Method and apparatus for evaluating muscle health of individual

Abstract

The present application provides an apparatus for evaluating muscle health of an individual, the muscle health evaluating apparatus comprising: at least one processor; and at least one memory including a computer program code, wherein the at least one memory and the computer program code are configured such that, through the at least one processor, the muscle health evaluating apparatus acquires health information of a plurality of individuals, derives a plurality of muscle health evaluation variables from the health information through a machine learning model, derives a muscle health evaluation equation in consideration of a weight of each of the plurality of muscle health evaluation variables, acquires information about the plurality of muscle health evaluation variables of a specific individual, and evaluates muscle health of the specific individual by using the information about the plurality of muscle health evaluation variables of the specific individual and the muscle health evaluation equation.

Description

Method and device for evaluating the muscle health of an individual

[0001] The present invention relates to a method and device for evaluating the muscle health of an individual.

[0002] Skeletal muscles, which constitute the body, are attached to bones and play an important role in maintaining our body's posture, generating force, and sustaining daily life. Furthermore, muscles are the largest organ, accounting for half of body weight, and the human body is composed of over 600 muscles. Muscles play primary roles in movement, exertion, breathing, balance, weight control, and as the main storehouse of body protein.

[0003] Meanwhile, natural aging, prolonged bed rest due to serious injuries, and immobilization (casting) can lead to a decrease in muscle mass and function, which play an important role in maintaining daily life.

[0004] While muscle mass may vary individually depending on strength training and nutritional status, it reaches its peak during adolescence and young adulthood and then gradually declines. After the age of 60, the decline in both muscle mass and strength accelerates, negatively impacting independent living, including posture maintenance, and increasing the risk of disability due to falls and fractures. This can lead to increased mortality and morbidity rates associated with comorbid diseases. However, research results indicate that maintaining a healthy lifestyle while incorporating proper nutrition and resistance training can significantly reduce the risk of various diseases associated with age-related muscle decline.

[0005] Sarcopenia is defined as a phenomenon in which physical function declines along with muscle strength due to a decrease in skeletal muscle mass caused by aging, and its severity is being highlighted as it has recently been listed in the ICD-10-CM (Clinical Modification) and assigned a disease code (M62.84). However, the measurement tests (pharmacological history, sppb test) and DEXA (Dual-energy X-ray absorptiometry) radiography (or CT) currently used as diagnostic tools for sarcopenia have the problem of being very cumbersome and uneconomical.

[0006] To address this problem, affordable portable body composition analyzers have made it possible to easily measure muscle mass at home; however, since measuring muscle mass alone makes it impossible to assess or determine muscle function, this is an incomplete method for evaluating muscle health.

[0007] Recently, the development of personalized health management solutions based on big data and applying advanced technologies such as cloud computing, AI, blockchain, the metaverse, and NFTs has been gaining attention, and the development of algorithms capable of conveniently measuring an individual's muscle health status is actively underway.

[0008] Therefore, there is an urgent need for technology that can measure comprehensive muscle health status by utilizing individual health and lifestyle information without the need for specific locations or facilities, and which is practically applicable.

[0009] To evaluate muscle health, not only muscle mass but also functional aspects must be assessed and reflected. Therefore, measuring muscle mass and cross-sectional area requires expensive medical equipment, while measuring muscle function requires various physical fitness testing devices. In particular, measuring muscle function may require active participation from the subject, enabling them to exert their full potential.

[0010] Therefore, this method aims to overcome these difficulties and present a method for evaluating an individual's muscle health by utilizing simple anthropometric measurements or health checkup indicators.

[0011] The present invention provides a muscle health evaluation device comprising at least one processor and at least one memory including computer program code, wherein the at least one memory and the computer program code are configured such that the muscle health evaluation device acquires health information of a plurality of individuals through the at least one processor, derives a plurality of muscle health evaluation variables from the health information through a machine learning model, derives a muscle health evaluation formula by considering the weight of each of the plurality of muscle health evaluation variables, acquires information on the plurality of muscle health evaluation variables of a specific individual, and evaluates the muscle health of the specific individual using the information on the plurality of muscle health evaluation variables of the specific individual and the muscle health evaluation formula.

