showing a screen for setting an analysis mode for a measurement instruction.

The screen of fig. 10 is displayed on a display unit provided in the analysis unit 300, for example. In the screen of fig. 10, each row corresponds to a measurement instruction identified by the object number. In the screen of fig. 10, a check box for setting the AI analysis mode and a check box for setting the calculation process analysis mode are provided for each measurement instruction. The user operates the check box to select which of the AI analysis and the calculation processing analysis is to be performed by the analysis unit 300 for each measurement instruction, and operates the setting button. Thus, the analysis pattern is stored in the storage unit of the analysis unit 300 in association with the measurement instruction.

The analysis mode associated with each measurement instruction is not limited to being set by the user via the analysis unit 300, and may be set in advance by a host computer or the like when setting the measurement instruction.

Returning to fig. 9, in step S32, the analysis unit 300 determines whether the waveform data determined in step S31 is the waveform data of the measurement instruction of the AI analysis target. When the determined waveform data is the waveform data of the measurement instruction to be AI-analyzed (yes in S32), in step S14, the analysis unit 300 performs AI analysis on the waveform data of the measurement instruction. On the other hand, when the determined waveform data is the waveform data of the measurement instruction to be analyzed by the calculation process (S32: no), in step S16, the analysis unit 300 performs the calculation process analysis on the waveform data of the measurement instruction.

Fig. 11 is a flowchart showing an example of performing analysis according to an analysis mode of the apparatus.

In fig. 11, steps S41 and S42 are added instead of steps S12 and S13, respectively, and step S15 is deleted, as compared with fig. 6. The following describes the change point with respect to fig. 6.

In step S41, the analysis unit 300 refers to the rule including the analysis mode of the analysis unit 300, and determines which of the waveform data to be subjected to AI analysis and the waveform data to be subjected to calculation processing analysis is the waveform data acquired in step S11 based on the referenced rule. When the AI analysis mode is set in the rule, all waveform data is the object of AI analysis, and when the calculation process analysis mode is set in the rule, all waveform data is the object of calculation process analysis.

Fig. 12 is an exemplary view schematically showing a screen for setting an analysis mode of the analysis unit 300.

The screen of fig. 12 is displayed on a display unit provided in the analysis unit 300, for example. The screen of fig. 12 includes a check box for setting the AI analysis mode and a check box for setting the calculation process analysis mode for the analysis unit 300. The user operates the check box to select which of the AI analysis and the calculation processing analysis is to be performed by the analysis unit 300, and operates the setting button. Thereby storing the rule in the storage section of the analysis unit 300.

Returning to fig. 11, in step S42, the analysis unit 300 determines whether the waveform data determined in step S41 is waveform data that is the object of AI analysis. When the determined waveform data is the data to be AI-analyzed (yes in S42), that is, when the analysis mode of the analysis unit 300 is the AI-analysis mode, the analysis unit 300 performs AI-analysis on the waveform data in step S14. On the other hand, when the determined waveform data is the waveform data to be analyzed by the calculation process (no in S42), that is, when the analysis mode of the analysis unit 300 is the calculation process analysis mode, the analysis unit 300 performs the calculation process analysis on the waveform data in step S16.

Fig. 13 is a flowchart showing an example of performing analysis according to the category of the measurement instruction.

In fig. 13, steps S51 and S52 are added instead of steps S12 and S13, respectively, and step S15 is deleted, as compared with fig. 6. The following describes the change point with respect to fig. 6.

In step S51, the analysis unit 300 refers to the rule including the analysis pattern corresponding to the type of the measurement instruction, and determines which of the waveform data to be subjected to AI analysis and the waveform data to be subjected to calculation processing analysis is the waveform data acquired in step S11 based on the type of the measurement instruction and the referenced rule. The types of measurement instructions include "Normal" (Normal) corresponding to a Normal measurement such as a primary test, "retest" (Rerun) corresponding to a retest in which a measurement item similar to the primary test is set, and "return" (Reflex) corresponding to a retest after the primary test is changed from the measurement item. In the rule, any one of the AI analysis mode and the calculation processing analysis mode is set for each type of measurement instruction.

Fig. 14 is an exemplary view schematically showing a screen for setting an analysis mode for each type of measurement instruction.

The screen of fig. 14 is displayed on a display unit provided in the analysis unit 300, for example. The screen of fig. 14 includes a check box for setting the AI analysis mode and a check box for setting the calculation processing analysis mode for each type of measurement instruction (regular, retest, return). The user operates a check box to select which of the AI analysis and the calculation processing analysis is to be performed for each type of measurement instruction, and operates a setting button. Thereby, the rule is stored in the storage unit of the analysis unit 300.

The analysis mode associated with each type of measurement instruction is not limited to being set by the user via the analysis unit 300, and may be set in advance by the host computer or the like according to the type of measurement instruction.

Returning to fig. 13, in step S52, the analysis unit 300 determines whether the waveform data determined in step S51 is waveform data to be AI-analyzed. When the determined waveform data is the waveform data to be subjected to AI analysis (yes in S52), that is, when the analysis mode corresponding to the type of the measurement instruction is the AI analysis mode, in step S14, the analysis unit 300 performs AI analysis on the waveform data. On the other hand, when the determined waveform data is waveform data to be analyzed by calculation processing (S52: no), that is, when the analysis mode corresponding to the type of the measurement instruction is the calculation processing analysis mode, in step S16, the analysis unit 300 performs calculation processing analysis on the waveform data.

Fig. 15 is a flowchart showing an example of performing analysis according to the types of measurement items and measurement instructions.

In fig. 15, step S61 is added instead of step S12, as compared with fig. 6. The following describes the change point with respect to fig. 6.

In step S61, the analysis unit 300 refers to a rule for selecting an analysis job, and determines waveform data to be subjected to AI analysis and waveform data to be subjected to calculation processing analysis for the waveform data acquired in step S11 based on the measurement item and the type of the measurement instruction.

Fig. 16 is an explanatory diagram schematically showing a screen for setting AI analysis or calculation processing analysis for each measurement item and each category of measurement instruction.

The screen of fig. 16 is displayed on a display unit provided in the analysis unit 300, for example. In the screen of fig. 16, as in fig. 8, a check box for setting AI analysis and a check box for setting calculation processing analysis are provided for each measurement item, and as in fig. 14, a check box for setting any one of AI analysis and calculation processing analysis for each category (regular, retest, return) of a measurement instruction is provided. The user operates the check box of the upper list, selects which of the AI analysis and the calculation process analysis is to be performed for each measurement item, operates the check box of the lower list, selects which of the AI analysis and the calculation process analysis is to be performed for each category of the measurement instruction, and operates the setting button. Thereby, the rule is stored in the storage unit of the analysis unit 300.

When setting is made as shown in fig. 16, in the case where the kind of the measurement instruction is "normal", the analysis unit 300 determines the waveform data acquired by the measurement unit 400 as the object of the calculation processing analysis. For example, when the type of measurement instruction is "normal", calculation processing analysis is performed on all measurement items based on the measurement instruction, regardless of the setting of analysis for each measurement item. In addition, when the type of the measurement instruction is "retest" or "return", the analysis unit 300 sets the waveform data acquired by the measurement unit 400 as the object of the AI analysis or the calculation process analysis according to the analysis setting set for each measurement item. For example, when the types of measurement instructions are "retest" and "return", measurement items related to Nucleated Red Blood Cells (NRBC) and BASO are the subjects of AI analysis, and other measurement items are the subjects of calculation processing analysis.

Returning to fig. 15, in step S13, the analysis unit 300 determines whether or not the waveform data determined in step S61 includes data to be AI-analyzed. If there is waveform data to be AI-analyzed (yes in S13), in step S14, the analysis unit 300 performs AI analysis on the waveform data to be AI-analyzed determined in step S61.

Next, in step S15, the analysis unit 300 determines whether or not waveform data to be analyzed in the calculation process is included in the waveform data determined in step S61. When there is waveform data to be subjected to the calculation process analysis (yes in S15), in step S16, the analysis unit 300 performs the calculation process analysis on the waveform data to be subjected to the calculation process analysis determined in step S61.

Fig. 17 is a flowchart showing an example of deciding whether AI analysis is necessary or not based on a flag obtained by calculation processing analysis.

In fig. 17, steps S71 to S74 are added after step S11, and steps S12 to S16 are deleted, as compared with fig. 6. The following describes the change point with respect to fig. 6.

In step S71, the analysis unit 300 performs calculation processing analysis on the waveform data acquired in step S11, and sets a flag indicating an abnormality related to the analyte in the subject based on the result of the calculation processing analysis. The flag is, for example, a flag indicating that a predetermined abnormal cell is detected, a flag indicating that the count value of a predetermined blood cell is an abnormal value, or the like. In step S72, the analysis unit 300 refers to a rule including whether AI analysis is performed for the analysis result of the flag.

Fig. 18 is an explanatory diagram schematically showing a screen for setting AI analysis for each flag of the analysis result.

The screen of fig. 18 is displayed on a display unit provided in the analysis unit 300, for example. The screen of fig. 18 includes a check box for setting AI analysis for each flag given to the analysis result obtained by calculation processing analysis. When the sample analyzer 4000 is a blood cell analyzer, the markers given to the analysis result by the calculation processing analysis include cytopenia, abnormal cells, and the like. The user operates a check box to select whether AI analysis is performed for each flag, and operates a setting button. Thereby, the rule is stored in the storage unit of the analysis unit 300.

When a check box for a flag is checked, AI analysis is performed on waveform data corresponding to the analysis result. In the example shown in fig. 18, the check boxes are checked for the analysis results of the blast/abnormal lymphocyte, the blast, the abnormal lymphocyte, and the atypical lymphocyte, and when the markers for the presence of these blood cells are set by the calculation processing analysis, the AI analysis is performed on these blood cells.

Returning to fig. 17, in step S73, the analysis unit 300 determines whether the subject is an AI analysis target based on the flag given to the analysis result acquired in step S71 and the rule referred to in step S72. When the subject is the object of AI analysis (yes in S73), in step S74, the analysis unit 300 performs AI analysis on each waveform data. For example, in the case where a flag indicating that a parent cell is detected is obtained by calculation processing analysis, AI analysis is performed on each waveform data acquired in step S11 according to the rule illustrated in fig. 18.

On the other hand, when the subject is not the AI-analysis object (S73: no), the analysis unit 300 skips step S74.

Fig. 19 is a flowchart showing an example of performing AI analysis on a specific analyte classified by calculation processing analysis.

In fig. 19, steps S81 to S84 are added after step S11, and steps S12 to S16 are deleted, as compared with fig. 6. The following describes the change point with respect to fig. 6.

In step S81, the analysis unit 300 performs calculation processing analysis on the waveform data acquired in step S11, classifying the analytes. In step S82, the analysis unit 300 refers to a rule including whether AI analysis is performed for each category of analyte.

Fig. 20 is an explanatory diagram schematically showing a screen for setting whether or not AI analysis is performed for each category of analyte.

The screen of fig. 20 is displayed on a display unit provided in the analysis unit 300, for example. The screen of fig. 20 includes a check box for setting AI analysis for each type of analyte. In the case where the sample analyzer 4000 is a blood cell analyzer, the categories classified according to the calculation processing analysis include eosinophils, neutrophils, lymphocytes, monocytes, and the like. The user operates the check box to select whether AI analysis is performed for each category of analyte, and operates the setting button. Thereby, the rule is stored in the storage unit of the analysis unit 300.

When a check box for a category of an analyte is checked, AI analysis is performed on waveform data classified into the category. In the example shown in fig. 20, AI analysis was performed on monocytes and lymphocytes.

Returning to fig. 19, in step S83, analysis unit 300 determines waveform data corresponding to the analytes classified into the specific classifications (for example, monocytes and lymphocytes in the case of the rule shown in fig. 20) based on the categories of the analytes classified in step S81 and the rule referred to in step S82. In step S84, the analysis unit 300 performs AI analysis on the determined waveform data.

Fig. 21 is a diagram illustrating a classification method using calculation process analysis and AI analysis, which is performed in the process shown in fig. 19.

In the calculation processing analysis, a scatter diagram having two representative values calculated from waveform data as axes is used. For example, as shown in fig. 20, when it is set to perform AI analysis on monocytes and lymphocytes, in step S81, a plot (plot) of the surrounding area of the broken line corresponding to the monocytes and lymphocytes on the scattergram is determined by calculation processing analysis. Then, in step S83, waveform data corresponding to the plot of the surrounding area is determined, and in step S84, AI analysis is performed on the determined waveform data.

Returning to fig. 19, in step S17, the analysis unit 300 provides the analysis results of the calculation processing analysis and the AI analysis. At this time, the analysis unit 300 replaces the analysis result of the category as the AI analysis object among the analysis results obtained by the calculation processing analysis in step S81 with the analysis result obtained by the AI analysis and supplies it. Further, the analysis result of the calculation processing analysis may be provided together with the analysis result of the AI analysis.

Fig. 22 is a flowchart showing an example of performing AI analysis in the case where a specific classification is made by calculation processing analysis.

In fig. 22, steps S91 to S95 are added after step S11, and steps S12 to S16 are deleted, as compared with fig. 6. The following describes the change point with respect to fig. 6.

In step S91, the analysis unit 300 performs calculation processing analysis on the waveform data acquired in step S11, classifying the analytes. In step S92, the analysis unit 300 refers to a rule including whether AI analysis is performed on a specific class of analytes, for example, cells that are not present in the peripheral blood of a healthy person.

Fig. 23 is an explanatory diagram schematically showing a screen for setting whether or not AI analysis is performed on a specific class of analytes.

The screen of fig. 23 is displayed on a display unit provided in the analysis unit 300, for example. The screen of fig. 23 includes a check box for setting AI analysis for each type of specific analyte. In the case where the sample analyzer 4000 is a blood cell analyzer, the classification of the sample according to the calculation processing analysis may include a blast cell, an abnormal lymphocyte, a atypical lymphocyte, an immature granulocyte, and the like. The user operates the check box to select whether AI analysis is performed for each category of analyte, and operates the setting button. Thereby, the rule is stored in the storage unit of the analysis unit 300.

When a check box for a category of a specific analyte is checked, AI analysis is performed on waveform data classified into the category. In the example shown in fig. 23, AI analysis was performed on blast cells, abnormal lymphocytes, and atypical lymphocytes.

Returning to fig. 22, in step S93, analysis unit 300 determines whether or not an analyte classified into a specific class (for example, a blast cell, an abnormal lymphocyte, and a atypical lymphocyte in the case of the rule shown in fig. 23) is detected based on the class of the analyte classified in step S91 and the rule referred to in step S92. When an analyte classified as a specific class is detected (yes in S93), in step S94, the analysis unit 300 determines waveform data corresponding to the analyte classified as the specific class. In step S95, the analysis unit 300 performs AI analysis on the determined waveform data.

In step S17, the analysis unit 300 provides analysis results of the calculation processing analysis and the AI analysis. At this time, the analysis unit 300 replaces the analysis result of the category as the AI analysis object among the analysis results obtained by the calculation processing analysis in step S91 with the analysis result obtained by the AI analysis and supplies it. Further, the analysis result of the calculation processing analysis may be provided together with the analysis result of the AI analysis.

Embodiment 4

Embodiment 4 shows a detailed configuration example in which the flow cytometry-based analysis of an analyte is performed in the analyte analyzer 4000, and the calculation processing analysis and the AI analysis are performed in parallel.

As the sample to be measured by the sample analyzer 4000 of embodiment 4, a biological sample collected from a subject can be cited. The test object may include, for example, peripheral blood such as venous blood and arterial blood, urine, and body fluids other than blood and urine. Body fluids other than blood and urine may include, for example, bone marrow fluid, ascites, hydrothorax, cerebrospinal fluid, and the like. Hereinafter, a body fluid other than blood and urine may be referred to simply as "body fluid". The blood sample is not limited as long as it can count the number of cells and determine the type of cells. The blood is preferably peripheral blood. For example, as blood, there can be mentioned peripheral blood collected using an anticoagulant such as ethylenediamine tetraacetate sodium or potassium salt, heparin sodium, or the like. Peripheral blood may be obtained from arteries or veins.

In the present embodiment, the cell type to be determined is based on the cell type classified based on morphology, and is different depending on the type of the subject. When the subject is blood and the blood is collected from a healthy person, the cell type to be determined in the present embodiment includes, for example, nucleated cells such as nucleated red blood cells and white blood cells, erythrocytes, platelets, and the like. Nucleated cells include, for example, neutrophils, lymphocytes, monocytes, eosinophils, and basophils. Neutrophils include, for example, split nuclear neutrophils and rod nuclear neutrophils. When blood is collected from an unhealthy person, the nucleated cells may include at least one selected from the group consisting of immature granulocytes and abnormal cells, for example. Such cells are also included in the cell types to be determined in the present embodiment. The immature granulocytes may include cells such as posterior myeloid cells, myeloblasts, anterior myeloblasts, and myeloblasts.

In addition, the nucleated cells may contain abnormal cells not contained in the peripheral blood of a healthy person, in addition to normal cells. Examples of abnormal cells are cells, such as tumor cells, which appear when a predetermined disease is developed. In the case of the hematopoietic system, the predetermined disease may be, for example, a disease selected from the group consisting of myelodysplastic syndrome, acute myeloblastic leukemia, acute promyelocytic leukemia, acute myelomonocytic leukemia, acute monocytic leukemia, erythroleukemia, acute megakaryoblastic leukemia, acute myelogenous leukemia, acute lymphoblastic leukemia, chronic myelogenous leukemia, chronic lymphocytic leukemia, malignant lymphoma such as hodgkin's lymphoma and non-hodgkin's lymphoma, and multiple myeloma.

Further, examples of the abnormal cells include erythroblasts such as lymphoblasts, plasma cells, atypical lymphocytes, reactive lymphocytes, pre-erythroblasts, basophilic erythroblasts, multi-stained erythroblasts, orthochromatic erythroblasts, pre-megaerythroblasts, basophilic megaerythroblasts, multi-stained megaerythroblasts and orthochromatic megaerythroblasts, which are nucleated erythrocytes, and megakaryocytes including micromegakaryocytes (micromegakaryocells).

