Immunodynamics generation device and immunodynamics generation method

Abstract

The present invention provides an immunodynamics generation device that leads to the management of a health condition, and early detection, early treatment, and optimal treatment of a disease. This immunodynamics generation device comprises at least one immunodynamics processing unit, a storage unit that stores at least clinical data, and a display unit. The immunodynamics processing unit implements control to acquire basic clinical features and immune subgroup dynamics information on the basis of the clinical data, carry out monitoring simultaneous to the formation of a classification model by implementing self-learning on the basis of the immune subgroup dynamics information and the classification model, and compare the overall relationship between the basic clinical features and the immune subgroup dynamics information on the basis of the classification model, thereby evaluating the immunodynamics information of a subject, and simultaneous thereto, displaying the evaluation results of the immunodynamics information on the display unit.

Description

Immune dynamics generation device and immune dynamics generation method

[0001] The present invention relates to the technical field of immune dynamics generation, and in particular to an immune dynamics generation device and an immune dynamics generation method.

[0002] Following the release of the Opdivo immune checkpoint inhibitor drug, which was developed after Kyoto University Distinguished Professor Tasuku Honjo was awarded the Nobel Prize in Physiology or Medicine in 2018, immune cell therapy, a fourth cancer treatment, and minimally invasive radiation cancer treatment MIS therapy (Minimum Invasive Surgery) have gradually come into the spotlight, with the number of treatment methods rapidly increasing. What these have in common is that they are treatment methods that focus on autoimmunity (hereinafter referred to as immunotherapy).

[0003] In immunotherapy, the study of immune dynamics is essential. The study of immune dynamics is not only useful for the early detection of disease, but can also be used as an indicator for the prevention of disease and the evaluation of the therapeutic effects and side effects of such treatment. Therefore, it is becoming an essential item for the early detection and treatment of disease in modern medical practice.

[0004] For example, when evaluating the effectiveness of clinical cancer treatment, it is considered most advantageous for immunotherapy to evaluate imaging diagnosis (changes in cancer burden (cancer size)) at a certain point in time (especially before and after a change in treatment) while simultaneously evaluating the immune dynamics at that time. The imaging diagnosis referred to here is, for example, the results of a cancer volume measurement test using PET-CT, as shown in Figure 5. The immune dynamics is, for example, immune dynamics information obtained based on the results of a blood count and white blood cell image of general blood components, as shown in Figure 6.

[0005] There are a great many immune cells that control immune functions in the blood, but they are generally derived from lymphoid stem cells and myeloid stem cells, and other blood cell components such as red blood cells and platelets are also born from the same type of hematopoietic stem cell (pluripotent stem cell) as immune cells. Lymphoid stem cells differentiated from hematopoietic stem cells redifferentiate into helper T cells, αβ killer T cells, γδ T cells, B cells, NKT cells, and NK cells. Myeloid stem cells differentiated from hematopoietic stem cells redifferentiate into macrophages, granulocyte neutrophils (more than 90%), eosinophils, basophils, red blood cells, and platelets.

[0006] Recently, with regard to immune dynamics, the inventors have proposed that immune subgroup dynamic information be considered, including the number of B cells, the number of NK cells, the number of NKT cells, and the number and proportion of immune lymphocyte subgroups including at least the number of helper T cells, the number of killer T cells, the number of NK cells, the number of NKT cells, and the number of B cells (Patent Document 1).

[0007] Japanese Patent Application Laid-Open No. 2021-189081

[0008] However, while there are currently methods for comparing changes in cancer images (changes in cancer burden) or generating immune dynamics information, there are no objective criteria for clinical treatment that simultaneously consider both. Furthermore, even in the general public, there are no objective criteria for simultaneously comparing and considering one's own health status and immune dynamics information. Particularly in today's society, where infectious diseases such as COVID-19 are rampant, there is an urgent need to establish objective criteria for evaluating clinical treatment while taking immune dynamics into account. In other words, there is an urgent need for a system that can objectively assess immune dynamics based on lymphocyte subpopulations, including helper T cells, αβ killer T cells, γδ T cells, B cells, NKT cells, and NK cells.

[0009] In order to overcome the problems of the existing technologies described above, the present inventors have conducted extensive research into the influence of immune dynamics on cancer treatment and other such treatments. As a result, they have succeeded in finding an integrated criterion for objectively evaluating health status and other factors by simultaneously comparing and considering clinical data and immune dynamics, and the results of this research are now disclosed.

[0010] To achieve the above objectives, the present invention provides the following technical solutions:

[0011] The immune dynamics generation device provided by the present invention comprises at least one immune dynamics processing unit, a memory unit for storing at least clinical data, and a display unit. The immune dynamics processing unit acquires clinical basic characteristics and immune subgroup dynamics information based on the clinical data, and performs self-learning based on the immune subgroup dynamics information and the classification model to form and monitor a classification model. Based on the classification model, the immune dynamics information of the subject is evaluated by comparing the overall relationship between the clinical basic characteristics and the immune subgroup dynamics information, and simultaneously controls the display of the evaluation result of the immune dynamics information on the display unit.

[0012] Preferably, the clinical data includes at least health checkup item data for normal individuals and health checkup item data for non-normal individuals.

[0013] Preferably, the immune subpopulation dynamic information includes the number of B cells, the number of NK cells, the number of NKT cells, and the numbers and proportions of immune lymphocyte subpopulations including at least the number of helper T cells, the number of killer T cells, the number of αβ killer-T cells, the number of γδT cells, the number of NK cells, the number of NKT cells, and the number of B cells.

[0014] Preferably, the immune dynamics processing unit obtains the clinical basic features, the immune subgroup dynamics information, and the classification model based on the health checkup item data of the normal individual, and obtains normal immune status information by comparing the overall relationship between the clinical basic features and the immune subgroup dynamics information.

[0015] Preferably, the immune dynamics processing unit obtains the clinical basic features, the immune subgroup dynamics information and the classification model based on the health checkup item data of the abnormal individual, and obtains abnormal immune status information by comparing the overall relationship between the clinical basic features and the immune subgroup dynamics information.

[0016] Preferably, the immune dynamics processing unit obtains the clinical basic features, the immune subgroup dynamics information, and the classification model based on the clinical data including cancer cell data, and obtains cancer incidence information by comparing the overall relationship between the clinical basic features and the immune subgroup dynamics information based on the abnormal immune state information.

[0017] Preferably, the immune dynamics processing unit determines that the cancer cells are in an aggravated state when the ratio of the cancer incidence information is greater than one.

[0018] Preferably, the immune dynamics processing unit determines that the cancer cells are in a life-prolonged state when the ratio of the cancer incidence information is 1.

[0019] Preferably, the immune dynamics processing unit determines that the cancer cells are in an improved state when the ratio of the cancer incidence information is less than 1.

[0020] The present invention also provides a method for generating immune dynamics, which includes the steps of acquiring clinical data of a subject, obtaining immune subpopulation dynamics information of the subject, and monitoring the average data and abnormal data with changes of a large number of normal samples calculated using the basic data and lymphocyte subpopulation data stored in a memory unit, and evaluating the immune dynamics of the subject by comparing the overall relationship between the clinical data and the immune subpopulation dynamics information based on the average data and abnormal data.

[0021] According to the present invention, it is possible to provide a method and device for generating immune dynamics that establishes integrated criteria for objective evaluation of immune dynamics based on clinical data, etc., and that allows for immediate and accurate understanding of the state of immune strength and the progression of a disease, etc. In particular, by utilizing the method and device for generating immune dynamics of the present invention, it is possible to manage health conditions, detect diseases early, treat them early, and optimize treatment, thereby improving and preventing serious diseases, based on integrated criteria for objective evaluation of immune dynamics.

[0022] FIG. 1 is a conceptual diagram of an immune dynamics generation device in a specific embodiment of the present invention. FIG. 2 is a conceptual diagram showing the configuration of a memory unit of an immune dynamics generation device. FIG. 3 is a Gomperzian tumor growth curve in which cancer cell growth is broadened in time units. FIG. 4 is a diagram showing an example of a classification extraction table of this embodiment. FIG. 5 is a diagram showing cancer volume measurement test results by PET-CT. FIG. 6 is a diagram showing test results of blood count and white blood cell image of general blood components. FIG. 7 is a diagram showing an example of the functional configuration of a learning mechanism of the present invention. FIG. 8 is a flowchart showing a learning process procedure for forming a classification model. FIG. 9 is a diagram showing an example of learning data of this embodiment. FIG. 10 is a flowchart showing a process procedure for performing immune dynamics evaluation processing. FIG. 11 is a diagram showing basic data of Example 1. FIG. 12 is a diagram showing the number of each component of Example 1 calculated based on the above mathematical formulas 1 and 2. FIG. 1 is a diagram showing the number of granulocytes, the number of monocytes, the number of lymphocytes, the number of helper T cells, the number of killer T cells, the number of αβ killer T cells, the number of γδ T cells, the number of NK cells, the number of NKT cells, and the number of B cells for classification data in Example 1. FIG. 2 is a diagram showing the proportions of granulocytes, monocytes, and lymphocytes in Example 1. FIG. 3 is a diagram showing the proportions of helper T cells, killer T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells, and B cells in Example 1. FIG. 4 is a diagram showing an example of a classification model in Example 1. FIG. 5 is a diagram showing an example of comparing an evaluation model and a reference model in Example 1. FIG. 1 is a diagram showing the number of granulocytes, the number of monocytes, the number of lymphocytes, the number of helper T cells, the number of killer T cells, the number of αβ killer-T cells, the number of γδ T cells, the number of NK cells, the number of NKT cells, and the number of B cells for the basic data of Example 2 and the classification data of Example 2. FIG. 2 is a diagram showing the proportions of granulocytes, monocytes, and lymphocytes of Example 2. FIG. 3 is a diagram showing the proportions of helper T cells, killer T cells, αβ killer-T cells, γδ T cells, NK cells, NKT cells, and B cells of Example 1. FIG. 4 is a diagram showing an example of a classification model of Example 2. FIG. 10 shows the number of granulocytes, monocytes, lymphocytes, helper T cells, killer T cells, αβ killer-T cells, γδ T cells, NK cells, NKT cells, and B cells used in the basic data of Example 3 and the classification data of Example 3.Fig. 1 is a diagram showing the proportions of granulocytes, monocytes, and lymphocytes in Example 3. Fig. 2 is a diagram showing the proportions of helper T cells, killer T cells, αβ killer-T cells, γδ T cells, NK cells, NKT cells, and B cells in Example 3. Fig. 3 is a diagram showing an example of comparing the evaluation model and the basic model in Example 3.

