Osteoporosis risk prediction system based on an ultrasonic osteoporosis risk prediction model

Abstract

The application discloses an osteoporosis risk prediction system based on an ultrasonic osteoporosis risk prediction model. The system uses the currently recognized osteoporosis diagnosis standard in the world as a control standard, uses the results of ultrasonic examination technology to detect the ultrasonic penetration of a temporal window and the thickness of the skull of the temporal window, combines the related factors associated with osteoporosis, analyzes and predicts the relationship with osteoporosis, constructs different osteoporosis risk prediction models, draws nomograms, establishes a scoring table, uses the computer programming processing system function of the osteoporosis risk prediction system, respectively matches the prediction models suitable for different basic data tuples, respectively calculates, obtains the scores of different basic data tuples and the predicted osteoporosis probability according to the scoring table, and outputs the highest score and the highest osteoporosis probability after comparison by an output program, so that the prediction accuracy is good.

Description

Technical Field

[0001] This invention relates to the field of osteoporosis risk prediction technology, and in particular to an osteoporosis risk prediction system based on an ultrasound osteoporosis risk prediction model. Background Technology

[0002] Osteoporosis is a systemic disease characterized by decreased bone mass and strength, altered bone fibrous structure, and increased fracture risk. The main indicator of bone mass is bone mineral density (BMD). Osteoporosis is common in middle-aged and elderly people, especially postmenopausal women. With increasing life expectancy, the incidence of osteoporosis and related fractures has risen significantly, leading to increased disability and mortality rates, and imposing a heavy economic burden on families and society. Therefore, early prediction of osteoporosis risk is of great importance for early diagnosis and treatment of osteoporotic osteoporosis and prevention of osteoporotic fractures.

[0003] Currently, there has been no breakthrough in measuring bone strength—a key concept in the definition of osteoporosis—making it difficult to meet the clinical needs for non-invasive and convenient methods. Therefore, bone mineral density (BMD) is used as a substitute indicator for bone strength. BMD measurement technology mainly utilizes the principle that X-rays attenuate to varying degrees as they pass through different media to analyze bone mineral content, bone density, and body composition. Currently, commonly used BMD measurement techniques include quantitative computed tomography (QCT). Theoretically, QCT can measure any part of the body with accurate data. However, in practice, the vast majority of measurements are concentrated in the lumbar spine (L1-3), and the measurement is affected by vertebral fat. With increasing age, fat content increases, and due to the "partial volume effect," the measured BMD value is lower than the actual value, making bone loss appear more significant than it actually is. Dual-energy X-ray ablation (DEXA) can address the fat issue, but it involves high radiation doses and can reduce precision, so it is rarely used clinically. QCT is not affected by bone size and should be suitable for children, but its high radiation dose has limited its widespread application.

[0004] Dual-energy X-ray absorptiometry (DXA) uses two X-ray images obtained from X-rays of different energies. Energy subtraction is used to eliminate the influence of soft tissue on bone mineral density measurement. Through instrument data processing and solving equations, the bone mineral density value of the measured site is determined. The radiation dose received by the human body is low, making it a standard for diagnosing osteoporosis. The World Health Organization (WHO) also recommends hip and lumbar spine bone mineral density measurements using dual-energy X-ray absorptiometry as an objective diagnostic standard for osteoporosis.

[0005] Quantitative Ultrasound Bone Measurements (QUS) works by utilizing the attenuation of ultrasound waves as they pass through body tissues. The amount of attenuation depends on tissue characteristics, and in bone tissue, it depends on the bone's elasticity model and bone mineral density. Bone mineral density can be calculated by measuring changes in ultrasound velocity of sound (SOS), broadband ultrasound attenuation (BUA), and bone quality index (BQI). The reliability of these parameters and their correlation with other methods require further investigation. Ultrasound bone densitometers are less expensive and radiation-free than X-ray bone densitometers, making them suitable for screening and general populations. They also offer valuable reference and guidance for children's physiological development and for the prevention of bone damage and fracture risks in the elderly. Summary of the Invention

[0006] The purpose of this invention is to provide an osteoporosis risk prediction system based on an ultrasound osteoporosis risk prediction model.

[0007] Therefore, the technical solution of the present invention is as follows:

[0008] An osteoporosis risk prediction system based on an ultrasound osteoporosis risk prediction model, the osteoporosis risk prediction system includes a functional module configured with transcranial detection of intracranial arteries, used to determine the ultrasound permeability of the temporal window via ultrasound.

[0009] It is equipped with a temporal window skull thickness measurement module, which is used to measure skull thickness using two-dimensional linear array B-mode ultrasound technology;

[0010] It is equipped with a computer programming processing system module for osteoporosis risk prediction.

[0011] Furthermore, the method includes configuring a pulsed wave ultrasound probe for determining the ultrasound penetration of the temporal window via ultrasound, wherein the probe frequency is ≤1.6MHz, the probe diameter is ≤15.6mm, and the probe detection depth is ≥134mm;

[0012] The method includes configuring a high-frequency linear array probe for measuring the skull thickness of the temporal window using ultrasound. The highest frequency value of the probe is ≥12MHz to ensure resolution of the skull thickness of the temporal window.

