KR102097188B1 - Prediction method for dementia using neuropsychological test and prediction system for dementia using neuropsychological test - Google Patents

Abstract

The method for predicting the development of dementia using a neuropsychological test includes the steps of a computer device receiving the subject's age and the subject's neuropsychological test data, and the computer device classifying the type of mild cognitive impairment based on the neuropsychological test data. And inputting the age and the type of mild cognitive impairment into a dementia prediction model using a nomogram, and the computer device generating dementia prediction information according to an output result of the dementia prediction model.

Description

Prediction method and prediction system of dementia using neuropsychological examination {PREDICTION METHOD FOR DEMENTIA USING NEUROPSYCHOLOGICAL TEST AND PREDICTION SYSTEM FOR DEMENTIA USING NEUROPSYCHOLOGICAL TEST}

The technique described below relates to a technique for predicting the development of dementia in a subject using a neuropsychological test.

Mild cognitive impairment is an intermediate stage between normal and dementia. Patients with mild cognitive impairment are high-risk patients with a high risk of developing dementia. It is known that 10-15% of patients with mild cognitive impairment progress to dementia after 1 year, and about 80% convert to dementia after 6 years. However, about 20% of patients with mild cognitive impairment appear to be converted to normal or maintain mild cognitive impairment. In other words, there is no exact standard for the possibility of patients with mild cognitive impairment converting to dementia.

Korean Registered Patent No. 10-1437569

Recently, studies have been conducted to more accurately predict the possibility of dementia conversion such as image analysis and amyloid PET. However, amyloid PET has the disadvantages of high cost of examination, image analysis such as brain cortical thickness analysis, which is time consuming and not clinically intuitive. The technique described below provides a technique for determining the risk of dementia based on neuropsychological test results for patients with mild cognitive impairment.

The method for predicting the development of dementia using a neuropsychological test includes the steps of a computer device receiving the subject's age and the subject's neuropsychological test data, and the computer device classifying the type of mild cognitive impairment based on the neuropsychological test data. And inputting the age and the type of mild cognitive impairment into a dementia prediction model using a nomogram, and the computer device generating dementia prediction information according to an output result of the dementia prediction model.

The dementia outbreak prediction system using a neuropsychological test is an input device that receives the subject's age and the subject's neuropsychological test data, and stores the predicted model of dementia that receives the age and type of mild cognitive impairment and generates dementia likelihood information. And a server that classifies the types of mild cognitive impairment based on the device and the neuropsychological data, and inputs the age and the types of mild cognitive impairment into the dementia prediction model using a nomogram to generate dementia prediction information.

The technique described below predicts the possibility of simple and very rapid dementia conversion using neuropsychological tests performed on patients with mild cognitive impairment.

1 is an example of demographic information and clinical information for aMCI subjects.
2 is an example of analysis of clinical and neuropsychological characteristics related to dementia conversion.
3 is an example of a nomogram for Model 1 shown in FIG. 2. That is, Figure 3 is an example of a nomogram for a patient without the APEO genotype.
FIG. 4 is an example of a graph showing a relationship between predicted values determined according to the nomogram of FIG. 3 and possibility of dementia conversion.
5 is another example of the analysis of clinical and neuropsychological characteristics associated with dementia conversion.
6 is an example of a nomogram for Model 2 shown in FIG. 5.
7 is an example of a graph showing a relationship between a predicted value determined according to the nomogram of FIG. 6 and a possibility of dementia conversion.
8 is a result of performing verification on the dementia prediction model.
9 is an example of a computer device for predicting the occurrence of dementia.
10 is an example of a system for predicting the occurrence of dementia.
11 is an example of a block diagram for a dementia prediction model.

The technique described below may be applied to various changes and may have various embodiments, and specific embodiments will be illustrated in the drawings and described in detail. However, this is not intended to limit the technology described below to specific embodiments, and should be understood to include all modifications, equivalents, or substitutes included in the spirit and scope of the technology described below.

Terms such as first, second, A, B, etc. may be used to describe various components, but the components are not limited by the above terms, and only for the purpose of distinguishing one component from other components Used only. For example, the first component may be referred to as a second component, and similarly, the second component may be referred to as a first component without departing from the scope of the technology described below. The term and / or includes a combination of a plurality of related described items or any one of a plurality of related described items.

In the terminology used herein, a singular expression should be understood to include a plurality of expressions, unless clearly interpreted differently in context, and terms such as “comprises” describe features, counts, steps, operations, and components described. It is to be understood that it means that a part or a combination thereof is present, and does not exclude the presence or addition possibility of one or more other features or numbers, step operation components, parts or combinations thereof.

