CN115602324A

CN115602324A - Osteoporosis risk prediction system based on ultrasonic osteoporosis risk prediction model

Info

Publication number: CN115602324A
Application number: CN202211181743.4A
Authority: CN
Inventors: 赵玮; 段洁雨; 王浩宇; 刘占辉; 曹新格; 王鹏飞; 姚贝文
Original assignee: Hebei Chengying Information Technology Co ltd
Current assignee: FIRST HOSPITAL OF HEBEI MEDICAL UNIVERSITY
Priority date: 2022-09-27
Filing date: 2022-09-27
Publication date: 2023-01-13

Abstract

The invention discloses an osteoporosis risk prediction system based on an ultrasonic osteoporosis risk prediction model, which takes the currently accepted osteoporosis diagnosis standard in the world as a comparison standard, detects the obtained result of ultrasonic penetrability of a temporal window and the thickness value of the skull of the temporal window by using an ultrasonic inspection technology, combines the relevant factors with osteoporosis, analyzes the relation with the osteoporosis prediction, constructs different osteoporosis risk prediction models, draws a nomogram, establishes an assignment table, utilizes the functions of a computer programming processing system of the osteoporosis risk prediction system, respectively matches the prediction models suitable for different basic data tuples to be applied, respectively calculates, obtains the scores of different basic data tuples and the predicted osteoporosis probability according to the assignment table, outputs the highest score and the highest osteoporosis probability after being compared by an output program, and has good prediction accuracy.

Description

Osteoporosis risk prediction system based on ultrasonic osteoporosis risk prediction model

Technical Field

The invention relates to the technical field of population osteoporosis risk prediction, in particular to an osteoporosis risk prediction system based on an ultrasonic osteoporosis risk prediction model.

Background

Osteoporosis is a disease of systemic disease characterized by a decrease in bone mass and strength, a change in fibrous structure of bone tissue and an increase in the risk of fracture. The main index reflecting Bone mass is Bone Mineral Density (BMD). Osteoporosis is common in the elderly, especially postmenopausal women. Along with the prolonging of the life of human beings, the incidence rate of osteoporosis and related fracture thereof obviously rises, causing disability rate and fatality rate to increase, and also bringing heavy economic burden to families and society, so that the early prediction of the risk of osteoporosis has very important significance for early diagnosis and treatment of osteoporosis and prevention of osteoporotic fracture.

At present, the determination of bone strength, which is a key concept in the definition of osteoporosis, has no breakthrough progress, and is difficult to meet the clinical noninvasive, simple and easy requirements, and the bone density is taken as a substitute index of the bone strength. The bone mineral density measuring technology is a non-invasive measuring method for analyzing the bone mineral content, the bone mineral density and the body constitution components of a human body by mainly utilizing the principle that X rays are attenuated in different degrees through different media. The currently common bone density measurement technology is quantitative CT (quantitative computed tomography, QCT), in theory, QCT can measure any part of the whole body, measurement data is accurate, but in practice, most of QCT is concentrated on lumbar vertebrae (L1-3), measurement is affected by vertebral body fat, fat content is increased along with the increase of age, BMD value measured by partial volume effect is lower than actual BMD value, bone loss is more obvious than actual bone loss, fat problem can be solved by using dual energy measurement, but the dose of radiation is large, precision is reduced, and clinical application is less. QCT is not affected by bone size and should be suitable for children, but cannot be widely used due to the large dose of radiation.

A dual-energy X-ray absorption method (DXA) adopts two X-ray images obtained from X-rays with different energies, eliminates the influence of soft tissues on bone density measurement by an energy subtraction method, and measures the bone density value of the skeleton at the measured part by instrument data processing and a solution equation. The human body is subjected to low radiation dose, which becomes the standard for diagnosing osteoporosis, and the World Health Organization (WHO) also recommends hip and lumbar vertebrae density measured by a dual-energy X-ray bone densitometer as the objective diagnosis standard for osteoporosis.

The principle of Quantitative Ultrasound (QUS) is that Ultrasound waves attenuate through body tissue by an amount related to the tissue properties, whereas the attenuation of Ultrasound in Bone tissue depends on the elastic model of the Bone and the Bone density. Bone density can be calculated by measuring the change of the Speed of Sound (SOS) of the calcaneus Bone, broadband Ultrasonic Attenuation (BUA), bone Quality Index (BQI) and the like. The credibility of the parameters and the relevance of other methods are needed to be further researched. Compared with an X-ray bone densitometer, the ultrasonic bone densitometer has the advantages of low cost and no radiation, and is suitable for screening for general investigation. Has reference value and guidance value for preventing the physiological development of children, the bone damage of the old and the fracture risk.

Disclosure of Invention

The invention aims to provide an osteoporosis risk prediction system based on an ultrasonic osteoporosis risk prediction model.