[0013] Furthermore, the computer-programmed processing system module for osteoporosis risk prediction includes:

[0014] The predictive factor input system module is used to input basic data settings, including temporal window ultrasound penetration results, skull thickness values, and data on osteoporosis-related factors such as gender, age, body mass index, and calcium-phosphorus product. The system decomposes the input data into a basic data tuple combining temporal window ultrasound penetration results with gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors, and a basic data tuple combining skull thickness values ​​with gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors.

[0015] The model building and training module is used for building and training ultrasound osteoporosis risk prediction models. Its built-in data results include: univariate statistical analysis results of temporal window ultrasound penetration and its correlation with factors such as gender, age, body mass index, and calcium-phosphorus product osteoporosis; multivariate logistic regression analysis results; nomograms and scoring tables of different ultrasound prediction models; and receiver operating characteristic (ROC) curves for osteoporosis. It also includes: univariate statistical analysis results of temporal window skull thickness and its correlation with factors such as gender, age, body mass index, and calcium-phosphorus product osteoporosis; multivariate logistic regression analysis results; nomograms and scoring tables of different ultrasound prediction models; and receiver operating characteristic (ROC) curves for osteoporosis.

[0016] The data allocation calculation and result output module is used to combine the input temporal window ultrasound penetration results with the basic data tuples of gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors, and the skull thickness value with the basic data tuples of gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors. Using the pre-programmed allocation calculation program in the device, it matches the prediction model suitable for different basic data tuples and calculates the scores and predicted osteoporosis probabilities for different basic data tuples in the existing prediction models 1-4 according to the scoring table. The output program compares the scores and outputs the highest score and the highest osteoporosis probability.

[0017] Furthermore, the model building and training module utilizes the basic data collection and analysis function to increase the number of prediction model nodal plots and optimize the prediction model based on the continuously updated example database.

[0018] Furthermore, the method for constructing the ultrasound osteoporosis risk prediction model in the model construction and training functional module is as follows:

[0019] Step 1: Collect ultrasound probe data through the temporal window to detect intracranial blood vessels, and determine the ultrasound penetration results of the temporal window and the skull thickness value detected by the high-frequency linear array probe through the temporal window.

[0020] Step 2: Using the temporal window ultrasound penetration results and temporal window skull thickness values, plot the receiver operating characteristic curves (ROCs) for osteoporosis.

[0021] Step 3: Incorporate the results of temporal window ultrasound penetration and the temporal window skull thickness values, and construct predictive models 1-4 for osteoporosis related to gender, age, body mass index, and calcium-phosphorus product. Plot nomograms, establish scoring tables, and plot receiver operating characteristic curves (ROCs) for osteoporosis.

[0022] Furthermore, the temporal window of the ultrasound refers to the area above the cheekbone, between the lateral orbital margin and the tragus, and the skull thickness is the value measured between the two layers of strong echoes of the cortical bone at the thinnest point of the temporal window.

[0023] The criteria for judging the non-penetrating nature of temporal window ultrasound refer to the inability to detect any signals from intracranial arteries;

[0024] The criteria for judging the permeability of temporal window ultrasound refer to the ability to successfully detect the signal of any intracranial artery.

[0025] Furthermore, the system also has basic transcranial Doppler intracranial artery detection and examination functions, as well as two-dimensional B-mode ultrasound superficial organ and blood vessel detection functions.

[0026] Compared with existing technologies, this invention provides an osteoporosis risk prediction system based on an ultrasound osteoporosis risk prediction model. It incorporates the results of temporal window ultrasound penetration and temporal window skull thickness values, along with predictive factors related to osteoporosis such as gender, age, body mass index, and calcium-phosphorus product, to construct prediction models one through four. Nomograms are plotted, scoring tables are established, and receiver operating characteristic (ROC) curves for osteoporosis are generated. This invention offers a novel approach to osteoporosis risk prediction.

[0027] The system inputs temporal window ultrasound penetration results, skull thickness values, and baseline data on osteoporosis-related factors such as gender, age, body mass index, and calcium-phosphorus product. The system decomposes the input data into two sets: one combining ultrasound penetration results with baseline data on osteoporosis-related factors (gender, age, body mass index, and calcium-phosphorus product), and the other combining skull thickness values ​​with baseline data on osteoporosis-related factors (gender, age, body mass index, and calcium-phosphorus product). Using a pre-programmed allocation calculation program, the system matches different prediction models suitable for each baseline data set. It calculates scores and predicted osteoporosis probabilities for each baseline data set within existing prediction models one through four, based on a scoring table. The output program compares these scores and outputs the highest score and the highest osteoporosis probability. The system demonstrates good biocompatibility and high accuracy in prediction results. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of temporal window skull thickness measurement.

[0029] Figure 2 ROC curves for temporal window ultrasound penetration and osteoporosis in 89 subjects;

[0030] Figure 3 ROC curves for temporal window skull thickness and osteoporosis in 89 subjects;

[0031] Figure 4 Multivariate logistic regression analysis of osteoporosis-related factors between temporal window ultrasound penetration and osteoporosis in 89 subjects;

[0032] Figure 5 Nomogram 1 was constructed to predict the temporal window ultrasound penetration, age, and calcium-phosphorus product of 89 subjects.