Prior to the detailed description of the drawings, it is intended to clarify that the division of components in this specification is only divided by the main functions of each component. That is, two or more components to be described below may be combined into one component, or one component may be divided into two or more for each subdivided function. In addition, each of the components to be described below may additionally perform some or all of the functions of other components in addition to the main functions in charge of the components, and some of the main functions of each of the components are different. Needless to say, it may also be carried out in a dedicated manner.

In addition, in performing the method or the operation method, each of the processes constituting the method may occur differently from the specified order, unless a specific order is explicitly described in context. That is, each process may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

The techniques described hereinafter describe the risk of dementia based on the results of neuropsychological tests (hereinafter referred to as neuropsychological tests) for patients with amnestic mild cognitive impairment (hereinafter referred to as "aMCI patients"). It is a judgment technique. The technique described below uses a model (hereinafter referred to as a dementia prediction model) that predicts a dementia risk according to a certain test and test results for an actual aMCI patient. The dementia prediction model is built using the relationship between psychological test results for patients and the risk of dementia.

Hereinafter, experimental conditions and processes in which the researcher has established a dementia prediction model will be described first. First, the environment or conditions for analysis, such as the population and evaluation methods used to establish a dementia prediction model, will be described.

Subject

The aMCI patient is a patient who meets the following Peterseon criteria. Common conditions that aMCI patients have include (i) subjective memory complaints by the patient or by a peripheral person (caregiver), and (ii) a standard deviation of Korean version MMSE (Mini-Mental State Examination) of -1.0 or higher (which is normal general recognition) Possible), (iii) normal activities of daily living (ADL) (which means possible daily living), (iv) standard deviation of verbal or visual memory test results is less than -1.0 (objective) Evaluation of memory dysfunction), (v) no symptoms of dementia.

All subjects had prior related studies (e.g., Kim MJ, Im K, Lee JM, Park A, Chin J, Kim GH, et al. Cortical thinning in verbal, visual, and both memory-predominant mild cognitive impairment. Alzheimer Disease & Associated Disorders 2011a; 25: 242-9.) Were subjected to specific clinical interviews using semi-structured questionnaires, neurological examinations and a comprehensive neuropsychological battery. ADL was evaluated by SI Seoul Instrumental ADL (SI ADL) described in the previous study (Ku HM KJ, Kwon EJ, Kim SH, Lee HS, Ko H. A study on the reliability and validity of Seoul-Instrumental Activities of Daily Living (S -IADL) .J Korean Neuropsychiatr Assoc 2004; 43: 189-99.).

Meanwhile, the following patients were excluded from the subjects. Patients with secondary cognitive impairment identified by tests such as vitamin B12 / folic acid measurements, syphilis serology, thyroid function tests, and structural structural lesions of the brain such as MRI were excluded from the subjects. In addition, patients with dementia symptoms such as Lewy bodies (DLB), frontotemporal lobar degeneration (FTLD), and progressive supranuclear palsy were excluded from the subjects.

Neuropsychological test method

All subjects were subjected to a standardized neuropsychological test called Soul Neuropsychological Screening Battery (SNSB). SNSB includes tests for attention, language, Gersmann syndrome, praxis, spatiotemporal function, language and visual memory, and frontal executive functions.

Of these, the tests indicated by the test results are as follows. Digit span forward / backward, Korean-Boston Naming Test (K-BNT) for evaluating language function, ray-austeris complex figure to measure spatiotemporal composition and memory Re-Osterrieth Complex Figure Test (RCFT), Seoul Verbal Learning Test (SVLT), Controlled Oral Word Association Test (COWAT), Color and Stroop test and MMSE for word reading evaluation.

The test results are adjusted to take into account the age and educational level of the subject. If the standard deviation of the adjusted test results was less than -1.0, it was judged as abnormal. In the RCFT, scores of delayed memory items were used as a measure of visual memory impairment. In the SVLT, the scores of delayed memory items were used as a measure of language memory future. The score of K-BNT was used as a measure of language function. The score of RCFT was used as a measure to evaluate spatiotemporal cognitive function. The frontal lobe management function is evaluated using three tests: exercise performance test (contrasting program, Go / no-go, fist-edge-palm, alternating hand movement, alternative square and triangle and Luria loop), COWAT and stroop test. do. If at least two of the three tests described above indicate a disorder, the frontal lobe management function is determined to be a disorder.