[0023] Hereinafter, an embodiment of an immune dynamics generating apparatus and generating method according to the present invention will be described with reference to the drawings.

[0024] 1 , the immune dynamics generating device 1 includes an immune dynamics processing unit 2, a memory unit 3, an information input unit 4, and a display unit 5. The memory unit 3 is connected to an external information source 6 in a communicable state via a network 7. The immune dynamics processing unit 2 includes a learning mechanism 21 and an immune dynamics evaluation unit 22.

[0025] The immune dynamics processing unit 2 may be, for example, a workstation or a personal computer, or may be a server computer connected to them via a network. The immune dynamics processing unit 2 has the immune dynamics processing program and learning program of the present invention installed. The immune dynamics processing program and learning program are stored in an externally accessible state in a storage device of a server computer connected to the network 7 or in a network storage, and are downloaded and installed into the immune dynamics processing unit 2 used by a doctor or subject upon request. Alternatively, the programs may be recorded on a recording medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory) and distributed, and then installed into the immune dynamics processing unit 2 from the recording medium.

[0026] The immune dynamics processing program of the present invention has at least an immune dynamics basic model formation function, a learning function, and an immune dynamics evaluation function. For example, based on the immune dynamics processing program, the immune dynamics processing unit 2 forms an immune dynamics basic model using related clinical data as the numerator and related immune subgroup dynamic information as the denominator, in accordance with the correlation between them, so that the ratio between them becomes 1.

[0027] The memory unit 3 is composed of hardware (circuits, dedicated logic, etc.), software (such as software running on a general-purpose computer system or a dedicated machine), etc., and has the function of storing various information. For example, the memory unit 3 may be a hard disk drive (HDD), a solid state drive (SSD), a flash memory, etc. Alternatively, the memory unit 3 may be a computer that saves and manages various data and is equipped with a large-capacity external storage device and database management software. The memory unit 3 communicates with other devices via a wired or wireless network. For example, the memory unit 3 is connected to an external information source 6 in a communicable state via the network 7, and transmits and receives basic data on immune dynamics, lymphocyte subpopulation data, etc. Furthermore, various data, including immune dynamics evaluation results generated by the immune dynamics processing unit 2, are acquired directly or via a network, and stored and managed in a recording medium such as a large-capacity external storage device. The storage format of various data and communication between devices via a network are based on protocols such as DICOM (Digital Imaging and Communication in Medicine).

[0028] The information input unit 4 may be input devices such as a keyboard and mouse, or a device having an information input function connected to a network. The display unit 5 has a function of displaying analytical information such as the immunodynamic evaluation results, and may be a general liquid crystal display such as a touch panel or display panel. A touch panel display integrating the display unit 5 and the information input unit 4 may also be used. The network 7 provides wired or wireless communication with an external information source 6.

[0029] The external information source 6 is a terminal used by companies, medical personnel, medical institutions, general users, etc. to send and receive immune data such as basic data and lymphocyte subpopulation data to the memory unit 3, and may be a mobile terminal such as a personal computer, laptop computer, smartphone, or tablet terminal, or a wearable terminal such as a head-mounted display such as smart glasses or a smart watch. There may be multiple companies, medical personnel, medical institutions, general users, etc.

[0030] As shown in FIG. 2 , the memory unit 3 stores at least basic data 31, a classification extraction table 33, a classification model box 34, and an immune dynamics table 35. The basic data 31 includes at least normal health checkup item data and abnormal health checkup item data. The normal health checkup item data includes at least health checkup data of normal individuals. The abnormal health checkup item data includes diagnostic data of individuals infected with infectious diseases such as influenza, cancer cell data of various cancer patients, and various other disease condition diagnostic data. Information and data formats according to the present invention include, but are not limited to, various formats such as images, tables, numerical values, and text. The memory unit 3 also stores a learning program and an immune dynamics processing program. The basic data 31 and the classification extraction table 33 together form the clinical data of the present invention. The memory unit 3 may also store immune subgroup dynamics information (lymphocyte subgroup data 32).

[0031] Health checkup and diagnosis items include, for example, blood test information, urine test information, diagnostic imaging information (X-ray, MRI, CT scan, etc.), biopsy result information, etc. Blood test information includes, for example, information obtained by analyzing blood components, such as white blood cell count, the percentage of basophils in white blood cells, the percentage of eosinophils, the percentage of neutrophils, the percentage of lymphocytes, the percentage of monocytes, CD3 antigen, CD56 antigen (and / or CD16 antigen), CD4 antigen, and CD8 antigen.

[0032] The cancer cell data of various cancer patients includes information on the amount of cancer cells in the body and information on the increase or decrease of the amount of cancer cells in the body, etc. The amount of cancer cells in the body and information on the increase or decrease of the amount of cancer cells in the body may be obtained based on the results of biochemical markers, imaging techniques, histological examinations, liquid biopsies, molecular biological techniques, etc., or may be obtained based on a Gomperzian tumor growth curve.

[0033] Biochemical markers are a method of indirectly estimating the amount of cancer cells by measuring the levels of specific biochemical markers released into the blood and other body fluids by specific types of cancer. For example, prostate-specific antigen (PSA) is known as a marker for prostate cancer.

[0034] Imaging technology is a method of estimating the extent and volume of cancer from images obtained using imaging technologies such as CT scans, MRIs, and PET scans. It helps to visually identify the location and size of cancerous tissue within the body.

[0035] Histology is the direct observation of the presence and amount of cancer cells in tissue samples obtained by biopsy or surgery under a microscope. This method is essential for diagnosing and staging cancer.

[0036] Liquid biopsy is a method for detecting cancer cells and their DNA fragments from a blood sample. Liquid biopsy is minimally invasive and is useful for monitoring the progression of cancer.

[0037] Molecular biology techniques involve using PCR (polymerase chain reaction) and other molecular techniques to detect gene mutations and expression patterns specific to cancer cells.

[0038] The Gomperzian tumor growth curve is a broad curve that shows the growth of cancer cells over time, as shown in Figure 3. Generally, the Gomperzian tumor growth curve presents an S-shaped curve, with the vertical axis representing the amount of cancer cell growth and the horizontal axis representing time.

[0039] When a single cancer cell develops in the body, it is invisible to the naked eye, but as time passes, the cancer cell gradually divides and increases, eventually forming a mass of cancer cells. When cancer cells first develop, they grow slowly, but when the number of cancer cells in the body reaches a certain amount, they grow rapidly. According to previous reports, when cancer cells grow and reach 100 (10 ２ ), the rate of proliferation increases rapidly, reaching 100 million (10 ９ ) masses of cancer cells are formed, and then they have the property of slowly growing. The Gomperzian tumor growth curve is a curve that demonstrates the nature of this cancer cell growth change.

[0040] With the current medical standards, cancer cells number about 100 million (10 ９) and is diagnosed as "early cancer." At this point, the size of the tumor is about 1 cm in diameter and weighs about 1 gram. After the cancer cells have multiplied to about 100 million, the growth rate slows down relatively. Cancer cells continue to grow from 100 million until the number of cells reaches 100 billion (10 １２ ) reaches the limit of human life.

[0041] The lymphocyte subpopulation data 32 includes immune cell basic data, immune dynamics pattern data, immune dynamics model data, etc. The immune cell basic data includes at least the white blood cell count, and the quantities and proportions of basophils, eosinophils, neutrophils, lymphocytes, monocytes, CD3 antigen, CD56 antigen (and / or CD16 antigen), CD4 antigen, CD8 antigen, etc. in the white blood cells. The lymphocytes include at least helper T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells, B cells, etc. The immune cell basic data, immune dynamics pattern data, and immune dynamics model data included in the lymphocyte subpopulation data 32 may be obtained based on the technical solution disclosed in Japanese Patent No. 6796737 depending on the actual situation.

[0042] The classification model box 34 is a database in which the classification model 600 is stored. The classification model 600 receives as input feature quantities and type quantities extracted from input data by referring to the classification extraction table 33 and vectorized according to weighting. The classification model 600 repeatedly adjusts the function values ​​such as type quantities or immune dynamic feature quantities so that the ratio between the function values ​​such as type quantities (e.g., the quantity of the basic feature 341 in this embodiment) and the immune dynamic feature quantities (e.g., the quantity of the immune dynamic feature 39 in this embodiment) falls within the overall condition index range. The classification model 600 determines the final type quantities and final feature quantities, which are then output as classification results. In this embodiment, the final type quantities are the sum of the basic features 341 and the basic feature weights 381, which will be described later, and the final feature quantities are the sum of the immune dynamic features 39 and the immune dynamic weights 391. Note that in this embodiment, the input data is information composed of the basic data 31 or the basic data 31 and the limbic cell subpopulation data 32 as components. The feature quantities represent characteristics of limbic cell subpopulations, and an example of such a feature quantity is the number of killer T cells. The type quantity is a quantity that represents the characteristics of the subject's classification, such as the number of cancer cells in the case of a cancer patient. The overall condition index is a numerical value or a numerical range that characterizes the relationship between the feature quantity and the type quantity ratio, and an example of a subject is a range in which the ratio of the feature quantity to the type quantity is based on 1.

[0043] Next, the classification model learned by the learning mechanism 21 of this embodiment will be described. FIG. 7 shows an example of the functional configuration of the learning mechanism 21 in the learning phase. As shown in FIG. 7, the learning mechanism 21 includes a data acquisition unit 210, a data classification unit 211, and a learning unit 212. The data acquisition unit 210 has the function of acquiring basic data 31 and limba cell subgroup data 32 from the memory unit 3. The data classification unit 211 has the function of extracting types and features corresponding to the basic data 31 and limba cell subgroup data 32 acquired by the data acquisition unit 210 from the classification extraction table 33. The learning unit 212 has the function of performing machine learning based on the basic data 31 and limba cell subgroup data 32 acquired by the data acquisition unit 210 and the type and feature extracted by the data classification unit 211, and the function of forming a classification model 600 and storing it in the classification model box 34.