[0033] Figure 6 ROC curves for osteoporosis in Nomogram 1 of 89 subjects;

[0034] Figure 7 Osteoporosis correction curves for Nomogram 1 in 89 subjects;

[0035] Figure 8 Osteoporosis determination curves for Nomogram 1 in 89 subjects;

[0036] Figure 9 Multivariate logistic regression analysis of osteoporosis-related factors in temporal window skull thickness and osteoporosis in 89 subjects;

[0037] Figure 10 Nomogram 2 is a predictive model 2 constructed from the temporal window skull thickness, age, and calcium-phosphorus product of 89 subjects.

[0038] Figure 11 ROC curves for osteoporosis in Nomogram 2 of 89 subjects;

[0039] Figure 12 Osteoporosis correction curves for Nomogram 2 in 89 subjects;

[0040] Figure 13 Osteoporosis determination curves for Nomogram 2 in 89 subjects;

[0041] Figure 14 A comparison of ROC curves for osteoporosis using Nomogram 1 and Nomogram 2 in 89 subjects.

[0042] Figure 15ROC curves for temporal window ultrasound penetration and osteoporosis in 58 female subjects;

[0043] Figure 16 ROC curves for temporal window skull thickness and osteoporosis in 58 female subjects;

[0044] Figure 17 Multivariate logistic regression analysis of osteoporosis-related factors associated with temporal window ultrasound penetration in 58 female subjects.

[0045] Figure 18 Nomogram 3 was constructed to predict temporal window ultrasound penetration and age in 58 female subjects.

[0046] Figure 19 ROC curves for osteoporosis in Nomogram 3 of 58 female subjects;

[0047] Figure 20 Osteoporosis-corrected curves for Nomogram 3 in 58 female subjects;

[0048] Figure 21 Osteoporosis determination curves for Nomogram 3 in 58 female subjects;

[0049] Figure 22 Multivariate logistic regression analysis of osteoporosis-related factors in temporal window skull thickness and osteoporosis in 58 female subjects.

[0050] Figure 23 ROC curves for osteoporosis in Nomogram 4 of 58 female subjects;

[0051] Figure 24 Osteoporosis-corrected curves for Nomogram 4 in 58 female subjects;

[0052] Figure 25 Osteoporosis determination curves for Nomogram 4 in 58 female subjects;

[0053] Figure 26 A comparison of ROC curves for osteoporosis using Nomogram 3 and Nomogram 4 in 58 female subjects. Detailed Implementation

[0054] The present invention will be further described below with reference to specific embodiments and accompanying drawings, but the following embodiments are by no means intended to limit the present invention.

[0055] This invention provides an osteoporosis risk prediction system based on an ultrasound osteoporosis risk prediction model, wherein the osteoporosis risk prediction analysis system includes a functional module configured with transcranial intracranial artery detection for determining the ultrasound permeability of the temporal window via ultrasound.

[0056] Equipped with a temporal window skull thickness measurement module, used to measure skull thickness using two-dimensional linear array ultrasound technology. The measurement method is as follows: Figure 1 As shown;

[0057] It is equipped with a computer programming processing system module for osteoporosis risk prediction.

[0058] It should be noted that the osteoporosis risk prediction system is equipped with a pulsed wave ultrasound probe for determining the transmissibility of the temporal window ultrasound via ultrasound. The probe frequency is ≤1.6MHz, the probe diameter is ≤15.6mm, and the probe detection depth is ≥134mm.

[0059] The osteoporosis risk prediction system is equipped with a high-frequency linear array probe for measuring the skull thickness of the ultrasound temporal window. The probe has a maximum frequency of ≥12MHz to ensure resolution of the skull thickness of the temporal window.

[0060] Further explanation is needed regarding the computer programming processing system module for osteoporosis risk prediction, which includes a predictive factor input system module for inputting basic data settings, including temporal window ultrasound penetration results, skull thickness values, and data on osteoporosis-related factors such as gender, age, body mass index, and calcium-phosphorus product. The system decomposes the input data into a basic data tuple combining temporal window ultrasound penetration results with gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors, and a basic data tuple combining skull thickness values ​​with gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors.

[0061] The model building and training module is used for building and training ultrasound osteoporosis risk prediction models. Its built-in data results include: univariate statistical analysis results of temporal window ultrasound penetration and its correlation with factors related to gender, age, body mass index, and calcium-phosphorus product osteoporosis; multivariate logistic regression analysis results; nomograms and scoring tables of different ultrasound prediction models; and osteoporosis subject operating characteristic curves. It also includes univariate statistical analysis results of temporal window skull thickness and its correlation with factors related to gender, age, body mass index, and calcium-phosphorus product osteoporosis; multivariate logistic regression analysis results; nomograms and scoring tables of different ultrasound prediction models; and osteoporosis subject operating characteristic curves. The model building and training module utilizes basic data collection and analysis functions, based on a continuously updated example database, to increase the number of prediction model nomograms and optimize the prediction models.