Classification of aMCI subjects

Based on the above test results, aMCI subjects were classified into several types. (i) First, aMCI subjects were classified according to the degree of memory impairment. In the delayed memory test, subjects with a standard deviation of scores between -1.0 and -1.5 were classified as early stage aMCI (E-aMCI) patients. Subjects with a standard deviation of scores below -1.5 in the language or visual memory test were classified as late-stage aMCI (L-aMCI) patients. (ii) Second, subjects were classified into three groups based on the pattern of memory impairment. Subjects with only visual memory impairment were classified as visual aMCI patients. Subjects with only speech memory impairment were classified as language aMCI patients. In addition, subjects with visual and speech memory impairment were classified as visual and verbal aMCI patients. (iii) Third, subjects were classified according to the multiplicity of related cognitive impairment areas. Subjects with memory impairment only were classified as single area aMCI patients. In case of having other cognitive disorders (such as language disorders and spatio-temporal cognitive disorders) along with memory disorders, they were classified as multi-domain aMCI patients.

Dementia diagnosis

Diagnosis of dementia was determined based on the Diagnostic and Statistical Manual of Mental Disorders (DSM) -IV, evidence of cognitive impairment confirmed by neuropsychological examination, and social and / or occupational disability identified by ADL disorder. Diagnosis of Alzheimer's Disease (AD) potential was performed using criteria provided by the National Institute of Neurological and Communicative Disorders and Stroke (NINCDS) and Alzheimer's Disease and Related Disorders Association (ADRDA). Furthermore, the expert personnel reviewed the interview results of the subjects and the neuropsychological test results of each aMCI patient, and confirmed the diagnosis of dementia. Dementia diagnosis results are used to build a dementia prediction model. In addition, the diagnosis results of dementia can be used to verify the dementia prediction model. Diagnosis of dementia is performed by subjecting the subject to neuropsychological examination for a period of time.

Hereinafter, an example of constructing a dementia prediction model according to the results of a neuropsychological test will be described. The dementia prediction model uses a logical regression analysis having information about the subject and test results of the subject as variables. Therefore, first, a regression analysis model must be generated using test results. Then, a new model is derived using the variables used in the regression model. Briefly, the valid variables are determined from the variables used in the regression model. Valid variables may be any variable in the regression equation. The dementia prediction model assigns a certain value to each valid variable of the regression model. The value assigned here is to convert each variable to a certain score for risk estimation. A nomogram can be used to convert each variable to a certain score. Eventually, depending on the value of each valid variable, each variable has a specific score. The final risk can be determined by summing the scores of all variables.

The dementia prediction model provides results for the likelihood of developing dementia within a period of time. The actual experiment provided an assessment of the risk of developing dementia within 3 years. The dementia prediction model basically uses aMCI classification according to the above-mentioned neuropsychological test as a variable. Furthermore, other variables can be considered. For example, the dementia prediction model may additionally use the subject's gender, age, education level, body mass index (BMI), presence or absence of Apolipoprotein E (APOE) gene. Hereinafter, the statistical analysis used and the risk model establishment process will be described.

Statistical analysis

The statistical analysis and analysis environment used are briefly described. Dementia and non-dementia have statistical differences in demographic and clinical characteristics. Statistical differences between dementia and non-dementia transducers can be analyzed using independent t and χ ² tests. As is well known, the independent t test is a method for testing the mean difference between two independent samples, and the χ ² test is a method for determining the correlation (or independence) between two samples.

Consecutive variables that did not satisfy the linearity assumption were grouped into two or more categories. The relationship between potential predictors and results was verified using a logical regression model with a single variable and a logical regression model with multiple variables.

The stepwise method was used for variable selection. Variables must have a significance level of 0.25 to enter the model. The variable remains in the model only when it maintains a significance level of 0.15 in the model. Multicollinearity was confirmed using a VIF (variance inflation factor). Multiple comparisons were revised using the Bonferroni technique.

Statistical analysis used STATA version 13.0, SAS version 9.4 (SAS Institute, Cary, NC) and R 3.3.2 (Austria, Vienna, http://www.R-project.org/). Nomograms were derived using the results of multivariate analysis by RMS 3.2.2 (http://www.r-project.org).

Now, the process of analyzing the risk based on the results of the neuropsychological test for the above-described subject will be described.

test results

The test was performed on 338 aMCI patients. 1 is an example of demographic information and clinical information for aMCI subjects. In FIG. 1, a p value refers to a general significance probability used in statistical analysis.

1 shows demographic information such as the subject's age, education level or duration, and female ratio. 222 out of 338 were women (65.7%). The mean age of aMCI subjects was 71.6 years (range 43 to 88 years, standard deviation 7.3).