In the case where the test object is urine, the cell type to be determined in the present embodiment may include, for example, epithelial cells such as transitional epithelium and squamous epithelium, erythrocytes, leukocytes, and the like. Examples of the abnormal cells include fungi such as bacteria, filamentous fungi and yeasts, and tumor cells.

In the case where the subject is a body fluid containing no normal blood component such as ascites, hydrothorax, cerebrospinal fluid, etc., the cell type may include, for example, erythrocytes, leukocytes, and macropytes. The large cells mentioned here mean cells which are exfoliated from the peritoneum of the body cavity inner membrane or viscera and larger than leukocytes, and are conforming to, for example, mesothelial cells, tissue cells, tumor cells, and the like.

In the case where the subject is bone marrow fluid, the cell type to be determined in the present embodiment may include mature blood cells and immature blood cell line cells as normal cells. Mature blood cells include nucleated cells such as nucleated erythrocytes and leukocytes, erythrocytes, platelets and the like. Nucleated cells such as leukocytes include, for example, neutrophils, lymphocytes, plasma cells, monocytes, eosinophils, and basophils. Neutrophils include, for example, split nuclear neutrophils and rod nuclear neutrophils. Immature blood cell line cells include, for example, hematopoietic stem cells, immature granulocyte line cells, immature lymphocyte line cells, immature monocyte line cells, immature erythrocyte line cells, megakaryocyte line cells, mesenchymal cells, and the like. Immature granulocytes can include cells such as posterior myeloid cells, anterior myeloid cells, myeloblasts, and the like. Immature lymphocyte lineage cells include, for example, lymphoblasts and the like. Immature monocyte lineage cells include monocyte-forming cells, and the like. The immature erythrocyte cells include, for example, nucleated erythrocytes such as pre-erythroblasts, basophils, multi-dye erythroblasts, positive dye erythroblasts, pre-megaerythroblasts, basophils, multi-dye megaerythroblasts, and positive dye megaerythroblasts. Megakaryocyte lineage cells include, for example, megakaryoblasts and the like.

Examples of the abnormal cells that may be contained in the bone marrow include hematopoietic tumor cells selected from the group consisting of myelodysplastic syndrome, acute myeloblastic leukemia, acute promyelocytic leukemia, acute myelomonocytic leukemia, acute monocytic leukemia, erythroleukemia, acute megakaryoblastic leukemia, acute myelogenous leukemia, acute lymphoblastic leukemia, chronic myelogenous leukemia, chronic lymphocytic leukemia, and other leukemias, malignant lymphomas such as hodgkin's lymphoma and non-hodgkin's lymphoma, and multiple myeloma, and metastatic tumor cells of malignant tumors that are generated in organs other than the bone marrow.

The signals obtained from the analytes (e.g., cells and formed components) in the sample illustrated above are exemplified by forward scattered light signals, side scattered light signals, and fluorescence signals, which are analog optical signals obtained by irradiating the cells flowing in the flow cell with light, but there are no particular limitations on the signals as long as they can characterize the analytes and classify the analytes by the respective types.

The optical signal is preferably an optical signal obtained as an optical response by irradiating the cell with light. The optical signal may include at least one selected from a signal based on light scattering, a signal based on light absorption, a signal based on transmitted light, and a signal based on fluorescence.

The signal based on light scattering may include a scattered light signal generated by light irradiation and an optical loss signal generated by light irradiation. The scattered light signal characterizes the analyte in the subject according to the light-receiving angle of the scattered light with respect to the direction of travel of the illuminating light. The forward scattered light signal is used to calculate a representative value indicative of the size of the analyte. In the case where the analyte in the subject is a cell, the side scatter light signal is used to calculate a representative value indicative of the complexity of the cell nucleus.

"forward" of forward scattered light means the direction of travel of the light emitted from the light source. As the "forward direction", a forward low angle in which the light receiving angle is around 0 ° to 5 ° and/or a forward high angle in which the light receiving angle is around 5 ° to 20 ° in the case where the angle of the irradiation light is set to 0 ° may be included. The "lateral direction" is not limited as long as it does not coincide with the "forward direction". As the "lateral direction", in the case where the angle of the irradiation light is set to 0 °, a light receiving angle may be included in the vicinity of 25 ° to 155 °, preferably in the vicinity of 45 ° to 135 °, more preferably in the vicinity of 90 °. Fluorescence of the present embodiment is detected at the same light receiving angle as the side-scattered light.

The signal based on light scattering may comprise polarized light or depolarization as a component of the signal. For example, by receiving scattered light generated by irradiating light to an analyte in a subject through a polarizing plate, only the scattered light polarized at a specific angle can be received. Further, by allowing light to pass through the polarizing plate and irradiate the analyte in the subject, and allowing the generated scattered light to pass through the polarizing plate and transmit only polarized light having a different angle from the polarizing plate for irradiation and receive the scattered light, only depolarized scattered light can be received.

The optical loss signal represents a loss amount based on the light reception amount in the case where the light reception amount of the light receiving section is reduced due to light irradiation to the analyte and scattering. The optical loss signal is preferably obtained as an optical loss of the irradiation light in the optical axis direction (axial optical loss). The optical loss signal can be expressed as a ratio of the light receiving amount of the measurement sample flowing through the flow cell when the light receiving amount of the light receiving portion is 100% in a state where the measurement sample does not flow through the flow cell. The axial light loss is used to calculate a representative value representing the size of the analyte as is the case with forward scattered light signals, whereas the resulting signal is different in the case where the cell is transmissive and in the case where the cell is not transmissive.

The fluorescence-based signal may be fluorescence excited by irradiation of light to the analyte labeled with the fluorescent substance or may be autofluorescence generated from a non-stained analyte. When the analyte in the sample is a cell, the fluorescent substance may be a fluorescent dye that binds to a nucleic acid or a membrane protein, or may be a labeled antibody obtained by modifying an antibody that binds to a specific protein of the cell with a fluorescent dye.

The optical signal may be acquired by irradiating light to the analyte in the subject and capturing image data of the irradiated analyte. Image data can be obtained by capturing each analyte flowing through the flow path of the flow cell with a capturing element such as a TDI camera or a CCD camera. Alternatively, an object or a measurement sample containing cells may be coated, scattered, or spotted on a slide glass, and image data of the cells may be acquired by photographing the slide glass with a photographing element.

The signal obtained from the analyte in the subject is not limited to an optical signal, and may be an electrical signal obtained from a cell. The electrical signal may be, for example, a direct current applied to the flow cell, and a change in impedance due to the flow of the analyte in the flow cell is used as the electrical signal. The electrical signal thus obtained is used to calculate a representative value reflecting the volume of the analyte. Alternatively, the electrical signal may be a change in impedance upon application of a radio frequency to the analyte flowing in the flow cell. The electrical signal thus obtained is used to calculate a representative value reflecting the conductivity of the analyte.

As the signal obtained from the analyte in the subject, at least two or more of the above signals may be combined. By combining a plurality of signals, the characteristics of the analyte can be analyzed from multiple aspects, and cell classification can be performed with higher accuracy. The combination may be a combination of a plurality of optical signals, for example, at least two of a forward scattered light signal, a side scattered light signal, and a fluorescent signal, or a combination of scattered light signals having different angles, for example, a low-angle scattered light signal and a high-angle scattered light signal. Or the optical signal and the electric signal may be combined, and the kind and the number of the combined signals are not particularly limited.

In the present embodiment, the AI algorithm 60 used for AI analysis is, for example, a deep learning algorithm. The deep learning algorithm is one of artificial intelligence algorithms and is composed of a neural network including a plurality of intermediate layers. The data input to the neural network is processed through a number of matrix operations. From the digital data obtained by a/D converting the analog optical signal illustrated in fig. 2 described above, waveform data corresponding to each analyte is obtained, and the obtained waveform data is input to the AI algorithm 60 and analyzed. For example, the AI algorithm 60 is used to classify the type of analyte corresponding to the input waveform data.

In the present embodiment, the types of analytes in the subject are not limited to classification by the AI algorithm 60. The type of each analyte may be determined based on the result of recognizing, as a pattern, the signal intensities of the obtained plurality of time points related to each analyte, for each analyte, which are obtained at a plurality of time points during which the analyte passes through a predetermined position in the flow path. The pattern may be identified as a numerical pattern of signal intensities at a plurality of points in time, or may be identified as a shape pattern when the signal intensities at a plurality of points in time are plotted as a graph. In the case of recognition as a numerical pattern, the class of the analyte can be determined by comparing the numerical pattern of the analyte with the numerical patterns that are already known classes. As a comparison of the numerical pattern of the analyte with the numerical pattern of the control, for example, spearman ordering correlation, z-score, etc. can be used. By comparing the pattern of the graph shape of the analyte with the pattern of the graph shape that is already a known class, the class of analyte can be determined. As a comparison of the pattern of the graph shape of the analyte with the pattern of the graph shape already of a known class, geometric pattern matching may be used, for example, and a feature Descriptor represented by SIFT Descriptor (SIFT Descriptor) may also be used.

(structural example)

A configuration example of the sample analyzer 4000 in the case where the measurement unit 400 includes an FCM detection unit (a detection unit based on flow cytometry) for measuring a sample (for example, a blood sample, a urine sample, a body fluid, or a bone marrow fluid) will be described.

Fig. 24 is a block diagram showing the structure of the measurement unit 400.

As shown in fig. 24, the measurement unit 400 includes an FCM detection unit 410 for detecting an analyte in a subject, an analog processing unit 420 for processing an analog optical signal output from the FCM detection unit 410, a device mechanism unit 430, a sample preparation unit 440, a subject aspiration unit 450, and a measurement unit control unit 460.

The specimen suction unit 450 sucks the specimen from the specimen container, for example, and discharges the sucked specimen to a reaction container (for example, a reaction chamber or a reaction test tube). The sample preparation unit 440 sucks, for example, a reagent for preparing a measurement sample, and discharges the reagent to a reaction container containing a sample. A test sample is prepared by mixing a test object and a reagent in a reaction vessel. The device mechanism 430 includes a mechanism in the measurement unit 400.

Fig. 25 is a diagram schematically showing the configuration of an optical system of the FCM detection unit 410.

The light emitted from the light source 4111 is irradiated to an analyte in the measurement sample passing through the flow cell (sheath flow cell) 4113 via the irradiation lens system 4112. Thereby, scattered light and fluorescence are emitted from the analyte flowing in the flow cell 4113.

The wavelength of light emitted from the light source 4111 is not particularly limited, and a wavelength suitable for exciting the fluorescent dye is selected. As the light source 4111, for example, a semiconductor laser light source, an argon laser light source, a gas laser light source such as a helium-neon laser, a mercury arc lamp, or the like can be used. In particular, a semiconductor laser light source is very inexpensive compared with a gas laser light source, and is thus preferable.

The forward scattered light generated from the analyte in the flow cell 4113 is received by the light receiving element 4116 through the condenser lens 4114 and the pinhole portion 4115. The light receiving element 4116 is, for example, a photodiode. The side-scattered light generated from the analyte in the flow cell 4113 is received by the light receiving element 4121 through the condenser lens 4117, the dichroic mirror 4118, the bandpass filter 4119, and the pinhole 4120. The light receiving element 4121 is, for example, a photodiode. Fluorescence generated from the analyte in the flow cell 4113 is received by the light receiving element 4122 through the condenser lens 4117 and the dichroic mirror 4118. The light receiving element 4122 is, for example, an avalanche photodiode. Further, photomultiplier tubes may be used as the light receiving elements 4116, 4121, 4122.

The analog light receiving signals (optical signals) output from the light receiving elements 4116, 4121, 4122 are input to the analog processing section 420 via the amplifiers 4151, 4152, 4153, respectively.

The analog processing unit 420 performs processing such as noise removal and smoothing on the optical signal input from the FCM detection unit 410, and outputs the processed optical signal to the a/D conversion unit 461.

Returning to fig. 24, measurement unit control unit 460 includes an a/D conversion unit 461, an IF (interface) unit 462, a bus 463, and IF units 464 and 465.

The a/D converter 461 converts the analog optical signal outputted from the analog processor 420 from the start of measurement to the end of measurement of the measurement sample into digital data. When a plurality of optical signals (for example, optical signals corresponding to the forward scattered light intensity, the side scattered light intensity, and the fluorescence intensity, respectively) are generated from 1 measurement sample, the a/D converter 461 converts each optical signal into digital data from the start of measurement to the end of measurement. For example, as shown in fig. 25, 3 kinds of optical signals (forward scattered light signal, side scattered light signal, and fluorescent signal) are input to the a/D converter 461 via a plurality of signal transmission paths 420a corresponding to each other. The a/D converter 461 converts the optical signals input from the plurality of signal transmission paths 420a into digital data, respectively. Each of the signal transmission paths 420a is configured to transmit an analog optical signal as a differential signal, for example.

The a/D conversion section 461 compares the signal level of the optical signal with a predetermined threshold value, and samples the optical signal having a signal level greater than the threshold value. The a/D conversion section 461 samples the analog optical signal at a predetermined sampling rate (for example, sampling at 1024 points at 10 nanosecond intervals, sampling at 128 points at 80 nanosecond intervals, sampling at 64 points at 160 nanosecond intervals, or the like). The a/D converter 461 generates digital data (waveform data) of forward scattered light signals, digital data (waveform data) of side scattered light signals, and digital data (waveform data) of fluorescent signals for each analyte by performing sampling processing on 3 kinds of optical signals corresponding to each analyte, for example. Each digital data (waveform data) corresponds to each analyte in the subject.

The a/D converter 461 gives an index to each generated waveform data. The generated waveform data is, for example, digital data corresponding to each of N analytes contained in 1 part of the subject. Thus, 3 kinds of waveform data are generated for each analyte corresponding to 3 kinds of optical signals (forward scattered light signal, side scattered light signal, and fluorescence signal).

The waveform data generated by the a/D converter 461 is sent to the analysis unit 300 via the IF units 462 and 465 and the bus 463. The analysis unit 300 controls the device mechanism 430, the sample preparation unit 440, and the sample aspiration unit 450 via the IF units 464 and 465 and the bus 463.

Fig. 26 is a block diagram showing the structure of the analysis unit 300.

The analysis unit 300 includes a processor 3001, a RAM 3017, a bus 3003, a storage unit 3004, an IF unit 3006, a display unit 3011, and an operation unit 3012. The analysis unit 300 includes, for example, a personal computer. The analysis unit 300 is connected to the measurement unit 400 via the IF unit 3006.

The processor 3001 includes, for example, a CPU. The processor 3001 executes a program developed from the storage portion 3004 to the RAM 3017. RAM 3017 is a so-called main memory. The processor 3001 analyzes the waveform data acquired by the measurement unit 400 by executing the analysis program. The processor 3001 executes a control program to control the analysis unit 300 and the measurement unit 400.

The storage unit 3004 includes, for example, a Hard Disk Drive (HDD) and a Solid State Drive (SSD). The storage unit 3004 stores waveform data received from the measurement unit 400, programs for controlling the analysis unit 300 and the measurement unit 400, and programs for analyzing the waveform data. The program for analyzing waveform data is configured to analyze waveform data based on the above-described calculation processing analysis and AI analysis. The storage 3004 stores rules for specifying waveform data to be analyzed by the AI analysis and the calculation process, and rules for selecting analysis operations.

The display unit 3011 includes, for example, a liquid crystal display. The display unit 3011 is connected to the processor 3001 via a bus 3003 and an IF unit 3006. The display unit 3011 displays, for example, the analysis result obtained by the measurement unit 400.

The operation unit 3012 is configured of a pointing device including a keyboard, a mouse, and a touch panel, for example. By operating the operation unit 3012, a user such as a doctor or a test technician can input a measurement instruction to the object analysis device 4000, and can input a measurement instruction based on the measurement instruction. The user can also input an instruction to display the analysis result by operating the operation unit 3012. The analysis results include, for example, numerical results based on analysis, graphs, charts, flag information given to the subject, and the like.

Fig. 27 is a block diagram showing a configuration of measurement unit 400 in the case where sample analyzer 4000 performs counting and sorting of blood cells in a blood sample.

The measurement unit 400 of fig. 27 includes, in addition to the configuration of fig. 24, an RBC/PLT detecting unit 4101, an HGB detecting unit 4102, analog processing units 4201 and 4202, and a/D converting units 4611 and 4612.

The RBC/PLT detecting unit 4101 is a resistive detecting unit, and performs measurement of blood cells by a sheath flow DC detection method based on an RBC/PLT measurement sample. The HGB detecting unit 4102 measures hemoglobin by the SLS-hemoglobin method based on the hemoglobin measurement sample. The data obtained by a/D converting the analog signals acquired from the RBC/PLT detecting unit 4101 and the HGB detecting unit 4102 are the calculation processing analysis targets. The RBC/PLT detector 4101 counts the number of red blood cells and platelets in the blood sample. The hemoglobin amount in the blood test object is acquired by using data based on the HGB detecting unit 4102.

The data obtained by a/D converting the analog signals acquired from the RBC/PLT detecting unit 4101 and the HGB detecting unit 4102 may be the AI analysis target. The data based on the RBC/PLT detecting unit 4101 and the HGB detecting unit 4102 may be analyzed by AI analysis and calculation processing. This reduces the load on the analysis unit 300 for processing data.

Fig. 28 is a block diagram showing the structures of the specimen suction unit 450 and the sample preparation unit 440 in the measurement unit 400 of fig. 27.

The specimen suction unit 450 includes a nozzle 451 for sucking a blood specimen (for example, whole blood) from the blood collection tube TB, and a pump 452 for applying negative and positive pressures to the nozzle. The nozzle 451 is inserted into the blood collection tube TB by being moved up and down by the device mechanism 430 (see fig. 27). When negative pressure is applied to the pump 452 in a state where the nozzle 451 is inserted into the blood collection tube TB, the blood sample is sucked through the nozzle 451. The device mechanism 430 may further include a manual member for stirring the blood collection tube TB upside down before sucking the TB blood from the blood collection tube.