[0012] In addition, the present invention provides a method for evaluating muscle health performed in a device for evaluating muscle health, comprising the steps of: acquiring health information of a plurality of individuals; deriving a plurality of muscle health evaluation variables that can be reflected in a muscle health evaluation formula from the health information through a machine learning model; deriving a muscle health evaluation formula by considering the weights of each of the plurality of muscle health evaluation variables; acquiring information on the plurality of muscle health evaluation variables of a specific individual; and evaluating the muscle health of the specific individual using the information on the plurality of muscle health evaluation variables of the specific individual and the muscle health evaluation formula.

[0013] Through this facility, the muscle health of an individual can be evaluated using simple anthropometric measurements or health checkup indicators.

[0014] FIG. 1 shows a block diagram of an apparatus according to the present invention.

[0015] Figure 2 shows a flowchart of the method according to the present invention.

[0016] Figure 3a shows the health and physical fitness standards for Koreans (adults).

[0017] Figure 3b shows the health and physical fitness standards for Koreans (elderly).

[0018] FIG. 4 is a diagram illustrating the training process of a machine learning modeling process according to one embodiment of the present invention.

[0019] FIG. 5 shows an optimal machine learning model derived according to one embodiment of the present invention.

[0020] FIG. 6 shows the result of calculating a muscle health score using an evaluation formula derived according to one embodiment of the present invention.

[0021] Figure 7 shows the AUC-ROC results of a development dataset according to one embodiment of the present invention.

[0022] Figure 8 shows the AUC-ROC results of a verification dataset according to one embodiment of the present invention.

[0023] Embodiments of the present invention are described below with reference to the attached drawings to enable those skilled in the art to easily implement the invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification are denoted by similar reference numerals.

[0024] Throughout this specification, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0025] The terms “step” or “step of” used throughout this specification do not mean “step for”.

[0026] Throughout this specification, the description "A and / or B" means "A or B, or A and B".

[0027] Throughout the entire specification, the term “part” includes a unit realized by hardware, a unit realized by software, and a unit realized using both. Additionally, one unit may be realized using two or more hardware, and two or more units may be realized by one hardware.

[0028] Throughout this specification, the term "object" refers to a subject for which muscle health is to be evaluated, and preferably may be a person. Here, the term "person" may include Koreans of other groups, or foreigners such as Chinese, Japanese, or American nationals, although the present embodiment used data obtained from a specific group of Koreans.

[0029] The functions realized by the components described herein may be implemented in a general-purpose processor, a specific-purpose processor, an integrated circuit, an Application Specific Integrated Circuit (ASIC), a Central Processing Unit (CPU), a circuit, and / or a combination thereof, which are programmed to realize the described functions. A processor may include transistors or other circuits and is considered to be a circuit or a processing circuit. A processor may be a programmed processor that executes a program stored in memory.

[0030] In this specification, circuits, parts, units, and means are hardware programmed to realize or perform the described functions. Such hardware may be any hardware disclosed in this specification or any hardware known to be programmed or perform the described functions.

[0031] If the hardware is a processor considered to be of a circuit type, the circuit, the part, means, or unit is a combination of the hardware and the software used to constitute the hardware and / or processor.

[0032] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings. However, the present invention may not be limited to these embodiments and drawings.

[0033] The first aspect of the present invention provides a muscle health evaluation device comprising at least one processor and at least one memory including computer program code, wherein the at least one memory and the computer program code are configured such that the muscle health evaluation device acquires health information of a plurality of individuals through the at least one processor, derives a plurality of muscle health evaluation variables from the health information through a machine learning model, derives a muscle health evaluation formula by considering the weight of each of the plurality of muscle health evaluation variables, acquires information on the plurality of muscle health evaluation variables of a specific individual, and evaluates the muscle health of the specific individual using the information on the plurality of muscle health evaluation variables of the specific individual and the muscle health evaluation formula.

[0034] Furthermore, the second aspect of the present invention provides a method for evaluating muscle health performed in a device for evaluating muscle health, comprising the steps of: acquiring health information of a plurality of individuals; deriving a plurality of muscle health evaluation variables that can be reflected in a muscle health evaluation formula from the health information through a machine learning model; deriving a muscle health evaluation formula by considering the weights of each of the plurality of muscle health evaluation variables; acquiring information on the plurality of muscle health evaluation variables of a specific individual; and evaluating the muscle health of the specific individual using the information on the plurality of muscle health evaluation variables of the specific individual and the muscle health evaluation formula.