Meanwhile, FIG. 1 shows clinical information on aMCI classification according to a subject's neuropsychological test result. As described above, the aMCI classification is divided into three groups. The aMCI classification is classified as (i) early stage early stage (aMCI) or late stage (aMCI), and (ii) visual aMCI (Visual), language aMCI (Verbal) or visual and language aMCI (Both). Classified, and (iii) single-region aMCI (Single) or multi-region aMCI (Multiple) patients. 1 shows classification results for a subject. In addition, Figure 1 shows information about whether the subject has the APOE genotype.

The subject is tracked for a period of time to monitor whether dementia develops, and if dementia is invented, it is defined as a transducer state. The mean follow-up period in the experiment was 2.92 years. Follow-up refers to the ongoing diagnosis of dementia in subjects who have undergone neuropsychological examination. Of the total subjects, 208 (61.5%) were diagnosed with a transducer within a 3-year period. The converter refers to the conversion from the non-dementia stage to the dementia stage. A state that has not been converted to the dementia stage is called a non-converter. Non-converted subjects remained stable (87), returned to normal (27), or converted to non-aMCI (16) during the follow-up period. Non-aMCI refers to subjects outside the conditions of aMCI described above.

As described above, the variables for establishing the dementia prediction model may use aMCI classification, age, gender, education level, body mass index (BMI), presence or absence of Apolipoprotein E (APOE) gene according to neuropsychological test results. According to the results of the follow-up test in FIG. 1, variables related to the development of dementia are age, neuropathy, and classification of aMCI patients and the presence of APOE4. In particular, L-aMCI patients, visual and verbal aMCI patients, and multi-domain aMCI patients were more likely to have dementia conversion compared to E-aMCI patients, visual or verbal aMCI patients and single-region aMCI patients, respectively. Meanwhile, the level of education and BMI did not appear to have a significant effect on the development of dementia.

Derivation of nomogram and risk score

The nomogram was prepared using a logical regression model with multiple variables. The nomogram was constructed by assigning a constant value to each variable using the beta coefficient of the regression model. The value calculated through the nomogram is the risk score. The nomogram assigns a constant value to each variable used in regression analysis. At this time, different values are determined through the nomogram depending on the importance of the risk assessment. The risk score is the sum of the total values determined according to the values of each variable using the nomogram. The risk of dementia can be evaluated according to the final risk score.

Hereinafter, a process for deriving a nomogram will be described. The data of the subjects used in the nomogram derivation process are shown in Table 1 below. It is classified based on the APOE genotype. Model 1 is data for subjects without the APOE genotype, and model 2 is data for subjects with the APOE genotype. In Table 1, AD conversion refers to dementia conversion.

Model 1 (without APOE)
N = 222 Model 2 (without APOE)
N = 167 Age 71.41 ± 7.3 71.54 ± 7.3 Sex (female) 141 (63.5%) 109 (65.2%) Education (years) 9.79 ± 5.3 10.36 ± 5.3 BMI 23.79 ± 3.4 23.7 ± 3.2 AD conversion 130 (58.5%) 99 (59.3%

Correction of risk prediction model

In order to verify the risk prediction model, the probability of predicting the risk using the actual observation probability and the nomogram can be compared. In order to improve the accuracy of the risk prediction model, a constant correction value may be assigned such that the result value output from the risk prediction model matches or is similar to the actual observation result.

Training data and verification data for model building

In order to build a model, the data were divided into two. Here, the data refers to test results for the subject. One data is for training (training data), and the other data is for verification (verification data). A part of the entire subject data was used as training data, and the other part was used as verification data. In addition, a predictive model was constructed for each of the patients with the APOE genotype and those without the corresponding genotype (74% of the subjects had the APOE genotype in the experiment). The model for the APOE genotype was built using 222 training data and 116 validation data. Models without the APOE genotype were constructed using 174 training data and 75 validation data.

2 is an example of analysis of clinical and neuropsychological characteristics related to dementia conversion. FIG. 2 is an example of analysis for a subject who does not have the APOE genotype, Model 1. The left side of FIG. 2 is an example of univariable analysis, and the right side is an example of multivariable analysis.

Factors in the table of FIG. 2 refer to variables related to dementia conversion. OR is the odds ratio, and the cross ratio has a 95% confidence interval (CI). The p value means a general significance probability used in statistical analysis. A point means a specific value converted into a score for each element.

The age was divided into 4 groups (49 to 66.2 years, 66.2 to 72.5 years, 72.5 to 76.8 years, and 76.8 to 88 years). Of the 4 groups, the 49-66.2-year-old group was used as a reference value (Reference, Ref) to convert the scores for the remaining groups. P value decreases with age. Therefore, it can be said that age affects dementia conversion, and the probability of dementia conversion increases as the age increases.