[0062] The data allocation calculation and result output module is used to combine the input temporal window ultrasound penetration results with the basic data tuples of gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors, and the skull thickness value with the basic data tuples of gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors. Using the pre-programmed allocation calculation program in the device, it matches the prediction model suitable for different basic data tuples and calculates the scores and predicted osteoporosis probabilities for different basic data tuples in the existing prediction models 1-4 according to the scoring table. The output program compares the scores and outputs the highest score and the highest osteoporosis probability.

[0063] The model building and training module utilizes basic data collection and analysis functions to increase the number of prediction model nodal plots and optimize the prediction model as the number of embodiments increases.

[0064] The method for constructing the ultrasound osteoporosis risk prediction model in the model construction and training module is as follows:

[0065] Step 1: Collect ultrasound probe data through the temporal window to detect intracranial blood vessels, determine the ultrasound penetration of the temporal window, and measure the skull thickness of the temporal window using a high-frequency linear array probe.

[0066] Step 2: Using the temporal window ultrasound penetration results and temporal window skull thickness values, plot the receiver operating characteristic curves (ROCs) for osteoporosis.

[0067] Step 3: Incorporate the results of temporal window ultrasound penetration and the temporal window skull thickness values, and construct predictive models 1-4 for osteoporosis related to gender, age, body mass index, and calcium-phosphorus product. Plot nomograms, establish scoring tables, and plot receiver operating characteristic curves (ROCs) for osteoporosis.

[0068] The temporal window of the ultrasound refers to the area above the cheekbone, between the lateral orbital margin and the tragus, and the skull thickness is the value measured between the two layers of strong echoes of the cortical bone at the thinnest point of the temporal window.

[0069] The criteria for judging the non-penetrating nature of temporal window ultrasound refer to the inability to detect any signals from intracranial arteries;

[0070] The criteria for judging the permeability of temporal window ultrasound refer to the ability to successfully detect the signal of any intracranial artery.

[0071] The system combines basic transcranial Doppler intracranial artery detection and examination functions with two-dimensional B-mode ultrasound for superficial organ and blood vessel detection.

[0072] Example:

[0073] The study included 89 participants. The inclusion criteria were those with clinical symptoms of cerebrovascular disease and complete data. Those with severe liver or kidney dysfunction were excluded. Among them, there were 58 females and 31 males.

[0074] Diagnostic criteria for osteoporosis:

[0075] 1. Spinal fragility fracture;

[0076] 2. Quantitative computed tomography (QCT): Mean absolute bone mineral density (BMD) < 80 mg / cm³; 3. Dual-energy X-ray absorptiometry (DXA): DXA measures the T-score, where T-score = (Measured value - Peak bone mineral density of normal young adults of the same race and sex) / Standard deviation (SD) of peak bone mineral density of normal young adults of the same race and sex. Osteoporosis: T-score ≤ -2.5 SD.

[0077] Non-osteoporosis includes both osteopenia and normal bone mass.

[0078] Referring to industry-recognized or internationally accepted standards or guidelines, when grouping relevant risk factors with normal value range standards during statistical analysis, they are divided into normal group and abnormal group. The abnormal group includes data that are greater than or less than the normal value range.

[0079] The criteria for grouping body mass index (BMI) are as follows: According to WHO standards, the unit is kg / m². Normal is defined as 18.5 ≤ BMI < 25, underweight as BMI < 18.5, overweight as BMI ≥ 25, and obese. Individuals are divided into a normal group and an abnormal group. The normal group is defined as 18.5 ≤ BMI < 25, and the abnormal group is defined as BMI < 18.5. A BMI ≥ 25 includes obese individuals.

[0080] Grouping criteria for calcium-phosphorus product: Unit: mg / dl. Normal value: 35 < calcium-phosphorus product < 40. Abnormal value: calcium-phosphorus product ≤ 35, calcium-phosphorus product ≥ 40. The normal group is 35 < calcium-phosphorus product < 40, and the abnormal group includes calcium-phosphorus product ≤ 35 and calcium-phosphorus product ≥ 40.

[0081] Calcium and phosphorus were determined using a Beckman Coulter AU5800 fully automated biochemical analyzer, with calcium assay kits (azoarsine III method) and phosphorus assay kits (phosphomolybdate method).

[0082] The statistical results of the 89 examinees are as follows:

[0083] A table showing the relationship between baseline data and osteoporosis was drawn for 89 subjects (Table 1).

[0084] Table 1. Baseline data of 89 subjects.

[0085]

[0086] The table above shows that, apart from gender, age, penetrability, BMI, calcium-phosphorus product, and skull thickness were all statistically associated with osteoporosis (P<0.05).

[0087] ROC curves for temporal window ultrasound penetration and osteoporosis were plotted in 89 subjects. The area under the curve (AUC) was 0.738. Figure 2 ).

[0088] ROC curves were plotted for temporal window skull thickness and osteoporosis in 89 subjects, with an AUC of 0.707. Figure 3 ).