[0044] In addition, in this embodiment, "learning" the classification model refers to learning the classification model 600 using training data that combines the basic data 31 and the limbic cell subpopulation data 32. In this embodiment, learning the classification model 600 refers to generating the classification model by learning using training data based on a machine learning model.

[0045] FIG. 8 shows one embodiment of the machine learning model of this embodiment. The data acquisition unit 210 acquires basic data 31 or the basic data 31 and limba cell subgroup data 32 from the storage unit 3, and classifies and combines the basic data 31 or the basic data 31 and limba cell subgroup data 32 according to the characteristics and properties of the basic data 31, to form classification data. The data classification unit 211 extracts corresponding elements in the classification extraction table 33 as basic features 341 (an example of clinical basic features) based on the acquired basic data 31 and limba cell subgroup data 32. FIG. 9 shows an example of training data. In this embodiment, as an example, as shown in FIG. 9, the data classification unit 211 associates a predetermined ID with the training data. For example, if the ID associated with the training data is "00001" to "09999," the data classification unit 211 forms the training data from a combination of basic data 31 and limba cell subgroup data 32 associated with an ID within the range of "00001" to "09999." 9 , a combination of basic data 31 including type and diagnosis data, basic features 341, and lymphocyte subgroup data 32, each having an ID of "00001," is formed as training data. The data classifying unit 211 vectorizes the extracted basic features 341 using weights assigned to the basic features 341. The data classifying unit 211 outputs the vectorized basic features 341, the basic data 31 acquired from the storage unit 3, and the lymphocyte subgroup data 32 to the learning unit 212. The data classifying unit 211 may also output the basic features 341 and the basic data 31 acquired from the storage unit 3 as basic elements 38 (an example of clinical basic features) and the lymphocyte subgroup data 32 as immune dynamic features 39 (an example of immune subgroup dynamic information) to the learning unit 212.

[0046] 9, for example, a combination of characterized basic data 31, basic features 341, and limb cell subgroup data 32 relating to a healthy individual is generated as training data with an ID of "00001." A combination of characterized basic data 31, basic features 341, and limb cell subgroup data 32 relating to an influenza patient is generated as training data with an ID of "00002."

[0047] The learning unit 212 learns the classification model 600 based on the basic data 31, basic features 341, and limba cell subpopulation data 32, which are learning data input from the data classification unit 211. The learning unit 212 of this embodiment receives the basic data 31, basic features 341, and limba cell subpopulation data 32 as input, and learns the classification model 600 so that the output ratio becomes the integrated therapeutic index. Note that any machine learning method may be used as long as it can perform classification, such as well-known Support Vector Machine or Random Forest. The learning unit 212 stores the learned classification model 600 in the classification model box 34.

[0048] When the learning mechanism 21 learns the classification model 600, the immune dynamics processing unit 2 executes a learning program, thereby performing the learning process shown in Fig. 8. The learning process shown in Fig. 8 is the learning model of this embodiment. The learning process is performed, for example, at a predetermined timing for performing learning, or when a user issues an instruction to start execution via the information input unit 4 or when other conditions are met.

[0049] 8, the data acquisition unit 210 acquires the basic data 31, or the basic data 31 and the limbic cell subpopulation data 32, from the storage unit 3, as described above, and forms classification data. An example of the formation of classification data will be introduced below.

[0050] As an example, the basic data 31 includes at least the white blood cell count, and the quantities and proportions of basophils, eosinophils, neutrophils, lymphocytes, monocytes, CD3 antigens, CD56 antigens (and / or CD16 antigens), CD4 antigens, and CD8 antigens among the white blood cells.

[0051] Specifically, the basic data 31 includes a list of test results of a blood count and a white blood cell image of blood components of a typical subject, i.e., blood test information such as the white blood cell count, the percentage of basophils in white blood cells, the percentage of eosinophils, the percentage of neutrophils, the percentage of lymphocytes, and the percentage of monocytes, as shown in Fig. 6. In Fig. 6, "WBC" is the number of white blood cells in 1 µL, and in the items of the white blood cell image (white blood cell differential information), "Baso" is the percentage of basophils in white blood cells, "Eosinophil" is the percentage of eosinophils, "Neutro" is the percentage of neutrophils, "Lympho" is the percentage of lymphocytes, and "Mono" is the percentage of monocytes.

[0052] The basic data 31 also includes analysis results of the CD numbers of surface antigens of white blood cells, such as CD3 antigen, CD56 antigen (and / or CD16 antigen), CD4 antigen, and CD8 antigen (hereinafter sometimes referred to as analysis information of CD classification).

[0053] The lymphocyte cell subpopulation data 32 includes the number of B cells, the number of NK cells, the number of NKT cells, and the numbers and proportions of immune lymphocyte subpopulations including at least the number of helper T cells, the number of killer T cells, the number of αβ killer-T cells, the number of γδT cells, the number of NK cells, the number of NKT cells, and the number of B cells.

[0054] The limba cell subpopulation data 32 is acquired by either externally acquiring the data or by calculating the data acquiring unit 210. When the data acquiring unit 210 calculates the limba cell subpopulation data 32, the data is calculated based on the following steps.

[0055] In the first step, the total number of white blood cells (immune cells) and the total number (count) of each white blood cell subpopulation (immune cell subpopulation) in the subject's body are calculated based on information related to the white blood cell count, the percentage of basophils (%), eosinophils (%), neutrophils (%), lymphocytes (%), and monocytes (%) among white blood cells.

[0056] Specifically, the total number of white blood cells (immune cells) and white blood cell subpopulations (immune cell subpopulations: granulocytes (eosinophils, basophils, neutrophils), monocytes, lymphocytes (helper T cells, killer T cells, B cells, NK cells, NKT cells)) are calculated using a high-speed calculation program loaded into the data acquisition unit 210. The high-speed calculation program includes at least the following calculation formulas (1) to (5). The total number of white blood cells is calculated using calculation formula (1) if the subject is male, and using calculation formula (2) if the subject is female. (Formula 1)

[0057] Next, the data acquisition unit 210 detects the total number of T cells in the lymphocyte subpopulation (lymphocyte subpopulation) based on the detected total number of lymphocytes and the analysis information of the CD classification.

[0058] Specifically, the total number of T cells in the lymphocyte subpopulation is calculated using a high-speed calculation program incorporated into the data acquisition unit 210, which includes the following calculation formula (6).

[0059] Next, the data acquisition unit 210 calculates the total number of helper T cells and the total number of killer T cells based on the calculated total number of T cells and the analysis information of the CD classification.

[0060] The total number of helper T cells and the total number of killer T cells are calculated using a high-speed calculation program incorporated into the data acquisition unit 210, which includes the following calculation formulas (7) and (8).

[0061] Next, the data acquisition unit 210 detects the total number of non-T cells other than T cells based on the calculated total number of T cells and total number of lymphocytes and the analysis information of the CD classification.

[0062] The total number of non-T cells is calculated using a high-speed calculation program incorporated into the data acquisition unit 210 and including the following calculation formula (9).

[0063] Next, the data acquisition unit 210 detects the total number of B cells, the total number of NK cells, and the total number of NKT cells based on the calculated total number of non-T cells and the total number of T cells and the analysis information of the CD classification.

[0064] The total number of B cells, the total number of NK cells, and the total number of NKT cells are calculated using a high-speed calculation program incorporated into the data acquisition unit 210 and including the following calculation formulas (10), (11), and (12). (Formula 2)

[0065] The number of each component in 1 μl can be calculated using the above formulas (1) to (12). The data acquisition unit 210 stores the calculated data in the storage unit 3.

[0066] In addition, the data acquisition unit 210 outputs the lymphocyte cell subgroup data 32 and basic data 31, which include the number of B cells, the number of NK cells, the number of NKT cells, and the numbers and proportions of immune lymphocyte subgroups including at least the number of helper T cells, the number of killer T cells, the number of αβ killer-T cells, the number of γδT cells, the number of NK cells, the number of NKT cells, and the number of B cells, to the data classification unit 211 as classification data.

[0067] In the next step S2, the data classifying unit 211, as described above, refers to the classification extraction table 33 and extracts the learning type and basic feature 341 corresponding to the classification data acquired from the data acquiring unit 210. The data classifying unit 211 outputs the acquired classification data, the extracted learning type, and basic feature 341 to the learning unit 212 as learning data. That is, the data classifying unit 211 outputs the basic feature 341 and the basic data 31 acquired from the memory unit 3 as basic elements 38, and the limba cell subgroup data 32 as immune dynamics features 39 to the learning unit 212. The basic feature 341 is shown as 34 points or more in the basic feature data of FIG. 4, but the minimum score for the basic feature 341 is 34 points, and points are added depending on the specific items in the basic data 31 and limba cell subgroup data 32. 11, the total of the highest scores for the main items is 34, so the minimum score for the basic feature 341 is assumed to be 34. In the present invention, the learning mechanism 21 can determine the minimum score for the basic feature 341 depending on the learning situation. The classification extraction table 33 of this embodiment is merely one example of the present invention, and the learning mechanism 21 may form the classification extraction table 33 by learning depending on the immune situation.

[0068] In the next step S3, the learning unit 212 learns the classification model 600 or the basic model 601 based on the above-mentioned learning data, and in step S4, determines whether the classification model 600 or the basic model 601 has been trained.

[0069] Here, the learning of the classification model 600 involves performing calculations for adjusting and comparing related elements, with the basic feature 341 as the numerator and the immune dynamic feature 39 as the denominator, so that the ratio between the two is 1. The adjustment and comparison of related elements may involve adjusting weights, which will be described later, to calculate associations.

[0070] In step S4, if the learned result satisfies the reference range, i.e., "Yes" in FIG. 8 , the learning unit 212 determines to end the learning process and forms the classification model 600. In step S4, if the learned result does not satisfy the reference range, i.e., "No" in FIG. 9 , the learning unit 212 returns to step S1 and performs the learning process again. As an example, the learning unit 212 of this embodiment adopts a predetermined learning process termination condition when the classification model 600 has been trained using a predetermined amount of training data, specifically, when the processes of steps S1 to S3 have been performed a predetermined number of times. Until the predetermined reference range is satisfied, the determination in step S4 is negative, and the process returns to step S1, and the processes of steps S1 to S3 are repeated. On the other hand, if the predetermined learning process termination condition is satisfied, the determination in step S4 is positive. If the determination in step S4 is positive, the learning process shown in FIG. 8 ends.