[0035] The first and second aspects of the present invention share the same technical concept.

[0036] FIG. 1 shows a block diagram of a device according to the present invention. Referring to FIG. 1, the muscle health assessment device (1) of the present invention may include at least one processor (10) and at least one memory (20) (storage device / storage) containing computer program code.

[0037] At least one memory (20) and computer program code can enable a muscle health evaluation device (1) to first acquire health information of multiple entities through at least one processor (10), derive multiple muscle health evaluation variables from the health information through a machine learning model, and derive a muscle health evaluation formula by considering the weight of each of the multiple muscle health evaluation variables.

[0038] According to one embodiment of the present invention, the term "multiple entities" means "multiple Koreans" when, for example, the entity is a Korean, and in this case, the term "health information of multiple entities" means "health information of multiple Koreans." Specifically, according to the embodiment below, the health information of multiple entities may be data from the "National Fitness 100 Project," but is not limited thereto.

[0039] According to one embodiment of the present invention, the health information includes information that can be a muscle health evaluation variable. According to the embodiment below, the health information includes body mass index (weight (kg) / height (m²) 2 It may include, but is not limited to, gender (male / female), age (years), total cholesterol (mg / dL), fasting blood glucose (mmol / L), body fat percentage (%), triglycerides (mg / dL), body fat mass (kg), HDL cholesterol (mg / dL), LDL cholesterol (mg / dL)), drinking status (abstinent, 0; drinking, 1), waist circumference (cm), smoking status (non-smoking, 0; quitting smoking, 1; smoking, 2), and muscle percentage by body weight (left arm (%), right arm (%), torso (%), right leg (%), left leg (%)).

[0040] According to one embodiment of the present invention, the machine learning model may be a generalized linear model, and a plurality of muscle health evaluation variables may additionally be derived from the health information through an Elastic net algorithm or a Lasso algorithm.

[0041] In this context, regression analysis refers to a technique for modeling the correlation between multiple independent variables and a single dependent variable. Generalized linear regression, used in generalized linear models, refers to modeling f(y), obtained by transforming the dependent variable into an appropriate function, as a linear combination of independent variables and regression coefficients.

[0042] According to one embodiment of the present invention, a plurality of muscle health evaluation variables may be derived by setting health information as an independent variable and the muscle health level of a plurality of individuals (muscle health upper group, lower group) as a dependent variable and using a generalized linear regression algorithm.

[0043] Additionally, the Elastic Net and Lasso algorithms are used as regularization techniques to exclude variables with low weights. Multiple muscle health assessment variables may be derived by passing through a generalized linear regression algorithm and then removing variables with low weights using the Elastic Net and Lasso algorithms.

[0044] According to one embodiment of the present invention, the plurality of muscle health evaluation variables may include age, gender, body mass index, fasting blood glucose, HDL cholesterol, LDL cholesterol, body fat percentage, right arm muscle percentage per body weight, left arm muscle percentage per body weight, and right leg muscle percentage per body weight.

[0045] According to one embodiment of the present invention, a muscle health evaluation device (1) can derive a muscle health evaluation formula by considering the weight of each of a plurality of muscle health evaluation variables. Here, the weight may refer to the degree of correlation (slope) between an independent variable and a dependent variable in regression analysis.

[0046] According to one embodiment of the present invention, a muscle health evaluation device (1) can evaluate the muscle health of a specific individual using a derived muscle health evaluation formula. Specifically, the muscle health evaluation device (1) can obtain information on a plurality of muscle health evaluation variables of a specific individual and evaluate the muscle health of a specific individual using the plurality of muscle health evaluation variable information and the muscle health evaluation formula.

[0047] The features of the muscle health evaluation device (1) described above also apply to the muscle health evaluation method according to another embodiment of the present invention shown in FIG. 2.

[0048] Example 1. Derivation of Muscle Health Assessment Formula

[0049] According to one embodiment of the present invention, in order to derive a muscle health evaluation formula, data from the 'National Fitness 100 Project' organized by the Ministry of Health and Welfare and the Korea Sports Promotion Foundation was obtained.

[0050] The 'National Fitness 100 Project' is a project that measures the diverse physical fitness and health status of Koreans to present fitness levels reflecting health characteristics by life cycle, aiming to prevent disease and improve mental health and quality of life. It involves conducting demographic surveys, blood and body composition tests, cardiorespiratory fitness tests, muscle strength tests, and flexibility tests.