The sample preparation unit 440 includes a WDF sample preparation unit 440a, a RET sample preparation unit 440b, a WPC sample preparation unit 440c, a PLT-F sample preparation unit 440d, and a WNR sample preparation unit 440e. The sample preparation units 440a to 440e each include a reaction chamber for mixing a sample and a reagent (e.g., a hemolyzing agent and a staining solution). The sample preparation sections 440a to 440e are used in WDF channel, RET channel, WPC channel, PLT-F channel, and WNR channel, respectively.

Here, the analyte analyzer 4000 includes a plurality of measurement channels corresponding to a plurality of measurement samples to be prepared, respectively. The analyte analyzer 4000 includes, for example, a WDF channel, a RET channel, a WPC channel, a PLT-F channel, and a WNR channel. The WDF channel is a channel for detecting neutrophils, lymphocytes, monocytes, and eosinophils. RET channels are channels used to detect reticulocytes. The WPC channel is a channel for detecting abnormal cells of a blast cell and lymphocyte line. The PLT-F channel is a channel for detecting platelets. The WNR channel is a channel for detecting leukocytes other than basophils, and nucleated erythrocytes.

The hemolytic agent container containing a hemolytic agent which is a reagent corresponding to the measurement channel and the staining solution container containing a staining solution are connected to the sample preparation sections 440a to 440e via flow paths. For example, a hemolyzing agent container containing a WDF hemolyzing agent (for example, lysercell WDF II, manufactured by hsienkang corporation) as a reagent for WDF measurement and a staining solution container containing a WDF staining solution (for example, fluorocell WDF, manufactured by hsienkang corporation) are connected to the WDF sample preparation unit 440a via a flow path. Although the configuration in which 1 sample preparation unit is connected to the hemolytic agent container and the staining solution container is exemplified here, the 1 sample preparation unit is not necessarily connected to both the hemolytic agent container and the staining solution container, and 1 reagent container may be shared by a plurality of sample preparation units. The sample preparation unit and the reagent container are not necessarily connected by a flow path, and may have the following structure: the reagent is sucked from the reagent container by the nozzle, and the nozzle is moved to discharge the sucked reagent from the nozzle into the reaction chamber of the sample preparation section.

By the horizontal and vertical movement by the device mechanism 430, the nozzle 451 that suctions the blood sample is positioned above the reaction chamber of the sample preparation unit corresponding to the measurement instruction among the sample preparation units 440a to 440 e. In this state, when positive pressure is applied to the pump 452, the blood sample is discharged from the nozzle 451 to the corresponding reaction chamber. The sample preparation unit 440 supplies a corresponding hemolyzing agent and staining solution to the reaction chamber from which the blood sample is discharged, and mixes the blood sample, the hemolyzing agent, and the staining solution in the reaction chamber to prepare a measurement sample.

A WDF measurement sample is prepared in the WDF sample preparation unit 440a, a RET measurement sample is prepared in the RET sample preparation unit 440b, a WPC measurement sample is prepared in the WPC sample preparation unit 440c, a PLT-F measurement sample is prepared in the PLT-F sample preparation unit 440d, and a WNR measurement sample is prepared in the WNR sample preparation unit 440 e. The prepared measurement sample is supplied from the reaction chamber to the FCM detection unit 410 via the flow path, and the flow cytometry-based cell measurement is performed.

The measurement channels (WDF, RET, WPC, PLT-F, WNR) correspond to measurement items included in measurement instructions. For example, WDF channels correspond to measurement items related to the classification of leukocytes, RET channels correspond to measurement items related to reticulocytes, PLT-F channels correspond to measurement items related to platelets, and WNR channels correspond to measurement items related to the number of leukocytes and nucleated erythrocytes. The measurement sample prepared in the measurement channel is measured by the FCM detector 410.

The measurement result of the RBC/PLT detecting unit 4101 corresponds to a measurement item related to the number of red blood cells. The measurement result of the HGB detecting unit 4102 corresponds to a measurement item related to the hemoglobin amount.

Fig. 29 is a block diagram showing another configuration of the sample preparation section 440 shown in fig. 28.

In the example shown in fig. 29, the configuration of the measurement channel of the sample preparation unit 440 is changed according to the contribution of the AI analysis and the calculation process analysis. Specifically, as compared with the sample preparation unit 440 of fig. 28, the sample preparation unit 440 of fig. 29 adds a WDF sample preparation unit 440a for a WDF channel and a reagent (WDF hemolytic agent and WDF dye solution) connected to the WDF sample preparation unit 440a instead of the WNR sample preparation unit 440e for a WNR channel and a reagent (WNR hemolytic agent and WNR dye solution) connected to the WNR sample preparation unit 440 e. That is, the sample preparation unit 440 of fig. 29 includes two sets of a WDF sample preparation unit 440a and a reagent corresponding to the WDF channel. The sample preparation unit 440 may include 3 or more groups of WDF sample preparation units 440a and reagents corresponding to the WDF channels.

In the case where the sample preparation unit 440 is configured as shown in fig. 29, the classification of basophils and nucleated erythrocytes performed on the WNR channel is performed on the WDF channel. The analysis unit 300 classifies the medium granulocytes, lymphocytes, monocytes, eosinophils, basophils and nucleated erythrocytes by performing AI analysis on waveform data obtained from a measurement sample prepared on a WDF channel. In this case, for example, waveform data corresponding to neutrophils, lymphocytes, monocytes, eosinophils, basophils, and nucleated erythrocytes among waveform data obtained by measurement of the WDF channel is used as teacher data in advance to learn the AI algorithm 60. Thus, an AI algorithm 60 capable of classifying neutrophils, lymphocytes, monocytes, eosinophils, basophils, and nucleated erythrocytes is generated from the waveform data of the WDF channel.

According to the configuration of fig. 29, for example, measurement samples of different samples can be prepared in parallel in the reaction chambers of the plurality of WDF sample preparation units 404 a. Thus, measurement of WDF channels for different subjects can be performed in parallel.

In the configuration of fig. 29, the reaction chamber and the reagent corresponding to the original measurement channel (WNR channel) are replaced with the reaction chamber and the reagent corresponding to the subsequent measurement channel (WDF channel). In this case, the analysis of the original measurement channel needs to be performed by the analysis of the subsequent measurement channel.

In the configuration of fig. 29, the analysis of the original measurement channel (WNR channel) is performed by AI analysis based on waveform data of the subsequent measurement channel (WDF channel). This allows the original measurement channel (WNR channel) to be replaced with a subsequent measurement channel (WDF channel). Thus, the number of measurement channels to be added can be increased without increasing the total number of measurement channels provided in the analyte analyzer 4000. By increasing the number of WDF channels, it is possible to measure different samples in parallel with a plurality of WDF channels, and throughput of measurement by the WDF channels is improved. Further, by sharing the AI analysis and the calculation processing analysis, it is possible to obtain a significant effect that the computational load required for the AI analysis is reduced and the throughput of the subject processing is improved.

In the process described with reference to fig. 7, an example in which the analysis is performed by either the AI analysis or the calculation process analysis according to the measurement item is shown, but it is also possible to determine which of the AI analysis and the calculation process analysis is performed according to the measurement channel.

Fig. 30 is a flowchart showing an example of performing analysis according to a measurement channel.

In fig. 30, step S101 is added instead of step S12, as compared with fig. 6. The following describes the change point with respect to fig. 6.

In step S101, the analysis unit 300 refers to a rule including which of the AI analysis and the calculation process analysis is performed based on the measurement channel, and determines waveform data to be subjected to the AI analysis and waveform data to be subjected to the calculation process analysis for the waveform data acquired in step S11 based on the referenced rule.

Fig. 31 is an explanatory diagram schematically showing a screen for setting AI analysis or calculation processing analysis for each measurement channel. The measurement channel illustrated in fig. 31 is related to a blood cell analysis device.

The screen of fig. 31 is displayed on a display unit provided in the analysis unit 300, for example. The screen of fig. 31 includes a check box for setting AI analysis and a check box for setting calculation processing analysis for each measurement channel. The user operates a check box to select which of the AI analysis and the calculation processing analysis is to be performed for each measurement channel, and operates a setting button. Thereby, the rule is stored in the storage unit of the analysis unit 300.

Returning to fig. 30, in the case where the waveform data to be AI-analyzed is included in the waveform data determined in step S101 (yes in S13), in step S14, the analysis unit 300 performs AI analysis on the waveform data to be AI-analyzed determined in step S101. When the waveform data to be analyzed by the calculation process is included in the waveform data determined in step S101 (yes in step S15), in step S16, the analysis unit 300 performs the analysis by the calculation process on the waveform data to be analyzed by the calculation process determined in step S101.

In the case of fig. 31, since the WDF channel is set as the object of AI analysis, analysis of a measurement item (for example, leukocyte class) corresponding to the WDF channel is performed by AI analysis on waveform data obtained from a measurement sample prepared on the WDF channel. On the other hand, since the other channel is set as the object of the calculation processing analysis, the analysis of the measurement items corresponding to the other channel is performed by the calculation processing analysis with respect to the waveform data obtained from the measurement sample prepared on the other channel.

In the case where the WDF channel is set as the target of the AI analysis, the AI analysis may be performed on all measurement items corresponding to the WDF channel, or the AI analysis may be performed on some of the measurement items corresponding to the WDF channel, or the calculation process analysis may be performed on the remaining measurement items.

< example of method for analyzing analyte in analyte >

Next, a method of generating training data 75 and a method of analyzing waveform data will be described with reference to examples shown in fig. 32 to 35.

< waveform data >

Fig. 32 is a schematic diagram for explaining waveform data used in the present analysis method.

As shown in the upper diagram of fig. 32, when a measurement sample prepared from an analyte containing an analyte a flows into the flow cell 4113 and light is irradiated to the analyte a flowing in the flow cell 4113, forward scattered light is generated in the forward direction with respect to the traveling direction of the light. Similarly, side scattered light and fluorescence are generated in the lateral direction with respect to the traveling direction of the light. The forward scattered light, the side scattered light, and the fluorescent light are received by the light receiving elements 4116, 4121, 4122, respectively, and output signals corresponding to the light receiving intensities. Thereby, analog optical signals representing changes in the signals with time are output from the light receiving elements 4116, 4121, 4122, respectively. The optical signal corresponding to the forward scattered light is referred to as a "forward scattered light signal", the optical signal corresponding to the side scattered light is referred to as a "side scattered light signal", and the optical signal corresponding to the fluorescence is referred to as a "fluorescence signal". The optical signal is input to the a/D converter 461 and converted into digital data.

The diagram in the middle of fig. 32 is a diagram schematically showing conversion into digital data by the a/D conversion section 461. Here, the analog optical signal is directly input to the a/D converter 461. The level of the optical signal may be converted into digital data as it is, or may be subjected to noise removal, baseline correction, normalization, or the like as appropriate.

As shown in the middle diagram of fig. 32, the a/D converter 461 samples the forward scattered light signal, the side scattered light signal, and the fluorescent signal from the analog optical signals input from the light receiving elements 4116, 4121, and 4122, with a point of time when the level of the forward scattered light signal exceeds a predetermined threshold value being the start point and a point of time when the level of the forward scattered light signal is below the predetermined threshold value being the end point. Using the waveform from the start point to the end point, digital waveform data corresponding to 1 analyte is acquired. The a/D conversion section 461 samples each optical signal at a predetermined sampling rate (for example, sampling at 1024 points at 10 ns intervals, sampling at 128 points at 80 ns intervals, sampling at 64 points at 160 ns intervals, or the like).

In this case, for convenience, the start point and the end point are set for the analog optical signal to acquire the waveform data, but as described above, the start point and the end point may be set for the digital data to acquire the waveform data after the optical signal is converted into the digital data.

The lower diagram of fig. 32 is a diagram schematically showing waveform data obtained by sampling. By sampling, as waveform data corresponding to 1 analyte, matrix data (one-dimensional array data) having values representing analog signal levels at a plurality of time points as numbers as elements can be obtained. The a/D converter 461 generates waveform data of forward scattered light, waveform data of side scattered light, and waveform data of fluorescence for each analyte. The a/D converter 461 repeatedly generates waveform data until the number of acquired analytes reaches a predetermined number or a predetermined time elapses from the flow cell 4113 of the subject flow. Thus, digital data composed of waveform data of N analytes contained in 1 part of the subject was obtained. The sampled data set for each analyte (e.g., a set of 1024 digital values every 10 nanoseconds from t=0 ns to t=10240 ns) corresponds to waveform data.

An index for identifying each analyte may be given to each waveform data generated by the a/D conversion section 461. For example, integers 1 to N are given to the indices in the order of the generated waveform data, and the same index is given to the waveform data of the forward scattered light, the waveform data of the side scattered light, and the waveform data of the fluorescence obtained from the same analyte. By assigning the same index to the waveform data corresponding to the same analyte, the AI algorithm 60 described later can analyze the waveform data of forward scattered light, the waveform data of side scattered light, and the waveform data of fluorescence corresponding to each analyte as 1 group to classify the types of the analytes.

< generation of training data >

Fig. 33 is a schematic diagram showing an example of a method for generating training data used for training the AI algorithm 50 for determining the type of the analyte in the subject.

By measuring an analyte based on flow cytometry, an optical signal 70a corresponding to forward scattered light, an optical signal 70b corresponding to side scattered light, and an optical signal 70c corresponding to fluorescence are obtained from the analyte. Waveform data 72a, 72b, 72c corresponding to the analyte are acquired based on the optical signals 70a, 70b, 70c, respectively. As the training data 75, for example, the following data can be used: the analysis of the analytes in the test object measured by the flow cytometry is performed by the calculation processing, and the result is judged to be the waveform data 72a, 72b, 72c of the analytes of the specific class with high probability.

An example of the sample analyzer 4000 which is used as a blood cell counting device for analyzing a blood sample will be described below.

The operator measures a blood sample with the FCM detector 410, and accumulates waveform data of forward scattered light, side scattered light, and fluorescence of each analyte contained in the sample. Next, the operator classifies analytes (cells) in the test object into a population of neutrophils, lymphocytes, monocytes, eosinophils, basophils, immature granulocytes, and abnormal cells, for example, based on the peak value of the waveform data based on the side scattered light and the peak value of the waveform data based on the fluorescence. The operator obtains training data 75 by assigning a label value 77 corresponding to the classified cell type to the waveform data of the cell. Since the training data 75 is generated for each class of cells, as shown in fig. 34, the tag value 77 differs according to the class of cells.

At this time, the operator obtains the mode, average value, or median of the peaks of the waveform data of the cells included in the population of neutrophils based on the side scattered light and fluorescence, identifies the representative cells based on these values, and assigns a tag value "1" corresponding to the identified waveform data of the cells.

The method of generating the training data 75 is not limited to this, and for example, an operator may collect only specific cells by a cell sorter (cell sorter), measure the cells by flow cytometry, and assign a label value to the obtained waveform data to obtain the training data 75.

The waveform data 72a, 72b, 72c are combined with a tag value 77 representing the class of cells as the data source. The training data 75 includes 3 pieces of waveform data (waveform data based on the optical signals 70a, 70b, 70 c) corresponding to each cell in a state of being associated with the waveform data. The training data 75 is then input to the AI algorithm 50.

< overview of deep learning >

An outline of training of the neural network will be described with reference to fig. 33.

The AI algorithm 50 is comprised of a neural network including multiple intermediate layers. The neural network in this case is, for example, a convolutional neural network having a convolutional layer. The number of nodes of the input layer 50a in the neural network corresponds to the number of elements of the array contained in the waveform data 72a, 72b, 72c of the input training data 75. The number of elements of the array is equal to the sum of the number of elements of the waveform data 72a, 72b, 72c of the forward scattered light, the side scattered light, and the fluorescence corresponding to 1 analyte.

In the example of fig. 33, the waveform data 72a, 72b, 72c each include 1024 elements, and thus the number of nodes of the input layer 50a is 1024×3=3072. The waveform data 72a, 72b, 72c is input to the input layer 50a of the neural network. The label value 77 of each waveform data of the training data 75 is input to the output layer 50b of the neural network, and the neural network is trained. The intermediate layer 50c is located between the input layer 50a and the output layer 50 b.

< method of analyzing waveform data >

Fig. 35 is a diagram schematically showing a method of analyzing waveform data of an analyte in a subject using the AI algorithm 60.

In the analysis method of waveform data illustrated in fig. 35, an analyte is measured by flow cytometry, and an optical signal 80a corresponding to forward scattered light, an optical signal 80b corresponding to side scattered light, and an optical signal 80c corresponding to fluorescence are obtained from the analyte. Waveform data 82a, 82b, 82c corresponding to the analyte are acquired based on the optical signals 80a, 80b, 80c, respectively. Then, analysis data 85 including waveform data 82a, 82b, 82c is generated.

It is preferable that at least the acquisition conditions of the analysis data 85 and the training data 75 are the same. The acquisition conditions are conditions for measuring an analyte in an analyte by flow cytometry, and include, for example, conditions for preparing a measurement sample, a flow rate at which the measurement sample is caused to flow through a flow cell, an intensity of light irradiated to the flow cell, and an amplification factor of a light receiving portion that receives scattered light and fluorescence. The acquisition conditions also include a sampling rate at which the analog optical signal is a/D converted.

The analysis data 85 includes 3 pieces of waveform data (waveform data based on the optical signals 80a, 80b, 80 c) corresponding to the respective analytes in a state that these waveform data are associated. The analysis data 85 is then input to the trained AI algorithm 60. The AI algorithm 60 is composed of a neural network including multiple intermediate layers.