[0071] In this way, in the learning phase, the learning mechanism 21 executes the learning process shown in FIG. 8, and the learned classification model 600 is stored in the classification model box 34.

[0072] Next, the evaluation phase of this embodiment will be described. Fig. 10 shows an example of the functional configuration of the immunodynamic evaluation unit 22 in the evaluation phase. When evaluating input data using the classification model 600, the CPU of the immunodynamic evaluation unit 22 executes the evaluation program, thereby functioning as a data acquisition unit 210 and a data classification unit 211.

[0073] When the immunodynamic evaluation unit 22 performs an immunodynamic evaluation process (hereinafter, sometimes simply referred to as an evaluation process), the learning mechanism 21 executes an evaluation program, thereby performing the immunodynamic evaluation process shown in Fig. 10. The process shown in Fig. 10 is the immunodynamic evaluation process of this embodiment. The immunodynamic evaluation process is performed, for example, at a predetermined timing for performing the evaluation process, or when a user issues an instruction to start execution via the information input unit 4 or when other conditions are met.

[0074] 10 , the data acquiring unit 210 may acquire the basic data 31 related to the classification model 600 from the storage unit 3 as the evaluation basic data 41. In this case, the data acquiring unit 210 may output the basic data 31 related to the classification model 600 to the data classification unit 211 as the evaluation basic data 41.

[0075] 10 , the data acquisition unit 210 acquires the evaluation basic data 41 of the user who will actually perform the immunodynamic evaluation from the storage unit 3 or the information input unit 4. In this embodiment, the user of the classification model 600 may acquire, as the evaluation basic data 41, basic data 31 relating to the results of a new blood test or the like conducted after the classification model 600 is formed.

[0076] Next, the data acquisition unit 210 forms evaluation data based on the evaluation basic data 41. As an example of forming evaluation data, the evaluation data is formed in a method similar to the classification data forming method in step S1 of FIG.

[0077] That is, the data acquisition unit 210 forms evaluation lymphocyte subpopulation data 42 including the number of B cells, the number of NK cells, the number of NKT cells, and the numbers and proportions of immune lymphocyte subpopulations including at least the number of helper T cells, the number of killer T cells, the number of αβ killer-T cells, the number of γδT cells, the number of NK cells, the number of NKT cells, and the number of B cells, based on the evaluation basic data 41, in a manner similar to the classification data formation method of step S1 in Figure 8 above. Then, the evaluation lymphocyte subpopulation data 42 and the evaluation basic data 41 are output as evaluation data to the data classification unit 211. Note that, when the data acquisition unit 210 acquires the evaluation lymphocyte subpopulation data 42 simultaneously with the evaluation basic data 41 in the initial stage of data acquisition, the data acquisition unit 210 may directly output the evaluation lymphocyte subpopulation data 42 and the evaluation basic data 41 to the data classification unit 211 as evaluation data.

[0078] In step S11, the data classification unit 211 acquires a classification model 600 of the same user from the classification model box 34 as a basic model 601 based on the evaluation basic data 41 and the evaluation limbus cell subpopulation data 42. Then, the data classification unit 211 outputs the basic model 601, the evaluation basic data 41, and the evaluation limbus cell subpopulation data 42 to the learning unit 212. If there is no classification model 600 of the same user, a reference model 602 is acquired as the basic model 601. Here, the reference model 602 is a classification model formed by the learning mechanism 21 based on the average data of a large number of normal healthy individuals. The learning mechanism 21 of the present invention can form the reference model 602 through the above-mentioned learning phase. Furthermore, the learning mechanism 21 can form the reference model 602 for each disease category shown in FIG. 4 based on an embodiment of the present invention.

[0079] In the next step S12, the learning unit 212 compares the overall relationship regarding immune dynamics between the evaluation basic data 41 and the evaluation lymphocyte subpopulation data 42 based on the basic model 601 or the reference model 602.

[0080] As an example, a comparison of the overall relationship regarding immune dynamics is performed by comparing the basic data 31 and limbal cell subgroup data 32 related to the basic model 601 with the evaluation basic data 41 and evaluation limbal cell subgroup data 42 to determine whether or not the specified range is met.

[0081] That is, the learning unit 212 compares the basic data 31 and limba cell subpopulation data 32 (hereinafter sometimes referred to as basic immune dynamics data) relating to the above-mentioned basic model 601 with the evaluation basic data 41 and evaluation limba cell subpopulation data 42 (hereinafter sometimes referred to as evaluation immune dynamics data), and if it determines that the specified range is satisfied, i.e., in the case of "Yes" in FIG. 10, it determines to end the comparison process. If the learning unit 212 compares the basic data 31 and limba cell subpopulation data 32 relating to the above-mentioned basic model 601 with the evaluation basic data 41 and evaluation limba cell subpopulation data 42 and determines that the specified range is not satisfied, i.e., in the case of "No" in FIG. 10, it returns to step S10 and performs the comparison process again. As an example, the learning unit 212 of this embodiment may adopt, as the predetermined comparison process end condition, when the comparison process is performed using a predetermined number of evaluation data, specifically, when the processes of steps S10 to S12 have been performed a predetermined number of times. Until the specified range is satisfied, the determination in step S12 becomes negative, and the process returns to step S10, and the processes in steps S10 to S12 are repeated. On the other hand, if a predetermined comparison process termination condition is satisfied, the determination in step S12 becomes positive, and the calculations in steps S1 to S4 of the learning phase are performed to calculate a classification model 600 related to the evaluation basic data 41 and the evaluation limba cell subpopulation data 42. Then, the calculated classification model 600 is used as an evaluation model 603, and the process proceeds to step 13. In other words, if the determination in step S12 becomes positive, the calculation results of the ratios of the basic features and the immunodynamic features obtained based on the evaluation basic data 41 and the evaluation limba cell subpopulation data 42 form the evaluation model 603 (hereinafter, may be referred to as immunodynamic evaluation data).

[0082] On the other hand, if it is determined in step S12 that the specified range is not ultimately satisfied, the calculations in steps S1 to S4 of the learning phase are performed to calculate a new classification model 600. In this case, the newly calculated classification model 600 is used as an evaluation model 603, and the process proceeds to step S13. Then, in step S13, immunodynamic evaluation is performed using the reference model 602 as the basic model 601 and the newly calculated classification model 600 as the evaluation model 603.

[0083] In step S13, the learning unit 212 compares the evaluation model 603 with the basic model 601. If the evaluation model 603 is smaller than the calculation result of the basic model 601, the learning unit 212 evaluates that the immune dynamics have improved. If the evaluation model 603 is larger than the calculation result of the basic model 601, the learning unit 212 evaluates that the immune dynamics have worsened. If the evaluation model 603 is equal to the calculation result of the basic model 601, the learning unit 212 evaluates that there is no change in the immune dynamics. The learning unit 212 stores the comparison result between the calculation result of the evaluation model 603 and the basic model 601 in the classification model box 34, and at the same time displays the comparison result on the display unit 5, thereby ending the immune dynamics evaluation process.

[0084] 10, the evaluation of lymphocyte subpopulation data 42 is calculated in step S10 and sequentially acquired, but the present invention is not limited to this. For example, the evaluation of lymphocyte subpopulation data 42 may be acquired by directly inputting it by the user in data acquisition step S10.

[0085] In this way, in the evaluation phase, the learning mechanism 21 executes the immune dynamics evaluation process shown in Figure 10, whereby a comparative evaluation of the evaluation basic data 41 and the evaluation limbic cell subgroup data 42 using the basic model 601 is performed, and the comparison results are associated with the basic model 601 and stored in the classification model box 34, and at the same time, are displayed on the display unit 5.

[0086] The immune dynamics generation device 1 of the present invention comprises at least one immune dynamics processing unit 2, a memory unit 3 for storing at least clinical data, and a display unit 5. The immune dynamics processing unit 2 acquires basic clinical features and immune subgroup dynamics information based on the clinical data, and performs self-learning based on the basic clinical features and immune subgroup dynamics information to form and monitor a classification model. Based on the classification model, the immune dynamics information of the subject is evaluated by comparing the overall relationship between the basic clinical features and the immune subgroup dynamics information, and the evaluation result of the immune dynamics information is displayed on the display unit 5.

[0087] Furthermore, the immune dynamics generating device 1 of the present invention can acquire at least normal immune status information, abnormal immune status information, and cancer incidence rate information through the above-mentioned functions, thereby establishing an integrated judgment standard that can objectively evaluate immune dynamics, and providing an immune dynamics generating method and device that can instantly and accurately grasp the state of immunity and the progression of diseases, etc.

[0088] In Example 1, an example of acquiring immune dynamics information related to a normal immune state based on health checkup item data of a normal individual will be described. That is, in Example 1, the function of the immune dynamics processing unit 2 will be described based on the formation of a basic model and immune dynamics evaluation related to a healthy individual.

[0089] In Example 1, first, as a process corresponding to step S1, the data acquisition unit 210 acquires basic data 31 of the subject (user) stored in the storage unit 3. Then, based on the acquired basic data 31, in accordance with the process of step S1 described above, the data acquisition unit 210 calculates lymphocyte subpopulation data 32 including the number of B cells, the number of NK cells, the number of NKT cells, and the numbers and proportions of immune lymphocyte subpopulations including at least the number of helper T cells, the number of killer T cells, the number of αβ killer-T cells, the number of γδ T cells, the number of NK cells, the number of NKT cells, and the number of B cells, related to the basic data 31. Note that if the lymphocyte subpopulation data 32 is stored in the storage unit 3, or if the user inputs the lymphocyte subpopulation data 32 via the information input unit 4, the above calculation processing operation may not be necessary. Then, the data acquisition unit 210 outputs the lymphocyte subpopulation data 32 and the basic data 31 to the data classification unit 211 as classification data.

[0090] FIG. 11 is a diagram showing an example of the immune dynamics table 35. FIG. 12 is a diagram showing basic data 31 of Example 1. FIG. 13 is a diagram showing the number of each component of Example 1 calculated based on the above-mentioned Equations 1 and 2. FIG. 14 is a diagram showing the number of granulocytes, the number of monocytes, the number of lymphocytes, the number of helper T cells, the number of killer T cells, the number of αβ killer T cells, the number of γδ T cells, the number of NK cells, the number of NKT cells, and the number of B cells for classification data of Example 1. FIG. 15 is a diagram showing the proportions of granulocytes, monocytes, and lymphocytes of Example 1. FIG. 16 is a diagram showing the proportions of helper T cells, killer T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells, and B cells of Example 1.