Therefore, the technical scheme of the invention is as follows:

an osteoporosis risk prediction system based on an ultrasonic osteoporosis risk prediction model comprises a functional module which is configured with transcranial detection intracranial arteries and used for judging ultrasonic penetrability of a temporal window through an ultrasonic temporal window;

a temporal window skull thickness measuring module is configured and used for measuring the skull thickness by using a two-dimensional linear array B ultrasonic inspection technology;

a computer programmed processing system module configured with osteoporosis risk prediction.

Furthermore, the method comprises the steps of configuring a pulse wave ultrasonic probe for judging the ultrasonic penetrability of a temporal window through an ultrasonic temporal window, wherein the frequency of the probe is less than or equal to 1.6MHz, the diameter of the probe is less than or equal to 15.6mm, and the detection depth of the probe is more than or equal to 134mm;

the method comprises the step of configuring a high-frequency linear array probe for measuring the skull thickness value of an ultrasonic temporal window, wherein the highest frequency value of the probe frequency is more than or equal to 12MHz, and the resolution ratio of the probe frequency to the skull thickness of the temporal window is ensured.

Further, the computer programming processing system module for osteoporosis risk prediction comprises

The system comprises a prediction factor input system module, a basic data setting module and a basic data analysis module, wherein the prediction factor input system module is used for inputting basic data settings, and the basic data settings comprise a temporal window ultrasonic penetration result, a skull thickness value and data of factors related to osteoporosis, such as sex, age, body mass index and calcium phosphorus product, and the system decomposes the input data into a basic data tuple of the temporal window ultrasonic penetration result combined with the sex, age, body mass index and calcium phosphorus product osteoporosis and a basic data tuple of the skull thickness combined with the sex, age, body mass index and calcium phosphorus product osteoporosis;

the model building and training functional module is used for building and training an ultrasonic osteoporosis risk prediction model, and a built-in data result comprises the following steps: the method comprises the steps of respectively obtaining single-factor statistical analysis results and multiple Logistic regression analysis results of factors related to osteoporosis by temporal window ultrasonic penetrability and sex, age, body mass index and calcium phosphorus product, constructing different ultrasonic prediction models, drawing a nomogram and an established result of a scoring table, and drawing a working characteristic curve (ROC) result of an osteoporosis subject; the method comprises the following steps of (1) respectively constructing single-factor statistical analysis results of factors related to osteoporosis by the product of the thickness of the skull of a temporal window, gender, age, body mass index and calcium and phosphorus, multiple Logistic regression analysis results, drawing nomograms by constructing different ultrasonic prediction models, establishing assignment table results and drawing working characteristic curve results of osteoporosis subjects;

and the data distribution calculation and result output module is used for combining the input temporal window ultrasonic penetrability result with the data of the basic data tuple of the factors related to sex, age, body mass index and calcium phosphorus product osteoporosis and the data of the basic data tuple of the factors related to skull thickness, sex, age, body mass index and calcium phosphorus product osteoporosis, respectively matching prediction models suitable for different basic data tuples by using a distribution calculation program preset in equipment, respectively calculating the data in the existing prediction models 1-4, obtaining the scores of different basic data tuples and the predicted osteoporosis probability according to the assignment table, and outputting the highest score and the highest osteoporosis probability after comparing by an output program.

Furthermore, the model building and training function module utilizes the basic data collection and analysis function and increases the number of the alignment charts of the prediction models and optimizes the prediction models on the basis of the continuously updated embodiment database.

Further, the method for constructing the ultrasonic osteoporosis risk prediction model of the model construction and training functional module is as follows

Step 1: collecting an intracranial blood vessel detected by an ultrasonic probe through a temporal window, and judging an obtained ultrasonic penetrability result of the temporal window and a skull thickness value of the temporal window detected by a high-frequency linear array probe;

step 2: respectively drawing a work characteristic curve of the osteoporosis testee by using the ultrasonic penetrability result of the temporal window and the thickness value of the temporal window skull;

and step 3: and (3) building a prediction model 1-a prediction model 4 by taking the results of the ultrasonic penetrability of the temporal window and prediction factors related to osteoporosis by the product of the thickness value of the skull of the temporal window and the sex, the age, the body mass index and the calcium and phosphorus, drawing a nomogram, establishing an assignment table, and drawing a working characteristic curve of the osteoporosis testee.

Further, the temporal window of the ultrasound refers to the area above the zygomatic bone, between the lateral orbital margin and the tragus, and the thickness of the skull is measured between two cortical bone hyperechoes at the thinnest part of the temporal window;

the judgment standard of the incapability of penetration of the temporal window ultrasound refers to the incapability of detecting any intracranial artery signal;

the judgment standard of the ultrasonic penetrability of the temporal window refers to the signal that any intracranial artery can be successfully detected.

Furthermore, the system has the basic functions of transcranial Doppler intracranial artery detection and examination, two-dimensional B-ultrasound superficial organ and blood vessel detection.

Compared with the prior art, the osteoporosis risk prediction system based on the ultrasonic osteoporosis risk prediction model is provided, the prediction factors related to osteoporosis by respectively taking the result of temporal window ultrasonic penetrability and the thickness value of the temporal window skull and the product of sex, age, body weight index and calcium and phosphorus are used for constructing the first prediction model to the fourth prediction model, drawing a nomogram, establishing an assignment table and drawing a working characteristic curve of a subject with osteoporosis. The invention provides a novel osteoporosis risk prediction idea.

