Prediction tool for auxiliary diagnosis of coronary heart disease through combination of traditional Chinese medicine and western medicine and evaluation table

Abstract

The invention provides a prediction tool for auxiliary diagnosis of coronary heart disease through combination of traditional Chinese medicine and western medicine and a score table, particularly relates to a coronary heart disease noninvasive auxiliary diagnosis score table based on four diagnosis information of traditional Chinese medicine, and belongs to the technical field of medical diagnosis. The method focuses on a non-invasive prediction model of the coronary heart disease for the first time, and solves the problems of strong invasiveness and unintegration of traditional Chinese medicine characteristics in the prior art. Core predictive variables are screened by collecting four diagnosis information (including intraocular angular blood vessel thickening, blood stasis expression and other traditional Chinese medicine sign characteristic clinical expressions) of coronary heart disease patients, and a Logistic regression model is constructed and converted into a score table. The scale comprises indexes such as age, combined diseases, typical angina pectoris, abnormal angular blood collaterals in eyes, blood stasis and kidney deficiency, and the coronary heart disease is diagnosed when the total score is greater than or equal to 4. According to the method, the model for predicting the coronary heart disease is constructed by utilizing a statistical method, and the model is converted into the score table which is easy to operate through the statistical method, so that clinical popularization and application are facilitated, doctors are assisted to make accurate and efficient clinical decisions, and clinical doctors are guided to make accurate and effective treatment schemes.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of medical diagnosis, and particularly relates to a prediction tool for auxiliary diagnosis of coronary heart disease by a combination of traditional Chinese medicine and western medicine. BACKGROUND

[0003] The current diagnosis method of coronary heart disease has certain limitations. The gold standard for the diagnosis of coronary heart disease is coronary angiography, but coronary angiography is an invasive examination with problems such as high cost, radiation risk, contrast nephropathy risk, etc., and is not suitable for large-scale screening. In addition, traditional Chinese medicine relies on the experience of doctors and lacks objective and quantitative diagnostic criteria, making it difficult to achieve standardized promotion. Although there are prediction models for coronary heart disease, existing prediction models for coronary heart disease (such as the updated Diamond-Forrester model, UDFM) are mostly based on western population data, and their applicability to Chinese population is questionable, and they do not include information on traditional Chinese medicine diagnosis methods. Therefore, it is an urgent problem to be solved in clinical practice to establish a non-invasive detection tool and method for diagnosing coronary heart disease, which has great application value. Currently, western medicine models (such as the UDFM model) rely on indicators such as age, gender, and chest pain type, and have a low AUC value (about 0.779); while traditional Chinese medicine researches are mostly focused on syndrome classification, and lack quantitative prediction models based on four diagnostic information, and are difficult to meet the precise diagnosis needs of large numbers of patients.

[0004] Therefore, it is urgent to establish a new coronary heart disease prediction data analysis system to provide doctors with a more standardized and reproducible detection information analysis method, in order to improve the diagnostic accuracy of coronary heart disease, optimize clinical decision-making, and reduce the risk of misdiagnosis and missed diagnosis, thereby providing more precise medical services for patients. SUMMARY

[0005] In view of the problems in the prior art, the present application provides a prediction tool for auxiliary diagnosis of coronary heart disease by a combination of traditional Chinese medicine and western medicine. The present application uses statistical methods to construct a model for predicting coronary heart disease, and converts the model into an easy-to-operate scoring scale by statistical methods, which is convenient for clinical application, thereby guiding clinicians to develop precise and effective treatment plans.

[0006] In order to achieve the above-mentioned application purposes, the present application provides the following technical solutions.

[0007] The present application discloses a coronary heart disease prediction model, characterized in that the model comprises the following variables: age (male ≥ 45 years old or female ≥ 55 years old), ≥ 3 kinds of comorbidities, typical angina pectoris, blood stasis performance ≥ 2 items, kidney deficiency performance ≥ 2 items.

[0008] Further, the combined diseases are at least 3 of hypertension, hyperlipidemia, diabetes, hyperuricemia, ischemic cerebrovascular disease, and peripheral atherosclerosis.

