Computer implementation method and system for evaluating coronary artery stenosis degree based on data

Abstract

The invention belongs to the crossing field of biomedical engineering and computer data processing, particularly relates to the technical field of ultrasonic medical data processing, and discloses a computer implementation method and system for evaluating the coronary artery stenosis degree based on data. Collected echocardiogram data is used as a core analysis object, target coronary blood vessel positioning processing is carried out on color Doppler blood flow imaging data through a computer, and computer-readable multi-dimensional clinical data and a preset coronary flow velocity influence factor database are fused; and performing dynamic optimization of a linear regression equation and subject working characteristic curve analysis by a computer to realize positioning correlation and quantitative calculation of related degree parameters of coronary artery stenosis. Through database dynamic iteration and parameter correction, interference of individual differences on a processing result is effectively avoided, a brand new computer technical scheme is provided for noninvasive accurate analysis of coronary artery disease related data, and the safety and universality of data processing are remarkably improved.

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of biomedical engineering and clinical diagnosis, specifically to a computer-based method for assessing the degree of coronary artery stenosis based on data. Background Technology

[0002] Coronary artery disease is one of the leading causes of death worldwide. Clinical studies have shown that coronary artery stenosis can lead to abnormal blood perfusion in distal myocardial tissue, thereby triggering a series of cardiovascular events. Therefore, obtaining parameters related to coronary artery stenosis is of great significance for clinical decision-making.

[0003] Numerous basic and clinical studies in coronary hemodynamics have clearly demonstrated that when coronary artery stenosis occurs, the blood flow at the lesion site undergoes compensatory acceleration. This characteristic hemodynamic change provides crucial theoretical and practical basis for indirectly obtaining coronary artery stenosis-related parameters through ultrasound data. Currently, commonly used clinical methods for detecting coronary artery blood flow velocity and analyzing parameters largely rely on invasive procedures or manual analysis. These methods suffer from drawbacks such as significant trauma, high risk of radiation exposure, complex operation, low data processing efficiency, and significant interference from individual differences. They are unsuitable for screening relevant data in routine populations and for long-term follow-up monitoring of postoperative patients.

[0004] In recent years, with the iterative updates of echocardiography equipment hardware and the development of computer data processing technology, the resolution of ultrasound images and the sensitivity of data detection have been substantially improved, providing technical support for the accurate extraction and analysis of coronary hemodynamic parameters. In the field of coronary stenosis-related parameter analysis, researchers have explored various detection methods using echocardiography. However, previous methods mostly focused on the detection operation itself, lacking automated and standardized computer processing schemes for the acquired ultrasound data, and failing to fully integrate multi-dimensional clinical structured data for individualized correction. This resulted in insufficient accuracy and universality of data processing, making it difficult to meet the needs of large-scale clinical data processing and efficient application. Against this backdrop, developing non-invasive, safe, accurate, and efficient computerized analysis methods based on acquired echocardiography data to fill the technological gap in clinical data processing has become a key issue urgently needing to be addressed in current cardiovascular imaging data processing and clinical practice. Summary of the Invention

[0005] The purpose of this invention is to provide a computer-based method and system for assessing the degree of coronary artery stenosis based on data, so as to solve the technical defects of existing procedures, such as large invasiveness, high risk of radiation exposure, complex operation and difficulty in direct quantitative localization.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0007] A computer-based method for assessing the degree of coronary artery stenosis based on data uses acquired echocardiographic data as the core analysis object. The method employs a computer to perform target coronary artery localization processing on color Doppler flow imaging (CDFI) data from the echocardiographic data, extracts pulsed Doppler (PW) coronary blood flow velocity parameters from the echocardiographic data, and combines these velocity parameters with a pre-set computer-readable database of factors influencing coronary blood flow velocity. Through standardized statistical analysis performed by the computer, the method achieves the localization, correlation, and quantitative calculation of parameters related to the degree of coronary artery stenosis.

[0008] In a further embodiment, the target coronary vessels for computer-based localization processing include the ultrasound image data regions corresponding to the left anterior descending coronary artery, the left circumflex coronary artery, the right main coronary artery, and their respective branches.