The method comprises the steps of inputting a temporal window ultrasonic penetrability result, a skull thickness value and basic data of osteoporosis-related factors such as sex, age, body weight index and calcium phosphorus product, decomposing the input data into an ultrasonic penetrability result, a basic data tuple combining sex, age, body weight index and calcium phosphorus product osteoporosis-related factors and a basic data tuple combining sex, age, body weight index and calcium phosphorus product osteoporosis-related factors by a system, respectively matching prediction models suitable for different basic data tuples by using a preset programmed distribution calculation program in equipment, respectively calculating the basic data tuples from the first prediction model to the fourth prediction model, obtaining scores of different basic data tuples and predicted osteoporosis probability according to a scoring table, and outputting the highest score and the highest osteoporosis probability after comparison by an output program, wherein the biological safety is good, and the accuracy of the prediction result is high.

Drawings

FIG. 1 is a schematic view of thickness measurement of temporal window skull

FIG. 2 is ROC curve of ultrasound penetration and osteoporosis of temporal window of 89 examined persons

FIG. 3 shows ROC curve of skull thickness and osteoporosis of temporal window of 89 patients

FIG. 4 is a multivariate Logistic regression analysis of osteoporosis for 89 cases of factors associated with osteoporosis and temporal window ultrasound permeability of examined persons

FIG. 5 is a Nomogram (Nomogram 1) of 89 subjects whose temporal window ultrasonic penetration, age, calcium phosphorus product and prediction model 1 were constructed

FIG. 6 is the osteoporosis ROC curve of Nomogram1 for 89 subjects

FIG. 7 is an osteoporosis correction curve of Nomogram1 of 89 subjects

FIG. 8 is an osteoporosis determination curve of Nomogram1 of 89 subjects

FIG. 9 is a multivariate Logistic regression analysis of osteoporosis of related factors of temporal window skull thickness and osteoporosis of 89 subjects

FIG. 10 is a Nomogram (Nomogram 2) of a prediction model 2 constructed from the thickness, age, calcium-phosphorus product of the skull of the temporal window of 89 subjects

FIG. 11 is an osteoporosis ROC curve of Nomogram2 of 89 subjects

FIG. 12 is an osteoporosis correction curve of Nomogram2 in 89 cases of subjects

FIG. 13 is an osteoporosis determination curve of Nomogram2 in 89 cases of subjects

FIG. 14 is a ROC curve comparing Nomogram1 and Nomogram2 osteoporosis of 89 subjects

FIG. 15 is a ROC curve of ultrasound penetration of temporal window and osteoporosis for 58 female subjects

FIG. 16 shows ROC curve of skull thickness and osteoporosis of temporal window of 58 female examinees

FIG. 17 is a multiple Logistic regression analysis of osteoporosis for factors associated with temporal window ultrasound penetration and osteoporosis for 58 female subjects

FIG. 18 is a Nomogram of 58 female subjects for temporal window ultrasound penetration, age building prediction model 3 (Nomogram 3)

FIG. 19 is the osteoporosis ROC curve for Nomogram3 in 58 female subjects

FIG. 20 is a graph showing the osteoporosis correction curves of Nomogram3 in 58 female subjects

FIG. 21 is a graph showing the osteoporosis determination curve of Nomogram3 in 58 female subjects

FIG. 22 is a multivariate Logistic regression analysis of osteoporosis for correlation factors of skull thickness and osteoporosis of 58 female subjects in temporal window

Fig. 23 is a Nomogram (Nomogram 4) for constructing a prediction model 4 from skull thickness, age, and BMI of 58 female subjects.

FIG. 24 is the osteoporosis ROC curve for Nomogram4 in 58 female subjects

FIG. 25 is a graph showing the osteoporosis correction curves of Nomogram4 in 58 female subjects

FIG. 26 is a graph showing the osteoporosis determination curve of Nomogram4 in 58 female subjects

FIG. 27 is a ROC curve comparing Nomogram3 and Nomogram4 osteoporosis in 58 female subjects

Detailed Description

The present invention will be further described with reference to the following specific examples and drawings, but the following examples are by no means intended to limit the present invention.

The invention provides an osteoporosis risk prediction system based on an ultrasonic osteoporosis risk prediction model, wherein the osteoporosis risk prediction analysis system comprises a functional module which is configured with a transcranial detection intracranial artery and used for judging the ultrasonic penetrability of a temporal window through an ultrasonic temporal window;

a temporal window skull thickness measuring module is configured for measuring the skull thickness by using a two-dimensional linear array B ultrasonic inspection technology, and the measuring method is shown in figure 1;

It should be noted that the osteoporosis risk prediction system is provided with a pulse wave ultrasonic probe for judging the ultrasonic penetrability of a temporal window through an ultrasonic temporal window, the frequency of the probe is less than or equal to 1.6MHz, the diameter of the probe is less than or equal to 15.6mm, and the probe detection depth is more than or equal to 134mm;

the osteoporosis risk prediction system is provided with a high-frequency linear array probe for measuring the skull thickness value of an ultrasonic temporal window, and the highest frequency value of the probe frequency is more than or equal to 12MHz, so that the resolution ratio of the thickness of the skull of the temporal window is ensured.