[0009] Further, the kidney deficiency is at least 2 of the following: obvious forgetfulness, tinnitus, increased nocturia (≥2 times / night), weak sensation of urination, urinary incontinence, tooth loss (≥2 teeth), white hair (≥50%), baldness, multiple facial age spots (≥3), and corneal senile ring.

[0010] Further, the blood stasis is at least 2 of the following: tongue with blood stasis points / blood stasis spots, sublingual collateral vessels thickening / entanglement / beads, red / purple palms, and large thenar purple / purple veins (spring and summer).

[0011] Further, the model formula = 2.5×age variable + 1.5×combined disease variable + 2.0×angina variable + 2.5×inner canthus variable + 1.5×blood stasis variable + 1.0×kidney deficiency variable.

[0012] The application further discloses a coronary heart disease prediction score scale, characterized in that the scale comprises variable data according to any one of the above and a formula for scoring.

[0013] The application further discloses a coronary heart disease auxiliary prediction system, comprising: an input module for collecting the variable data according to any one of the above; a processing module for calculating a score based on the regression model formula; an output module for displaying the model prediction score.

[0014] The application further discloses an electronic device comprising a memory and a processor, wherein the memory stores executable instructions, and the instructions are executed by the processor to realize the calculation logic of the score scale.

[0015] The application further discloses a computer readable storage medium storing a program, wherein the program is executed by a processor to realize the calculation of the model.

[0016] The application further discloses an application of the model or the scale in preparing a coronary heart disease prediction tool.

[0017] Compared with the prior art, the application has the following beneficial effects.

[0018] (1) For the first time, a diagnosis method with Chinese medicine characteristics (such as blood collateral vessels thickening in the inner canthus, large thenar purple / purple veins, etc.) is introduced into the coronary heart disease prediction model.

[0019] (2) The prediction model of the combination of traditional Chinese and western medicine is constructed, and good diagnostic efficiency is embodied in external verification, and is significantly better than the western medicine diagnostic prediction model UDFM, and the diagnostic accuracy is significantly improved.

[0020] (3) The non-invasive and low-cost evaluation tool is developed, a simple and non-invasive reference tool is provided for the screening of coronary heart disease high-risk population in community or primary hospital, and the tool has important clinical significance for the secondary prevention of coronary heart disease.

[0021] (4) The construction method of the quantitative prediction model of the application establishes the model of the combination of traditional Chinese and western medicine scoring scale for predicting coronary heart disease from the aspects of simplicity, rapidness and quantitative evaluation, the model is simple and intuitive, the prediction accuracy is high, the model can guide clinicians to formulate accurate and efficient diagnosis and treatment plans, and has good clinical application and promotion prospect. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 Hosmer-Lemeshow goodness-of-fit test.

[0023] Figure 2 UDFM and traditional Chinese medicine prediction model ROC curve comparison.

[0024] Figure 3 UDFM and the combination of traditional Chinese and western medicine scoring scale ROC curve comparison (external verification). DETAILED DESCRIPTION

[0025] The application will be further described in detail below with specific examples. However, it should not be understood that the scope of the above-mentioned subject matter of the application is limited to the following examples, and any technology realized based on the content of the application belongs to the scope of the application.

[0026] Unless otherwise specified, the reagents and materials used in the application are commercially available.

[0027] Example 1: Development of the prediction tool of the combination of traditional Chinese and western medicine scoring scale for assisting in the diagnosis of coronary heart disease.

[0028] 1. Data source.

[0029] Literature research: The ancient medical books, ancient medical cases, modern medical cases and characteristic diagnosis method literatures in “Chinese Medical Dictionary” are included, the four diagnostic information related to coronary heart disease (such as chest pain, tongue appearance, pulse appearance, etc.) is extracted, and the basis for formulating the clinical epidemiology questionnaire is provided. After literature research, we summarized and sorted 78 symptoms related to coronary heart disease, 30 tongue appearances, 29 pulse appearances, 27 other signs, and 88 characteristic diagnosis methods, a total of 252 information items. After repeated demonstration by experts, 139 information items were finally determined, including 9 items of personal information, 29 items of medical history, 42 items of life and living, 27 items of symptoms, and 32 items of signs.

[0030] Clinical epidemiology investigation: According to the results of literature research, the questionnaire was developed. Four diagnostic information of 138 cases of coronary heart disease patients and 125 cases of non-coronary heart disease patients were collected as research data sources. The questionnaire is as follows: .