When the analysis data 85 is input to the input layer 60a of the neural network constituting the AI algorithm 60, classification information 82 relating to the category of the analyte corresponding to the analysis data 85 is output from the output layer 60 b. The intermediate layer 60c is located between the input layer 60a and the output layer 60 b. Classification information 82 includes probabilities that the analyte meets each of a plurality of categories. Further, the highest-probability class is determined as the class to which the analyte belongs, and a tag value 83 which is an identifier indicating the class and an analysis result 84 which is a character string indicating the class are output.

In the example of fig. 35, since the probability that the analyte type corresponding to the analysis data 85 is neutrophils is highest, a "1" is output as the label value 83, and text data such as "neutrophils" is output as the analysis result 84. The AI algorithm 60 may output the tag value 83 and the analysis result 84, or another computer program may output the most preferable tag value 83 and analysis result 84 based on the probability calculated by the AI algorithm 60.

The method of analyzing waveform data of the examples shown in fig. 19 to 21 will be described with reference to fig. 32 and 35.

In the case of the examples shown in fig. 19 to 21, first, the analysis unit 300 analyzes the acquired waveform data by calculation processing. After that, the analysis unit 300 performs AI analysis on waveform data corresponding to the predetermined class of cells (monocytes and lymphocytes in the example of fig. 19 to 21) classified by the calculation process analysis.

When classified as a predetermined cell by calculation processing analysis, the cell is identified, for example, with an index of waveform data of fig. 32. In this way, the waveform data classified into monocytes and lymphocytes by the calculation process analysis is determined by using the index in the AI analysis. In accordance with the example of fig. 35, the analysis unit 300 performs AI analysis on the waveform data determined based on the index. The analysis unit 300 inputs the waveform data determined by the index, for example, to the AI algorithm 60, and the AI algorithm 60 is learned to be able to classify monocytes and lymphocytes in more detail.

The method of analyzing waveform data described with reference to fig. 29 will be described with reference to fig. 32 and 35.

In the analysis method described with reference to fig. 29, for example, the analysis unit 300 performs classification and counting of Nucleated Red Blood Cells (NRBC) and classification and counting of BASO, and further performs classification and counting of eosinophils, neutrophils, lymphocytes, and monocytes by AI analysis of waveform data obtained by measurement of WDF channels. In the case of this example, the AI algorithm 60 learns to enable classification of nucleated erythrocytes, basophils, eosinophils, neutrophils, lymphocytes and monocytes using waveform data. By using such an AI algorithm 60, the WNR channel can be replaced with a WDF channel.

Fig. 36 is a flowchart showing an example of performing AI analysis on waveform data acquired at a WDF channel.

In step S111, the measurement unit 400 acquires an optical signal from a measurement sample prepared on the WDF channel, and acquires waveform data from the acquired optical signal. In step S112, the analysis unit 300 performs AI analysis on the waveform data acquired in step S111. In step S113, the analysis unit 300 provides the analysis result of the waveform data of the WDF channel and the analysis result of the waveform data of the other channels together. For example, it is determined how to share analysis of waveform data of other channels with AI analysis and calculation processing analysis based on any of the rules exemplified in the above embodiments.

In another analysis method based on the structure shown in fig. 29, the analysis unit 300 performs, for example, classification and counting of nucleated red blood cells and classification and counting of basophils by AI analysis of waveform data obtained in the WDF channel. The analysis unit 300 performs calculation processing analysis on waveform data corresponding to cells that are not classified into any of nucleated erythrocytes and basophils, and performs classification and counting of eosinophils, neutrophils, lymphocytes, and monocytes. In the case of this example, the AI algorithm 60 learns to be able to classify analytes into nucleated erythrocytes, basophils, and other analytes, for example, based on waveform data.

The analysis unit 300 performs calculation processing analysis on waveform data corresponding to cells that are not classified as either of nucleated red blood cells and basophils. For example, peaks of waveform data corresponding to cells that are not classified as either nucleated red blood cells or basophils are extracted, and cell types are classified based on a two-dimensional graph (scatter diagram) generated from peaks corresponding to side scattered light and peaks corresponding to fluorescence. For example, based on the two-dimensional graph, cells are classified into any of eosinophils, neutrophils, lymphocytes, monocytes, and other cells. Cells other than eosinophils, neutrophils, lymphocytes and monocytes are classified as fragments (Debris) in a two-dimensional graph-based analysis, for example.

Fig. 37 is a flowchart showing an example of classifying nucleated erythrocytes and basophils by AI analysis and classifying others by computational processing analysis based on waveform data acquired in the WDF channel.

In step S121, the measurement unit 400 acquires an optical signal from a measurement sample prepared on the WDF channel, and acquires waveform data from the acquired optical signal. In step S122, the analysis unit 300 performs AI analysis on the waveform data acquired in step S121. From this, nucleated erythrocytes and basophils were classified. In step S123, the analysis unit 300 determines waveform data corresponding to cells classified into neither nucleated red blood cells nor basophils.

In step S124, the analysis unit 300 performs calculation processing analysis on the waveform data determined in step S123. Thus, lymphocytes, monocytes, eosinophils and neutrophils were classified. In step S125, the analysis unit 300 provides the analysis result of the waveform data of the WDF channel and the analysis result of the waveform data of the other channels together.

In other analysis methods based on the structure shown in fig. 29, the analysis unit 300 performs classification and counting of lymphocytes, classification and counting of monocytes, classification and counting of eosinophils, and classification and counting of neutrophils or basophils, for example, by computational processing analysis of waveform data obtained at WDF channels. In the classification and counting of neutrophils or basophils, for example, cells classified as either neutrophils or basophils are counted. Next, the analysis unit 300 performs AI analysis on waveform data corresponding to lymphocytes, monocytes, eosinophils, and cells classified as neither neutrophils nor basophils, and cells classified as either neutrophils or basophils. Thus, analytes are classified into nucleated erythrocytes, basophils, and other cells.

For example, the count results of the cells classified as basophils by the AI analysis are subtracted from the count results of the cells classified as either neutrophils or basophils by the calculation process analysis, and the count results of the neutrophils and basophils are calculated, respectively. Cells that are classified by AI analysis as neither nucleated red blood cells nor basophils are classified, for example, as fragments (Debris).

Fig. 38 is a flowchart showing an example of performing AI analysis on neutrophils/basophils determined by analysis of a calculation process on a WDF channel.

In step S131, the measurement unit 400 acquires an optical signal from a measurement sample prepared on the WDF channel, and acquires waveform data from the acquired optical signal. In step S132, the analysis unit 300 performs calculation processing analysis on the waveform data acquired in step S131. Thus, lymphocytes, monocytes, eosinophils, and a group consisting of neutrophils and basophils are classified. In step S133, the analysis unit 300 determines waveform data corresponding to the following (1) and (2): (1) Lymphocytes, monocytes, eosinophils, and cells classified as neither neutrophils nor basophils, (2) cells classified as neutrophils or basophils.

In step S134, the analysis unit 300 performs AI analysis on the waveform data determined in step S133. Thus, neutrophils and basophils were classified. In step S135, the analysis unit 300 provides the analysis result of the waveform data of the WDF channel and the analysis result of the waveform data of the other channels together.

Embodiment 5

Embodiment 5 shows a detailed configuration example in which calculation processing analysis and AI analysis are performed in a sample analyzer 4000 for analyzing the clotting ability of a blood sample.

As the sample to be measured by the sample analyzer 4000 of embodiment 5, a biological sample collected from a subject can be cited. The test substance may include, for example, whole blood, plasma, and the like. The analyte analyzer 4000 of embodiment 5 analyzes the presence or absence of an abnormality caused by an interfering substance in the analyte based on a coagulation method, a synthetic matrix method, an immunoturbidimetry method, a coacervation method, a chemiluminescent enzyme immunoassay (CLEIA method), or the like. For example, as in the configuration example of embodiment 1 shown in fig. 1, the analyte analyzer 4000 of embodiment 5 includes a measurement unit 400 and an analysis unit 300.

(structural example)

Fig. 39 is a block diagram schematically showing the structure of measurement unit 400 according to embodiment 5.

In comparison with the measurement unit 400 shown in fig. 24, the measurement unit 400 of fig. 39 includes a detection unit 470 instead of the FCM detection unit 410, and also includes a control unit 466.

The detection unit 470 includes a light source unit 471 and a detection block 476. The light source 471 includes, for example, a halogen lamp. The light source 471 is configured to emit, for example, 660nm light for coagulation time measurement, 405nm light for synthetic matrix measurement, and 800nm light for immunonephelometry measurement. The sample preparation unit 440 mixes a blood coagulation reagent with a sample to prepare a measurement sample. The detection unit 470 irradiates a measurement sample composed of a blood coagulation reagent and a sample with light from the light source unit 471, and detects light transmitted through the sample. The detection unit 470 may irradiate the measurement sample with light from the light source unit 471, and detect light scattered by the object.

The control unit 466 includes, for example, an FPGA. Control unit 466 is connected to analysis unit 300 via bus 463 and IF unit 465. The control unit 466 controls each unit of the measurement unit 400 based on the instruction from the analysis unit 300.

Fig. 40 is a side view schematically illustrating the assay performed by the detection block 476.

The detection block 476 includes a holding portion 472, an optical fiber 473, a condensing lens 474, and a light receiving portion 475.

The reaction container C1 containing a measurement sample prepared from the sample corresponding to the measurement item and the object is held by the holding unit 472. Thereby, the measurement sample is allowed to stand. Light emitted from the light source 471 (see fig. 39) is guided to the condenser lens 474 by the optical fiber 473. The condensing lens 474 condenses the light from the optical fiber 473 on the reaction container C1. The transmitted light condensed in the reaction container C1 and passing through the measurement sample in the reaction container C1 is received by the light receiving unit 475. The light receiving portion 475 is, for example, a photodiode. The light receiving section 475 outputs an optical signal based on the intensity of the received transmitted light.

Returning to fig. 39, the analog processing unit 420 processes the analog optical signal output from the light receiving unit 475 (see fig. 40) and outputs the processed signal to the a/D conversion unit 461. The a/D converter 461 converts the analog optical signal into digital. As described above, the digital data of the optical signal after being digitally converted is the solidification waveform data as shown in fig. 4. The control unit 466 transmits the acquired solidification waveform data to the analysis unit 300.

Fig. 41 is a flowchart showing an analysis example of embodiment 5. In embodiment 5, the processor 3001 (see fig. 26) of the analysis unit 300 performs calculation processing analysis and AI analysis on the solidification waveform data.

In step S141, the measurement unit 400 acquires an optical signal at the detection unit 470, and acquires coagulation waveform data from the acquired optical signal. In step S142, the analysis unit 300 performs calculation processing analysis on the solidification waveform data acquired in step S141. For example, as described with reference to fig. 4, the analysis unit 300 acquires the time (T-T2) required for the absorbance of the coagulation waveform data to decrease by 50% as a result of representing the time until the blood sample is coagulated.

In step S143, the analysis unit 300 performs AI analysis on the solidification waveform data acquired in step S141. Thus, the analysis unit 300 obtains whether or not there is an abnormality related to measurement based on the feature amount extracted from the solidification waveform data by the AI algorithm 60. The analysis unit 300 determines the possibility of occurrence of a non-specific reaction based on the presence or absence of an abnormality related to measurement.

In step S144, the analysis unit 300 provides the result indicating the time until the blood sample is coagulated acquired in step S142 and the result indicating the presence or absence of abnormality related to measurement acquired in step S143.

In fig. 41, the AI analysis is always performed in step S143, but the analysis unit 300 may perform the processing of step S143 based on a rule set in advance indicating whether the AI analysis is necessary.

The sample analyzer 4000 of embodiment 5 is a blood coagulation measuring device that optically measures a turbidity change of a measurement sample caused by coagulation of a blood sample, but is not limited thereto. For example, the following coagulation measuring device may be used: the change in the amplitude movement of the steel ball in the measurement sample caused by the viscosity change of the measurement sample due to coagulation of the blood sample is measured by using the high frequency receiving frequency emitted from the high frequency generating coil. The sample analyzer 4000 of embodiment 5 is a blood coagulation measuring device, but may be an immunoassay device, a biochemical measuring device, or a gene measuring device.

Embodiment 6

In embodiment 6, a configuration example of an analyte analyzer 4000 including a host processor and a parallel processing processor is shown. In embodiment 6, parallel processing is performed on waveform data in the parallel processing processor 3002, and information related to each type of analyte is generated based on the result of the parallel processing.

According to embodiment 6, even in the case of analyzing a huge volume of data from several hundred megabytes to several gigabytes per 1 subject, processing related to waveform data can be executed in parallel by a parallel processing processor provided separately from the host processor. Therefore, for example, even when a large volume of data is processed by the AI algorithm 60, the processing of the data is completed in the object analysis device 4000. Thus, it is unnecessary to transmit data to the analysis server storing the AI algorithm 60 via, for example, the internet or an intranet. Therefore, according to embodiment 6, it is possible to improve the accuracy of classifying analytes in a subject and to maintain the processing capacity of the subject analyzing device 4000 high without transmitting large-capacity data from the subject analyzing device 4000 to the analyzing server and acquiring the analysis result returned from the analyzing server.

The structure of the analyte analyzer 4000 according to embodiment 6 will be described with reference to fig. 42 and 43. In the configuration example shown in fig. 42 and 43, the measurement unit 400 includes an FCM detection unit 410 for measuring an object (for example, a blood object, a urine object, a body fluid, and a bone marrow fluid).

Fig. 42 is a block diagram showing the structure of a sample analyzer 4000 according to embodiment 6.

The analyte analyzer 4000 according to embodiment 6 includes a measurement unit 400 and an analysis unit 300 provided inside the measurement unit 400. In comparison with measurement unit 400 of embodiment 4 shown in fig. 24, measurement unit 400 of embodiment 6 omits IF units 462, 464, 465 and bus 463. The analysis unit 300 according to embodiment 6 is connected to the a/D conversion unit 461, the device mechanism unit 430, the sample preparation unit 440, the sample suction unit 450, and the computer 301 disposed outside the measurement unit 400 in the measurement unit 400.

Fig. 43 is a block diagram showing the structure of analysis unit 300 according to embodiment 6.

In comparison with the analysis unit 300 of embodiment 4 shown in fig. 26, the analysis unit 300 of embodiment 6 includes a parallel processing processor 3002, a bus controller 3005, and IF units 462 and 464.

The parallel processing processor 3002 is configured to be able to process arithmetic processing based on the AI algorithm 60 instead of a main processor (master processor). The TAT required for AI analysis can be improved by using a parallel processing processor 3002 adapted to process the matrix operations performed by the AI algorithm 60. However, although the TAT is improved with the parallel processing processor 3002, the computer load required for AI analysis increases if the data amount of the analysis object becomes large. In contrast, as described above, by sharing data analysis based on the calculation processing analysis and the AI analysis, it is possible to reduce the computer load and achieve an improvement in the inspection efficiency.

The processor 3001 uses the parallel processing processor 3002 to perform analysis processing of waveform data by the AI algorithm 60. That is, the processor 3001 performs AI analysis based on the waveform data of the AI algorithm 60 by executing the analysis software 3100. The analysis software 3100 is used to analyze waveform data corresponding to an analyte in the subject based on the AI algorithm 60.

The analysis software 3100 may be stored in the storage unit 3004. In this case, the processor 3001 executes the analysis software 3100 stored in the storage 3004 to perform AI analysis based on the waveform data of the AI algorithm 60.

In the present embodiment, for example, AI analysis is performed by the processor 3001 and the parallel processing processor 3002, and calculation processing analysis is performed by the processor 3001 without using the parallel processing processor 3002.

The processor 3001 is, for example, a CPU (Central Processing Unit ). The processor 3001 may be, for example, core i9 (Core i 9), core i7 (Core i 7), core i5 (Core i 5), ryzen 9, ryzen 7, ryzen 5, ryzen 3, etc. manufactured by intel corporation.

The processor 3001 controls the parallel processing processor 3002. The parallel processing processor 3002 performs parallel processing related to, for example, matrix operations according to the control of the processor 3001. That is, the processor 3001 is a master processor of the parallel processing processor 3002, and the parallel processing processor 3002 is a slave processor (slave processor) of the processor 3001. The processor 3001 is also referred to as a host processor or main processor (main processor). The processor 3001 performs matrix operations based on the AI algorithm 60 by parallel processing by the parallel processing processor 3002.

The parallel processing processor 3002 performs a plurality of arithmetic processing in parallel as at least a part of processing related to analysis of waveform data. The parallel processing processor 3002 is, for example, a GPU (Graphics Processing Unit ), an FPGA (Field Programmable Gate Array, field programmable gate array), an ASIC (Application Specific Integrated Circuit ). In the case where the parallel processor 3002 is an FPGA, the parallel processor 3002 may be programmed with, for example, an arithmetic process related to the trained AI algorithm 60. In the case where the parallel processor 3002 is an ASIC, for example, a circuit for executing an arithmetic process related to the trained AI algorithm 60 may be embedded in the parallel processor 3002, or a programmable module may be incorporated in addition to such an embedded circuit.

As the parallel processing processor 3002, for example, geForce, quadro, TITAN, jetson manufactured by NVIDIA corporation, or the like can be used. If the Jetson series is, jetson Nano, jetson Tx2, jetson Xavier, jetson AGX Xavier, for example, may be used.

The processor 3001 performs, for example, calculation processing related to control of the measurement unit 400. The processor 3001 performs, for example, calculation processing related to control signals transmitted and received among the device mechanism unit 430, the sample preparation unit 440, and the specimen suction unit 450. Further, the processor 3001 performs, for example, calculation processing related to transmission and reception of information between the computers 301.

The computer 301 has, for example, a function of displaying the analysis result transmitted from the analysis unit 300 based on the processing of the processor 3001. The computer 301 transmits a measurement instruction to the analysis unit 300, for example. For example, a measurement instruction is sent from the host computer to the computer 301. The user can also input a measurement instruction via an input device of the computer 301.