[0009] In a further embodiment, the blood flow velocity parameters extracted by the computer include peak systolic velocity (PSV), peak diastolic velocity (PDV), mean velocity (MFV), and blood flow velocity integral (VTI). The core analytical parameter is the peak diastolic velocity (PDV) at the lesion site in the ultrasound data corresponding to the suspected coronary artery stenosis area.

[0010] In a further embodiment, the extraction conditions for pulsed Doppler (PW) related data in the echocardiographic data are as follows: the sampling volume is placed in the center of the coronary artery lumen, the sampling volume size is 2.0-5.0 mm, the angle between the ultrasound beam and the blood flow direction is ≤60°, the blood flow velocity spectrum is continuously recorded for at least 3 stable cardiac cycles during data acquisition, and the computer performs interference signal and artifact removal on the data before using it for parameter analysis.

[0011] In a further embodiment, the quantitative calculation of the coronary artery stenosis-related parameter based on peak diastolic velocity (PDV) employs a simple linear regression equation, defined as Y = aX + b, where: Y is the coronary artery stenosis-related parameter (unit: %, ranging from 0% to 100%), X is the peak diastolic velocity at the site of coronary artery stenosis (unit: cm / s), a is the regression coefficient, and b is a constant term. This equation is obtained by fitting clinical case data, with a fitted sample size ≥ 50 cases, and the goodness of fit R² ≥ 0.75, with a validation set prediction accuracy ≥ 85%.

[0012] In a further embodiment, the computer implementation method specifically includes the following steps:

[0013] (1) Data input and preprocessing

[0014] ① The computer receives the acquired echocardiogram data (including CDFI data and PW data) and synchronous surface electrocardiogram data through the data input module. The echocardiogram data must meet the preset data quality standards (no severe motion artifacts and qualified image resolution).

[0015] ② The computer performs preliminary preprocessing on the received echocardiogram data, including format standardization, image noise reduction, and invalid data removal, to ensure that the data meets the requirements for subsequent analysis;

[0016] ③ The computer calls the multi-dimensional clinical data integration module to import the computer-readable structured clinical data corresponding to the patient (covering data related to factors affecting coronary flow velocity such as age, blood pressure, and heart rate), and performs data classification and verification.

[0017] (2) Target coronary artery localization processing

[0018] ① The computer automatically analyzes the preprocessed CDFI data. Based on the preset coronary artery anatomy and course algorithm model, and combined with imaging section data such as short axis and long axis, it tracks and locates the image regions corresponding to the target coronary arteries (left anterior descending artery, circumflex artery, right coronary artery trunk and branches).

[0019] ② The computer identifies abnormal data signs in CDFI data, such as local blood flow acceleration, turbulence, and blood flow signal aliasing, and marks the image coordinates corresponding to suspected narrow areas to provide a location basis for subsequent parameter extraction.

[0020] (3) Extraction of blood flow velocity parameters

[0021] ① The computer automatically locates the PW data sampling volume to the area with the most significant color blood flow aliasing based on the marked coordinates of the suspected narrow area, ensuring that the data analysis range corresponding to the sampling volume completely includes the target blood flow signal and does not exceed the image boundary corresponding to the blood vessel lumen.

[0022] ② Based on preset parameter extraction rules, the computer automatically measures and records parameters such as peak systolic velocity (PSV), peak diastolic velocity (PDV), mean velocity (MFV), and blood flow velocity integral (VTI) from PW data. Each parameter is extracted three times, and the average value is taken as the final analysis data.

[0023] (4) Statistical analysis and quantitative calculation

[0024] ① Data preprocessing: The computer standardizes the extracted flow rate parameters and imported clinical data. Quantitative data are expressed as mean ± standard deviation. The description states that the Shapiro-Wilk test is used to verify the normality of the data, and the Levene test is used to assess the homogeneity of variance.

[0025] ② Intergroup comparison: When comparing two groups of continuous data, if the data meet the normal distribution and have homogeneous variances, the computer uses the independent samples t test; if they meet the normal distribution but have unequal variances, the Welch corrected t test is used; count data are expressed as frequency (n%).