[0031] 2. Model construction.

[0032] (1) Variable screening.

[0033] Firstly, through single factor analysis, by chi-square test, there were 49 items with significant differences in the distribution of coronary heart disease group and non-coronary heart disease group (P<0.05), among which 11 items in life, 6 items in family history, disease history and menstrual history, 14 items in symptoms and 18 items in signs. See Table 1-Table 4. P

[0034] Table 1 Comparison of life distribution between coronary heart disease group and non-coronary heart disease group .

[0035] Table 2 Comparison of family history, disease history and menstrual history distribution between coronary heart disease group and non-coronary heart disease group .

[0036] Table 3 Comparison of symptoms distribution between coronary heart disease group and non-coronary heart disease group .

[0037] Table 4 Comparison of signs distribution between coronary heart disease group and non-coronary heart disease group .

[0038] Then, the experts were organized again to discuss, according to the clinical practice experience and TCM theory, further screening, combining 49 items with significant differences in the distribution of coronary heart disease group and non-coronary heart disease group in single factor analysis and gender, age, see Table 5.

[0039] Table 5 Selected variables and definitions .

[0040] (2) Model construction.

[0041] ​The 13 indicators with statistical significance in the above single factor analysis were used as independent variables, and whether coronary heart disease occurred was used as dependent variable to construct a binary Logistic regression equation. The results showed that at the 0.05 test level, the independent variables of age, combined diseases ≥ 3, typical angina pectoris, blood vessels thickening in the inner canthus, red color, extension, blood stasis manifestations ≥ 2, and kidney deficiency manifestations ≥ 2 were statistically significant in the association with the dependent variable coronary heart disease. According to the regression coefficient and estimated value, they were positively correlated with the occurrence of coronary heart disease and were independent risk factors for the occurrence of coronary heart disease. See Table 6 for details.

[0042] The independent variables were estimated by maximum likelihood value (Maximum Likelihood Estimate, MLE), and according to the mathematical prediction formula of the regression model: P = 1 / 1 + exp (-Z), the mathematical formula of the prediction model of this study was developed: P = 1 / {1+exp[-(2.396×age male ≥ 45, female ≥ 55)+1.435×(combined diseases ≥ 3)+1.981×typical angina pectoris+2.354×blood vessels thickening in the inner canthus, red color, extension+1.561×(blood stasis manifestations ≥ 2)+1.266×(kidney deficiency manifestations ≥ 2)-4.302)]}, i.e. ln(P / 1-P) = 2.396×(age male ≥ 45 years, female ≥ 55 years)+1.435×(combined diseases ≥ 3)+1.981 typical angina pectoris+2.354 blood vessels thickening in the inner canthus, red color, extension+1.561×(blood stasis manifestations ≥ 2)+1.266×(kidney deficiency manifestations ≥ 2)-4.302.

[0043] Omnibus test X of binary Logistic regression model 2 =217.840, P <0.001, the maximum likelihood value is 146.113, the COX & Snell R2 value is 0.563, the Nagelkerke R2 value is 0.752, and the R2 value is <1, indicating that the Logistic regression model obtained has statistical significance.

[0044] Table 6 Logistic regression analysis results .

[0045] 3. Model presentation.

[0046] The cut-off value of -0.5025 is brought into the regression model, when {ln(p / 1-p) = 2.396 x (age ≥ 45 years old for men, ≥ 55 years old for women) + 1.435 x (comorbidities ≥ 3) + 1.981 x typical angina pectoris + 2.354 x blood vessels thickening, red color and prolongation at inner canthus + 1.561 x (stasis manifestations ≥ 2) + 1.266 x (kidney deficiency manifestations ≥ 2) - 4.302} ≥ -0.5025, it can be diagnosed as coronary heart disease.

[0047] For convenience of calculation, the formula is adjusted according to rounding off, Y = 2.5 x (age ≥ 45 years old for men, ≥ 55 years old for women) + 1.5 x (comorbidities ≥ 3) + 2 x typical angina pectoris + 2.5 x blood vessels thickening, red color and prolongation at inner canthus + 1.5 x (stasis manifestations ≥ 2) + 1 x (kidney deficiency manifestations ≥ 2). When Y ≥ 4.0, it can be diagnosed as coronary heart disease. According to the adjusted formula, the prediction result does not change, which indicates that the above adjustment has no effect on the logical relationship of the prediction model.