The processor 3001 executes processing related to, for example, reading out program data from the storage unit 3004, expanding the program to the RAM 3017, and transmitting/receiving data between the RAM 3017. The above-described respective processes executed by the processor 3001 are required to be executed in a predetermined order, for example. For example, when the processes required for controlling the device mechanism 430, the sample preparation unit 440, and the specimen suction unit 450 are A, B and C, respectively, it may be required to execute the processes in the order of B, A, C. Since the processor 3001 executes sequential processing depending on the order as described above in many cases, the processing speed is not necessarily increased even if the number of arithmetic units (sometimes referred to as "processor cores", "cores" or the like) is increased.

On the other hand, the parallel processing processor 3002 performs a regular and large number of calculation processes such as operations on matrix data including a large number of elements. In the present embodiment, the parallel processing processor 3002 executes parallel processing in which at least a part of the processing of analyzing waveform data in accordance with the AI algorithm 60 is parallelized. The AI algorithm 60 includes, for example, a large number of matrix operations. The AI algorithm 60 may include at least 100 matrix operations, and may include at least 1000 matrix operations, for example.

The parallel processing processor 3002 has a plurality of arithmetic units each capable of simultaneously performing matrix operations. That is, the parallel processing processor 3002 can execute matrix operations by each of the plurality of operation units in parallel as parallel processing. For example, the matrix operation included in the AI algorithm 60 can be divided into a plurality of operation processes that are independent of each other in order. The arithmetic processing divided in this way can be executed in parallel by a plurality of arithmetic units. These arithmetic units are sometimes referred to as "processor cores", "cores", and the like.

By executing such parallel processing, the arithmetic processing of the entire object analysis device 4000 can be made faster. Processing such as matrix operations included in the AI algorithm 60 is sometimes referred to as "single instruction multiple data processing" (SIMD: single Instruction Multiple Data), for example. The parallel processing processor 3002 is adapted to such SIMD operations, for example. Such parallel processing processor 3002 is sometimes referred to as a vector processor.

As described above, the processor 3001 is adapted to perform various and complex processes. On the other hand, the parallel processing processor 3002 is adapted to execute a large number of processes of finalization in parallel. By performing a large number of processes of finalization in parallel, TAT required for calculation processing is shortened.

Further, the object of parallel processing performed by the parallel processing processor 3002 is not limited to matrix operation. For example, when the parallel processing processor 3002 performs the learning process on the AI algorithm 50, a differential operation or the like related to the learning process may be an object of the parallel processing.

The number of arithmetic units of the processor 3001 is, for example, two cores (number of cores: 2), four cores (number of cores: 4), eight cores (number of cores: 8). On the other hand, the parallel processing processor 3002 may have, for example, at least 10 arithmetic units (kernel number: 10), and execute 10 matrix operations in parallel. The parallel processing processor 3002 sometimes has several tens of arithmetic units. In addition, the parallel processing processor 3002 may have, for example, at least 100 arithmetic units (kernel number: 100) and may execute 100 matrix operations in parallel. The parallel processing processor 3002 also sometimes has hundreds of arithmetic units. In addition, the parallel processing processor 3002 may have, for example, at least 1000 arithmetic units (kernel number: 1000) and may execute 1000 matrix operations in parallel. The parallel processing processor 3002 sometimes also has thousands of arithmetic units.

Fig. 44 is a block diagram showing another configuration of a sample analyzer 4000 according to embodiment 6. The sample analyzer 4000 shown in fig. 44 counts and classifies blood cells in a blood sample.

As compared with the sample analyzer 4000 of fig. 42, the sample analyzer 4000 of fig. 44 includes the RBC/PLT detecting unit 4101, HGB detecting unit 4102, analog processing units 4201 and 4202, and a/D converting units 4611 and 4612, which are similar to those of fig. 27. The sample preparation unit 440 of fig. 44 is configured in the same manner as the sample preparation unit 440 shown in fig. 28 or 29.

Fig. 45 is a diagram showing a configuration example of the parallel processing processor 3002.

The parallel processing processor 3002 includes a plurality of arithmetic units 3200 and a RAM 3201. The operation units 3200 each perform an operation process of matrix data in parallel. The RAM 3201 stores data related to arithmetic processing performed by the arithmetic unit 3200. RAM 3201 is memory having a capacity of at least 1 gigabyte. RAM 3201 may be memory having a capacity of 2 gigabytes, 4 gigabytes, 6 gigabytes, 8 gigabytes, or more than 10 gigabytes. The arithmetic unit 3200 acquires data from the RAM 3201, and performs arithmetic processing. The arithmetic unit 3200 is sometimes referred to as a "processor core," "core," or the like.

Fig. 46 to 48 schematically show mounting examples of the parallel processing processor 3002.

In the example shown in fig. 46, the processor 3001 is mounted on a substrate 3301. The parallel processor 3002 is mounted on a graphics board 3300, and the graphics board 3300 is connected to the board 3301 via a connector 3310. The processor 3001 is connected to the parallel processing processor 3002 via a bus 3003. In the example shown in fig. 47, the parallel processing processor 3002 is directly mounted on the board 3301, and is connected to the processor 3001 via the bus 3003. In the example shown in fig. 48, the processor 3001 and the parallel processing processor 3002 are provided as a single body. In this case, the parallel processing processor 3002 is built in the processor 3001 mounted on the substrate 3301.

Fig. 49 is a diagram showing another mounting example of the parallel processing processor 3002.

In the example shown in fig. 49, the parallel processor 3002 is mounted on the measurement unit 400 by an external device 3400 connected to the measurement unit 400. The parallel processing processor 3002 is mounted on an external device 3400 as a USB apparatus, for example. The external device 3400 is connected to the bus 3003 via the IF unit 467, and the parallel processing processor 3002 is mounted on the object analysis device 4000. The USB device may be a small device such as a USB dongle (dongle), for example. The IF section 467 is, for example, a USB interface having a transfer speed of several hundred Mbps, and more preferably a USB interface having a transfer speed of several Gbps to several tens of Gbps or more. As the external device 3400 to which the parallel processing processor 3002 is attached, for example, neural Compute Stick (nerve computation stick 2) manufactured by Intel corporation can be used.

The plurality of parallel processing processors 3002 can be mounted on the object analysis device 4000 by connecting the plurality of USB devices mounted with the parallel processing processors 3002 to the IF unit 467. Regarding the parallel processing processor 3002 mounted on 1 USB device, the number of the operation units 3200 may be small compared with that of GPUs or the like, and therefore, the number of cores can be increased by adding a plurality of USB devices connected to the measurement unit 400.

Next, an outline of the arithmetic processing performed by the parallel processing processor 3002 based on the control of the analysis software 3100 running on the processor 3001 will be described with reference to fig. 50 to 52.

Fig. 50 is a diagram showing a configuration example of the parallel processing processor 3002 that executes arithmetic processing.

The parallel processing processor 3002 has a plurality of arithmetic units 3200 and a RAM 3201. The processor 3001 executing the analysis software 3100 issues a command to the parallel processing processor 3002 to cause the parallel processing processor 3002 to execute at least a part of the arithmetic processing required when analyzing waveform data with the AI algorithm 60. The processor 3001 issues instructions to the parallel processing processor 3002 to perform arithmetic processing related to analysis of waveform data based on the AI algorithm 60.

All or at least a part of the waveform data is stored in the RAM 3017. The data stored in the RAM3017 is transferred to the RAM 3201 of the parallel processing processor 3002 by, for example, DMA (Direct Memory Access ). The plurality of arithmetic units 3200 of the parallel processing processor 3002 each execute arithmetic processing for data stored in the RAM 3201 in parallel. The plurality of arithmetic units 3200 each acquire necessary data from the RAM 3201, and execute arithmetic processing. The data corresponding to the operation result is stored in the RAM 3201 of the parallel processing processor 3002. Data corresponding to the operation result is transferred from the RAM 3201 to the RAM3017 by, for example, DMA.

Fig. 51 is a diagram showing an outline of a matrix operation performed by the parallel processing processor 3002.

When analyzing waveform data in accordance with the AI algorithm 60, calculation of the product of the matrix (matrix operation) is performed. The parallel processing processor 3002 performs a plurality of arithmetic processing related to matrix arithmetic, for example, in parallel.

The upper diagram of fig. 51 shows the calculation formula of the product of the matrix. In the calculation formula, a matrix c is obtained by the product of a matrix a of n rows and n columns and a matrix b of n rows and n columns. As illustrated in the upper diagram of fig. 51, the calculation formula is described by a multi-level loop syntax. The lower diagram of fig. 51 shows an example of arithmetic processing performed in parallel by the parallel processing processor 3002. The calculation formula shown in the lower diagram of fig. 51 can be divided into, for example, n×n arithmetic processes, which are the number of combinations of the circulation variable i of the 1 st layer and the circulation variable j of the 2 nd layer. The thus divided arithmetic processing is independent of each other, and thus can be executed in parallel.

Fig. 52 is a conceptual diagram showing that a plurality of arithmetic processes illustrated in the lower diagram of fig. 51 are executed in parallel with the parallel processing processor 3002.

As shown in fig. 52, each of the plurality of arithmetic processing is allocated to any arithmetic unit among the plurality of arithmetic units 3200 provided in the parallel processing processor 3002. The arithmetic units 3200 each execute the assigned arithmetic processing in parallel with each other. That is, the operation units 3200 each simultaneously execute the divided operation processing.

The calculation performed by the parallel processing processor 3002 illustrated in fig. 51 and 52 obtains, for example, information on the probability that the cells corresponding to the waveform data belong to each of the plurality of cell types. The processor 3001 executing the analysis software 3100 performs analysis on the cell type of the cell corresponding to the waveform data based on the operation result.

The operation of the probability that the analyte in the subject belongs to each of the plurality of classification categories may be performed by a processor different from the parallel processing processor 3002. For example, the calculation result of the parallel processing processor 3002 may be transferred from the RAM 3201 to the RAM3017, and the processor 3001 may calculate information on the probability that the analyte corresponding to each waveform data belongs to each of the plurality of classification categories based on the calculation result read out from the RAM 3017. The calculation result of the parallel processing processor 3002 may be transferred from the RAM 3201 to the analysis unit 300, and the processor mounted in the analysis unit 300 may calculate information on probabilities that analytes corresponding to the respective waveform data belong to respective classification categories among the plurality of classification categories.

The processing shown in fig. 51 and 52 is applied to, for example, an arithmetic processing (also referred to as a filter processing) related to the convolution layer in the AI algorithm 60.

Fig. 53 is a diagram schematically showing an outline of the arithmetic processing related to the convolutional layer.

In the upper diagram of fig. 53, waveform data obtained based on forward scattered light is shown as waveform data input to the AI algorithm 60. The waveform data of the present embodiment is one-dimensional matrix data as shown in fig. 32. More simply, waveform data is array data in which elements are arranged in a row. Here, for convenience of explanation, the number of elements of waveform data is assumed to be n (n is an integer of 1 or more). A plurality of filters are shown in the upper diagram of fig. 53. The filter is generated using a learning process of the AI algorithm 50. The plurality of filters are each one-dimensional matrix data representing characteristics of the waveform data. The filter shown in the upper diagram of fig. 53 is matrix data of 1 row and 3 columns, but the number of columns is not limited to 3. The characteristics corresponding to the cell type associated with the waveform data are calculated by performing matrix operations on the waveform data and the respective filters input to the AI algorithm 60.

The lower diagram of fig. 53 shows an outline of the matrix operation of the waveform data and the filter. The matrix operation is performed while shifting each filter one by one with respect to each element of the waveform data. The calculation of the matrix operation is performed by the following (expression 1).

[ mathematics 1]

In (formula 1), the subscript of x is a variable indicating the row number and column number of waveform data. The subscript of h is a variable representing the row number and column number of the filter. In the case of the example shown in fig. 53, the waveform data is one-dimensional matrix data, and the filters are matrix data of 1 row and 3 columns, so l=1, m=3, p=0, q=0, 1, 2, i=0, j=0, 1, … …, and n-1.

The parallel processing processor 3002 performs the matrix operation represented by (formula 1) in parallel by each of the plurality of operation units 3200. Based on the arithmetic processing performed by the parallel processing processor 3002, classification information relating to the class of each analyte in the subject is generated. The generated classification information is used to generate and display a test result of the subject based on the classification information.

As shown in fig. 42 and 43, the computer 301 is connected to the processor 3001 via the IF unit 3006 and the bus 3003, and is capable of receiving the analysis results of the processor 3001 and the parallel processing processor 3002. The IF unit 3006 is, for example, a USB interface. The computer 301 receives the analysis result of the analysis unit 300 via the IF unit 3006, and displays the analysis result on a display device of the computer 301.

The computer 301 may be provided with an operation section constituted by a pointing device including a keyboard, a mouse, or a touch panel. By operating the operation unit, a user such as a doctor or a test technician can input a measurement instruction to the object analysis device 4000, and can input a measurement instruction in accordance with the measurement instruction. The user can input an instruction to display the test result to the computer 301 via the operation unit. The user can operate the operation unit to browse various information related to the test result, such as numerical results based on analysis, graphs, charts, and flag information given to the test object.

< operation of analyte analysis device >

The analysis operation of the sample analyzer 4000 on the sample will be described with reference to fig. 54 to 56.

Fig. 54 is a flowchart showing the analysis operation of the analysis unit 300 and the measurement unit 400.

In step S200, the processor 3001 of the analysis unit 300 instructs the measurement unit 400 to perform measurement when receiving a measurement instruction. For example, the analysis unit 300 controls the operations of the respective detection units (FCM detection unit 410, RBC/PLT detection unit 4101, HGB detection unit 4102), the specimen suction unit 450, and the sample preparation unit 440 of the measurement unit 400 by the instruction to the measurement unit 400. The measurement unit 400 starts measuring the subject in response to the instruction from the analysis unit 300.

In step S300, the specimen suction unit 450 sucks the specimen from the blood collection tube according to the measurement instruction from the analysis unit 300, and discharges the sucked specimen to the reaction chamber. The measurement instruction from the analysis unit 300 includes information of a measurement channel required to be measured according to the measurement instruction. The specimen suction unit 450 discharges the specimen to the reaction chamber of the corresponding measurement channel based on the measurement channel information included in the measurement instruction.

In step S301, the sample preparation unit 440 prepares a measurement sample according to a measurement instruction from the analysis unit 300. Specifically, the sample preparation unit 440 supplies a reagent (hemolytic agent and staining solution) to the reaction chamber from which the sample is discharged based on the information of the measurement channel included in the measurement instruction, and mixes the sample and the reagent. Thus, measurement samples (for example, WDF measurement sample, RET measurement sample, WPC measurement sample, PLT-F measurement sample, WNR measurement sample) were prepared.

The sample preparation unit 440 supplies a reagent to the reaction chamber from which the sample is discharged, and mixes the sample and the reagent to prepare an RBC/PLT measurement sample. The sample preparation unit 440 supplies a reagent to the reaction chamber from which the sample is discharged, and mixes the sample and the reagent to prepare a hemoglobin measurement sample.

In step S302, the FCM detection unit 410 measures the prepared measurement sample in response to the measurement instruction from the analysis unit 300. Specifically, the device mechanism 430 conveys the measurement sample in the reaction chamber of the sample preparation unit 440 to the FCM detection unit 410 in response to a measurement instruction from the analysis unit 300. The measurement sample supplied from the reaction chamber flows into the flow cell 4113, and is irradiated with a laser beam by the light source 4111 (see fig. 25). When the analyte contained in the measurement sample passes through the flow cell 4113, the analyte is irradiated with light, and forward scattered light, side scattered light, and fluorescence generated from the analyte are detected by the light receiving elements 4116, 4121, and 4122, respectively, and an analog optical signal corresponding to the light receiving intensity is output. The optical signal is processed by the analog processing unit 420 and then output to the a/D conversion unit 461.

The RBC/PLT detecting unit 4101 performs measurement of blood cells by a sheath flow DC detection method based on the RBC/PLT measurement sample. The HGB detecting unit 4102 measures hemoglobin by the SLS-hemoglobin method based on the hemoglobin measurement sample. The analog signal detected by the RBC/PLT detecting unit 4101 is processed by the analog processing unit 4201, and then outputted to the a/D converting unit 4611, and the analog signal detected by the HGB detecting unit 4102 is processed by the analog processing unit 4202 and then outputted to the a/D converting unit 4612 (see fig. 27).

In step S303, as described above, the a/D conversion section 461 generates digital data by sampling the analog optical signal at a predetermined rate, and generates waveform data corresponding to each analyte based on the digital data. The waveform data generated by the a/D conversion section 461 is directly transferred to the RAM by, for example, DMA transfer without passing through the processor 3001 of the analysis unit 300. Thus, waveform data based on the forward scattered light signal, waveform data corresponding to the side scattered light, and waveform data corresponding to fluorescence acquired from the analyte are taken into the RAM 3017.

The a/D converter 4611 generates digital data by sampling an analog signal from the RBC/PLT detector 4101 at a predetermined rate. The a/D conversion section 4612 generates digital data by sampling an analog signal from the HGB detection section 4102 at a predetermined rate. These digital data can also be fetched into the RAM 3017.

In step S201, the processor 3001 of the analysis unit 300 performs AI analysis on the waveform data using the AI algorithm 60, and performs calculation processing analysis on a representative value corresponding to the characteristic of the analyte in the waveform data. The sharing of AI analysis and calculation process analysis is as described above. Thereby, the analytes in the test object are classified. The process of AI analysis in step S201 will be described later, and the processor 3001 acquires, for example, classification information 82 of each analyte in the subject, a tag value 83, and an analysis result 84 (refer to fig. 35) as a result of the process using the parallel processing processor 3002.

In step S202, the processor 3001 analyzes the tag value 83 and the analysis result 84 using the program stored in the storage 3004, and generates a test result of the test object. In step S202, for example, the number of analytes is counted for each class of analytes based on the label value 83 and the analysis result 84 of each analyte.