[0091] The data acquiring unit 210 acquires the basic data 31 shown in Fig. 12 from the storage unit 3. Alternatively, the subject (user) may directly input the basic data 31 to the data acquiring unit 210 using the information input unit 4. The data acquiring unit 210 calculates the number of each component in the blood as shown in Fig. 13 using Equations 1 and 2 based on the acquired basic data 31 in accordance with the processing of step S1 described above.

[0092] Next, the data acquisition unit 210 can obtain the lymphocyte subpopulation data 32 of the present invention by readjusting the number of each component in the blood shown in Fig. 13. In Example 1, the component corresponding to the lymphocyte subpopulation data 32 among the components shown in Fig. 13 is multiplied by 15.5, which is the hemoglobin (Hb) value of the subject, to calculate the numbers of granulocytes, monocytes, lymphocytes, helper T cells, killer T cells, αβ killer-T cells, γδ T cells, NK cells, NKT cells, and B cells used in the classification data, as shown in Fig. 14.

[0093] Then, based on the numbers of granulocytes, monocytes, lymphocytes, helper T cells, killer T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells and B cells used in the classification data shown in FIG. 14 , the proportions of granulocytes, monocytes and lymphocytes of Example 1 shown in FIG. 15 and the proportions of helper T cells, killer T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells and B cells of Example 1 shown in FIG. 16 are calculated.

[0094] The lymphocyte subgroup data 32 of Example 1 is composed of the numbers of granulocytes, monocytes, lymphocytes, helper T cells, killer T cells, αβ killer-T cells, γδ T cells, NK cells, NKT cells, and B cells used in the classification data shown in Figure 14; the proportions of granulocytes, monocytes, and lymphocytes of Example 1 shown in Figure 15; and the proportions of helper T cells, killer T cells, αβ killer-T cells, γδ T cells, NK cells, NKT cells, and B cells of Example 1 shown in Figure 16.

[0095] The data acquisition unit 210 outputs the limbic cell subgroup data 32 and the basic data 31 to the data classification unit 211 as classification data.

[0096] 4 in accordance with step S2 described above, the data classification unit 211 refers to the classification extraction table 33 as shown in Fig. 4 and extracts the learning type and basic feature 341 corresponding to the classification data acquired from the data acquisition unit 210. The data classification unit 211 outputs the acquired classification data and the extracted learning type and basic feature 341 to the learning unit 212 as learning data.

[0097] Specifically, the data classification unit 211 determines the learning type and basic feature 341 based on the subject's age, sex, health condition, type of diagnostic data, and characteristic immune elements. The subject in Example 1 is a healthy man in his 50s, with diagnostic data from a blood test and a CD4 count of 400 or higher, which is similar to the condition of a healthy person in the classification extraction table 33. Therefore, the data classification unit 211 extracts the learning type of the subject in Example 1 as healthy person and 34 points as the basic feature 341 as part of the learning data.

[0098] The data classification unit 211 outputs the limbic cell subgroup data 32, the basic data 31, the healthy individual as the learning type, and 34 points as the basic features 341 to the learning unit 212 as learning data.

[0099] The learning unit 212 learns the classification model 600 based on the learning data in accordance with the above-mentioned steps S3 and S4, and then judges the trained classification model 600 based on the evaluation phase of FIG.

[0100] The learning unit 212 first learns the classification model 600. In the first embodiment, the learning unit 212 performs calculations for adjusting and comparing the weights (basic feature weight 381, immune dynamics weight 391) using the basic feature 341 as the numerator and the immune dynamics feature 39 as the denominator so that the ratio between the two is 1.

[0101] Specifically, the learning unit 212 can use 34 points included in the training data as the basic features 341. The learning unit 212 forms basic feature weights 381, immune dynamic features 39, and immune dynamic weights 391 by self-learning based on the basic data 31 and the lymphocyte subpopulation data 32 included in the training data and with reference to the immune dynamics table 35 of Fig. 11 .

[0102] Fig. 11 shows an example of the immune dynamics table 35. The immune dynamics table 35 of the present invention includes immune dynamics main items and immune dynamics sub-items. The main items shown in Fig. 11 are examples of immune dynamics main items, and the sub-items shown in Fig. 11 are examples of immune dynamics sub-items.

[0103] The main immune dynamics items include the number of helper T cells, the ratio of granulocytes:monocytes:lymphocytic cells, the ratio of activated T cells (HLA-DR / CD38%), Th1 immune strength, and other immune cell indicators.

[0104] In this embodiment, the number of helper T cells is divided into six levels, from level 1 to level 6, and classified by evaluation score. For example, when the number of helper T cells is 500 cells / μl or more, it is evaluated as level 6, the highest score of 12 points. When the number of helper T cells is 400 to 500 cells / μl, it is evaluated as level 5, with 10 points. When the number of helper T cells is 300 to 400 cells / μl, it is evaluated as level 4, with 8 points. When the number of helper T cells is 200 to 300 cells / μl, it is evaluated as level 3, with 4 points. When the number of helper T cells is 100 to 200 cells / μl, it is evaluated as level 2, with 2 points. When the number of helper T cells is less than 100 cells / μl, it is evaluated as level 1, with 0 points.

[0105] The most ideal ratio of granulocytes to monocytes to lymphocytes is 60:5:35. Therefore, if the ratio of granulocytes to monocytes to lymphocytes is approximately 60-65:5-10:30 or more, the highest score of 8 is given. If the ratio of granulocytes to monocytes to lymphocytes is approximately 65-75:5-10:20-30, the score is 6. If the ratio of granulocytes to monocytes to lymphocytes is approximately 65-85:5-10:10-20, the score is 4. If the ratio of granulocytes to monocytes to lymphocytes is approximately 85 or more:5-10:10 or less, the score is 0.

[0106] The proportion of activated T cells (HLA-DR / CD38%) is scored as 6 points if it is 15% or more, 4 points if it is 10-15%, 2 points if it is 5-10%, and 0 points if it is less than 5%.

[0107] The Th1 immune strength is evaluated as 4 points if it is greater than 1, 2 points if it is 1, and 0 points if it is less than 1. Similarly, other immune cell indicators can be classified into levels in a similar manner, evaluated by scores, and added to the immune dynamics table 35.

[0108] The immunodynamic subitems include indicators that affect immunodynamics, such as Alb level, ChE level, Hb level, Ccr level, hyaluronic acid level, presence or absence of liver dysfunction, changes in tumor markers, and changes in HLA-DR versus CD38.

[0109] The Alb value can be scored as 2 if it is greater than 3.8 g / dl, 1 if it is between 3.2 and 3.8 g / dl, and 0 if it is less than 3.2 g / dl. The ChE value can be scored as 2 if it is greater than 300 U / l, 1 if it is between 250 and 300 U / l, and 0 if it is less than 250 U / l. The Hb value can be scored as 2 if it is greater than 10 g / dl, 1 if it is between 8 and 10 g / dl, and 0 if it is less than 8 g / dl. The Ccr value can be scored as 2 if it is greater than 80 ml / m, 1 if it is between 50 and 80 ml / m, and 0 if it is less than 50 ml / m. Other values, such as hyaluronic acid, can also be included. For example, the hyaluronic acid value can be scored as 2 if it is less than 50 ng / ml, and 0 if it is greater than 50 ng / ml. Liver dysfunction can also be scored. If there is liver dysfunction, the score can be 2, and if there is no liver dysfunction, the score can be 0.

[0110] Furthermore, depending on the state of immune dynamics evaluation, immune dynamics main items can be used as required items and sub-items as optional items. The learning unit 212 of this embodiment forms basic features 341 and immune dynamics features 39 by self-learning with reference to required items corresponding to the immune dynamics table 35 based on learning data including the basic data 31 and the lymphocyte subpopulation data 32 acquired by the data acquisition unit 210.

[0111] For example, in Example 1, the learning unit 212 determines that the helper T cells are 492 / μL, which corresponds to Level 5 of the immune dynamics table 35, and can be evaluated as 10 points. Next, the learning unit 212 compares the ratios of granulocytes, monocytes, and lymphocytes shown in Figure 15 with the ideal value of 60:5:35%. As shown in Figure 15, in Example 1, the ratios of granulocytes, monocytes, and lymphocytes are 56:8:37%, which corresponds to Level 4 of the immune dynamics table 35, and can be evaluated as 8 points.

[0112] In Example 1, the helper T cells are 492 / μL, which corresponds to level 5 in the immune dynamics table 35 and can be evaluated as 10 points, so the Th1 immune strength can be assumed to be greater than 1. Therefore, the learning unit 212 determines that the Th1 immune strength is greater than 1, which corresponds to level 3 in the immune dynamics table 35 and can be evaluated as 4 points.

[0113] In Example 1, the helper T cells are 492 / μL, which corresponds to level 5 in the immune dynamics table 35 and can be evaluated as 10 points, so the activated T cell proportion can be assumed to be greater than 15%. Therefore, the learning unit 212 determines that the activated T cell proportion is 15%, which corresponds to level 4 in the immune dynamics table 35 and can be evaluated as 6 points.

[0114] In Example 1, the helper T cells were 492 / μL, which corresponds to level 5 of the immune dynamics table 35 and can be evaluated as 10 points, so it can be assumed that the proportions of the core elements of the limbic cell subpopulation data 32 are normal, as shown in Figure 16. Therefore, the proportions of the core elements of the limbic cell subpopulation data 32, namely, helper T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells, and B cells, can be determined to be level 3 of the other immune cell indicators in the immune dynamics table 35 and can be evaluated as 4 points.

[0115] That is, when the helper T cells are at level 5 or higher in the immune dynamics table 35, the evaluation of the activated T cell proportion, Th1 immune strength, and other immune cell indices can be considered to be at the highest level for each index and can be used in the learning calculation. In this case, sub-items do not need to be taken into account.

[0116] Alternatively, if the levels of all three items (helper T cells; granulocyte, monocyte, and lymphocyte ratios; and activated T cell proportion) are level 4 or higher, the evaluation of Th1 immune strength and other immune cell indicators will be considered to be the highest level of each indicator and used in the learning calculation. In this case, sub-items do not need to be taken into account.