It should be further explained that the computer programming processing system module for osteoporosis risk prediction comprises a prediction factor input system module for inputting basic data settings, including data of temporal window ultrasound penetrability result, skull thickness value and osteoporosis associated factor sex, age, body mass index and calcium phosphorus product, and the system decomposes the input data into a temporal window ultrasound penetrability result combined with a basic data tuple of sex, age, body mass index and calcium phosphorus product osteoporosis associated factor and a basic data tuple of skull thickness value combined with sex, age, body mass index and calcium phosphorus product osteoporosis associated factor;

the model building and training functional module is used for building and training an ultrasonic osteoporosis risk prediction model, and a built-in data result comprises the following steps: the method comprises the steps of respectively obtaining single-factor statistical analysis results and multiple Logistic regression analysis results of factors related to the temporal window ultrasonic penetrability and osteoporosis by sex, age, body mass index and calcium phosphorus product, constructing different ultrasonic prediction models to draw nomograms and established assigning table results, and drawing the working characteristic curve results of osteoporosis subjects; the method comprises the steps of respectively constructing single-factor statistical analysis results and multiple Logistic regression analysis results of factors related to the thickness of a temporal window skull and related factors of gender, age, body mass index, calcium phosphorus product and osteoporosis, constructing different ultrasonic prediction models to draw a nomogram and an established assigning table result, and drawing a working characteristic curve result of an osteoporosis subject; the model building and training function module utilizes the basic data collection and analysis function and takes the constantly updated embodiment database as the basis to increase the number of the alignment charts of the prediction model and optimize the prediction model.

The model building and training function module utilizes basic data collection and analysis functions, increases the number of nomograms of the prediction model and optimizes the prediction model along with the increase of embodiments;

the method for constructing the ultrasonic osteoporosis risk prediction model of the model construction and training functional module comprises the following steps

Step 1: collecting an intracranial blood vessel detected by an ultrasonic probe through a temporal window, and judging an ultrasonic penetrability result of the temporal window and a skull thickness value of the temporal window detected by a high-frequency linear array probe;

and 3, step 3: and (3) building a prediction model 1-a prediction model 4 by taking the results of the ultrasonic penetrability of the temporal window and prediction factors related to osteoporosis by the product of the thickness value of the skull of the temporal window and the sex, the age, the body mass index and the calcium and phosphorus, drawing a nomogram, establishing an assignment table, and drawing a working characteristic curve of the osteoporosis testee.

The temporal window of the ultrasound refers to the area above the zygomatic bone, between the lateral orbital margin and the tragus, and the thickness of the skull is measured between two layers of cortical bone strong echoes at the thinnest part of the temporal window;

the judgment standard of the incapability of penetration of temporal window ultrasound refers to the incapability of detecting any intracranial artery signal;

The system has the basic functions of transcranial Doppler intracranial artery detection and examination and two-dimensional B-ultrasonic superficial organ and blood vessel detection.

The embodiment is as follows:

89 patients with cerebrovascular disease clinical symptoms and complete data are selected as the selected criteria, and patients with serious liver function and renal dysfunction are excluded, wherein 58 female patients and 31 male patients are selected.

Diagnostic criteria for osteoporosis:

1. brittle fracture of the spine;

2. quantitative CT (quantitative computed tomography, QCT) mean bone Density Absolute BDM<80mg/cm ³ (ii) a 3. dXA (dual-energy X-ray absorption, DXA) measures T-value (T-Score), T-value = (measured value-same race identity normal youth peak bone density)/same race identity normal youth peak bone density Standard Deviation (SD). Osteoporosis: the T-value is less than or equal to-2.5 SD.

Non-osteoporosis includes osteopenia and normotensive bone mass.

In statistical analysis of the groupings, the risk factors associated with the criteria having a normal range of values are divided into normal and abnormal groups, with the abnormal groups including data greater and less than the normal range of values, according to industry or internationally recognized standards or guidelines.

Body Mass Index (BMI) grouping criteria: according to WHO standard, unit kg/m ² Normal BMI is more than or equal to 18.5 and less than 25, too low BMI is less than 18.5, overweight BMI is more than or equal to 25, and obesity occurs. The body mass index is divided into a normal group and an abnormal group, wherein the normal group has a BMI of more than or equal to 18.5 and less than 25, the abnormal group has a BMI of less than 18.5, and the group with the BMI of more than or equal to 25 includes obesity patients.