VVB refers to the result of classifying patients based on the pattern of disability in neuropsychological examination. VVB has visual aMCI patients (Visual), language aMCI patients (Verbal) and time and language aMCI patients (Both). Visual aMCI patients were used as reference values to convert the scores for the remaining groups. Disability patterns are also interpreted as factors affecting dementia transformation. Furthermore, it is interpreted that patients with verbal aMCI (Verbal) and patients with time and language aMCI (Both) are more likely to have dementia transformation than patients with one disorder.

EL refers to the result of classifying patients according to the degree of memory impairment in neuropsychological examination. EL has early stage aMCI patients (Early) and late stage aMCI patients (Late). The scores for the remaining groups were converted using the initial stage aMCI patient as a reference value. The degree of memory impairment is also interpreted as a factor influencing dementia conversion. Furthermore, it is interpreted that late stage aMCI patients (Late) are more likely to have dementia transformation than early stage aMCI patients (Early).

SM means the result of classifying patients according to the multiplicity of the cognitive impairment area. SM has single domain aMCI patients (Single) and multi domain aMCI patients (Multiple). SM calculated the scores for the remaining groups using a single-region aMCI patient as a reference value. The multiplicity of the cognitive impairment area is also interpreted as a factor influencing dementia transformation. Furthermore, it is interpreted that a multi-domain aMCI patient (Multiple) has a higher probability of dementia transformation than a single-domain aMCI patient (Single).

On the other hand, the results of multivariate analysis also show roughly the same results as those of single variable analysis.

3 is an example of a nomogram for Model 1 shown in FIG. 2. That is, Figure 3 is an example of a nomogram for a patient without the APEO genotype. The bottom of FIG. 3 shows an example of each item used in each nomogram and a score according to the value of the item. In FIG. 3, Age is an item for converting scores according to the age of the subject. VVB1 is an item for converting scores for cases classified only as visual aMCI (Visual) or language aMCI (Verbal). If only the visual or language aMCI is present (value of 1 in VVB1), a score of 14 is assigned. VVB2 is an item for converting scores for cases classified as visual and language aMCI (Both). In case of having visual and language aMCI (value of 1 in VVB2), a score of 35 is assigned. EL is an item for converting scores for early stage aMCI (Early) and late stage aMCI (Late). The late stage aMCI (value of 2 in EL) assigns a score of 18 points. SM is an item for converting scores for single-region aMCI (Single) and multi-region aMCI (Multiple). The multi-domain aMCI (value of 2 in SM) assigns a score of 31 points. The total score is the sum of the scores of each item of Age, VVB1, VVB2, EL and SM. The linear predicted value is a predicted value for the possibility of dementia conversion. The final prediction value is derived from the overall score corresponding to the linear prediction value.

FIG. 4 is an example of a graph showing a relationship between predicted values determined according to the nomogram of FIG. 3 and possibility of dementia conversion. 4 shows the likelihood of actual aMCI patients converting to dementia according to the linear predicted values. On the other hand, qualitatively specific possibilities can be judged as having a high possibility of dementia conversion. For example, if the possibility of dementia conversion ≤ 0.2 is low, the probability of dementia conversion is 0.2, the probability of dementia conversion is 0.4, the probability is normal, and if the possibility of dementia conversion> 0.4, it can be determined as high probability.

5 is another example of the analysis of clinical and neuropsychological characteristics associated with dementia conversion. FIG. 5 is an example of analysis for a subject with the APOE genotype, Model 2. The left side of FIG. 5 is an example of univariable analysis, and the right side is an example of multivariable analysis.

Factors in the table of FIG. 5 refer to variables related to dementia conversion. OR is the odds ratio, and the cross ratio has a 95% confidence interval (CI). The p value means a general significance probability used in statistical analysis. A point means a specific value converted into a score for each element.

Unlike the FIG. 3, age is displayed for one continuous type. Since the P value is 0.0038 (univariate analysis) /0.0004 (multivariate analysis), it is interpreted that age is also a factor influencing dementia transformation. It can be said that the age affects the dementia transformation, and as the age increases, the probability of dementia transformation increases.

VVB refers to the result of classifying patients based on the pattern of disability in neuropsychological examination. VVB has visual aMCI patients (Visual), language aMCI patients (Verbal) and time and language aMCI patients (Both). Visual aMCI patients were used as reference values to convert the scores for the remaining groups. Disability patterns are also interpreted as factors affecting dementia transformation. Furthermore, it is interpreted that patients with verbal aMCI (Verbal) and patients with time and language aMCI (Both) are more likely to have dementia transformation than patients with one disorder.