[0117] The learning unit 212 forms basic features 341 by referring to the basic elements 38 and immune dynamic features 39. The subject in Example 1 is a healthy man in his 50s, and the diagnostic data is from a blood test with a CD4 count of 400 or more, which is similar to the situation of a healthy person in the classification extraction table 33. The data classification unit 211 extracts the learning type of the subject in Example 1 as healthy person and a score of 34 for basic features 341 as part of the learning data. In other words, if the subject is healthy, the sub-items in the immune dynamics table 35 do not need to be taken into consideration, so 34 is taken as the minimum score.

[0118] In this embodiment, the learning unit 212 may use the basic feature weight 381 as the numerator and / or the immune dynamic weight 391 as the denominator when learning to make the ratio of the basic feature 341 as the numerator and the immune dynamic weight 391 as the denominator so that the ratio becomes 1. In addition, in one learning calculation, either the basic feature weight 381 or the immune dynamic weight 391 may be used, or both may be used simultaneously.

[0119] In this embodiment, when the helper T cells are at level 5 or higher in the immune dynamics table 35, the learning unit 212 can select the immune dynamics weight 391 from 4 points to 22 points and perform self-learning. That is, when the subject type is healthy and the helper T cells are at level 5 or higher in the immune dynamics table 35, the learning unit 212 can consider the evaluations of the granulocyte, monocyte, and lymphocyte percentages, the activated T cell proportion, Th1 immune strength, and other immune cell indices to be at the highest level for each index, and can therefore determine the immune dynamics weight 391 within the range of the total score for the highest level. If actual data exists for any of the subject's granulocyte, monocyte, and lymphocyte percentages, the activated T cell proportion, Th1 immune strength, and other immune cell indices, that actual data may be used.

[0120] That is, if the helper T cell is at level 5 or above in the immune dynamics table 35 and the immune dynamics main item has only 10 points, the learning unit 212 can select 22 points as the immune dynamics weight 391 .

[0121] If the helper T cells are at level 5 or higher in the immune dynamics table 35, the proportions of granulocytes, monocytes, and lymphocytes are at level 4 in the immune dynamics table 35, and the immune dynamics main item is 18 points, the learning unit 212 can select 14 points as the immune dynamics weight 391. Also, if the helper T cells are at level 5 or higher in the immune dynamics table 35, the proportions of granulocytes, monocytes, and lymphocytes are at level 3 in the immune dynamics table 35, and the immune dynamics main item is 16 points, similarly to the above, the learning unit 212 can select 14 points as the immune dynamics weight 391.

[0122] That is, in the immune dynamics table 35, if the score extracted by the learning unit 212 as the immune dynamics main item is greater than the score corresponding to the specified level of each item, the score corresponding to the highest level of the next item can be excluded from the selection range score of the immune dynamics weight 391. The specified level here refers to level 5 of the immune dynamics table 35 in the case of helper T cells, and can be specified as the highest level of each item in the case of other items.

[0123] In this embodiment, when the helper T cell is at level 5 or higher in the immune dynamics table 35 , the learning unit 212 uses only the immune dynamics weight 391 and does not need to use the basic feature weight 381 .

[0124] In this embodiment, if the levels of all three items, namely, helper T cells; proportion of granulocytes, monocytes, and lymphocytes; and proportion of activated T cells, are level 4 or higher, the learning unit 212 uses only the immune dynamics weight 391 and does not need to use the basic feature weight 381.

[0125] In the first embodiment, the immune dynamic weight 391 can be selected from 4 to 22 points, so the learning unit 212 can use only the immune dynamic weight 391 and does not need to use the basic feature weight 381 .

[0126] In Example 1, the learning unit 212 selects 34 points for the basic feature 341 as the numerator, 18 points for the immune dynamics main item as the denominator, and selects an appropriate immune dynamics weight 391 from 4 points to 22 points so that the ratio between the two is 1, thereby performing self-learning. When the predetermined reference range in step S4 is satisfied, the classification model 600 is formed. The predetermined reference range in Example 1 is a range for determining whether the value of the immune dynamics weight 391 is maximum in relation to actual test data, etc. In Example 1, it is assumed that there is no actual data on the activated T cell proportion, Th1 immune strength, and other immune cells. Since helper T cells are at level 5 or higher in the immune dynamics table 35, the predetermined range can be set to 14 points, which is the sum of the highest evaluation scores for the activated T cell proportion, Th1 immune strength, and other immune cell indicators.

[0127] That is, in Example 1, it is assumed that the learning unit 212 performs self-learning so that the basic feature 341 has 34 points, and the total of the immune dynamic feature 39 and the immune dynamic weight 391 has 32 points. Specifically, since helper T cells are at level 5 or higher in the immune dynamic table 35, the proportions of granulocytes, monocytes, and lymphocytes are at level 4 in the immune dynamic table 35, and the immune dynamic main item has 18 points, it is considered that the learning unit 212 repeatedly performs S1 to S3 and performs self-learning optimization calculations so that the immune dynamic weight 391 has 14 points. Finally, when the score reaches 14 points, it is considered that the classification model 600 is formed and stored in the classification model box 34.

[0128] 17 , in Example 1, the learning unit 212 performs calculations based on the above elements, resulting in a basic feature 341 with a score of 34, a total of the immune dynamic feature 39 and the immune dynamic weight 391 with a score of 32, and a ratio of 1.06. The learning unit 212 then stores the calculation results shown in FIG. 17 as a classification model 600 (an example of normal immune status information) in the classification model box 34. The learning unit 212 may also store the classification model 600 in the classification model box 34 and simultaneously display it on the display unit 5.

[0129] Next, the evaluation phase of Example 1 is performed. In Example 1, if the memory unit 3 does not contain data related to the immune dynamics of the same user, in step S10 of FIG. 10 , the data acquisition unit 210 acquires the basic data 31 related to the classification model 600 from the memory unit 3 as the evaluation basic data 41. Also, in step S11 of FIG. 10 , since the memory unit 3 does not store a classification model for the same user, the reference model 602 is acquired as the basic model 601. In step S12, the basic immune dynamics data related to the reference model 602 is compared with the evaluation immune dynamics data related to the classification model 600, and if a positive determination is obtained, the classification model 600 may be output as the evaluation model 603. Then, in step S13, the learning unit 212 compares the evaluation model 603 with the reference model 602 to evaluate the immune dynamics. As shown in Figure 18, in the case of Example 1, the learning unit 212 determines that the immune status of the subject is below the normal immune status because the ratio of the evaluation model 603 is 1.06, which is greater than the ratio of the reference model 602, which is 1, and the ratio is greater than 1, and displays this on the display unit 5.

[0130] [Example 2] In Example 2, the function of the immune dynamics processing unit 2 will be described based on the formation of a basic model and immune dynamics evaluation for colorectal cancer patients. That is, in Example 2, an example of acquisition of immune dynamics information related to immune status information of abnormal individuals (abnormal immune status information) will be described. In Example 2, descriptions of parts similar to those in Example 1 will be omitted.

[0131] Fig. 19 is a diagram showing the number of granulocytes, the number of monocytes, the number of lymphocytes, the number of helper T cells, the number of killer T cells, the number of αβ killer T cells, the number of γδ T cells, the number of NK cells, the number of NKT cells, and the number of B cells for the basic data 31 of Example 2 and the classification data of Example 2. Fig. 20 is a diagram showing the proportions of granulocytes, monocytes, and lymphocytes of Example 2. Fig. 21 is a diagram showing the proportions of helper T cells, killer T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells, and B cells of Example 1.

[0132] In Example 2, the same processes as those in Example 1 are omitted here. The process in step S1 in Example 2 was carried out in the same manner as in Example 1.

[0133] The lymphocyte subgroup data 32 of Example 2 is composed of the numbers of granulocytes, monocytes, lymphocytes, helper T cells, killer T cells, αβ killer-T cells, γδ T cells, NK cells, NKT cells, and B cells used in the classification data shown in Figure 19; the proportions of granulocytes, monocytes, and lymphocytes of Example 2 shown in Figure 20; and the proportions of helper T cells, killer T cells, αβ killer-T cells, γδ T cells, NK cells, NKT cells, and B cells.

[0134] The data acquisition unit 210 outputs the above-mentioned lymphocyte subpopulation data 32 and the basic data 31 of Example 2 (basic data including cancer cell data) to the data classification unit 211 as classification data.

[0135] 4 in accordance with step S2 described above, the data classification unit 211 refers to the classification extraction table 33 as shown in Fig. 4 and extracts the learning type and basic feature 341 corresponding to the classification data acquired from the data acquisition unit 210. The data classification unit 211 outputs the acquired classification data and the extracted learning type and basic feature 341 to the learning unit 212 as learning data.

[0136] Specifically, the data classification unit 211 determines the learning type and basic feature 341 based on the subject's age, sex, health condition, type of diagnostic data, and characteristic immune elements. The subject in Example 2 is a man in his 50s with colon cancer, and the diagnostic data is from a blood test, and since the CD4 value is smaller than the CD8 value, the condition is similar to that of colon cancer in the classification extraction table 33. Therefore, the data classification unit 211 extracts colon cancer as the learning type and 34 points as the basic feature 341 as part of the learning data for the subject in Example 2.

[0137] The data classification unit 211 outputs the above-mentioned lymphocyte subpopulation data 32, basic data 31, colon cancer patient as a learning type, and 34 points as basic features 341 to the learning unit 212 as learning data.

[0138] The learning unit 212 learns the classification model 600 based on the learning data in accordance with the above-mentioned steps S3 and S4, and determines the trained classification model 600.

[0139] The learning unit 212 first learns the classification model 600. In the second embodiment, the learning unit 212 performs calculations for adjusting and comparing related elements, using the basic feature 341 as the numerator and the immune dynamic feature 39 as the denominator, so that the ratio between the two is 1.

[0140] Specifically, the learning unit 212 uses 34 points included in the training data as basic features 341. The learning unit 212 forms basic feature weights 381, immune dynamic features 39, and immune dynamic weights 391 by self-learning based on the basic data 31 and the lymphocyte subpopulation data 32 included in the training data, with reference to the immune dynamics table 35 in Fig. 11 .