[0048] Finally, the above model formula is presented in the form of a scoring scale.

[0049] 4. Evaluation of the model.

[0050] (1) Discrimination.

[0051] The discrimination is generally evaluated by the area under the ROC curve. The higher the AUC, the better the discrimination of the model for high-risk and low-risk populations. Generally, AUC < 0.6 is low discrimination, 0.6-0.75 is medium discrimination, and > 0.75 is high discrimination.

[0052] The results of this study show that the AUC of the integrated traditional Chinese and Western medicine diagnosis prediction model (hereinafter referred to as the "integrated traditional Chinese and Western medicine prediction model") is 0.951 > 0.75, and the 95% CI is 0.929-0.974, indicating that the discrimination ability of the prediction model is good. See Table 7 for details.

[0053] Table 7 Area under the curve (AUC) of the prediction model .

[0054] (2) Calibration.

[0055] The calibration ability of the model is evaluated by Hosmer-Lemeshow goodness-of-fit test. The smaller the chi-square value, the larger the corresponding P value, indicating that the calibration of the model is better.

[0056] The results of this study show that Hosmer-Lemeshow X²=3.198, P=0.921>0.05, suggesting that there was no statistical significance between the predicted value and the actual observed value of the integrated traditional Chinese and western medicine prediction model, and the calibration ability of the prediction model was good, as shown in Figure 1 .

[0057] (3) Comparison with existing models.

[0058] The integrated traditional Chinese and western medicine prediction model was compared with UDFM. Results: The sensitivity of the integrated traditional Chinese and western medicine prediction model in diagnosing coronary heart disease was 92.8%, the specificity was 83.2%, and the AUC was 0.951〔95%CI (0.929, 0.974)〕. The sensitivity of UDFM in diagnosing coronary heart disease was 97.1%, the specificity was 44.0%, and the AUC was 0.779〔95%CI (0.724, 0.833)〕; the difference in AUC between the two was statistically significant (Z=5.733, P P<0.05, as shown in Table 8, Figure 2 ).

[0059] Table 8 Comparison of AUC between UDFM and integrated traditional Chinese and western medicine prediction model .

[0060] Example Two: External validation of the integrated traditional Chinese and western medicine scoring scale as an auxiliary diagnostic tool for coronary heart disease.

[0061] From July 2023 to December 2023, 78 inpatients who visited the Department of Cardiology of the Affiliated Hospital of Liaoning University of Traditional Chinese Medicine due to chest pain / shortness of breath and were scheduled for coronary angiography were reselected. Using the cross-sectional epidemiological survey method, the patients were investigated face-to-face using the "Integrated Traditional Chinese and Western Medicine Auxiliary Diagnostic Scoring Scale for Coronary Heart Disease" (hereinafter referred to as the "Integrated Traditional Chinese and Western Medicine Scoring Scale"), and the results of coronary angiography or coronary CT were tracked. The prediction results of the integrated traditional Chinese and western medicine scoring scale and the prediction results of UDFM were compared with the results of coronary angiography or coronary CT to evaluate the accuracy of the integrated traditional Chinese and western medicine scoring scale.

[0062] 1. Area under the ROC curve (AUC).

[0063] The sensitivity of UDFM in diagnosing coronary heart disease was 91.3%, the specificity was 55.6%, and the AUC was 0.756〔95%CI (0.574, 0.938)〕; the sensitivity of the integrated traditional Chinese and western medicine scoring scale in diagnosing coronary heart disease was 91.3%, the specificity was 88.9%, and the AUC was 0.881〔95%CI (0.733, 1.000)〕. The difference in AUC between the two was statistically significant (Z=1.04, P<0.05, as shown in Table 9, Figure 2 ).

[0064] Table 9 AUC of UDFM and integrated traditional Chinese and western medicine score scale (external validation) .

[0065] 2. Net reclassification index (NRI).