[0051] Regarding the project data, initial data was obtained for 2,887 individuals, excluding missing values, out of a total of 3,782 adult and elderly men and women aged 19 or older who visited a national physical fitness certification center and completed physical fitness measurements and health checkups / surveys.

[0052] Muscle mass and muscle function were selected as indicators to assess the level of muscle health (upper and lower groups). For muscle mass, ① skeletal muscle mass (kg) was used; for muscle function, ② relative grip strength was used as an indicator of muscle strength, and ③ cross sit-ups were used as an indicator of muscle endurance; and for each indicator, the upper and lower groups were classified based on the paper "Development of Health Fitness Standards for Koreans" presented by the Korea Institute of Physical Fitness Promotion (Figs. 3a, 3b).

[0053] ① Skeletal muscle mass and ③ cross sit-ups were evaluated by dividing into 3 quartiles by gender and age group, with the lower 1 / 3 classified as the lower group and the remaining 2 / 3 as the upper group.

[0054] ② In the case of relative grip strength, the criteria were applied based on gender and age, and the results were determined by dividing into upper and lower groups.

[0055] Based on the previously defined judgment criteria, the proportion of each combination of the upper and lower groups of the three muscle function indicators was confirmed. For the application of logistic regression analysis, those judged to have all three indicators as normal were classified into the upper muscle health group (n=992, 34.4%), and those judged to have two or more indicators as deficient were classified into the lower muscle health group (n=830, 28.7%). The remaining data were excluded, and the data for the upper and lower muscle health groups were used as the final data to derive the muscle health evaluation formula according to the present invention.

[0056] The final data of a total of 1,822 people was divided into a development dataset (training set; 70% of the total data, 1,276 people) and a validation dataset (test set; 30% of the total data, 546 people). Then, the development dataset was used to derive a muscle health evaluation formula using the health information of multiple individuals, and the performance of the muscle health evaluation device (1) or method according to the present invention was evaluated in Example 2 below through ROC curve analysis using the validation dataset.

[0057] Specific information on the health of multiple individuals is as shown in Table 1 below.

[0058]

[0059] Using a generalized linear model, a regression model among machine learning algorithms, the health information in Table 1 above was set as independent variables, and modeling was performed to verify the correlation between two or more independent variables and the dependent variable, muscle health level (high muscle health group, low muscle health group). As modeling methods, Elastic net (selecting or removing multiple variables with high correlation) and Lasso (reducing the coefficient (a constant multiplied in front of the equation) by randomly selecting one of multiple correlated variables) were applied. After conducting 100 machine learning modeling iterations, the model that used the fewest features while exhibiting the lowest error was finally selected as the optimal model. A muscle health evaluation equation was derived by considering the multiple muscle health evaluation variables and their respective weights derived through the above modeling process.

[0060] FIG. 4 is a diagram illustrating the training process of a machine learning modeling process, and FIG. 5 is a diagram showing an optimal model. The evaluation formula according to an embodiment of the present invention, derived through FIG. 4 and FIG. 5, is as follows.

[0061] Y=(0.59)+(-1.53)*Age+(2.32)*Gender+(0.62)*Body Mass Index+(-0.11)*Fasting Blood Glucose+(0.14)*HDL Cholesterol+(0.09)*LDL Cholesterol+(-0.14)*Body Fat Percentage+(0.79)*Right Arm Muscle Percentage per Body Weight+(0.46)*Left Arm Muscle Percentage per Body Weight+(1.29)*Right Leg Muscle Percentage per Body Weight

[0062] Herein, the above mathematical formula is merely an example and does not limit the scope of the present invention.

[0063] As shown in Figure 6, the muscle health evaluation score was calculated using the final derived evaluation formula and the validation dataset, and it was confirmed that the upper muscle health group and the lower muscle health group were clearly distinguished through this score.

[0064] Example 2. ROC curve analysis

[0065] According to an embodiment of the present invention, as illustrated in FIGS. 7 to 8, the performance of the muscle health evaluation device (1) or method according to the present invention was evaluated through ROC curve analysis using the verification dataset (test set; 30% of the total data, 546 people) of Example 1.