[0089] Ultrasound penetration through the temporal window is strongly correlated with the thickness of the cranial temporal window, both of which are predictors of osteoporosis. Logistic regression is prone to multicollinearity, so the statistical analysis was performed separately from another item.

[0090] The age, penetrability, BMI, and calcium-phosphorus product of 89 subjects were used as independent variables, and osteoporosis was used as the dependent variable. These data were included in a multivariate logistic regression analysis for osteoporosis. Figure 4 The results showed that ultrasound penetration, age, and calcium-phosphorus product were predictive factors for osteoporosis (P<0.05).

[0091] The nomogram 1 (constructing prediction model 1 using ultrasound penetration, age, and calcium-phosphorus product as predictive factors for osteoporosis) is shown. Figure 5 ).

[0092] Table 2 is the osteoporosis scoring table for Nomogram 1.

[0093] Table 2 Osteoporosis Scoring Table for Nomogram 1

[0094] The osteoporosis ROC curve for Nomogram 1 is shown in the figure. Figure 6 The AUC is 0.848.

[0095] The osteoporosis correction curve for Nomogram 1 is shown below. Figure 7 The Hosmer-Lemeshow test showed a p-value of 0.986.

[0096] The osteoporosis determination curve for Nomogram 1 is shown in [reference needed]. Figure 8 The benefit range is [0.01-0.97].

[0097] The age, BMI, calcium-phosphorus product, and skull thickness of 89 subjects were used as independent variables, and osteoporosis was used as the dependent variable. These parameters were included in a multivariate logistic regression analysis for osteoporosis. Figure 9 The results showed that age, calcium-phosphorus product, and skull thickness were predictors of osteoporosis (P<0.05).

[0098] A nomogram 2 was constructed to predict osteoporosis using skull thickness, age, and the calcium-phosphorus product as predictive factors. Figure 10 Table 3 is the osteoporosis scoring table for Nomogram 2.

[0099] Table 3. Osteoporosis Scoring Table for Nomogram 2

[0100] The osteoporosis ROC curve for Nomogram 2 is shown below. Figure 11 The AUC is 0.829.

[0101] The osteoporosis correction curve for Nomogram 2 is shown below. Figure 12 The Hosmer-Lemeshow test showed a p-value of 0.853.

[0102] The osteoporosis determination curve for Nomogram 2 is shown in [reference needed]. Figure 13 The benefit range is [0.01-1].

[0103] A comparison of the osteoporosis ROC curves of Nomogram 1 and Nomogram 2 is shown in the figure. Figure 14

[0104] The statistical results of 58 female subjects out of 89 cases are as follows:

[0105] Of the 89 cases, 58 were female subjects. A table showing the relationship between baseline data and osteoporosis was drawn (Table 4).

[0106] Table 4. Baseline data of 58 female subjects.

[0107] The table above shows that, in addition to the calcium-phosphorus product, age, penetrability, BMI, and skull thickness are all statistically associated with osteoporosis (P<0.05).

[0108] ROC curves were plotted for temporal window ultrasound penetration and osteoporosis in 58 female subjects, with an AUC of 0.766 ( ). Figure 15 ).

[0109] ROC curves were plotted for temporal window skull thickness and osteoporosis in 58 female subjects. The AUC was 0.723. Figure 16 ).

[0110] Ultrasound penetration through the temporal window is strongly correlated with the thickness of the cranial temporal window, both of which are risk factors for osteoporosis. Logistic regression analysis showed multicollinearity. Therefore, the statistical analysis was conducted separately from another study.

[0111] The age, penetrability, BMI, and calcium-phosphorus product of 58 female subjects were used as independent variables, and osteoporosis was used as the dependent variable. These data were included in a multivariate logistic regression analysis for osteoporosis. Figure 17 The results showed that ultrasound penetration, age, and calcium-phosphorus product were predictive factors for osteoporosis (P<0.05).

[0112] A nomogram 3 was constructed to predict osteoporosis using ultrasound penetrability and age as predictive factors. Figure 18 Table 5 is the osteoporosis scoring table for Nomogram 3.

[0113] Table 5. Osteoporosis Scoring Table for Nomogram 3

[0114] The osteoporosis ROC curve for Nomogram 3 is shown below. Figure 19 The AUC is 0.847.

[0115] The osteoporosis correction curve for Nomogram 3 is shown below. Figure 20 The Hosmer-Lemeshow test p = 0.877.

[0116] The osteoporosis determination curve for Nomogram 3 is shown in [reference needed]. Figure 21The benefit range is [0.08-0.95].

[0117] The age, BMI, and skull thickness of 58 female subjects were used as independent variables, and osteoporosis was used as the dependent variable. These subjects were included in a multivariate logistic regression analysis for osteoporosis. Figure 22 The results showed that age (P=0.016), BMI (P=0.059), and skull thickness (P=0.059) were all positive. A nomogram (Nomogram 4) was constructed to build a predictive model for the osteoporosis predictors skull thickness, age, and BMI. Table 6 shows the osteoporosis scoring table for Nomogram 4.

[0118] Table 6. Osteoporosis Scoring Table for Nomogram 4

[0119] ROC curve for osteoporosis in Nomogram 4 Figure 23 The AUC is 0.838.