[0066] NRI is often used to compare the accuracy of the predictive ability of two models. Compared with ROC curve and AUC, NRI focuses on the change in the number of subjects correctly classified by two models at a certain set point. If NRI > 0, it is a positive improvement, indicating that the predictive ability of the new model is improved compared with the old model; if NRI < 0, it is a negative improvement, and the predictive ability of the new model decreases; if NRI = 0, it is considered that the new model has no improvement.

[0067] The results of this study showed that compared with UDFM, NRI = (c1-b1) / N1+ (b2-c2) / N2 = (65-0) / 69+ (3-9) / 9 = 0.275, indicating that the predictive ability of the integrated traditional Chinese and western medicine score scale is improved compared with UDFM.

[0068] 3. Integrated discrimination improvement (IDI).

[0069] IDI reflects the change in the difference between the predictive probabilities of two models, and can be used to reflect the overall improvement of the model. In general, the larger the IDI, the better the predictive ability of the new model. If IDI > 0, it is a positive improvement, indicating that the predictive ability of the new model is improved compared with the old model, if IDI < 0, it is a negative improvement, and the predictive ability of the new model decreases, if IDI = 0, it is considered that the new model has no improvement. The results of this study showed that IDI = (Pnew, events-Pold, events)-(Pnew, non-events-Pold, non-events) = (94.2-0.07)-(66.7-0) = 56.95, indicating that the predictive ability of the integrated traditional Chinese and western medicine score scale is improved compared with UDFM.

[0070] Example three, use of integrated traditional Chinese and western medicine auxiliary diagnosis and prediction score scale for coronary heart disease.

[0071] The score scale is as follows: Y = 2.5 x (age ≥ 45 years old for men and ≥ 55 years old for women) + 1.5 x (combined diseases ≥ 3) + 2 x typical angina pectoris + 2.5 x blood vessels thickening, red color and extension at the inner canthus + 1.5 x (stasis manifestations ≥ 2) + 1 x (kidney deficiency manifestations ≥ 2), when Y ≥ 4.0, it can be diagnosed as coronary heart disease.

[0072] .

[0073] The above descriptions are only the preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A coronary heart disease prediction model, characterized in that, The model includes the following variables: age (male ≥45 years or female ≥55 years), comorbidities ≥3, typical angina pectoris, thickened / reddened / prolonged blood vessels at the inner canthus of the eye, blood stasis manifestations ≥2, and kidney deficiency manifestations ≥2.

2. The model as described in claim 1, characterized in that, The comorbidities are defined as the presence of at least three of the following: hypertension, hyperlipidemia, diabetes, hyperuricemia, ischemic cerebrovascular disease, and peripheral arteriosclerosis.

3. The model as described in claim 1, characterized in that, The symptoms of kidney deficiency are defined as at least two of the following: significant forgetfulness, tinnitus like cicadas, increased urination at night (≥2 times / night), weak urination, urinary incontinence, tooth loss (≥2 teeth), graying hair (≥50% gray hair), baldness, multiple age spots on the face (≥3), and corneal arcus.

4. The model as described in claim 1, characterized in that, The blood stasis is defined as meeting at least two of the following criteria: petechiae / ecchymosis on the tongue, thickened / tortuous / beaded sublingual veins, red / purple palms, and purplish-blue / visible veins on the thenar eminence (in spring and summer).

5. The model as described in claim 1, characterized in that, Model formula = 2.5 × age variable + 1.5 × comorbidity variable + 2.0 × angina variable + 2.5 × inner canthus variable + 1.5 × blood stasis variable + 1.0 × kidney deficiency variable.

6. A coronary heart disease prediction scoring scale, characterized in that, The scale includes scoring based on the variable data according to any one of claims 1 to 4 and the formula according to claim 5.

7. A coronary artery disease auxiliary prediction system, comprising: Input module: used to collect the variable data as described in any one of claims 1 to 4; Processing module: Calculates the score based on the regression model formula of claim 5; Output module: Displays the model's predicted score.

8. An electronic device comprising a memory and a processor, the memory storing executable instructions which, when executed by the processor, implement the calculation logic of the rating scale of claim 6.

9. A computer-readable storage medium storing a program that, when executed by a processor, performs the calculation of the model of claim 5.

10. The application of the model according to any one of claims 1-5 or the scale according to claim 6 in the preparation of a tool for predicting coronary heart disease.

Images