EL refers to the result of classifying patients according to the degree of memory impairment in neuropsychological examination. EL has early stage aMCI patients (Early) and late stage aMCI patients (Late). The scores for the remaining groups were converted using the initial stage aMCI patient as a reference value. The degree of memory impairment is also interpreted as a factor influencing dementia conversion. Furthermore, it is interpreted that late stage aMCI patients (Late) are more likely to have dementia transformation than early stage aMCI patients (Early). On the other hand, in the multivariate analysis, there was no significance for EL, so the items were not displayed.

SM means the result of classifying patients according to the multiplicity of the cognitive impairment area. SM has single domain aMCI patients (Single) and multi domain aMCI patients (Multiple). SM calculated the scores for the remaining groups using a single-region aMCI patient as a reference value. The multiplicity of the cognitive impairment area is also interpreted as a factor influencing dementia transformation. Furthermore, it is interpreted that a multi-domain aMCI patient (Multiple) has a higher probability of dementia transformation than a single-domain aMCI patient (Single).

APOE is a factor depending on the presence or absence of the APOE4 genotype. It is interpreted that there is a higher probability of dementia conversion than patients without the APOE gene. If there is no corresponding genotype, an additional score for the case with the APOE4 genotype is assigned using the reference value.

On the other hand, the results of multivariate analysis also show roughly the same results as those of single variable analysis.

6 is an example of a nomogram for Model 2 shown in FIG. 5. That is, Figure 6 is an example of a nomogram for a patient who does not have the APEO genotype. The bottom of FIG. 6 shows an example of a score according to each item and the value of the item used in each nomogram. In FIG. 6, Age is an item for converting scores according to the age of the subject. VVB1 is an item for converting scores for cases classified only as visual aMCI (Visual) or language aMCI (Verbal). If only the visual or language aMCI is present (value of 1 in VVB1), a score of 23 is assigned. VVB2 is an item for converting scores for cases classified as visual and language aMCI (Both). A score of 47 is assigned for a visual and language aMCI (value of 1 in VVB2). SM is an item for converting scores for single-region aMCI (Single) and multi-region aMCI (Multiple). The multi-domain aMCI (value of 2 in SM) assigns a score of 24 points. APOE is an item for adding a score for aMCI having the APOE gene. If you have the APOE gene, a score of 40 is assigned. The total score is the sum of the scores of each item of Age, VVB1, VVB2, SM and APOE. The linear predicted value is a predicted value for the possibility of dementia conversion. The final prediction value is derived from the overall score corresponding to the linear prediction value.

7 is an example of a graph showing a relationship between a predicted value determined according to the nomogram of FIG. 6 and a possibility of dementia conversion. 7 shows the likelihood of actual aMCI patients converting to dementia according to the linear predicted values. On the other hand, qualitatively specific possibilities can be judged as having a high possibility of dementia conversion. For example, if the possibility of dementia conversion ≤ 0.2 is low, the probability of dementia conversion is 0.2, the probability of dementia conversion is 0.4, the probability is normal, and if the possibility of dementia conversion> 0.4, it can be determined as high probability.

8 is a result of performing verification on the dementia prediction model. Verification was carried out for Model 1 and Model 2, respectively. For verification, the performance of the dementia prediction model was first quantified by a C-index under the C (receiver operating characteristic) curve (using C-statistics). The accuracy of the predictive model was confirmed through internal validation and external validation. For internal verification, a bootstrapping technique, a K-fold cross validation technique, or the like can be used. In actual verification, bootstrapping technique and 10-times cross-validation were used. In Model 1, the internal verification was 0.75 or more, and the external verification was 0.742, and the prediction performance was excellent. Model 2 considering the APOE genotype has an internal verification of 0.8 or higher, and an external verification of 0.73, which has better predictive performance than Model 1.

Hereinafter, an example of an apparatus for performing dementia prediction using the above-described dementia prediction model will be described. 9 is an example of a computer device 100 for predicting the occurrence of dementia. The computer device 100 refers to a device such as a PC, laptop, smart device or server. The computer device 100 includes an input device 110, a computing device 120, a storage device 130, and an output device 140.

The input device 110 receives demographic information (eg, age, gender, etc.) of the subject and a neuropsychological test performed by the subject. Hereinafter, the results of the neuropsychological test performed by the subject are referred to as neuropsychological test data. The input device 110 is a device for inputting demographic information and previously performed neuropsychological data into the computer device 100 through communication or a separate storage device. Furthermore, the input device 110 may be an interface device (keyboard, mouse, touch screen, etc.) that receives a value for a neuropsychological test item performed by a subject through the computer device 100. In this case, the computer device 110 generates certain neuropsychological test data according to the input item value. The neuropsychological test is a standardized neuropsychological test called NSB described above.