[0120] The osteoporosis correction curve for Nomogram 4 is shown below. Figure 24 The Hosmer-Lemeshow test showed a p-value of 0.933.

[0121] The osteoporosis determination curve for Nomogram 4 is shown in [reference needed]. Figure 25 The benefit range is [0.08-1].

[0122] A comparison of the osteoporosis ROC curves of Nomogram 3 and Nomogram 4 is shown in the figure. Figure 26 .

[0123] The statistical results of the examples are analyzed as follows:

[0124] Multiple studies have shown that gender is associated with osteoporosis, especially that the risk of osteoporosis increases significantly after menopause in women. The results in Table 1 of this example show that the gender factor did not reach a statistical association with osteoporosis, which may be related to the sample size or the fact that the univariate analysis did not calibrate other confounding factors.

[0125] The calcium-phosphorus product reflects the state of calcium and phosphorus metabolism in the body that affects bone mass. The results in Table 4 of this example show that the calcium-phosphorus product did not reach a statistical association with osteoporosis, which may be related to the sample size or the fact that the univariate analysis did not calibrate other confounding factors.

[0126] The AUC of translucency and osteoporosis ROC in temporal window ultrasound of 89 subjects was 0.738. Figure 2 The AUC > 0.7 indicates that temporal window ultrasound penetration has a relatively good predictive accuracy for osteoporosis.

[0127] The AUC of temporal window skull thickness and osteoporosis ROC in 89 subjects was 0.707 ( Figure 3 The AUC > 0.7 indicates that temporal window skull thickness has a relatively good predictive accuracy for osteoporosis.

[0128] The AUC of translucency and osteoporosis ROC in 58 female subjects via temporal window ultrasound was 0.766. Figure 15 The AUC > 0.7 indicates that the prediction accuracy is relatively good.

[0129] The AUC of temporal window skull thickness and osteoporosis ROC in 58 female subjects was 0.723 ( Figure 16 The AUC > 0.7 indicates that the prediction accuracy is relatively good.

[0130] This invention demonstrates that temporal window ultrasound penetration and temporal window skull thickness are both statistically associated with osteoporosis and are independent predictors of osteoporosis.

[0131] In the statistical analysis of 89 subjects, the temporal window ultrasound penetration, age, body mass index, and calcium-phosphorus product, which are statistically associated factors for osteoporosis, were included in the multivariate logistic regression analysis of osteoporosis in the 89 subjects. Figure 4 Body mass index (BMI) is not a predictor of osteoporosis in multivariate analysis. It is related to the normal range standard referenced when grouping people in the statistical analysis. Different statistical results are obtained for normal groups and abnormal groups distinguished by different standards. Therefore, further detailed analysis is needed to study the relationship between each BMI group and osteoporosis.

[0132] In the statistical analysis of 89 subjects, factors statistically associated with osteoporosis, such as temporal window skull thickness, age, body mass index, and calcium-phosphorus product, were included in the multivariate logistic regression analysis for osteoporosis. Figure 9 Body mass index (BMI) is not a predictor of osteoporosis; its significance is related to the normal range standards referenced during statistical analysis grouping.

[0133] To explore the relationship between gender and osteoporosis, statistical analysis was conducted on 58 female subjects out of 89 participants in this study. Factors associated with osteoporosis, such as temporal window ultrasound penetration, age, and body mass index, were included in a multivariate logistic regression analysis for osteoporosis. Figure 17 Body mass index (BMI) is not a risk factor for osteoporosis, which may be related to sample size.

[0134] Statistical analysis was conducted on 58 female subjects out of 89 examinees. Factors associated with osteoporosis, including temporal window skull thickness, age, and body mass index, were included in a multivariate logistic regression analysis for osteoporosis. Figure 22The p-values ​​for skull thickness and osteoporosis, and body mass index (BMI) and osteoporosis, are both close to P=0.05, possibly related to sample size. The American Statistical Association (ASA) states that scientific conclusions and business or policy decisions should not be based solely on whether a p-value passes a specific threshold. Researchers should utilize many background factors to arrive at scientific inferences. If the sample size is small, a large effect may produce a less impressive p-value. This study, analyzing data from 58 women, is based on internal data from 89 subjects. This patented study has shown that temporal window skull thickness is an independent predictor of osteoporosis, and BMI is statistically associated with osteoporosis. Clinical practice has also found a correlation between BMI and osteoporosis; therefore, age, BMI, and skull thickness were included in the predictive model. Figure 4 (Nomogram 4)

[0135] The AUC of Nomogram 1 was 0.848, the benefit range of the determination curve was [0.01-0.97], and the calibration curve showed good fit, indicating good predictive ability for osteoporosis.

[0136] The AUC of Nomogram 2 was 0.829, the benefit range of the determination curve was [0.01-1], and the calibration curve showed good fit, indicating good predictive ability for osteoporosis.

[0137] The AUC of Nomogram 3 was 0.847, the benefit range of the determination curve was [0.08-0.95], and the calibration curve showed good fit, indicating good predictive ability for osteoporosis.