[0141] In Example 2, the learning unit 212 determines that the helper T cells are 127 / μL, which corresponds to Level 2 of the immune dynamics table 35, and can be evaluated as 2 points. Next, the learning unit 212 compares the ratio of granulocytes, monocytes, and lymphocytes shown in Figure 20 with the ideal value of 60:5:35%. As shown in Figure 20, in Example 2, the ratio of granulocytes, monocytes, and lymphocytes is 68:9:23%, which corresponds to Level 3 of the immune dynamics table 35, and can be evaluated as 6 points.

[0142] In Example 2, the helper T cells are 127 / μL, which corresponds to level 2 in the immune dynamics table 35 and can be evaluated as 2 points, so it can be assumed that the Th1 immune strength is less than 1. Therefore, the learning unit 212 determines that the Th1 immune strength is less than 1, which corresponds to level 1 in the immune dynamics table 35 and can be evaluated as 0 points.

[0143] In Example 2, the helper T cells are 127 / μL, which corresponds to Level 2 of the immune dynamics table 35 and can be evaluated as 2 points, so the activated T cell proportion can be assumed to be less than 5%. Therefore, the learning unit 212 determines that the activated T cell proportion is less than 5%, which corresponds to Level 1 of the immune dynamics table 35 and can be evaluated as 0 points.

[0144] In Example 2, the helper T cells were 127 / μL, which corresponds to Level 2 of the immune dynamics table 35 and can be evaluated as 2 points. Therefore, it can be assumed that the proportions of the core elements of the limbic cell subpopulation data 32 are abnormal, as shown in Figure 20. Therefore, the proportions of the core elements of the limbic cell subpopulation data 32, namely, helper T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells, and B cells, can be determined to be Level 1 of the other immune cell indicators in the immune dynamics table 35 and can be evaluated as 0 points.

[0145] That is, when the helper T cells are at level 2 or lower in the immune dynamics table 35, the evaluation of the activated T cell proportion, Th1 immune strength, and other immune cell indices can be considered to be at the lowest level for each index and can be used in the learning calculation. However, if actual values ​​for the activated T cell proportion, Th1 immune strength, and other immune cell indices are available, those actual values ​​may be used.

[0146] Alternatively, if the levels of all three items, helper T cells; granulocyte, monocyte, and lymphocyte ratios; and activated T cell proportion, are level 2 or lower, the evaluation of Th1 immune strength and other immune cell indices will be considered to be at the lowest level for each indicator and used in the learning calculation. However, if actual values ​​for activated T cell proportion, Th1 immune strength, and other immune cell indices are available, those actual values ​​may be used.

[0147] Furthermore, in blood tests for colon cancer, indicators that affect immune dynamics, such as changes in tumor markers and changes in HLA-DR versus CD38, are necessary indicators. In this case, indicators that affect immune dynamics, such as changes in tumor markers and changes in HLA-DR versus CD38, can be used for self-learning. In this case, these indicators are taken into consideration when forming the basic feature weight 381 and the immune dynamics weight 391. In Example 2, with reference to FIG. 11 , 6 points may be calculated as the basic feature weight 381, representing the score corresponding to changes in tumor markers and changes in HLA-DR versus CD38.

[0148] In Example 2, the tumor marker CA72-4 (6.9 U / m or less) was 14.9, and the anti-P53 antibody (1.3 U / m or less) was 22.73. Both were greater than the standard values, so they were classified as Level 1 in the immunodynamics table 35, and could be evaluated as 0 points.

[0149] In this embodiment, when the helper T cells are at level 2 or lower in the immune dynamics table 35, the learning unit 212 can select the basic feature weight 381 from 0 to 22 points and perform self-learning to form the basic feature weight 381.

[0150] The learning unit 212 performs self-learning to appropriately form the basic feature weight 381 and the immune dynamics weight 391 based on the actual test data in Example 2 and the immune dynamics table 35 of Figure 11. In Example 2, the sub-items in Figure 11 include test data for three items: tumor marker, HLA-DR(+)CD38, and HLA-DR(-)CD38. Therefore, data for these three items can be taken into consideration in the immune dynamics evaluation. Therefore, the basic feature weight 381 can be set to the sum of the highest scores for the three items: tumor marker, HLA-DR(+)CD38, and HLA-DR(-)CD38, so it is conceivable to select 6 points.

[0151] In Example 2, since the helper T cells are at level 2 or lower in the immune dynamics table 35, the proportions of granulocytes, monocytes, and lymphocytes are at level 3 in the immune dynamics table 35, and the immune dynamics main item is 8 points, it is conceivable that the learning unit 212 selects 0 points as the immune dynamics weight 391.

[0152] As shown in FIG. 22 , in Example 2, the learning unit 212 performs calculations based on the above elements, resulting in a total of 40 points for the basic feature 341 and the basic feature weight 381, and a total of 8 points for the immune dynamic feature 39 and the immune dynamic weight 391n, with a ratio of 5. The learning unit 212 then stores the calculation results shown in FIG. 22 in the classification model box 34 as a classification model 600. The learning unit 212 may also store the classification model 600 in the classification model box 34 and simultaneously display it on the display unit 5. The learning unit 212 performs processing in a manner similar to the processing in the evaluation phase of Example 1. The classification model 600 of Example 2 may be used as an evaluation model 603 and compared with a reference model 602 to acquire cancer incidence rate information. As a result of comparing the evaluation model 603 with the reference model 602, the ratio of the evaluation model 603 is 5, which is greater than the ratio of 1 for the reference model 602, and therefore the immune status of the subject is determined to be at risk, and this is displayed on the display unit 5. Furthermore, the evaluation model 603 of Example 2 may be compared with the classification model 600 of Example 1 (normal immune status information) to obtain cancer incidence rate information. That is, the ratio result of 5 for the evaluation model 603 of Example 2 may be compared with the ratio result of 1.06 for the classification model 600 of Example 1, and the ratio may be obtained as cancer incidence rate information.

[0153] Example 3 describes the function of the immune dynamics processing unit 2 based on the basic model formation and immune dynamics evaluation for the second examination of the same colorectal cancer patient as in Example 2. That is, Example 3 describes an example of acquiring immune dynamics information related to cancer incidence rate information.

[0154] Example 3 will be described along the evaluation phase of Fig. 10. In step S10, the data acquisition unit 210 acquires basic evaluation data 41 of the same user as in Example 2 from the storage unit 3 or the information input unit 4.

[0155] Next, the data acquisition unit 210 forms evaluation data based on the evaluation basic data 41. Hereinafter, as an example of forming evaluation data, evaluation data is formed by a method similar to the classification data formation method in step S1 of Fig. 8 above. The evaluation limba cell subgroup data 42 may be acquired simultaneously with the evaluation basic data 41 from the storage unit 3 or the information input unit 4. Then, the data acquisition unit 210 outputs the evaluation limba cell subgroup data 42 and the evaluation basic data 41 to the data classification unit 211 as evaluation data.

[0156] In step S11, the data classification unit 211 acquires the classification model 600 of the same user from the classification model box 34 as the basic model 601 based on the evaluation basic data 41 and the evaluation limba cell subgroup data 42. Then, the data classification unit 211 outputs the basic model 601, the evaluation basic data 41, and the evaluation limba cell subgroup data 42 to the learning unit 212.

[0157] In the next step S12, the learning unit 212 compares the overall relationship regarding immune dynamics between the evaluation basic data 41 and the evaluation lymphocyte subpopulation data 42 based on the basic model 601 or the reference model 602.

[0158] As an example, a comparison of the overall relationship regarding immune dynamics is performed by comparing the basic data 31 and limbal cell subgroup data 32 related to the basic model 601 with the evaluation basic data 41 and evaluation limbal cell subgroup data 42 to determine whether or not the specified range is met.

[0159] That is, the learning unit 212 compares the basic data 31 and limba cell subgroup data 32 relating to the basic model 601 described above with the evaluation basic data 41 and evaluation limba cell subgroup data 42, and if it determines that the specified range is satisfied, i.e., in the case of "Yes" in Fig. 10, it determines to end the comparison process. If the learning unit 212 compares the basic data 31 and limba cell subgroup data 32 relating to the basic model 601 described above with the evaluation basic data 41 and evaluation limba cell subgroup data 42 and determines that the specified range is not satisfied, i.e., in the case of "No" in Fig. 10, it returns to step S10 and performs the comparison process again. The predetermined comparison process end condition refers, for example, to a case where at least the evaluation type, health test type, etc. of the same user are the same, and the limba cell subgroup data are also within a similar range.

[0160] If a predetermined comparison processing termination condition is met, the determination in step S12 is affirmative, and the calculations in steps S1 to S4 of the learning phase are performed to calculate a classification model 600 related to the evaluation basic data 41 and the evaluation limbic cell subpopulation data 42. Then, the calculated classification model 600 is used as an evaluation model 603, and the process proceeds to step S13. That is, if the determination in step S12 is affirmative, the calculation results of the ratios of the basic features 48 and the immune dynamic features 49 obtained based on the evaluation basic data 41 and the evaluation limbic cell subpopulation data 42 form the evaluation model 603 (hereinafter, sometimes referred to as immune dynamic evaluation data). That is, in Example 3, specifically, the following processing is performed.

[0161] Fig. 23 is a diagram showing the numbers of granulocytes, monocytes, lymphocytes, helper T cells, killer T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells, and B cells for the basic evaluation data 41 of Example 3 and the classification data of Example 3. Fig. 24 is a diagram showing the proportions of granulocytes, monocytes, and lymphocytes of Example 3. Fig. 25 is a diagram showing the proportions of helper T cells, killer T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells, and B cells of Example 3.

[0162] In the third embodiment, the description of the same processes as those in the second embodiment will be omitted. The process of step S1 in the third embodiment was carried out in the same manner as in the first embodiment.

[0163] The evaluation lymphocyte subgroup data 42 of Example 3 is composed of the number of granulocytes, the number of monocytes, the number of lymphocytes, the number of helper T cells, the number of killer T cells, the number of αβ killer-T cells, the number of γδ T cells, the number of NK cells, the number of NKT cells and the number of B cells used in the classification data shown in Figure 23; the proportions of granulocytes, monocytes and lymphocytes of Example 3 shown in Figure 24; and the proportions of helper T cells, killer T cells, αβ killer-T cells, γδ T cells, NK cells, NKT cells and B cells.

[0164] The data acquisition unit 210 outputs the evaluation basic data 41 relating to the second test and the above-mentioned evaluation lymphocyte subpopulation data 42 to the data classification unit 211 as classification data.