Calcium-phosphorus product (calcium-phosphorus product) grouping criteria: in unit mg/dl, the normal value is more than 35 and less than 40 of the product of calcium and phosphorus, the abnormal value is that the product of calcium and phosphorus is less than or equal to 35, and the product of calcium and phosphorus is more than or equal to 40. The normal group is 35-40, and the abnormal group comprises 35-40.

The calcium and phosphorus detection adopts a Beckmann Coulter AU5800 full-automatic biochemical analyzer, a calcium determination kit (an azoarsenic III method) and a phosphorus determination kit (a phosphomolybdate method).

The statistics of 89 subjects are as follows:

89 subjects, plotted as a baseline data to osteoporosis (Table 1).

TABLE 1 89 base-line data table of examinees

The above table indicates that, in addition to gender, age, penetration, BMI, calcium phosphorus product, and skull thickness are all statistically correlated with osteoporosis (P < 0.05).

ROC curves for temporal window ultrasound penetration and osteoporosis were plotted for 89 subjects, with an area under the curve (AUC) of 0.738 (fig. 2).

An ROC curve of the skull thickness of the temporal window of 89 subjects and osteoporosis is plotted, and the AUC is 0.707 (figure 3).

The ultrasonic penetration temporal window and the thickness of the cranial temporal window have strong correlation and are prediction factors of osteoporosis, and Logistic regression has multiple collinearity, so that statistical analysis is performed separately from the other.

The age, penetration, BMI, calcium phosphorus product of 89 subjects as independent variables and osteoporosis as dependent variables were subjected to multiple Logistic regression analysis (FIG. 4), which showed that the product of ultrasound penetration, age, calcium phosphorus was the predictor of osteoporosis (P < 0.05).

The ultrasonic penetration, age, calcium phosphorus product of osteoporosis prediction factors were used to construct a Nomogram (Nomogram 1) of prediction model 1 (fig. 5).

Table 2 shows the osteoporosis assignment table of Nomogram 1.

TABLE 2 osteoporosis assignment table to Nomogram1

The osteoporosis ROC curve for Nomogram1 is shown in FIG. 6, with an AUC of 0.848.

The osteoporosis correction curve for nomogrm 1 is shown in fig. 7, hosmer-Lemeshow test P =0.986.

The osteoporosis determination curve of Nomogram1 is shown in fig. 8, with benefit ranging from 0.01 to 0.97.

The age, BMI, calcium phosphorus product, and skull thickness of 89 subjects were used as independent variables, and osteoporosis was used as a dependent variable, and the subjects were subjected to osteoporosis multivariate Logistic regression analysis (fig. 9), indicating that age, calcium phosphorus product, and skull thickness are predictors of osteoporosis (P < 0.05).

The skull thickness, age, calcium-phosphorus product of osteoporosis predictor was used to construct a Nomogram (Nomogram 2) of prediction model 2 (fig. 10).

Table 3 is an osteoporosis assignment table of Nomogram 2.

TABLE 3 osteoporosis assignment table for Nomogram2

The osteoporosis ROC curve for Nomogram2 is shown in FIG. 11, with an AUC of 0.829.

The osteoporosis correction curve for nomogrm 2 is shown in fig. 12, and the hosmer-Lemeshow test P =0.853.

The osteoporosis determination curve of Nomogram2 is shown in fig. 13, with benefit ranging from [0.01-1].

Comparison of the osteoporosis ROC curves for Nomogram1 and Nomogram2 is shown in FIG. 14

The statistics of 58 female subjects out of 89 were as follows:

58 female subjects out of 89, presented a baseline data to osteoporosis (Table 4).

TABLE 4 Baseline data sheet of 58 female subjects

The above table shows that, in addition to the calcium phosphorus product, age, penetration, BMI, and skull thickness are statistically correlated with osteoporosis (P < 0.05).

ROC curves for temporal window ultrasound penetration and osteoporosis were plotted for 58 female subjects, with AUC of 0.766 (fig. 15).

ROC curves were plotted for temporal window skull thickness versus osteoporosis for 58 female subjects, with AUC of 0.723 (fig. 16).

The thickness of the ultrasound-penetrating temporal window and the cranial temporal window have strong correlation and are risk factors of osteoporosis, and Logistic regression can have multiple collinearity. Thus, the statistical analysis is performed separately from the other

The age, penetration, BMI, calcium phosphorus product of 58 female subjects as independent variables and osteoporosis as dependent variables were subjected to multiple Logistic regression analysis (fig. 17), indicating that ultrasound penetration, age, calcium phosphorus product are predictors of osteoporosis (P < 0.05).

An alignment chart (Nomogram 3) of the prediction model 3 was constructed from the ultrasound permeability and age of the osteoporosis prediction factor (fig. 18).

Table 5 is an osteoporosis assignment table for nomogrm 3.

TABLE 5 osteoporosis assignment table to Nomogram3

The osteoporosis ROC curve for Nomogram3 is shown in FIG. 19, with an AUC of 0.847.

The osteoporosis correction curve of nomogrm 3 is shown in fig. 20, and the hosmer-Lemeshow test P =0.877.