[0026] ③ Correlation and Regression Analysis: When both variables satisfy a normal distribution, the computer uses Pearson product-moment correlation analysis to quantify the strength of the linear association between the variables; a linear regression equation (Y= aX + b) is constructed between the coronary artery stenosis correlation parameter (Y) and the peak diastolic velocity (X) through simple linear regression analysis.

[0027] ④ Predictive efficacy assessment: Receiver operating characteristic (ROC) curves were plotted by computer, the area under the curve (AUC) was calculated, and the optimal diagnostic cutoff value for peak diastolic velocity (PDV) was determined using the maximum Yoden index.

[0028] ⑤ Individualized correction: The computer accesses a database of factors affecting coronary flow velocity and performs individualized correction on the above calculation results based on the patient's structured clinical data;

[0029] ⑥ Results output: All statistical tests were two-sided, with P < 0.05 considered statistically significant; AUC ≥ 0.85 indicated good predictive efficacy. The computer combined the results of linear regression equation calculation to finally output the location association results (image coordinates corresponding to lesion branches and segments) and quantitative calculation results (stenosis correlation degree parameter %) of coronary artery stenosis-related parameters.

[0030] A computer-based system for assessing the degree of coronary artery stenosis based on data includes a data input module, a multi-dimensional clinical data integration module, and an intelligent data analysis and assessment module. The system uses any of the aforementioned computer implementation methods to process the acquired echocardiographic data and achieve the localization, correlation, and quantitative analysis of coronary artery stenosis-related parameters.

[0031] In a further embodiment, the intelligent data analysis and evaluation module includes core functional units:

[0032] ① A pre-defined database of linear regression equations based on peak diastolic velocity (PDV) (computer readable);

[0033] ② Dynamic database of optimal diagnostic cutoff values ​​for receiver operating characteristic (ROC) curves (computer readable);

[0034] ③ Multi-dimensional clinical data adaptation unit (supports structured data format conversion and association);

[0035] ④ Coronary artery flow velocity influencing factor correction unit (the computer executes an individualized correction algorithm); among them, the influencing factor database covers but is not limited to computer-readable structured data corresponding to physiological and pathological factors that may affect coronary blood flow velocity, such as age, gender, blood pressure, heart rate, blood lipid level, blood glucose concentration, degree of myocardial hypertrophy, cardiac function classification, etc. The system can perform targeted correction on the extracted blood flow velocity parameters through this database to make the calculation results more accurate.

[0036] In a further solution, the computer implementation system for evaluating the degree of coronary artery stenosis based on data further includes computer-compatible auxiliary components, including but not limited to a surface electrocardiogram synchronous data interface module (supporting data synchronization import), an ultrasonic data format adaptation component (compatible with the data formats of mainstream ultrasonic devices), a multi-source clinical data structured entry interface (supporting the import and structured conversion of electronic medical records and test report data), a computer-readable data storage and update module (supporting data security storage and database iteration), an analysis report generation module (automatically generating a standardized data report), an operation guidance module (providing a computer operation process guidance), etc. The auxiliary components need to meet the requirements of data processing security, compatibility, and data privacy protection. Among them, the data storage and update module supports regularly incorporating newly added clinical case data and structured data corresponding to newly discovered coronary artery flow velocity influencing factors to achieve dynamic iterative optimization of the influencing factor database and analysis algorithm.

[0037] In a further solution, the computer implementation system for evaluating the degree of coronary artery stenosis based on data can be configured into clinically acceptable application forms, including but not limited to a computer software module supporting ultrasonic devices, an independent data analysis computer terminal, an integrated computer processing system for ultrasonic devices, a remote data processing and evaluation platform, etc.; preferably, it is an application form integrated with a conventional echocardiogram device, adapting to the hardware configuration of existing clinical ultrasonic devices, without the need to additionally install dedicated hardware, and supporting real-time integration of relevant patient clinical structured data through a multi-source clinical data entry interface, combining with the influencing factor database to complete measurement parameter correction and equation optimization, continuously improving the accuracy of calculating coronary artery stenosis-related parameters, and reducing the clinical transformation cost.