[0138] The AUC of Nomogram 4 was 0.838, the benefit range of the determination curve was [0.08-1], and the calibration curve showed good fit, indicating good predictive ability for osteoporosis.

[0139] This invention demonstrates that the predictive models constructed from temporal window ultrasound penetration, temporal window skull thickness, and osteoporosis-related predictive factors all improve the ability to predict osteoporosis.

[0140] Figure 14 and Figure 26A comparison of the osteoporosis risk prediction model's ROC curves shows that Nomogram 2 has better specificity than Nomogram 1 in the high-specificity, low-sensitivity region, and Nomogram 2 has better sensitivity than Nomogram 1 in the high-sensitivity, low-specificity region; Nomogram 4 has better specificity than Nomogram 3 in the high-specificity, low-sensitivity region. Through the data allocation calculation and result output module of the osteoporosis risk prediction analysis system of this invention, using a pre-programmed allocation calculation program within the device, the four existing prediction models are matched, and the highest score and the highest osteoporosis probability are output after comparison, which can improve the accuracy of osteoporosis risk prediction.

[0141] This invention uses the currently recognized diagnostic criteria for osteoporosis as a control standard to develop an osteoporosis risk prediction system. It employs ultrasound detection technology to obtain temporal window ultrasound penetration and temporal window skull thickness values, combining these with related factors to analyze and predict the relationship with osteoporosis, constructing an osteoporosis risk prediction model to predict the risk of osteoporosis. The temporal window ultrasound penetration, temporal window skull thickness, and ROC curves for osteoporosis all show that they are independent predictive factors and can effectively predict the risk of osteoporosis. The prediction model constructed from ultrasound penetration, temporal window skull thickness, and related osteoporosis predictive factors all improved the AUC value under the osteoporosis ROC curve, indicating an improved ability to predict osteoporosis.

[0142] Osteoporosis is a systemic disease characterized by decreased bone mass and strength, with severe cases leading to fractures. Fractures commonly occur in areas of the body that bear significant stress, such as the upper limbs, radius, lumbar vertebrae, and upper femur. Dual-energy X-ray absorptiometry (DXA) can diagnose osteoporosis by measuring the hip bone, which is formed by the fusion of the ilium, ischium, and pubis. The ilium is a flat bone, as is the skull of the temporal window; osteoporosis can also occur in the temporal window skull. Analyzing relevant parameters of the temporal window skull can predict the risk of osteoporosis.

[0143] No literature was found reporting that using ultrasound examination technology to detect intracranial blood vessels, determine the ultrasound penetration of the temporal window, and measure the skull thickness of the temporal window as data to predict osteoporosis risk is a new approach to predicting osteoporosis risk.

[0144] The osteoporosis risk prediction and analysis system developed in this invention can determine the ultrasound penetration of the temporal window and measure the skull thickness of the temporal window. Prediction models 1-4 constructed in this invention can all accurately predict the risk of osteoporosis. This osteoporosis risk prediction and analysis system is simple and convenient to operate, has better biocompatibility than quantitative CT and dual-energy X-ray absorptiometry, and exhibits excellent predictive accuracy.

[0145] The osteoporosis risk prediction and analysis system studied in this invention has a model building and training module that can train and improve the system. Utilizing basic data collection and analysis functions, and with the increase of embodiments, based on the research of this invention, the relationship between temporal window ultrasound penetration, temporal window skull thickness and osteoporosis-related predictive factors is further refined and analyzed. The number of prediction model nomograms is increased and the prediction model is optimized, so that the predictive ability of the osteoporosis risk prediction system of this invention can better approach the diagnosis of osteoporosis.

[0146] In summary, this invention provides an osteoporosis risk prediction system based on an ultrasound osteoporosis risk prediction model. It determines the ultrasound penetration of the temporal window and measures the skull thickness within the temporal window. Utilizing a computer programming processing module, it inputs a basic data set including the temporal window ultrasound penetration result, skull thickness value, and factors associated with osteoporosis. Using a pre-programmed allocation calculation program within the device, it matches prediction models suitable for different basic data sets. Based on a scoring table, it obtains scores and predicted osteoporosis probabilities for different basic data sets. The output program compares these scores and outputs the highest score and the highest osteoporosis probability. The system offers good safety and high prediction accuracy.

[0147] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention, nor does it limit the invention to the specific implementation described. Any equivalent structural transformations made using the content of the present invention specification, or direct or indirect applications in the technical fields of other related products, are similarly included within the patent protection scope of the present invention.