The osteoporosis determination curve of Nomogram3 is shown in fig. 21, with benefit ranging from 0.08 to 0.95.

Age, BMI, skull thickness as independent variables and osteoporosis as dependent variables were included in multiple Logistic regression analysis of osteoporosis (fig. 22) for 58 female subjects, showing age P =0.016, BMI P =0.059, skull thickness P =0.059. A Nomogram (Nomogram 4) of a prediction model was constructed from the skull thickness, age, and BMI (FIG. 23) which are factors for osteoporosis prediction,

table 6 is an osteoporosis assignment table for nomogrm 4.

TABLE 6 osteoporosis assignment table to Nomogram4

The osteoporosis ROC plot 24 of Nomogram4, AUC was 0.838.

The osteoporosis correction curve of nomogrm 4 is shown in fig. 25, and the hosmer-Lemeshow test P =0.933.

The osteoporosis determination curve of Nomogram4 is shown in fig. 26, with benefit ranging from [0.08-1].

A graph comparing the osteoporosis ROC curves of Nomogram3 and Nomogram4 is shown in FIG. 27.

The example statistics were analyzed as follows:

multiple studies indicate that gender is associated with osteoporosis, and particularly, the risk of osteoporosis in women after menopause is obviously increased, and the results in table 1 in the embodiment show that gender factors do not achieve statistical association with osteoporosis, may be related to sample size, and may be related to single factor analysis without calibration of other confounding factors.

The calcium phosphorus product reflects the status of calcium phosphorus metabolism in vivo, which has an effect on bone mass, and the results in table 4 of this example show that failure to achieve a statistical association with osteoporosis may be related to sample size or may be related to single factor analysis without calibration of other confounding factors.

The AUC of the temporal window ultrasonic penetrability and osteoporosis ROC of 89 subjects is 0.738 (figure 2), and the AUC is more than 0.7, which shows that the temporal window ultrasonic penetrability has better prediction accuracy on osteoporosis.

The AUC of the thickness of the skull of the temporal window of 89 examined patients and the ROC of the osteoporosis is 0.707 (figure 3), and the AUC is more than 0.7, which shows that the thickness of the skull of the temporal window has better prediction accuracy on the osteoporosis.

The AUC of the ultrasonic penetration of the temporal window and the ROC of osteoporosis of 58 female examinees was 0.766 (fig. 15), and the AUC > 0.7, indicating a relatively good prediction accuracy.

The AUC of the skull thickness of the temporal window and the osteoporosis ROC of 58 female testees is 0.723 (figure 16), and the AUC is greater than 0.7, which shows that the prediction accuracy is relatively good.

The research of the invention shows that the ultrasonic penetrability of the temporal window and the thickness of the skull of the temporal window are statistically associated with osteoporosis and are independent prediction factors of osteoporosis.

In 89 subjects statistical analysis, ultrasonic penetrability, age, body mass index and calcium-phosphorus product of a temporal window which are related factors of osteoporosis statistics are included in multivariate Logistic regression analysis (figure 4) of 89 subjects about osteoporosis, the body mass index is not a prediction factor of osteoporosis in the multifactor analysis and is related to a normal value range standard which is referred when the statistical analysis is grouped, and a normal group and an abnormal group which are distinguished according to different standards obtain different statistical results, so that the relationship between people in each body mass index group and osteoporosis needs to be further refined, analyzed and researched.

In 89 subjects' statistical analyses, the skull thickness, age, body mass index, calcium phosphorus product of the temporal window, which is a statistically relevant factor for osteoporosis, was included in the multivariate Logistic regression analysis for osteoporosis (fig. 9), and the body mass index, which is not a predictive factor for osteoporosis, was correlated with the normal range criteria referenced in the statistical analysis cohort.

To investigate the relationship between gender and osteoporosis, statistical analysis 58 female subjects in 89 cases of this study included ultrasound penetration of the temporal window, age, and body mass index of factors associated with osteoporosis, which is not a risk factor for osteoporosis, in a multiple Logistic regression analysis (fig. 17) on osteoporosis, and which may be related to sample size.

Statistical analysis of 58 female subjects in 89 included skull thickness, age, body mass index of the temporal window, which is a factor associated with osteoporosis, in a multivariate Logistic regression analysis on osteoporosis (figure 22), possibly related to sample size, skull thickness and P-value for osteoporosis =0.059, body mass index and P-value for osteoporosis =0.059, both close to P =0.05, and the American Statistical Association (ASA) statement on Statistical significance and P-value, scientific conclusions and business or policy decisions should not be based solely on whether P-value passed a particular threshold, researchers should use many background factors to obtain scientific inferences, and if sample size was small, large effects may produce unappealing P-values. This set of 58 female data analyses was an internal data study of 89 subjects, who have shown that skull thickness of the temporal window is an independent predictor of osteoporosis, that BMI is statistically correlated with osteoporosis, and that in clinical practice BMI is found to be correlated with osteoporosis, thus incorporating age, BMI and skull thickness into the prediction model shown in fig. 4 (Nomogram 4).