The storage device 130 is a device for storing a dementia prediction model. The dementia prediction model may be data generated with constant programming code. The storage device 130 may store demographic information of the subject and neuropsychological data of the subject received from the input device 110. Also, the storage device 130 may store information on the possibility of dementia predicted by the dementia prediction model. The storage device 130 may store information (hereinafter referred to as classification information) for determining the aCMI classification described above based on neuropsychological data.

The computing device 120 determines the aCMI classification of the subject using the input neuropsychological data and classification information. The computing device 120 derives the possibility of dementia for the subject by inputting the demographic information of the subject and the classification of neural aCMI into the dementia prediction model stored in the storage device 130. The likelihood of dementia can be output as a constant score, probability (%) or qualitative value (low probability / high probability). The possibility of various types of dementia is hereinafter referred to as dementia prediction information.

The output device 140 is a device that outputs dementia prediction information in a certain form. The output device 140 includes at least one of a display device, an output device that outputs a document, and a communication device that delivers dementia prediction information to another device.

10 is an example of a system 200 for predicting the occurrence of dementia. The system 200 for predicting the occurrence of dementia is composed of various objects located in a network. The system 200 for predicting the occurrence of dementia in FIG. 10 includes a user terminal 210, a test DB 220, a judgment server 230, and a model DB 240.

The user terminal 210 is a device that receives and stores neuropsychological test data performed by a subject. The user terminal 210 may provide a neuropsychological test, and may generate neuropsychological test data according to a value input by a subject. Furthermore, the user terminal 210 receives and stores demographic information of the subject. The test DB 220 is a device that stores demographic information of a subject and neuropsychological data of a subject previously performed. The user terminal 210 and the test DB 220 correspond to an input device that delivers the subject's age and neuropsychological test data to the determination server 230.

The determination server 230 is a device that determines the predictability of dementia for the subject using demographic information and neuropsychological data of the subject. The determination server 230 receives demographic information and neuropsychological test data from the user terminal 210 or the test DB 220. That is, the system 200 for predicting the occurrence of dementia may use only one of the user terminal 210 or the inspection DB 220. The determination server 230 determines the aCMI classification of the subject using the received neuropsychological test data and classification information. The determination server 230 inputs the demographic information of the subject and the classification of neural aCMI to the dementia prediction model stored in the storage device 130 to generate dementia likelihood information for the subject. Although not illustrated in FIG. 10, the determination server 230 may transmit the dementia likelihood information to the user terminal 210, other user devices, and a separate server.

The model DB 240 is a device that stores information on the dementia prediction model described above. The model DB 240 corresponds to a storage device that stores a dementia prediction model. In addition, the model DB 240 may store the above-described classification information. Although illustrated as a separate object in FIG. 10, the model DB 240 may be integrated with the determination server 230.

11 is an example of a block diagram for the dementia prediction model 300. 11 is an example of a dementia prediction model 300 used in the computer apparatus 100 of FIG. 9 or the system 200 for predicting the occurrence of dementia of FIG. 10. The computer device 100 of FIG. 9 may store the dementia prediction model 300 in the storage device 130. The model DB 240 of FIG. 10 may store the dementia prediction model 300.

The dementia prediction model 300 is composed of a score estimation model 310 and a dementia likelihood evaluation model 320. The score calculation model 310 is a model that receives a value (a value of a variable) of an individual item for evaluating the possibility of dementia and calculates a specific score. As described above, the variables include demographic information (age, gender, etc.) of the subject and classification of aMCI. The aMCI classification is determined in advance using psychological examination data in the computer device 100 or the determination server 230.

The score calculation model 310 determines a specific score for all variables to be used, and finally outputs a value (total score) summing the scores for all variables. The score calculation model 310 may include at least one of a score function 311 and a score table 312. The score calculation model 310 may determine a score for each variable using at least one of the score function 311 and the score table 312. The score function 311 means a function for converting the value of the input variable into a specific score. The score function 311 may use different functions according to the type of the input variable. The score table 312 means a table for converting the value of the input variable into a specific score. The score table 312 may include a constant score table for each type of input variable. For example, when the age of the subject is input, the score estimation model 310 determines a constant score using the score function 311 or the score table 312. In addition, the score calculation model 310 determines a constant score using the score function 311 or the score table 312 when the subject's aMCI classification is input. The score calculation model 310 sums scores for all variables and outputs a total score. Meanwhile, the score function 311 and the score table 312 perform the functions of the nomogram of FIGS. 3A and 4A. The scoring model 310 includes individual models (information) for subjects without the APOE genotype (Model 1 in FIG. 3A) and subjects with the APOE genotype (Model 2 in FIG. 4A).