[0038] The present invention has the following beneficial effects:

[0039] (1) Computerized implementation, efficient and accurate data processing: The present invention is entirely based on computer automation to process the collected echocardiogram data, and through a standardized algorithm, it realizes target blood vessel positioning, parameter extraction, statistical analysis, and quantitative calculation, avoiding errors caused by manual operation, and significantly improving the efficiency and stability of data processing; at the same time, it integrates multi-dimensional clinical structured data for individualized correction, and the calculation accuracy rate ≥ 92%, solving the pain points of insufficient data processing accuracy and large interference of individual differences in traditional methods.

[0040] (2) Non-invasive and safe, with friendly data sources: This invention only processes the collected echocardiogram data by computer, without the need for additional invasive operations, contrast agents, or radiation exposure risks. It avoids the risks of complications such as trauma and infection caused by invasive operations. The data source has the advantages of being non-invasive and safe, and is suitable for ultrasound data processing of various populations, including ultrasound data of special populations who cannot tolerate invasive examinations.

[0041] (3) High universality and low clinical translation cost: The computer system of the present invention can be adapted to the data format of mainstream ultrasound equipment and supports multiple application forms (software modules, independent terminals, etc.). No additional dedicated hardware is required. It can be deployed and applied in primary medical institutions and large hospitals. At the same time, the operation process is standardized and the computer automatically completes data processing and report generation, which reduces the dependence on the professional level of operators, effectively reduces the cost of clinical data processing, and helps large-scale screening and long-term follow-up monitoring of coronary artery related data.

[0042] (4) Dynamic iteration to adapt to diverse data needs: The computer-readable coronary flow velocity influencing factor database constructed by this invention supports the continuous inclusion of newly added clinical case data and structured data corresponding to newly discovered influencing factors. Through data updates, the analysis algorithm and regression equation are dynamically optimized, which can adapt to the ultrasound data processing needs of different populations and different diseases, significantly improve the individualized accuracy of data processing, and provide more reliable data references for clinical practice. Attached Figure Description

[0043] Figure 1 Color Doppler flow imaging (CDFI) results of the proximal and mid-segment of the left anterior descending coronary artery (LAD).

[0044] Figure 2 Color Doppler flow imaging (CDFI) results of the mid-distal segment of the left anterior descending coronary artery (LAD).

[0045] Figure 3 Schematic diagram of transthoracic echocardiography-color Doppler flow imaging (TTE-CDFI) for detecting flow velocity in the LAD stenosis segment.

[0046] Figure 4 Another schematic diagram of transthoracic echocardiography-color Doppler flow imaging (TTE-CDFI) for detecting flow velocity in the narrowed segment of the LAD.

[0047] Figure 5 Scatter plot of the correlation between left anterior descending artery (LAD) peak diastolic velocity (LAD-PDV) and the degree of stenosis.

[0048] Figure 6 : Residual plot.

[0049] Figure 7 Actual vs. Predicted Values ​​Comparison Chart (I).

[0050] Figure 8 Actual vs. Predicted Values ​​Comparison Chart (II).

[0051] Figure 9 The ROC curve is used to determine whether there is moderate to severe stenosis of the LAD using LAD-PDV.

[0052] Figure 10 Use LAD-PDV to determine if there is a narrow ROC curve for LAD. Detailed Implementation

[0053] The technical solution of the present invention will be clearly and completely described below with reference to specific embodiments and experimental examples. However, those skilled in the art should understand that the embodiments are only used to illustrate the technical solution of the present invention and should not be regarded as limiting the scope of protection of the present invention. Based on the following embodiments, all other implementation schemes obtained by those skilled in the art without creative effort, such as implementation schemes obtained by modification, variation or simple substitution, should fall within the scope of protection of the present invention.

[0054] Unless otherwise specified, the experimental methods used in the following embodiments and experimental examples are conventional methods; the raw materials (including biological materials), reagents, culture media, instruments, etc. used are all commonly used in the field and commercially available to the public unless otherwise specified; the terms and abbreviations used have their conventional meanings in the field.