For example, in the case of an example of testing blood cells in a blood test object, if N pieces of classification information to which a tag value "1" indicating neutrophils is given are derived from 1 part of the test object, a count result of the number of neutrophils=n is obtained as a test result of the test object. The processor 3001 acquires a count result related to the measurement item corresponding to the measurement channel based on the analysis result 84, and stores the count result in the storage 3004 together with the identification information of the subject.

Here, the measurement item corresponding to the measurement channel is an item for which a count result is required in accordance with a measurement instruction. For example, the term "measurement item corresponding to the WDF channel" includes the measurement item of the leukocyte 5 class, that is, the number of monocytes, neutrophils, lymphocytes, eosinophils and basophils. The term "measurement item corresponding to RET channel" includes measurement items of the number of reticulocytes. The measurement items corresponding to PLT-F include measurement items of the number of platelets. The measurement item corresponding to WPC includes measurement items of the number of hematopoietic progenitor cells. The measurement items corresponding to WNR include measurement items of the numbers of white blood cells and nucleated red blood cells.

The counting result is not limited to the items requiring measurement (also referred to as reportable items) listed above, and may include counting results of other cells that can be measured on the same measurement channel. For example, in the case of WDF channel, as shown in fig. 34, immature Granulocytes (IG) and abnormal cells are included in the count result in addition to the classification of leukocytes 5.

Further, the processor 3001 analyzes the obtained count result to generate a test result of the test object, and stores the test result in the storage 3004. Analysis of the count result includes, for example, determination of whether the count result is within a normal value range, whether abnormal cells are not detected, whether a deviation from the previous test result is within an allowable range, and the like.

In step S203, the computer 301 displays the test result generated by the analysis unit 300 on the display unit.

Fig. 55 is a flowchart showing details of the AI analysis in step S201 of fig. 54.

Step S201 is executed by the processor 3001 according to the operation of the parsing software 3100.

In step S2010, the processor 3001 forwards the waveform data taken into the RAM 3017 in step S303 to the parallel processing processor 3002. As shown in fig. 50, waveform data is DMA-transferred from the RAM 3017 to the RAM 3201 by DMA transfer. At this time, the processor 3001 controls the bus controller 3005, for example, so that the waveform data is DMA-forwarded from the RAM 3017 to the RAM 3201.

In step S2011, the processor 3001 instructs the parallel processing processor 3002 to perform parallel processing for waveform data. The processor 3001 instructs to perform parallel processing, for example, by calling a kernel function (kernel function) of the parallel processing processor 3002. The processing performed by the parallel processing processor 3002 is described later with reference to fig. 56. The processor 3001, for example, instructs the parallel processing processor 3002 to perform matrix operations related to the AI algorithm 60. Waveform data corresponding to each analyte in the subject is input to the AI algorithm 60, respectively. The waveform data input to the AI algorithm 60 is operated on by the parallel processing processor 3002.

In step S2012, the processor 3001 receives the operation result performed by the parallel processing processor 3002. As shown in fig. 50, the operation result is DMA-forwarded from the RAM 3201 to the RAM 3017. In step S2013, the processor 3001 generates an analysis result of each analyte class based on the operation result of the parallel processing processor 3002.

Fig. 56 is a flowchart showing details of step S2011 of fig. 55.

Step S2011 is executed by the parallel processing processor 3002 based on the instruction of the processor 3001.

In step S2100, the processor 3001 executing the analysis software 3100 causes the parallel processing processor 3002 to execute the allocation arithmetic processing to the arithmetic unit 3200. The processor 3001 causes the parallel processing processor 3002 to execute allocation of arithmetic processing to the arithmetic unit 3200, for example, by calling a kernel function of the parallel processing processor 3002. As shown in fig. 52, for example, a matrix operation related to the AI algorithm 60 is divided into a plurality of operation processes, and each of the divided operation processes is distributed to the operation unit 3200. Waveform data corresponding to each analyte in the subject is input to the AI algorithm 60, respectively. Matrix operations corresponding to waveform data are divided into a plurality of operation processes and distributed to the operation unit 3200.

In step S2101, the arithmetic processing is processed in parallel by the plurality of arithmetic units 3200. An arithmetic process is performed on the plurality of waveform data. In step S2102, the operation result generated by parallel processing by the plurality of operation units 3200 is forwarded from the RAM 3201 to the RAM3017. As shown in fig. 50, the operation result is DMA-forwarded from the RAM 3201 to the RAM3017.

In step S201 of fig. 54, the processor 3001 of the analysis unit 300 can acquire an analysis result of a measurement item (for example, the number of red blood cells, the hematocrit value, etc.) corresponding to the RBC/PLT channel by using the AI algorithm 60 on digital data based on the analog signal from the RBC/PLT detecting section 4101. Further, the processor 3001 may acquire an analysis result of a measurement item (for example, hemoglobin amount or the like) for the HGB channel by using the AI algorithm 60 on digital data based on an analog signal from the HGB detecting section 4102.

Next, another configuration example of the object analysis device 4000 including the measurement unit 400 and the analysis unit 300 will be described with reference to fig. 57 and 58.

Fig. 57 is a block diagram showing another configuration of the measurement unit 400.

In the example shown in fig. 57, the analog optical signal processed by the analog processing unit 420 is transmitted to the analysis unit 300 via the connection port 421. The connection cable 4210 is connected to the connection port 421. The other structure shown in fig. 57 has the same structure and function as those of the measurement unit 400 of the above embodiment.

Fig. 58 is a block diagram showing another structure of the analysis unit 300.

In the example shown in fig. 58, the analysis unit 300 is connected to the measurement unit 400 via the IF unit 3006. The RAM3017 and the bus 3003 are transmission paths having a data transfer speed of several hundred MB/s or more, for example. Bus 3003 may be a transmission path having a data transfer speed of 1GB/s or more. Bus 3003 carries out data transfer based on, for example, the PCI-Express or PCI-X standard. The configuration of the processor 3001, the parallel processing processor 3002, the storage portion 3004, and the RAM3017, and the processing performed by them are the same as those described above.

The analysis unit 300 includes a connection port 3007, an a/D conversion unit 3008, and an IF unit 3009.

The connection port 3007 is connected to the connection port 421 of the measurement unit 400 via a connection cable 4210 (see fig. 57). The connection cable 4210 includes, for example, a number of transmission paths corresponding to the types of analog signals transmitted from the measurement unit 400 to the analysis unit 300. For example, the connection cable 4210 is formed of twisted pair wires, and has the number of pairs corresponding to the type of analog signal transmitted to the analysis unit 300. In order to reduce noise in signal transmission, the connection cable 4210 is preferably of a length of, for example, 1 meter or less.

The a/D conversion unit 3008 is connected to the connection port 3007. As described above, the a/D converter 3008 samples the analog optical signal output from the measurement unit 400, and generates waveform data corresponding to each analyte in the subject. The generated waveform data is stored in the storage unit 3004 or the RAM 3017 via the IF unit 3009 and the bus 3003. The transmission path from the connection port 3007 to the a/D conversion unit 3008 may have a number of wirings corresponding to the types of optical signals transmitted to the analysis unit 300.

The processor 3001 and the parallel processing processor 3002 execute arithmetic processing on waveform data stored in the storage unit 3004 or the RAM 3017. The analysis software 3100 running on the processor 3001 is the same as the analysis software 3100 shown in fig. 50. The processor 3001 generates classification information concerning the type of the analyte in the subject by executing the analysis software 3100 by the same operation as described above.

Next, another configuration example of the object analysis device 4000 including the measurement unit 400 and the analysis unit 300 will be described with reference to fig. 59 and 60.

Fig. 59 is a block diagram showing another configuration of the measurement unit 400.

The measurement unit 400 shown in fig. 59 includes an IF section 4631 for transmitting the waveform data generated by the a/D conversion section 461 to the analysis unit 300. The transmission path 4632 is connected to the IF section 4631. This other structure and function are similar to those of the measurement unit 400 described above.

The IF unit 4631 is an interface, for example, as a dedicated line having a communication band of 1 gigabit/sec or more. For example, IF section 4631 is an interface conforming to gigabit ethernet, USB 3.0, or Thunderbolt 3 (Thunderbolt 3). When the IF unit 4631 is a gigabit ethernet, the transmission path 4632 is a LAN cable. When the IF section 4631 is USB 3.0, the transmission path 4632 is a USB cable conforming to USB 3.0. The transmission path 4632 is, for example, a dedicated transmission path for transmitting digital data between the measurement unit 400 and the analysis unit 300.

Fig. 60 is a block diagram showing another structure of the analysis unit 300.

The analysis unit 300 shown in fig. 60 includes an IF unit 3010. The other structure and function are the same as those of the analysis unit 300 described above. The analysis unit 300 may be connected to the plurality of measurement units 400 via the plurality of IF units 3010 and the plurality of IF units 3006.

The analysis software 3100 running on the processor 3001 has the same functions as the analysis software 3100 described above. The analysis software 3100 analyzes the type of the analyte in the subject by the same operations as those described in the above-mentioned association.

In the configuration of fig. 59 and 60, the a/D converter 461 in the measuring unit 400 generates digital waveform data based on the analog optical signal generated in the FCM detector 410. The waveform data is transmitted to the analysis unit 300 via the IF unit 462, the bus 463, the IF unit 4631, and the transmission path 4632.

The measurement unit 400 and the analysis unit 300 are connected in a one-to-one manner, for example, via a transmission path 4632. The transmission path 4632 in this case is a transmission path that does not involve transmission of data related to devices other than the components (for example, the measurement unit 400 and the analysis unit 300) constituting the object analysis device 4000. The transmission path 4632 is, for example, a transmission path different from an intranet or the internet. Thus, even if the waveform data generated in the measuring unit 400 is transmitted to the analyzing unit 300, a bottleneck in the communication speed of the digital data transmission can be avoided.

Next, another configuration example of the sample analyzer 4000 will be described with reference to fig. 61 to 65.

Fig. 61 is a block diagram showing another configuration of the object analysis apparatus 4000.

In this configuration example, an analysis unit 600 is provided between the measurement unit 400 and the computer 301. That is, in the configuration of fig. 61 to 65, the analyte analyzer 4000 includes a measurement unit 400, a computer 301, and an analysis unit 600. The analysis unit 600 analyzes the type of the measured cells. As will be described later, the parallel processing processor 6002 of the present configuration example is mounted on the sample analyzer 4000 so as to be embedded in the analysis unit 600.

Fig. 62 is a block diagram showing another configuration of the measurement unit 400.

In comparison with the configuration of fig. 59, the measurement unit 400 of fig. 62 is configured such that the computer 301 is connected to the IF unit 465, and the analysis unit 600 is provided between the IF unit 4631 and the computer 301. The analysis unit 600 is communicably connected to the IF section 4631 and the computer 301. Further, the analysis unit 600 may be connected to a plurality of measurement units 400. The analysis unit 600 may be connected to a plurality of computers 301.

Fig. 63 is a block diagram showing the structure of the analysis unit 600.

The analysis unit 600 includes a processor 6001, a parallel processing processor 6002, a bus 6003, a storage unit 6004, a RAM 6005, and IF units 6006 and 6007. Each part of the analysis unit 600 is connected to a bus 6003.

The bus 6003 is a transmission path having a data transfer speed of several hundred MB/s or more, for example. Bus 3003 may be a transmission path having a data transfer speed of 1GB/s or more. Bus 3003 carries out data transfer based on, for example, the PCI-Express or PCI-X standard. The analysis unit 600 may be connected to the plurality of measurement units 400 via the plurality of IF units 6006. In the case where a plurality of measurement units 400 are provided, the analysis units 600 may be connected to the measurement units 400, respectively. In this case, for example, the plurality of measurement units 400 and the plurality of analysis units 600 are connected one to one, respectively.

Fig. 64 is a diagram showing a configuration example of a parallel processing processor 6002 that executes arithmetic processing.

The processor 6001 and the parallel processing processor 6002 have the same configuration and functions as the processor 3001 and the parallel processing processor 3002, respectively. The parallel processing processor 6002 includes a plurality of operation units 6200 and a RAM 6201. Analysis software 6100 that analyzes the type of analyte in the subject runs on the processor 6001. The analysis software 6100 running on the processor 6001 has the same function as the analysis software 3100 shown in fig. 50. The analysis software 6100 analyzes the type of the analyte in the subject in the same manner as the operation described in fig. 50. The analysis software 6100 transmits the classification information of the analyte in the subject to the computer 301 via the IF unit 6007.

Fig. 65 is a block diagram showing the structure of the computer 301.

The computer 301 of fig. 65 has the same configuration as the analysis unit 600 of fig. 63 in which the parallel processing processor 6002 is omitted. The computer 301 includes a processor 3501, a bus 3503, a storage unit 3504, a RAM 3505, and an IF unit 3506.

On processor 3501, parsing software 3100 may not be running. The computer 301 receives the analysis result of the analysis unit 600 via the IF unit 3506. The IF unit 3506 is, for example, ethernet or USB. The IF unit 3506 may be an interface capable of wireless communication.

In the configurations of fig. 62 to 65, the analog optical signal of the cell generated in the FCM detection unit 410 is converted into digital waveform data in the a/D conversion unit 461 in the measurement unit 400. The waveform data is transmitted to the analysis unit 600 via the IF unit 462, the bus 463, the IF unit 4631, and the transmission path 4632.

The IF unit 4631 is a dedicated interface for connecting the measurement unit 400 and the analysis unit 600 as described above, and connects the measurement unit 400 and the analysis unit 600 in a one-to-one manner. In other words, the transmission path 4632 is a transmission path that does not involve transmission of data related to devices other than the components (for example, the measurement unit 400 and the analysis unit 300) constituting the object analysis device 4000, for example. The transmission path 4632 is a transmission path different from an intranet and the internet. Thus, even if the waveform data generated in the measuring unit 400 is transmitted to the analyzing unit 600, a bottleneck in the communication speed of the waveform data transmission can be avoided.

In this case, steps S200 to S202 of fig. 54 are executed by the analysis unit 600, and step S203 is executed by the computer 301.

Next, another configuration example of the sample analyzer 4000 of fig. 61 will be described with reference to fig. 66 and 67. The analyte analyzer 4000 of this example includes a measurement unit 400, a computer 301, and an analysis unit 600.

In comparison with the configuration of fig. 57, the measurement unit 400 of fig. 66 is configured such that the computer 301 is connected to the IF unit 465, and the analysis unit 600 is provided between the connection port 421 and the computer 301. The analysis unit 600 is communicably connected to the connection port 421 and the computer 301. The measurement unit 400 transmits an analog optical signal to the analysis unit 600 via the connection cable 4210.

In comparison with the configuration of fig. 63, the analysis unit 600 of fig. 67 includes a connection port 6008 and an a/D conversion unit 6009 instead of the IF unit 6006.

The analog optical signal transmitted from the analysis unit 600 via the connection cable 4210 is input to the a/D conversion portion 6009 via the connection port 6008. The a/D converter 6009 generates waveform data from the optical signal by the same processing as the a/D converter 461.

The analysis unit 600 may be connected to a plurality of measurement units 400 via a plurality of connection ports 6008. In the case where a plurality of measurement units 400 are provided, the analysis units 600 may be connected to the measurement units 400, respectively. In this case, for example, the plurality of measurement units 400 and the plurality of analysis units 600 are connected one to one, respectively.

In the configurations of fig. 66 and 67, in step S303 of fig. 54, analysis section 600 generates waveform data based on the analog optical signal transmitted from measurement section 400. Steps S200 to S202 of fig. 54 are executed by the analysis unit 600, and step S203 is executed by the computer 301.

Next, another configuration example of the measurement unit 400 and the analysis unit 300 included in the analyte analysis device 4000 will be described with reference to fig. 68 and 69.

In comparison with the configuration of fig. 27, the measurement unit 400 of fig. 68 includes connection ports 421, 4211, 4212 instead of the a/D conversion units 461, 4611, 4612 and the IF unit 462. The analog optical signals acquired by the respective detection units are transmitted to the analysis unit 300 via the connection cables 4210, respectively.

In comparison with the configuration of fig. 58, the analysis unit 300 of fig. 69 has 3 groups of connection ports 3007, a/D conversion units 3008, and IF units 3009. The 3 connection ports 3007 are connected to the connection ports 421, 4211, 4212 in fig. 68, respectively.

In the configurations of fig. 68 and 69, in step S303 of fig. 54, analysis section 300 generates waveform data based on the analog optical signal transmitted from measurement section 400.

Next, another configuration example of the measurement unit 400 and the analysis unit 300 included in the analyte analysis device 4000 will be described with reference to fig. 70 and 71.

In comparison with the configuration of fig. 27, the measurement unit 400 of fig. 70 includes an IF section 4631. The a/D conversion sections 461, 4611, 4612 generate waveform data based on the analog optical signals acquired by the corresponding detection sections, respectively. Waveform data corresponding to each detection section is transmitted to the analysis unit 300 via the transmission path 4632.

In comparison with the configuration of fig. 60, the analysis unit 300 of fig. 71 includes 3 IF sections 3010. The 3 IF sections 3010 are connected to the transmission paths 4632 in fig. 70, respectively.

Next, another configuration example of the measurement unit 400 and the analysis unit 300 included in the analyte analysis device 4000 will be described with reference to fig. 72 and 73.

In comparison with the configuration of fig. 68, the measurement unit 400 of fig. 72 is configured such that the computer 301 is connected to the IF section 465, and the analysis unit 600 is disposed between the connection ports 421, 4211, 4212 and the computer 301. The analysis unit 600 is communicably connected to the connection ports 421, 4211, 4212 and the computer 301. The analysis unit 600 and the computer 301 are connected to be able to transmit and receive digital data.

In comparison with the configuration of fig. 67, the analysis unit 300 of fig. 73 includes 3 groups of connection ports 6008 and a/D conversion units 6009. The 3 connection ports 6008 are connected to the connection ports 421, 4211, 4212 in fig. 72, respectively.

Next, the data sizes of the waveform data and the digital data will be described.