Claims

1. An osteoporosis risk prediction system based on an ultrasound osteoporosis risk prediction model, characterized in that: The osteoporosis risk prediction system is equipped with a transcranial intracranial artery detection module, which is used to determine the ultrasound permeability of the temporal window via ultrasound. It is equipped with a temporal window skull thickness measurement module, which is used to measure skull thickness using two-dimensional linear array B-mode ultrasound technology; It is equipped with a computer programming processing system module for osteoporosis risk prediction; The computer programming processing system module uses the temporal window ultrasound penetration results and temporal window skull thickness values ​​to plot receiver operating characteristic curves (ROCs) for osteoporosis. It incorporates the temporal window ultrasound penetration results and temporal window skull thickness values ​​with predictive factors related to osteoporosis such as gender, age, body mass index, and calcium-phosphorus product to construct prediction models one through four, plots nomograms, and establishes scoring tables. The computer programming processing system module for osteoporosis risk prediction includes: The predictive factor input system module is used to input basic data settings, including temporal window ultrasound penetration results, skull thickness values, and data on osteoporosis-related factors such as gender, age, body mass index, and calcium-phosphorus product. The system decomposes the input data into a basic data tuple combining temporal window ultrasound penetration results with gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors, and a basic data tuple combining skull thickness values ​​with gender, age, body mass index, and calcium-phosphorus product osteoporosis-related factors. The model building and training module is used to build and train ultrasound osteoporosis risk prediction models, obtaining prediction models one through four. Its built-in data results include: univariate statistical analysis results of temporal window ultrasound penetration and its correlation with factors related to gender, age, body mass index, and calcium-phosphorus product osteoporosis; multivariate logistic regression analysis results; results of different constructed ultrasound prediction models, plotted nomograms, and established scoring tables; and results of osteoporosis subject operating characteristic curves. It also includes univariate statistical analysis results of temporal window skull thickness and its correlation with factors related to gender, age, body mass index, and calcium-phosphorus product osteoporosis; multivariate logistic regression analysis results; results of different constructed ultrasound prediction models, plotted nomograms, and established scoring tables; and results of osteoporosis subject operating characteristic curves. The data allocation calculation and result output module is used to combine the input temporal window ultrasound penetration results with basic data tuples related to osteoporosis factors such as gender, age, body mass index, and calcium-phosphorus product, and skull thickness values ​​with the basic data tuples related to osteoporosis factors such as gender, age, body mass index, and calcium-phosphorus product. Using a pre-programmed allocation calculation program within the device, it matches prediction models suitable for different basic data tuples. According to the scoring table, it obtains scores and predicted osteoporosis probabilities for different basic data tuples. The output program compares and outputs the highest score and the highest osteoporosis probability. Among them, prediction model 1 is a model constructed based on temporal window ultrasound penetration, age, and calcium-phosphorus product; prediction model 2 is a model constructed based on temporal window skull thickness, age, and calcium-phosphorus product; prediction model 3 is a model constructed based on temporal window ultrasound penetration and age for women; prediction model 4 is a model constructed based on temporal window skull thickness, age, and BMI for women.

2. The osteoporosis risk prediction system based on the ultrasound osteoporosis risk prediction model according to claim 1, characterized in that: Configure a pulsed wave ultrasound probe for determining the transmissibility of the temporal window ultrasound via ultrasound, with a probe frequency ≤1.6MHz, probe diameter ≤15.6mm, and probe detection depth ≥134mm; A high-frequency linear array probe is configured to measure the skull thickness of the temporal window using ultrasound. The highest frequency value of the high-frequency linear array probe is ≥12MHz to ensure resolution of the skull thickness of the temporal window.

3. The osteoporosis risk prediction system based on the ultrasound osteoporosis risk prediction model according to claim 1, characterized in that: The model building and training module utilizes basic data collection and analysis functions to increase the number of prediction model nodal plots and optimize the prediction model based on a continuously updated database of example implementations.

4. The osteoporosis risk prediction system based on the ultrasound osteoporosis risk prediction model according to claim 3, characterized in that: The method for constructing the ultrasound osteoporosis risk prediction model in the model construction and training module is as follows: Step 1: Collect the results of ultrasound penetration of the temporal window obtained by ultrasound probe through the temporal window to determine intracranial blood vessels, and the skull thickness value of the temporal window obtained by high-frequency linear array probe. Step 2: Using the temporal window ultrasound penetration results and temporal window skull thickness values, plot the receiver operating characteristic curves (ROCs) for osteoporosis. Step 3: Incorporate the results of temporal window ultrasound penetration and temporal window skull thickness values, and construct prediction models one through four with gender, age, body mass index, and calcium-phosphorus product, respectively. Plot a nomogram, establish a scoring table, and plot the subject operating characteristic curve for osteoporosis. Based on the scoring table, obtain the scores and predicted osteoporosis probabilities for different baseline data tuples. The output program compares these scores and outputs the highest score and the highest osteoporosis probability.

5. The osteoporosis risk prediction system based on the ultrasound osteoporosis risk prediction model according to claim 4, characterized in that, The temporal window of the ultrasound refers to the area above the cheekbone, between the lateral orbital margin and the tragus, and the skull thickness is the measurement between the two layers of strong echoes of the cortical bone at the thinnest point of the temporal window. The criterion for determining that the temporal window ultrasound cannot penetrate is that no signal from any intracranial artery can be detected. The criterion for judging whether the temporal window ultrasound can penetrate is that it can successfully detect the signal of any intracranial artery.

6. The osteoporosis risk prediction system based on the ultrasound osteoporosis risk prediction model according to claim 5, characterized in that: The system combines transcranial Doppler intracranial artery detection and examination functions with two-dimensional B-mode ultrasound for superficial organ and blood vessel detection.

Images