[0055] Example: A computer-based method and system for assessing the degree of coronary artery stenosis based on data;

[0056] This embodiment discloses a computer-based method for assessing the degree of coronary artery stenosis based on data, and also provides a computer-based system adapted to this method for assessing the degree of coronary artery stenosis based on data. The specific technical solution is as follows:

[0057] I. Core Technical Solution

[0058] The method uses acquired echocardiographic data as the core analysis object. It employs computer processing to locate the target coronary artery in color Doppler flow imaging (CDFI) data, extracts pulsed Doppler (PW) coronary blood flow velocity parameters, and combines these velocity parameters with a pre-defined computer-readable database of factors influencing coronary blood flow velocity. Through standardized statistical analysis performed by the computer, it achieves the location correlation and quantitative calculation of parameters related to the degree of coronary artery stenosis. The system includes a data input module, a multi-dimensional clinical data integration module, and an intelligent data analysis and evaluation module, completing the integrated analysis and processing of coronary artery stenosis-related parameters using the above method.

[0059] II. Detailed Implementation Steps

[0060] (1) Data input and preprocessing

[0061] (1.1) The computer receives the acquired echocardiogram data (including CDFI data and PW data) and synchronous surface electrocardiogram data through the data input module. The data format supports mainstream ultrasound data formats such as DICOM and BMP. The data quality must meet the following requirements: no serious motion artifacts and image resolution ≥ 512×512 pixels.

[0062] (1.2) The computer performs automated preprocessing on the received echocardiogram data: the median filtering algorithm is used for image noise reduction, the threshold segmentation algorithm is used to remove invalid data areas (such as image areas corresponding to air interference and equipment noise), and the ultrasound data of different formats are converted into standardized formats to ensure that the data meets the requirements of subsequent analysis.

[0063] (1.3) The computer calls the multi-dimensional clinical data integration module and imports the patient's corresponding clinical data (including age, gender, blood pressure, heart rate, blood lipid level, blood glucose concentration, degree of myocardial hypertrophy, NYHA classification of heart function, etc.) through the structured input interface. The data format is CSV or JSON structured format. The computer automatically classifies and stores the data and verifies its legality (such as removing abnormal values ​​that are outside the physiological range).

[0064] (2) Target coronary artery localization processing

[0065] (2.1) The computer loads the preset coronary vascular anatomy and course algorithm model, performs automated analysis on the preprocessed CDFI data, and combines the short-axis and long-axis section data of the heart to track and locate the image regions corresponding to the left anterior descending artery (LAD), circumflex artery (LCX), right main coronary artery (RCA) and its branches at all levels, and outputs the image coordinate range of the target blood vessel.

[0066] (2.2) The computer uses a blood flow signal feature recognition algorithm to identify abnormal data signs in CDFI data, such as local blood flow acceleration (flow velocity is more than 30% higher than the surrounding area), turbulence (disordered blood flow direction) and color aliasing, mark the image coordinates of suspected narrow areas, generate a location report and store it.

[0067] (3) Extraction of blood flow velocity parameters

[0068] (3.1) The computer automatically locates the analysis range corresponding to the PW data sampling volume to the area where the color blood flow aliasing is most significant, based on the coordinates of the marked suspected narrow area. The sampling volume size is set to 2.0-5.0 mm (which can be adaptively adjusted according to the diameter of the blood vessel lumen) to ensure that the data analysis range corresponding to the sampling volume completely includes the target blood flow signal and does not exceed the image boundary corresponding to the blood vessel lumen.

[0069] (3.2) The computer automatically adjusts the calculation parameters of the angle between the ultrasound beam and the blood flow direction to ensure that the angle is ≤60°. At least three stable cardiac cycles of blood flow velocity spectrum data are selected from the PW data to remove artifact signals caused by respiratory interference, body position changes, etc.