The AUC of the Nomogram1 is 0.848, the benefit range of the curve is determined to be [0.01-0.97], and the correction curve shows good fitting property and shows good prediction capability on osteoporosis.

The AUC of Nomogram2 is 0.829, the benefit range of the curve is determined to be [0.01-1], and the calibration curve shows good fitting property and better prediction capability on osteoporosis.

The AUC of Nomogram3 is 0.847, the benefit range of the curve is determined to be [0.08-0.95], and the correction curve shows good fitting property and better prediction capability on osteoporosis.

The AUC of Nomogram4 is 0.838, the benefit range of the curve is determined to be [0.08-1], and the correction curve shows good fitting performance and better prediction capability on osteoporosis.

The research of the invention shows that the prediction model established by the factors related to the ultrasonic penetrability of the temporal window, the thickness of the skull of the temporal window and the osteoporosis improves the capability of predicting the osteoporosis.

Comparing fig. 14 with the osteoporosis ROC curve of the osteoporosis risk prediction model of fig. 27, it can be seen that Nomogram2 has better specificity than Nomogram1 in the high specificity, low sensitivity region, and Nomogram2 has better sensitivity than Nomogram1 in the high sensitivity, low specificity region; nomogram4 has better specificity than Nomogram3 in the high specificity, low sensitivity region. Through the data distribution calculation and result output module of the osteoporosis risk prediction analysis system, the existing four prediction models are matched by using a distribution calculation program preset in equipment, and the highest score and the highest osteoporosis probability are output after comparison, so that the accuracy of osteoporosis risk prediction can be improved.

According to the osteoporosis risk prediction system, the currently accepted osteoporosis diagnosis standard in the world is used as a reference standard, the osteoporosis risk prediction system is researched, the acquired result of the ultrasonic penetrability of the temporal window and the thickness value of the skull of the temporal window are detected by an ultrasonic detection technology, the relationship between the information and the osteoporosis is analyzed and predicted by combining with the osteoporosis related factors, an osteoporosis risk prediction model is constructed, and the osteoporosis risk is predicted. The results of the ultrasonic penetrability of the temporal window, the thickness value of the skull of the temporal window and the ROC curve display of osteoporosis are independent prediction factors, and the risk of osteoporosis can be well predicted. The result of ultrasonic penetrability, the thickness of the skull of the temporal window and the prediction model constructed by the related osteoporosis prediction factors all improve the AUC value under the ROC curve of the osteoporosis, which shows that the capability of predicting the osteoporosis is improved.

Osteoporosis is a systemic disease manifested by reduction of bone mass and reduction of bone strength, and is a serious manifestation of fracture, which is easily generated at the upper limbs, the radius, the lumbar vertebrae, the upper ends of thighbones and the like, and the body bears large force. Dual energy X-ray absorption can be used to diagnose osteoporosis by measuring the hip bone, which is formed by the fusion of the ilium, ischium and pubis. Iliac bones are flat bones, the skull of the temporal window is flat bones, and osteoporosis can also occur in the skull of the temporal window. The risk of osteoporosis can be predicted by analyzing and researching the related parameters of the skull of the temporal window.

The method has the advantages that no literature report is found, and the method is a new idea for predicting the osteoporosis risk by detecting intracranial blood vessels through an ultrasonic examination technology to judge the ultrasonic penetrability of a temporal window and measuring the skull thickness of the temporal window as data for predicting the osteoporosis risk.

The osteoporosis risk prediction analysis system researched by the invention can judge the ultrasonic penetrability of the temporal window and measure the skull thickness of the temporal window, and the constructed prediction model 1-prediction model 4 can well predict the osteoporosis risk. The osteoporosis risk prediction analysis system researched by the patent is simple and convenient to operate, convenient and fast, has better biological safety than a quantitative CT (computed tomography) and a dual-energy X-ray absorption method, and has better prediction accuracy.

According to the osteoporosis risk prediction analysis system researched by the invention, the model construction and training function module has the function of training the improvement system, and the basic data collection and analysis functions are utilized, so that with the increase of the embodiments, based on the research of the invention, the relation between the ultrasonic penetrability of the temporal window and the skull thickness of the temporal window and the osteoporosis related prediction factors is further refined, analyzed and researched, the number of the prediction model nomograms is increased, the prediction model is optimized, and the prediction capability of the osteoporosis risk prediction system disclosed by the invention is better close to the diagnosis of osteoporosis.

In summary, the present invention provides an osteoporosis risk prediction system based on an ultrasound osteoporosis risk prediction model, which determines ultrasound penetrability of a temporal window and measures skull thickness of the temporal window, inputs basic data tuples including an ultrasound penetrability result of the temporal window, skull thickness values and factors associated with osteoporosis by using a computer programming processing system module function, respectively matches prediction models suitable for application of different basic data tuples by using a pre-programmed allocation calculation program in equipment, obtains scores and predicted osteoporosis probabilities of different basic data tuples according to a scoring table, and outputs a highest score and a highest osteoporosis probability after comparison by an output program, so that the system has good safety and high accuracy of the prediction result.