[0165] 4 in accordance with step S2 described above, the data classification unit 211 refers to the classification extraction table 33 as shown in Fig. 4 and extracts the learning type and basic feature 341 corresponding to the classification data acquired from the data acquisition unit 210. The data classification unit 211 outputs the acquired classification data and the extracted learning type and basic feature 341 to the learning unit 212 as learning data.

[0166] Since the subject in Example 3 is the same subject as the subject in Example 2, the data classification unit 211 extracts the learning type as colon cancer and the basic feature 341 as 34 points as part of the learning data, just like the subject in Example 2.

[0167] The data classification unit 211 outputs the evaluation lymphocyte subpopulation data 42, the evaluation basic data 41, colon cancer as the learning type, and 34 points as the basic feature 341 to the learning unit 212 as learning data.

[0168] The learning unit 212 learns the classification model 600 based on the learning data in accordance with the above-mentioned steps S3 and S4, and determines the trained classification model 600.

[0169] In the third embodiment, the learning unit 212 performs calculations for adjusting and comparing related elements, using the basic feature 341 as the numerator and the immune dynamic feature 39 as the denominator, so that the ratio between the two is 1.

[0170] The learning unit 212 uses 34 points included in the training data as basic features 341. The learning unit 212 forms basic feature weights 381, immune dynamic features 39, and immune dynamic weights 391 by self-learning based on the evaluation basic data 41 included in the training data and the evaluation lymphocyte subpopulation data 42, with reference to the immune dynamics table 35 in Fig. 11.

[0171] In Example 3, the learning unit 212 can evaluate the helper T cells as 78 cells / μL, which corresponds to Level 1 of the immune dynamics table 35, as 0 points. As shown in Figure 24, in Example 3, the ratio of granulocytes, monocytes, and lymphocytes is 80:7:13%, which corresponds to Level 3 of the immune dynamics table 35, as 4 points.

[0172] In Example 3, the helper T cells are 78 cells / μl, which corresponds to level 1 in the immune dynamics table 35 and can be evaluated as 0 points, so it can be assumed that the Th1 immune strength is less than 1. Therefore, the learning unit 212 determines that the Th1 immune strength is less than 1, which corresponds to level 1 in the immune dynamics table 35 and can be evaluated as 0 points.

[0173] In Example 3, the helper T cells are 78 cells / μL, which corresponds to Level 1 of the immune dynamics table 35 and can be evaluated as 0 points, so the activated T cell proportion can be assumed to be less than 5%. Therefore, the learning unit 212 determines that the activated T cell proportion is less than 5%, which corresponds to Level 1 of the immune dynamics table 35 and can be evaluated as 0 points.

[0174] In Example 3, the helper T cells were 78 / μL, which corresponds to Level 1 of the immune dynamics table 35 and can be evaluated as 0 points, so it can be assumed that the proportions of the core elements of the evaluated lymphocyte subpopulation data 42 are abnormal, as shown in Figure 23. Therefore, the proportions of the core elements of the evaluated lymphocyte subpopulation data 42, namely, helper T cells, αβ killer T cells, γδ T cells, NK cells, NKT cells, and B cells, can be determined to be Level 1 of the other immune cell indicators of the immune dynamics table 35 and can be evaluated as 0 points.

[0175] That is, when the helper T cells are at level 1 in the immune dynamics table 35, the evaluations of the activated T cell proportion, Th1 immune strength, and other immune cell indices can be considered to be at the lowest level for each index and can be used in the learning calculations. However, if actual values ​​for the activated T cell proportion, Th1 immune strength, and other immune cell indices are available, those actual values ​​may be used.

[0176] Alternatively, if the levels of all three items, helper T cells; granulocyte, monocyte, and lymphocyte ratios; and activated T cell proportion, are level 2 or lower, the evaluation of Th1 immune strength and other immune cell indices will be considered to be at the lowest level for each indicator and used in the learning calculation. However, if actual values ​​for activated T cell proportion, Th1 immune strength, and other immune cell indices are available, those actual values ​​may be used.

[0177] In Example 3, the tumor marker CA72-4 (6.9 U / m or less) was 14.6, and the anti-P53 antibody (1.3 U / m or less) was 13.97. Since both are greater than the standard values, they are classified as Level 1 in the immune dynamics table 35, and can be evaluated as 0 points.

[0178] That is, in Example 3, since helper T cells are at level 1 in the immune dynamics table 35, the proportions of granulocytes, monocytes, and lymphocytes are at level 2 in the immune dynamics table 35, and the immune dynamics main item is 4, it is conceivable that the learning unit 212 will select 0 points as the immune dynamics weight 391.

[0179] The learning unit 212 performs self-learning to appropriately form the basic feature weight 381 and the immune dynamics weight 391 based on the actual test data in Example 3 and the immune dynamics table 35 of Figure 11. In Example 3, as in Example 2, there is test data for three items in the sub-items of Figure 11: tumor marker, HLA-DR(+)CD38, and HLA-DR(-)CD38, and the data for these three items can be taken into consideration in the immune dynamics evaluation. Therefore, the basic feature weight 381 can be set to the sum of the highest scores for the three items: tumor marker, HLA-DR(+)CD38, and HLA-DR(-)CD38, so it is conceivable to select 6 points.

[0180] 26 , in Example 3, the learning unit 212 performs calculations based on the above elements, resulting in a total of 40 points for the basic elements 38 and basic feature weights 381, and a score of 4 for the immune dynamic feature 39, for a ratio of 10. The learning unit 212 then stores the calculation results shown in FIG. 26 in the classification model box 34 as a classification model 600. The learning unit 212 may also store the classification model 600 in the classification model box 34 and simultaneously display it on the display unit 5.

[0181] In this case, the learning unit 212 extracts the classification model 600 of Example 2 as the basic model 601. Then, the learning unit 212 compares the classification model 600 of Example 3 as the evaluation model 603 with the basic model 601. A typical comparison result of the present invention is that if the evaluation model 603 is larger than the basic model 601, it is determined that the immune status is worsening. If the evaluation model 603 is equal to the basic model 601, it is determined that there is no change in the immune status. If the evaluation model 603 is smaller than the basic model 601, it is determined that the immune status is improving. The learning unit 212 stores the above determination results in the classification model box 34 and simultaneously displays them on the display unit 5.

[0182] In Example 3, as shown in FIG. 26, the evaluation model 603 is larger than the basic model 601, and therefore the immune status is determined to be in a deteriorated state.

[0183] Although the present invention has been described above based on the best embodiment, those skilled in the art will understand that various changes, modifications, and substitutions can be made to these embodiments without departing from the principle and spirit of the present invention. The present invention is not limited to the specific embodiments disclosed herein, and other embodiments within the scope of the claims of this application are also included in the scope of protection of the present invention.

[0184] REFERENCE SIGNS LIST 1 immune dynamics generation device 2 immune dynamics processing unit 3 memory unit 4 information input unit 5 display unit 6 external information source 7 network 21 learning mechanism 210 data acquisition unit 211 data classification unit 212 learning unit 22 immune dynamics evaluation unit 31 basic data 32 immune subgroup dynamics information (limb cell subgroup data) 33 classification extraction table 34 classification model box 341 basic feature 35 immune dynamics table 38 basic element 600 classification model 601 basic model 602 reference model 603 evaluation model

Claims

1. An immune dynamics generation device comprising at least one immune dynamics processing unit, a memory unit for storing at least clinical data, and a display unit, wherein the immune dynamics processing unit acquires clinical basic characteristics and immune subgroup dynamics information based on the clinical data, forms a classification model by self-learning based on the immune subgroup dynamics information and the classification model, and simultaneously monitors the classification model, evaluates the immune dynamics information of the subject by comparing the overall relationship between the clinical basic characteristics and the immune subgroup dynamics information based on the classification model, and simultaneously controls the display of the evaluation results of the immune dynamics information on the display unit.

2. The immune dynamics generating device according to claim 1, wherein the clinical data includes at least health checkup item data for normal individuals and health checkup item data for abnormal individuals.

3. The immune dynamics generating device according to claim 1, wherein the immune subgroup dynamics information includes the number of B cells, the number of NK cells, the number of NKT cells, and the numbers and proportions of immune lymphocyte subgroups, including at least the number of helper T cells, the number of killer T cells, the number of αβ killer-T cells, the number of γδT cells, the number of NK cells, the number of NKT cells, and the number of B cells.

4. The immune dynamics generating device of claim 2, wherein the immune dynamics processing unit acquires the clinical basic features, the immune subgroup dynamics information, and the classification model based on the health checkup item data of the normal person, and obtains normal immune status information by comparing the overall relationship between the clinical basic features and the immune subgroup dynamics information.

5. The immune dynamics generating device of claim 2, wherein the immune dynamics processing unit acquires the clinical basic features, the immune subgroup dynamics information, and the classification model based on the health check item data of the non-normal person, and obtains abnormal immune state information by comparing the overall relationship between the clinical basic features and the immune subgroup dynamics information.

6. The immune dynamics generating device described in claim 5, wherein the immune dynamics processing unit acquires the clinical basic features, the immune subgroup dynamics information, and the classification model based on the clinical data including cancer cell data, and obtains cancer incidence rate information by comparing the overall relationship between the clinical basic features and the immune subgroup dynamics information based on the abnormal immune state information.

7. The immune dynamics generating device according to claim 6, wherein the immune dynamics processing unit determines that the cancer cells are in a worsening state when the ratio of the cancer incidence rate information is greater than 1.

8. The immune dynamics generating device according to claim 6, wherein the immune dynamics processing unit determines that the cancer cells are in a life-prolonging state when the ratio of the cancer incidence rate information is 1.

9. The immune dynamics generating device according to claim 6, wherein the immune dynamics processing unit determines that the cancer cells are in an improved state when the ratio of the cancer incidence rate information is less than 1.

10. A method for generating immune dynamics, comprising the steps of: acquiring clinical data of a subject; obtaining immune subgroup dynamics information of the subject; and monitoring average data of a large number of normal subjects and abnormal data with changes calculated using basic data and lymphocyte subgroup data stored in a memory unit, and evaluating the immune dynamics of the subject by comparing the overall relationship between the clinical data and the immune subgroup dynamics information based on the average data and the abnormal data.

Images