[0066] Muscle health scores were calculated by applying the derived evaluation formula to the entire development dataset and the validation dataset, and ROC curve analysis was performed to evaluate the ability to distinguish between the upper and lower muscle health groups. The AUC-ROC of the entire development dataset (training set) was derived as 0.939 (sensitivity 86.6%, specificity 85.5%, accuracy 86.1%). As a result of applying it to the validation dataset (test set), the AUC-ROC was derived as 0.948 (sensitivity 87.4%, specificity 87.2%, accuracy 87.4%). In other words, it was confirmed that the muscle health evaluation device (1) or method according to the present invention can accurately evaluate muscle health.

Claims

1. In a device for evaluating the muscle health of an individual, At least one processor; and It includes at least one memory containing computer program code, and The above at least one memory and the computer program code through the above at least one processor The above muscle health evaluation device Acquire health information of multiple entities, and Multiple muscle health evaluation variables are derived from the above health information through a machine learning model, and A muscle health evaluation formula is derived by considering the weight of each of the aforementioned multiple muscle health evaluation variables, and Acquiring information on the aforementioned multiple muscle health evaluation variables of a specific individual, and A muscle health evaluation device configured to evaluate the muscle health of the specific individual using the plurality of muscle health evaluation variable information of the specific individual and the muscle health evaluation formula.

2. In Paragraph 1, The above health information is, Body Mass Index (Weight (kg) / Height (m) 2 A muscle health assessment device comprising: )), gender (male / female), age (years), total cholesterol (mg / dL), fasting blood glucose (mmol / L), body fat percentage (%), triglycerides (mg / dL), body fat mass (kg), HDL cholesterol (mg / dL), LDL cholesterol (mg / dL)), drinking status (abstinence, 0; drinking, 1), waist circumference (cm), smoking status (non-smoking, 0; quitting smoking, 1; smoking, 2), and muscle percentage by body weight (left arm (%), right arm (%), torso (%), right leg (%), left leg (%)).

3. In Paragraph 1, The above machine learning model is a generalized linear model, and The above plurality of muscle health evaluation variables are, Additionally, a muscle health assessment device derived from the above health information through an Elastic net algorithm or a Lasso algorithm.

4. In Paragraph 1, The above plurality of muscle health evaluation variables are, A muscle health assessment device including age, gender, body mass index, fasting blood glucose, HDL cholesterol, LDL cholesterol, body fat percentage, right arm muscle percentage per body weight, left arm muscle percentage per body weight, and right leg muscle percentage per body weight.

5. In a method for evaluating muscle health performed in a device for evaluating muscle health, Step of acquiring health information of multiple entities; A step of deriving a plurality of muscle health evaluation variables that can be reflected in a muscle health evaluation formula from the above health information through a machine learning model; and A step of deriving a muscle health evaluation formula by considering the weight of each of the plurality of muscle health evaluation variables mentioned above; A step of obtaining information on the plurality of muscle health evaluation variables of a specific object; A step of evaluating the muscle health of the specific individual using the plurality of muscle health evaluation variable information of the specific individual and the muscle health evaluation formula; A muscle health assessment method including 6. In Paragraph 5, The above health information is, Body Mass Index (Weight (kg) / Height (m) 2 A muscle health assessment method including )), gender (male / female), age (years), total cholesterol (mg / dL), fasting blood glucose (mmol / L), body fat percentage (%), triglycerides (mg / dL), body fat mass (kg), HDL cholesterol (mg / dL), LDL cholesterol (mg / dL)), drinking status (abstinent, 0; drinking, 1), waist circumference (cm), smoking status (non-smoking, 0; quitting smoking, 1; smoking, 2), and muscle percentage by body weight (left arm (%), right arm (%), torso (%), right leg (%), left leg (%)).

7. In Paragraph 5, The above machine learning model is a generalized linear model, and The step of deriving the above plurality of muscle health evaluation variables is, A muscle health evaluation method comprising the step of additionally deriving muscle health evaluation variables from health information through an Elastic net algorithm or a Lasso algorithm.

8. In Paragraph 5, The above plurality of muscle health evaluation variables are, A muscle health assessment method that includes age, gender, body mass index, fasting blood glucose, HDL cholesterol, LDL cholesterol, body fat percentage, right arm muscle percentage per body weight, left arm muscle percentage per body weight, and right leg muscle percentage per body weight.

Images