The above description is only an embodiment of the present invention, and is not intended to limit the scope of the present invention, and the present invention is not limited to the specific embodiments described herein, and all equivalent structural changes made by using the content of the present specification, or the direct or indirect application to the technical field of other related products, are included in the scope of the present invention.

Claims

1. An osteoporosis risk prediction system based on an ultrasonic osteoporosis risk prediction model is characterized in that:

the osteoporosis risk prediction system comprises a functional module which is configured with transcranial detection intracranial arteries and used for judging the ultrasonic penetrability of a temporal window through an ultrasonic temporal window;

2. The ultrasound osteoporosis risk prediction model-based osteoporosis risk prediction system of claim 1, wherein: the method comprises the steps of configuring a pulse wave ultrasonic probe for judging the ultrasonic penetrability of a temporal window through an ultrasonic temporal window, wherein the frequency of the probe is less than or equal to 1.6MHz, the diameter of the probe is less than or equal to 15.6mm, and the detection depth of the probe is more than or equal to 134mm;

the method comprises the step of configuring a high-frequency linear array probe for measuring the skull thickness value of an ultrasonic temporal window, wherein the highest frequency value of the probe frequency is more than or equal to 12MHz, and the resolution ratio of the probe frequency to the thickness of the skull of the temporal window is ensured.

3. The ultrasound osteoporosis risk prediction model-based osteoporosis risk prediction system of claim 2, wherein: the computer programming processing system module for predicting the risk of osteoporosis comprises

The system comprises a prediction factor input system module, a data analysis module and a data analysis module, wherein the prediction factor input system module is used for inputting basic data settings, including a temporal window ultrasonic penetrability result, a skull thickness value and data of factors sex, age, body mass index and calcium phosphorus product associated with osteoporosis, and decomposing the input data into a basic data tuple of the temporal window ultrasonic penetrability result combined with the sex, age, body mass index and calcium phosphorus product osteoporosis related factors and a basic data tuple of the skull thickness value combined with the sex, age, body mass index and calcium phosphorus product osteoporosis related risks;

the model building and training functional module is used for building and training an ultrasonic osteoporosis risk prediction model, and a built-in data result comprises the following steps: single-factor statistical analysis results of factors respectively related to the temporal window ultrasonic penetrability and the osteoporosis by sex, age, body mass index and calcium phosphorus product, multiple Logistic regression analysis results, different ultrasonic prediction models, a collinear chart and an established assigning table result are constructed, and a working characteristic curve result of the osteoporosis subject is drawn; the method comprises the following steps of (1) respectively constructing single-factor statistical analysis results of factors related to osteoporosis by the product of the thickness of the skull of a temporal window, gender, age, body mass index and calcium and phosphorus, multiple Logistic regression analysis results, drawing nomograms by constructing different ultrasonic prediction models, establishing assignment table results and drawing working characteristic curve results of osteoporosis subjects;

and the data distribution calculation and result output module is used for combining the input temporal window ultrasonic penetrability result with data of basic data tuples of factors related to sex, age, body mass index and calcium phosphorus product osteoporosis and basic data tuples of factors related to skull thickness value, sex, age, body mass index and calcium phosphorus product osteoporosis, respectively matching prediction models suitable for different basic data tuples by using a distribution calculation program preset in equipment, respectively calculating the scores and the predicted osteoporosis probability of different basic data tuples in the existing prediction models I to IV according to the assignment table, and outputting the highest score and the highest osteoporosis probability after comparison by the output program.

4. The ultrasound osteoporosis risk prediction model-based osteoporosis risk prediction system of claim 3, wherein: the model building and training function module utilizes the basic data collection and analysis function and takes the constantly updated embodiment database as the basis to increase the number of the alignment charts of the prediction model and optimize the prediction model.

5. The ultrasound osteoporosis risk prediction model-based osteoporosis risk prediction system of claim 4, wherein: the method for constructing the ultrasonic osteoporosis risk prediction model of the model construction and training functional module comprises the following steps

and 2, step: respectively drawing a work characteristic curve of the osteoporosis testee by using the ultrasonic penetrability result of the temporal window and the thickness value of the temporal window skull;

and step 3: building prediction models I to IV by taking the results of the ultrasonic penetrability of the temporal window and prediction factors related to osteoporosis, wherein the thickness values of the skull of the temporal window are respectively related to sex, age, body mass index and calcium-phosphorus product, drawing a nomogram, establishing an assignment table, and drawing a work characteristic curve of the osteoporosis testee.

6. The system of claim 5, wherein the temporal window of ultrasound is the area above the zygomatic bone, between the lateral orbital margin and the tragus, and the thickness of the skull is measured between two cortical hyperechoes at the thinnest part of the temporal window;

7. The ultrasound osteoporosis risk prediction model-based osteoporosis risk prediction system of claim 6, wherein: the system has the basic functions of transcranial Doppler intracranial artery detection and examination and two-dimensional B-ultrasonic superficial organ and blood vessel detection.