In the present embodiment, for example, 1 analyte in the subject is sampled for each of the analog optical signals based on forward scattered light (FSC), the analog optical signals based on side scattered light (SSC), and the analog optical signals based on Fluorescence (FL).

Examples of the sampling rate include sampling at 1024 points at 10 nanosecond intervals, sampling at 128 points at 80 nanosecond intervals, and sampling at 64 points at 160 nanosecond intervals. The amount of data is for example 2 bytes per 1 sample. For FSC, SSC, FL, data of an amount corresponding to the sampling rate (2 bytes×1024=2048 bytes in the case of a rate of 1024 points) is acquired, respectively. The data amount is the data amount for each 1 analyte in the test object.

In 1 assay, for example, FSC, SSC, FL is determined for at least 100 analytes. In addition, FSC, SSC, FL is also sometimes measured for at least 1000 analytes in 1 measurement, for example. In addition, FSC, SSC, FL may be measured for, for example, about 10000 analytes to about 140000 analytes in 1 measurement. Thus, when the number of analytes measured in 1 measurement is 100000 and the sampling rate is 1024, the data amount of each digital data of FSC, SSC, FL is 2 bytes×1024×100000= 204,800,000 bytes, and FSC, SSC, FL is 614,400,000 bytes in total.

Further, FSC, SSC, FL was measured for each measurement channel. Thus, when the number of analytes measured in 1 measurement is 100000, the sampling rate is 1024, and the number of measurement channels is 5, the data amount of each FSC, SSC, FL is 2 bytes×1024×100000×5= 1,024,000,000 bytes, and FSC, SSC, FL is 3,072,000,000 bytes in total.

As described above, the volume of digital data is, for example, from several hundred megabytes to several gigabytes per 1 sample, and is at least 1 gigabyte depending on the number of analytes, the sampling rate, and the number of measurement channels.

According to the present embodiment, when analyzing a huge volume of digital data from several hundred megabytes to several gigabytes per 1 part of the subject, as described above, the analysis processing using the AI algorithm 60 is completed inside the subject analysis device 4000, and the digital data is not transmitted to the analysis server provided outside the subject analysis device 4000 via the internet or intranet. Thus, it is possible to avoid a decrease in processing capacity due to an increase in communication load generated when the digital data is transmitted from the object analysis device 4000 to the analysis server.

Embodiment 7

< Structure of waveform data analysis System >

Fig. 74 is a diagram schematically showing the configuration of the waveform data analysis system according to the present embodiment.

The measurement unit 400a has the same structure as the measurement unit 400 described above. The measurement unit 400a transfers a measurement sample prepared based on the sample to the flow cell 4113. The light source 4111 (see fig. 25) irradiates the measurement sample supplied to the flow cell 4113 with light, and the light receiving elements 4116, 4121, 4122 (see fig. 25) detect forward scattered light, side scattered light, and fluorescence generated from the analyte in the measurement sample. The measurement unit 400a generates waveform data from the optical signals based on the forward scattered light, the side scattered light, and the fluorescence outputted from the light receiving elements 4116, 4121, 4122, and transmits the generated waveform data to the deep learning device 100.

The deep learning device 100 is a vendor-side device. The deep learning device 100 receives training waveform data acquired by the measurement unit 400 a. The method for generating training waveform data is as described above. The AI algorithm 50 stored in the deep learning device 100 is a deep learning algorithm. The deep learning device 100 learns the AI algorithm 50 composed of the neural network before training using the training data, and provides the AI algorithm 60 trained using the training data to the user. The AI algorithm 60 composed of the learned neural network is supplied from the deep learning device 100 to the object analysis device 4000 via the recording medium 98 or the communication network 99. The recording medium 98 is a computer-readable, non-transitory, tangible recording medium such as a DVD-ROM, a USB memory, or the like.

The deep learning device 100 is constituted by, for example, a general-purpose computer, and performs deep learning processing based on a flowchart to be described later.

The analyte analysis device 4000 performs AI analysis on waveform data corresponding to the analyte using the AI algorithm 60 composed of the learned neural network.

< hardware Structure of deep learning device >

Fig. 75 is a block diagram showing the structure of the deep learning apparatus 100.

The deep learning device 100 includes a processing unit 10, an input unit 16, and an output unit 17.

The input unit 16 and the output unit 17 are connected to the processing unit 10 via the IF unit 15. The input unit 16 is an input device such as a keyboard and a mouse. The output unit 17 is a display device such as a liquid crystal display.

The processing unit 10 includes a CPU 11, a memory 12, a storage unit 13, a bus 14, an IF unit 15, and a GPU 19.

The CPU 11 performs data processing to be described later. The memory 12 is used as a work area for data processing. The storage unit 13 records a program and processing data to be described later. Bus 14 transfers data between the various sections. The IF unit 15 performs input/output of data to/from an external device. The GPU19 functions as an accelerator for arithmetic processing (for example, parallel arithmetic processing) performed by the auxiliary CPU 11. That is, in the following description, the meaning of the processing performed by the CPU 11 also includes the processing performed by the CPU 11 using the GPU19 as an accelerator. GPU19 has the same functions as parallel processing processors 3002 and 6002 described above. Furthermore, a chip for which calculation of the neural network is preferable may be used instead of the GPU 19. Examples of such chips include FPGA, ASIC, myriad X (Intel), and the like.

In order to perform the processing of each step described later with reference to fig. 77, the processing unit 10 records the program of the present embodiment and the AI algorithm 50 configured by the neural network before training in the storage unit 13 in advance in, for example, an execution format. The execution format is a format generated by, for example, a compiler transformed from a programming language. The processing unit 10 performs training processing of the AI algorithm 50 before training using the program recorded in the storage unit 13.

In the following description, unless otherwise specified, the processing performed by the processing section 10 means processing performed by the CPU 11 based on the program stored in the storage section 13 or the memory 12 and the AI algorithm 50. The CPU 11 temporarily stores necessary data (intermediate data during processing, etc.) in the memory 12 as a work area, and appropriately records data to be stored for a long period of time, such as calculation results, in the storage unit 13.

< hardware Structure of analysis device >

The subject analyzer 4000 (see fig. 74) processes waveform data based on an algorithm provided from the deep learning device 100 in the same manner as the above-described configuration. The object analysis device 4000 may also function as the deep learning device 100, and learn the AI algorithm 50 using training data. In this case, the deep learning device 100 is not required.

In order to perform the processing of each step described in the waveform data analysis processing described below, the subject analysis device 4000 records the program of the present embodiment and the AI algorithm 60 configured by the trained neural network in advance in, for example, the storage 3004 (see, for example, fig. 26) and the storage 6004 (see, for example, fig. 63) in an execution format. The subject analysis device 4000 performs processing using the program recorded in the storage unit 3004 and the AI algorithm 60.

The AI algorithm 60 recorded in the storage sections 3004, 6004 may be updated via a communication network. The deep learning device 100 transmits the AI algorithm 60 to the subject analysis device 4000 via a communication network (e.g., internet, intranet). The subject analysis device 4000 updates the AI algorithm 60 recorded in the storage sections 3004, 6004 with the received AI algorithm 60.

< functional Block and Process flow >

(deep learning process)

Fig. 76 is a functional block diagram of the deep learning device 100.

The processing unit 10A of the deep learning device 100 includes a training data generating unit 101, a training data input unit 102, and an algorithm updating unit 103. A program for causing a computer to execute the deep learning process is installed in the memory 13 or the memory 12 of the processing unit 10 shown in fig. 75, and the cpu 11 and the GPU 19 execute the program, thereby realizing each functional block of the processing unit 10A.

The training data Database (DB) 104 and the algorithm Database (DB) 105 are recorded in the memory 13 or the memory 12 of the processing unit 10 shown in fig. 75. The training waveform data 72a, 72b, 72c are acquired in advance by the measurement unit 400a, for example, and stored in the training data database 104 in advance. The AI algorithm 50 is stored in an algorithm database 105.

Fig. 77 is a flowchart showing the processing performed by the deep learning device 100.

The processing of steps S401, S404, and S406 in fig. 77 is executed by the training data generating section 101. The process of step S402 is performed by the training data input section 102. The processing of steps S403 and S405 is executed by the algorithm updating unit 103.

First, the processing unit 10A acquires training waveform data 72a, 72b, 72c. The training waveform data 72a, 72b, and 72c are waveform data based on forward scattered light, side scattered light, and fluorescence, respectively. The acquisition of the training waveform data 72a, 72b, 72c may be performed by an operator, for example, from the measurement unit 400a, from the recording medium 98, or via the communication network 99. When acquiring the training waveform data 72a, 72b, 72c, information about which cell type the training waveform data 72a, 72b, 72c indicates is also acquired. The information of the cell type may be associated with the training waveform data 72a, 72b, 72c, or may be input by the operator via the input unit 16.

In step S401, as shown in fig. 33, the processing unit 10A generates training data 75 from the training waveform data 72a, 72b, 72c and the label value 77. In step S402, the processing unit 10A inputs the training data 75 to the AI algorithm 50 to acquire a test result. The test results are accumulated each time a plurality of training data 75 is input to the AI algorithm 50.

In the analysis method of the cell type according to the present embodiment, since the convolutional neural network is used and the random gradient descent method is used, the processing unit 10A determines whether or not training results of a predetermined number of tests are accumulated in step S403. When a predetermined number of training results are accumulated (yes in S403), the processing unit 10A advances the process to step S404. On the other hand, when a predetermined number of training results have not been accumulated (S403: NO), the processing section 10A skips the processing of step S404.

When a predetermined number of training results are accumulated (yes in S403), the processing unit 10A updates the connection weight w of the neural network constituting the AI algorithm 50 using the training results accumulated in step S402 in step S404. In the method for analyzing a cell type according to the present embodiment, since the random gradient descent method is used, the connection weight w of the neural network is updated at a stage where training results of a predetermined number of times are accumulated. Specifically, the process of updating the connection weight w is a process of performing calculation by the gradient descent method, which will be described later, as shown in (expression 12) and (expression 13).

In step S405, the processing unit 10A determines whether or not the AI algorithm 50 is trained with a predetermined number of training data 75. When the AI algorithm 50 is trained with the predetermined number of training data 75 (yes in S405), the deep learning process ends. On the other hand, when the AI algorithm 50 is not trained with the predetermined number of training data 75 (S405: no), in step S406, the processing unit 10A fetches the other training waveform data 72a, 72b, 72c, and returns the process to step S401.

According to the above-described processing, the processing unit 10A trains the AI algorithm 50 to obtain the AI algorithm 60.

(construction of neural network)

The upper part of fig. 78 is a schematic diagram illustrating the configuration of the neural network constituting the AI algorithm 50. As described above, in the present embodiment, a convolutional neural network is used. The neural network of the AI algorithm 50 includes an input layer 50a, an output layer 50b, and an intermediate layer 50c between the input layer 50a and the output layer 50b, the intermediate layer 50c including a plurality of layers. The number of layers constituting the intermediate layer 50c is, for example, 5 or more, preferably 50 or more, and more preferably 100 or more.

In the neural network of the AI algorithm 50, a plurality of nodes 89 configured as layers are connected between layers. Thus, information propagates from the input layer 50a to the output layer 50b in only one direction as indicated by arrow D in the figure.

(operation at nodes)

The middle of fig. 78 is a schematic diagram showing the operation at each node 89. At each node 89, a plurality of inputs are received, and 1 output (z) is calculated. In the case of the example shown in the middle of fig. 78, node 89 receives 4 inputs. The total input (u) received by node 89 is represented, for example, by the following (equation 2). Here, in the present embodiment, one-dimensional matrix data is used as the training data 75 and the analysis data 85, and therefore, when the variable of the operation expression corresponds to the two-dimensional matrix data, a process of converting the variable into the one-dimensional matrix data is performed.

[ math figure 2]

u＝w ₁ x ₁ +w ₂ x ₂ +w ₃ x ₃ +w ₄ x ₄ +b … … (2)

A different weight is applied to each input. In (formula 2), b is a value called offset. The output (z) of the node is an output of a predetermined function f with respect to the total input (u) expressed by the following (formula 2), and is expressed by the following (formula 3). The function f is called an activation function.

[ math 3]

z=f (u) … … (3)

The lower part of fig. 78 is a schematic diagram showing the operation between nodes. In the neural network, the nodes 89 that output the result (z) represented by (formula 3) for the total input (u) of each node 89 represented by (formula 2) are arranged in layers. The output of node 89 of the previous layer becomes the input of node 89 of the next layer. In the example shown in the lower part of fig. 78, the output of the node 89a of the layer on the left side in the drawing becomes the input of the node 89b of the layer on the right side in the drawing. Each node 89b receives an output from node 89 a. Different weights are applied to the connections between nodes 89a and 89 b. When the outputs of the plurality of nodes 89a are x1 to x4, the inputs to the 3 nodes 89b are expressed by the following (equations 4-1) to 4-3).

[ mathematics 4]

u ₁ ＝w ₁₁ x ₁ +w ₁₂ x ₂ +w ₁₃ x ₃ +w ₁₄ x ₄ +b ₁ … … (4-1)

u ₂ ＝w ₂₁ x ₁ +w ₂₂ x ₂ +w ₂₃ x ₃ +w ₂₄ x ₄ +b ₂ … … (4-2)

u ₃ ＝w ₃₁ x ₁ +w ₃₂ x ₂ +w ₃₃ x ₃ +w ₃₄ x ₄ +b ₃ … … (4-3)

These (formulae 4-1) to (formula 4-3) are summarized as follows (formula 4-4). Here, i=1, … …, I, j=1, … …, J. I is the total number of inputs, J is the total number of outputs.

[ math 5]

When (expression 4-4) is applied to the activation function, an output represented by (expression 5) below can be obtained.

[ math figure 6]

z _j ＝f(u _j ) (j=1, 2, 3) … … (formula 5)

(activation function)

In the method for analyzing a cell type according to the embodiment, a normalized linear function (rectified linear unit function, modified linear unit function) is used as an activation function. The normalized linear function is represented by the following (equation 6).

[ math 7]

f (u) =max (u, 0) … … (formula 6)

(equation 6) is a function in which the portion of u <0 in the linear function z=u is set to u=0. In the example shown in the lower part of fig. 78, the output of the node where j=1 is represented by the following (equation 6).

[ math figure 8]

z ₁ ＝max((w ₁₁ x ₁ +w ₁₂ x ₂ +w ₁₃ x ₃ +w ₁₄ x ₄ +b ₁ )，0)

(learning of neural networks)

When the function expressed by the neural network is y (x: w), the function y (x: w) is changed when the parameter w of the neural network is changed. The function y (x: w) is adjusted in such a way that the neural network selects the more suitable parameter w for the input x, which is called training or learning of the neural network. Set to give groups of inputs and outputs of the functions expressed by the neural network. When the desired output for a certain input x is set to d, the group of inputs and outputs is given as { (x 1, d 1), (x 2, d 2), … …, (xn, dn) }. The set of groups denoted by (x, d) is referred to as training data. Specifically, as shown in fig. 33, the set of waveform data 72a, 72b, 72c is training data 75.

Learning of the neural network means that the weight w is adjusted so that the output y (xn: w) of the neural network at a given input xn is as close as possible to the output dn as shown in the following equation for any group (xn, dn) of inputs and outputs.

[ math figure 9]

y(x _n ：w)≈d _n

Error function (error function) refers to a measure of the proximity of a function expressed by a neural network to training data. The error function is also called loss function (loss function). The error function E (w) used in the method for analyzing a cell type according to the embodiment is represented by the following formula 7. The formula (7) is called cross entropy (cross entropy).

[ math figure 10]

A method of calculating the cross entropy of (formula 7) will be described. The output layer 50b of the neural network, i.e., the final layer of the neural network, used in the analysis method of cell type of the embodiment uses an activation function for classifying the input x into a limited number of categories according to the content. The activation function is called a normalized exponential function (softmax function) and is expressed as follows (equation 8). The output layer 50b is provided with the same number of nodes as the number of categories k. The total input u of each node K (k=1, … …, K) of the output layer L is the output uk according to the preceding layer L-1 ^(L) Given by the operator. Thus, the output of the kth node of the output layer is expressed as (formula 8) below.

[ mathematics 11]

(equation 8) is a normalized exponential function. The sum of the outputs y1, … …, yK determined by (expression 8) is always 1.

When the categories are denoted as C1, … …, CK, the output yK of node k of the output layer L (i.e., uk ^(L) ) Representing the probability that a given input x belongs to category CK. The input x is classified into a category in which the probability expressed by the following (expression 9) is maximum.

[ math figure 12]

In learning of the neural network, a function expressed by the neural network is regarded as a model of a posterior probability (posterior probability) of each type of purpose, and under such a probability model, a likelihood (likelihood) of the evaluation weight w with respect to the training data is selected, and the weight w that maximizes the likelihood is selected.

The target output obtained by the normalized exponential function of (expression 8) is set to 1 only when the output is of the correct category, and the output is set to 0 otherwise. When the target output is expressed in a vector format of dn= [ dn1, … …, dnK ], for example, when the correct category of the input xn is C3, only the target output dn3 is 1 and the other target outputs are 0. When encoding as such, the posterior distribution (posterior) is represented by the following (formula 10).

[ math 13]

The likelihood L (w) of the weight w for the training data { (xn, dn) } (n=1, … …, N) is represented by the following (formula 11). When taking the logarithm of the likelihood L (w) and inverting the sign, an error function of (equation 7) is derived.

[ math 14]

Learning means that the error function E (w) calculated based on the training data is minimized for the parameter w of the neural network. In the method for analyzing a cell type according to the embodiment, the error function E (w) is represented by (formula 7).

Minimizing the error function E (w) for the parameter w is the same meaning as finding the local minimum of the error function E (w). The parameter w is the weight of the connections between the nodes. The minimum point of the weight w is obtained by repeating the repeated calculation of the update parameter w with an arbitrary initial value as a starting point. As an example of such calculation, there is a gradient descent method (gradient descent method).