The dementia likelihood evaluation model 320 receives the total score output by the score calculation model 310. The dementia likelihood evaluation model 320 may include at least one of a risk function 321 and a risk table 322. The dementia likelihood evaluation model 320 calculates dementia likelihood information using at least one of the risk function 321 and the risk table 322. The score function 321 refers to a function that converts the input total score into specific dementia likelihood information. The score table 322 means a table for converting the input total score into specific dementia likelihood information. The score table 322 stores dementia likelihood information according to individual scores or a range of scores. The dementia likelihood evaluation model 320 provides individual models (information) for subjects without APOE genotype (dementia potential information based on Model 1) and subjects with APOE genotype (dementia potential information based on Model 2).

The drawings attached to the present embodiment and the present specification merely show a part of the technical idea included in the above-described technology, and are easily understood by those skilled in the art within the scope of the technical idea included in the above-described technical specification and drawings. It will be apparent that all examples and specific examples that can be inferred are included in the scope of the above-described technology.

100: computer device
110: input device
120: computing device
130: storage device
140: output device
200: dementia prediction system
210: user terminal
220: inspection DB
230: judgment server
240: model DB
300: dementia prediction model
310: scoring model
311: score function
312: score table
320: Dementia likelihood evaluation model
321: Risk function
322: Risk table

Claims (12)

An input device of the computer device receiving the subject's age and the subject's neuropsychological data;
Classifying the type of mild cognitive impairment based on the neuropsychological test data by the computing device of the computer device;
Inputting the age and the type of mild cognitive impairment into a dementia prediction model using a nomogram; And
The computing device includes the step of generating dementia prediction information according to the output result of the dementia prediction model,
The type of mild cognitive impairment is the first classification classified as early or late according to the degree of memory impairment, the second classification classified as visual impairment, language disorder, visual and language disorder according to the disability pattern, and a single domain according to the disability domain multiplicity A method for predicting the development of dementia using neuropsychological examination including a third classification classified as a disorder or a multi-domain disorder. According to claim 1,
The neuropsychological test is a method for predicting the development of dementia using a neuropsychological test, which is a Soul Neuropsychological Screening Battery (SNSB). delete According to claim 1,
The first classification is the development of dementia using a neuropsychological test including an initial mild cognitive impairment with a standard deviation of the delayed memory test score of -1.0 to -1.5 and a late mild cognitive impairment with a standard deviation of the delayed memory test score of less than -1.5. Prediction method. According to claim 1,
The second classification is dementia using neuropsychological test including visual mild cognitive impairment with only visual memory impairment, verbal mild cognitive impairment with only language memory impairment, and time-language mild cognitive impairment with visual memory impairment and speech memory impairment simultaneously. How to predict the onset. According to claim 1,
The third classification is a method for predicting the development of dementia using a neuropsychological test including a single-region mild cognitive impairment with only memory impairment and a multi-domain mild cognitive impairment with additional cognitive impairment along with memory impairment. According to claim 1,
The computing device prepares a regression analysis model using training data having the subject's age and neuropsychological data and a diagnosis result of dementia for the same subject, and allocates a certain score according to the value of the variable used in the regression analysis model. The dementia disease prediction method using a neuropsychological test further comprising the step of generating in advance the dementia prediction model including the nomogram. A computer-readable recording medium recording a program for executing a method for predicting the development of dementia using the neuropsychological test according to any one of claims 1, 2, and 4 to 7, in a computer. An input device that receives the subject's age and the subject's neuropsychological data;
A storage device for storing a model for predicting dementia that receives the type of age and mild cognitive impairment and generates dementia likelihood information; And
And a server for classifying mild cognitive impairment type based on the neuropsychological test data and inputting the age and the mild cognitive impairment type into the dementia prediction model using a nomogram,
The type of mild cognitive impairment is the first classification classified as early or late according to the degree of memory impairment, the second classification classified as visual impairment, language disorder, visual and language disorder according to the disability pattern, and a single domain according to the disability domain multiplicity A system for predicting the development of dementia using neuropsychological examination including a third classification classified as a disorder or a multi-domain disorder. The method of claim 9,
The neuropsychological test is a system for predicting the development of dementia using a neuropsychological test, which is a Soul Neuropsychological Screening Battery (SNSB). delete The method of claim 9,
The dementia prediction model extracts variables used in the regression model generated using training data having the subject's age and neuropsychological data and dementia diagnosis results for the same subject, and assigns a constant score to the variable values A system for predicting dementia development using a neuropsychological test including the nomogram.

Images