In the gradient descent method, a vector represented by the following (expression 12) is used.

[ math 15]

In the gradient descent method, the value of the current parameter w is repeated a plurality of times in the negative gradient direction (i.e.)) And (5) processing movement. When the current weight is reset to w ^(t) Let the weight after the shift be w ^(t+1) In this case, the operation by the gradient descent method is represented by the following (expression 13). The value t means the number of times the parameter w is moved.

[ math 16]

The sign shown in the following (formula 14) used in the (formula 13) is a constant that determines the magnitude of the update amount of the parameter w, and is called a learning coefficient.

[ math 17]

Epsilon … … (14)

By repeating the operation represented by (formula 13), the error function E (w ^(t) ) As the value t increases and decreases, the parameter w reaches a minimum point.

The calculation based on (expression 13) may be performed on all the training data (n=1, … …, N), or may be performed on only a part of the training data. The gradient descent method performed on only a part of the training data is called a random gradient descent method (stochastic gradient descent). The random gradient descent method is used in the method for analyzing a cell type of the embodiment.

Effect of the embodiment

The analyte analyzer 4000 includes: the measurement unit 400 includes an FCM detection unit 410 or a detection unit 470 (optical detection unit) for acquiring an optical signal from a subject; and an analysis unit 300 or 600 for analyzing the 1 st and 2 nd data corresponding to the optical signal. The analysis units 300 and 600 perform AI analysis (1 st analysis operation based on an artificial intelligence algorithm) on 1 st data in all acquired waveform data, and perform calculation processing analysis (2 nd analysis operation to process a representative value corresponding to a feature of an analyte) on 2 nd data in all acquired waveforms.

According to this configuration, by sharing the analysis processing of the data corresponding to the optical signal acquired from the subject by the AI analysis and the calculation processing analysis, the load of the analysis units 300, 600 as the processing data of the computer can be reduced as compared with the case where the data corresponding to the optical signal is analyzed by using only the artificial intelligence algorithm.

When the analyte analyzer 4000 is a blood cell analyzer or a urine analyzer, the 1 st and 2 nd data are digital data (waveform data) corresponding to the intensities of optical signals based on light generated from the respective analytes (cells, presence of components). The optical signal in this case is an analog signal output from the light receiving element based on the forward scattered light, the side scattered light, and the fluorescence. The optical signal is a signal that has regions corresponding to the respective analytes in the subject and reflects the presence of the analytes in the subject. Waveform data (1 st and 2 nd data) is generated corresponding to the region of the optical signal. In other words, the waveform data corresponds to an optical signal acquired during the irradiation position of the analyte by the light from the light source 4111. The representative value corresponding to the characteristic of the analyte is, for example, a peak value, an area, or a width value obtained from waveform data corresponding to the analyte (see fig. 3). The 1 st analysis operation and the 2 nd analysis operation refer to operations for determining the type of an analyte (cell, physical component).

When the sample analyzer 4000 is a blood coagulation measuring device, the 1 st and 2 nd data are digital data (coagulation waveform data) corresponding to the intensity of an optical signal based on transmitted light or scattered light. The optical signal in this case is an analog signal from the start of photometry to the end of photometry (for example, 180 seconds after the start) based on the intensity of the transmitted light or the scattered light. The optical signal may be an analog signal from the start of the coagulation reaction (timing T2 in fig. 4) to the end of the coagulation reaction (timing T3). The solidification waveform data (1 st and 2 nd data) is generated from the optical signal. The representative value corresponding to the characteristic of the analyte is, for example, a time (for example, T-T2) obtained from the coagulation waveform data when the intensity of the detected light satisfies a predetermined condition (for example, when the absorbance is 50%) (see fig. 4). The 1 st analysis operation is an operation for determining whether or not there is a possibility of a non-specific reaction, and the 2 nd analysis operation is an operation for determining the coagulation time.

The 1 st data and the 2 nd data may be the same data or completely different data. For example, when AI analysis is performed on nucleated erythrocytes and basophils and calculation processing analysis is performed on other leukocytes based on waveform data obtained by 1 measurement using the WDF channel, the 1 st data and the 2 nd data are the same data. Among the waveform data obtained by the two measurements, when the calculation processing analysis is performed on the waveform data based on the 1 st measurement and the AI analysis is performed on the waveform data based on the 2 nd measurement, the 1 st data and the 2 nd data are different data.

The representative value of the waveform data (the 2 nd data) as the processing object in the calculation processing analysis is determined based on the size of the waveform data (the 2 nd data). Specifically, representative values such as peak value, area, width, etc. are determined based on the size of the 2 nd data, and representative values such as time until the absorbance becomes 50%. Thus, the representative value can be smoothly determined.

In the case where the analyte analyzer 4000 is a blood cell analyzer or a urine analyzer, the optical signal has a region corresponding to each analyte in the analyte. In the calculation processing analysis, the analysis units 300, 600 determine representative values as the calculation processing analysis targets based on waveform data (data 2) corresponding to the respective areas of the optical signal. As described above, since the optical signal includes the regions corresponding to the respective analytes, the representative values such as the peak value, the area, and the width corresponding to the respective analytes can be smoothly determined based on the waveform data corresponding to the respective regions of the optical signal.

In the case where the analyte analyzer 4000 is a blood cell analyzer or a urine analyzer, the optical signal has a region corresponding to each analyte in the analyte. In the AI analysis, the analysis units 300, 600 input waveform data (1 st data) corresponding to each of the regions of the optical signal to the artificial intelligence algorithm. As described above, the optical signal includes the regions corresponding to the respective analytes, and thus AI analysis can be smoothly performed by inputting waveform data corresponding to the respective regions of the optical signal to the artificial intelligence algorithm.

As described above, when the optical signal has regions corresponding to the respective analytes in the subject, the measurement unit 400 acquires waveform data (data 1 and 2) based on a signal greater than a predetermined threshold corresponding to the intensity of the optical signal, as shown in the upper diagram of fig. 3. According to this structure, waveform data corresponding to each analyte can be accurately acquired.

As shown in fig. 6, the analysis unit 300 determines data (data 1) that is the object of AI analysis and data (data 2) that is the object of calculation processing analysis based on rules for determining data that is the object of AI analysis (analysis work 1) and calculation processing analysis (analysis work 2). With this configuration, it is possible to smoothly determine which of the 1 st analysis operation and the 2 nd analysis operation is used to analyze the data corresponding to the optical signal.

As shown in fig. 7, the analysis unit 300 determines data (data 1) to be an AI analysis object and data (data 2) to be a calculation processing analysis object from measurement items included in a measurement instruction for an object to be examined. According to this configuration, for example, the AI analysis can be used to analyze measurement items that are difficult to analyze with high accuracy by the calculation process analysis, and the calculation process analysis can be used to analyze normal measurement items. This can realize high-precision analysis and reduce the load on the analysis unit 300.

As shown in fig. 13, the analysis unit 300 determines data (data 1) to be an AI analysis object and data (data 2) to be a calculation processing analysis object according to the type of a measurement instruction for an object to be examined. According to this configuration, for example, it is possible to determine which of the AI analysis and the calculation process analysis is to be performed, based on the type of measurement instruction such as the normal measurement (normal), the re-measurement (re-measurement) in which the same measurement instruction is executed again, and the measurement (return) after the measurement instruction is reset, that is, based on the purpose of the measurement based on the measurement instruction.

As shown in fig. 11, the analysis unit 300 determines data (data 1) that is the object of AI analysis and data (data 2) that is the object of calculation processing analysis according to the analysis mode of the object analysis device 4000. According to this configuration, for example, by setting any one of the AI analysis mode and the calculation processing analysis mode in advance for the object analysis device 4000, the trouble of setting the analysis mode for each object and measurement item can be eliminated.

As shown in fig. 17 and 22, the analysis unit 300 determines whether or not the AI analysis (analysis operation 1) needs to be performed based on the analysis result of the calculation process analysis (analysis operation 2). According to this configuration, for example, in the case where an analysis result based on the calculation processing analysis requires further detailed analysis, AI analysis is performed, so that high-precision analysis can be performed.

As shown in fig. 17 and 22, the analysis unit 300 determines whether or not the AI analysis (analysis operation 1) needs to be performed, based on whether or not a predetermined analyte is detected in the subject by the calculation processing analysis (analysis operation 2). According to this structure, when, for example, a blast cell, an abnormal lymphocyte, a atypical lymphocyte, or the like is detected by calculation processing analysis, a further detailed test can be performed by AI analysis.

As shown in fig. 19, the analysis unit 300 analyzes waveform data (data 1) corresponding to the analytes classified into the predetermined category by the calculation process analysis (analysis work 2) with the calculation process analysis (analysis work 1). As shown in fig. 21, for example, according to the analysis result of the calculation processing analysis, the distribution areas of the cells classified into monocytes, lymphocytes, and the like are close. Therefore, by performing AI analysis on cells classified into monocytes, lymphocytes, and the like by calculation processing analysis, high-precision classification can be performed.

The amount of data of the representative value processed in the calculation processing analysis (analysis operation 2) is smaller than the waveform data (data 1) input to the AI algorithm 60 in the AI analysis (analysis operation 1). That is, in the calculation processing analysis, since the amount of data to be processed is small compared with the AI analysis, the load on the computer performing the analysis is small compared with the AI analysis. This can shorten the TAT (turnaround time) of the analysis of the measurement results.

The embodiments of the present invention can be modified in various ways within the scope of the technical idea shown in the claims.

Claims

1. An analyte analyzer for analyzing an analyte in an analyte, the analyte analyzer comprising:

a measuring unit including an optical detection unit for acquiring an optical signal from the object; and

an analysis unit that analyzes the 1 st data and the 2 nd data corresponding to the optical signal,

wherein the analysis unit

Performing an artificial intelligence algorithm-based analysis job 1 on the data 1,

and executing the 2 nd analysis work of processing the representative value corresponding to the characteristic of the analyte in the 2 nd data.

2. The analyte analysis device of claim 1, wherein,

the analysis unit determines the representative value based on the 2 nd data in the 2 nd analysis operation, and processes the determined representative value.

3. The analyte analysis device of claim 2, wherein,

the analysis unit determines the representative value based on the size of the 2 nd data in the 2 nd analysis operation.

4. The analyte analysis device according to claim 2 or 3, wherein,

The optical signal has regions corresponding to respective analytes in the subject,

the analysis unit determines the representative value based on the 2 nd data corresponding to each of the regions of the optical signal in the 2 nd analysis operation.

5. The analyte analysis device of claim 4, wherein,

the analysis unit determines a peak value of the 2 nd data in the region as the representative value.

6. The analyte analysis device according to any one of claims 1 to 5, wherein,

the analysis unit inputs the 1 st data corresponding to each of the regions of the optical signal to the artificial intelligence algorithm in the 1 st analysis operation.

7. The analyte analysis device according to any one of claims 4 to 6, wherein,

the measurement unit acquires the 1 st data and the 2 nd data based on a signal that is greater than a predetermined threshold corresponding to the intensity of the optical signal.

8. The analyte analysis device of claim 1, wherein,

the measurement unit acquires the representative value based on the optical signal,

The analysis unit processes the representative value acquired by the measurement unit in the analysis operation 2.

9. The analyte analysis device according to any one of claims 1 to 8, wherein,

the optical signal is a signal reflecting the presence of an analyte in the subject.

10. The analyte analysis device according to any one of claims 1 to 9, wherein,

the optical detection unit includes a light source, a flow cell, and a photodetector, irradiates the flow cell with light, and detects light generated from an analyte in the subject flowing through the flow cell.

11. The analyte analysis device of claim 10, wherein,

the 1 st data and the 2 nd data correspond to the optical signals acquired during the passage of the analyte through the illuminated location of the light.

12. The analyte analysis device according to any one of claim 1 to 3, wherein,

the optical detection unit includes a light source and a photodetector, irradiates the object to be detected with light, and detects light transmitted through or scattered by the object to be detected.

13. The analyte analysis device of claim 12, wherein,

The object to be tested is blood,

the measuring unit further includes a sample preparation unit for mixing a blood coagulation reagent with the sample,

the optical detection unit irradiates the sample mixed with the blood coagulation reagent with light, and detects light transmitted through or scattered by the sample.

14. The analyte analysis device of claim 13, wherein,

the 1 st data and the 2 nd data include data corresponding to the optical signal acquired from a timing indicating a start of coagulation of the object to a timing indicating an end of coagulation of the object.

15. The analyte analysis device according to claim 13 or 14, wherein,

the analysis unit determines the 2 nd data when the detected intensity of the light satisfies a predetermined condition as the representative value.

16. The analyte analysis device according to any one of claims 13 to 15, wherein,

the analysis unit determines the possibility of generating a non-specific reaction using the 1 st analysis job.

17. The analyte analysis device according to any one of claims 1 to 16, wherein,

the analysis unit analyzes the 1 st data using a convolution operation based on the artificial intelligence algorithm.

18. The analyte analysis device according to any one of claims 1 to 17, wherein,

the analysis unit analyzes the 1 st data using a matrix operation based on the artificial intelligence algorithm.

19. The analyte analysis device of claim 18, wherein,

the analysis unit performs matrix operations based on the artificial intelligence algorithm by parallel processing by a parallel processing processor.

20. The analyte analysis device of claim 19, wherein,

the analysis unit executes the 1 st analysis job through the parallel processing processor, and executes the 2 nd analysis job through a host processor of the parallel processing processor.

21. The analyte analysis device according to any one of claims 1 to 20, wherein,

the artificial intelligence algorithm is a deep learning algorithm.

22. The analyte analysis device according to any one of claims 1 to 21, wherein,

the analysis unit determines the 1 st data and the 2 nd data based on a rule for determining data that is an object of each of the 1 st analysis job and the 2 nd analysis job.

23. The analyte analysis device according to any one of claims 1 to 22, wherein,

The analysis unit determines the 1 st data and the 2 nd data based on measurement items included in a measurement instruction for the object to be tested.

24. The analyte analysis device according to any one of claims 1 to 23, wherein,

the analysis unit determines the 1 st data and the 2 nd data according to a type of a measurement instruction for the object to be tested.

25. The analyte analysis device according to any one of claims 1 to 24, wherein,

the analysis unit determines the 1 st data and the 2 nd data according to an analysis mode of the object analysis device.

26. The analyte analysis device according to any one of claims 1 to 25, wherein,

the analysis unit determines whether the 1 st analysis job needs to be executed according to the analysis result of the 2 nd analysis job.

27. The analyte analysis device according to any one of claims 1 to 26, wherein,

the analysis unit determines whether the 1 st analysis job needs to be executed according to whether a predetermined analyte is detected in the subject by the 2 nd analysis job.

28. The analyte analysis device according to any one of claims 1 to 27, wherein,

The analysis unit analyzes the 1 st data corresponding to the analytes classified into the predetermined category by the 2 nd analysis work using the 1 st analysis work.

29. The analyte analysis device according to any one of claims 1 to 28, wherein,

the amount of data of the representative value processed in the analysis work 2 is smaller than the 1 st data input to the artificial intelligence algorithm in the analysis work 1.

30. The analyte analysis device according to any one of claims 1 to 29, wherein,

further comprises a sample preparation unit for preparing a measurement sample based on the sample and the reagent,

the optical detection unit acquires the optical signal from the object to be detected contained in the measurement sample,

the analysis unit analyzes the 1 st data and the 2 nd data corresponding to the optical signals acquired from 1 measurement sample.

31. The analyte analysis device of claim 30, wherein,

the 1 st data and the 2 nd data are a plurality of data, respectively, and at least a part of the data are the same data as each other.

32. The analyte analysis device according to any one of claims 1 to 31, wherein,

the analysis unit analyzes the 1 st data and the 2 nd data, wherein the 1 st data and the 2 nd data correspond to respective optical signals of the optical signals respectively acquired from the plurality of measurement samples containing the object acquired from the same subject.

33. The analyte analysis device according to any one of claims 1 to 32, wherein,

the analysis unit analyzes the 1 st data and the 2 nd data, the 1 st data and the 2 nd data corresponding to respective optical signals of the optical signals acquired from the plurality of measurement samples including the subject different from each other.

34. The analyte analysis device of claim 32 or 33, wherein,

The reagents contained in the plurality of measurement samples are the same type of reagent as each other.

35. The analyte analysis device of claim 32 or 33, wherein,

the reagents contained in the plurality of measurement samples are different types of reagents from each other.

36. A method of analyzing an analyte in an analyte, the method comprising:

a step of acquiring an optical signal from the subject; and

an analysis step of analyzing the 1 st data and the 2 nd data corresponding to the optical signal,

wherein, in the analysis process,

37. A program for causing a computer to execute a process of analyzing an analyte in a subject, the program characterized in that,

comprises a process of analyzing the 1 st data and the 2 nd data corresponding to the optical signal obtained from the object,

wherein, in the process, the processing is performed,

38. An analyte analyzer for analyzing an analyte in an analyte, the analyte analyzer comprising:

an analysis unit that analyzes data corresponding to the optical signal,

wherein the analysis unit analyzes the data using an artificial intelligence algorithm based analysis 1 or an analysis 2 that processes a representative value corresponding to a characteristic of the analyte in the data according to an analysis mode of the data.

39. The analyte analysis device of claim 38, wherein,

the analysis mode may be selected for each measurement item included in a measurement instruction for the subject or for each type of measurement instruction for the subject.

40. A method of analyzing an analyte in an analyte, the method comprising:

a step of acquiring an optical signal from the subject; and

An analysis step of analyzing data corresponding to the optical signal,

wherein, in the analysis process,

according to the analysis mode of the data, the data is analyzed by using the 1 st analysis based on an artificial intelligence algorithm or the 2 nd analysis for processing the representative value corresponding to the characteristic of the analyte in the data.

41. A program for causing a computer to execute a process of analyzing an analyte in a subject, the program characterized in that,

comprising a process of analyzing data corresponding to an optical signal acquired from the subject,

wherein, in the process, the processing is performed,