[0070] (3.3) The computer automatically measures and records the peak systolic velocity (PSV), peak diastolic velocity (PDV), mean velocity (MFV), and blood flow velocity integral (VTI) based on the preset parameter extraction algorithm. Each parameter is extracted three times, and the average value is taken as the final analysis data. The data precision is retained to one decimal place.

[0071] (4) Statistical analysis and quantitative calculation

[0072] (4.1) Data preprocessing: The computer standardized the extracted flow rate parameters and imported clinical data. Quantitative data were expressed as mean ± standard deviation (SD). The normality was verified by the Shapiro-Wilk test and the homogeneity of variance was assessed by the Levene test.

[0073] (4.2) Correlation and Regression Analysis: The computer used Pearson product-moment correlation analysis to quantify the linear association between LAD-PDV and stenosis-related parameters, and constructed a linear regression equation (Y=aX+b) (Y is the coronary stenosis-related parameter %, X is the peak diastolic velocity at the lesion site in cm / s, a is the regression coefficient, and b is the constant term). The currently generated simple linear regression curve is shown in [reference needed]. Figure 5 It will continue to be iterated and optimized as new clinical data is added.

[0074] (4.3) Predictive efficacy evaluation: ROC curves were plotted by computer, and the optimal diagnostic cutoff value for LAD-PDV was determined by the maximum Yoden index. When the area under the curve (AUC) ≥ 0.85, it indicates that the model has good predictive efficacy (typical results are shown in Figure 1). Figures 9-10 (As shown).

[0075] (4.4) Individualized correction: The computer calls the database of factors affecting coronary flow velocity and uses a multivariate correction algorithm to perform individualized correction on the above calculation results based on the patient's structured clinical data (such as age, blood pressure, heart rate, etc.) to reduce the impact of individual differences on the calculation results.

[0076] (4.5) Results output: The computer automatically generates a standardized analysis report, which clarifies the location correlation results (image coordinates and annotations corresponding to lesion branches and segments) and quantitative calculation results (stenosis correlation degree parameter %) of coronary artery stenosis-related parameters. The report format supports editable formats such as PDF and Word, and can be directly exported or uploaded to the hospital information system.

[0077] III. System Composition and Parameter Limitations

[0078] (1) Core module functions

[0079] (1.1) Data input module: Supports the import and format conversion of acquired echocardiogram data (DICOM, BMP and other formats) and synchronous surface electrocardiogram data. It has a built-in data quality verification algorithm to automatically filter qualified data and remove invalid data.

[0080] (1.2) Multi-dimensional clinical data integration module: Provides structured input interface for multi-source data such as electronic medical records and test reports, supports data import in formats such as CSV and JSON, and realizes classification storage, association matching and legality verification of clinical data.

[0081] (1.3) Intelligent data analysis and evaluation module: It has a built-in linear regression equation database (computer readable), ROC curve optimal diagnostic cutoff value database (computer readable) and coronary flow velocity influencing factor correction unit (executes multivariate correction algorithm), which supports automatic data analysis, result calculation and dynamic database update (supports manual upload of new data or automatic synchronization update).

[0082] (2) Auxiliary components and application forms

[0083] (2.1) Auxiliary components: including a surface electrocardiogram synchronous data interface module (supports real-time data synchronous import), an ultrasound data format adaptation component (compatible with mainstream ultrasound equipment data formats), a computer-readable data storage and update module (adopts encrypted storage technology and supports data backup and recovery), and a standardized report generation module (automatically generates analysis reports with attached figures). The auxiliary components meet the requirements of medical data security specifications and privacy protection.

[0084] (2.2) Application form: It adopts the form of computer software module with ultrasound equipment, which can be directly installed on the computer terminal of existing conventional echocardiography equipment without the need for additional dedicated hardware; the software supports mainstream operating systems such as Windows and Linux, and is compatible with the hardware configuration of existing clinical ultrasound equipment, reducing the cost of clinical translation.

[0085] IV. Application Scenarios

[0086] The method and system described herein can be widely applied to screening coronary artery stenosis-related data, quantitative calculation of stenosis-related parameters, lesion localization correlation analysis, and preoperative data support and postoperative follow-up data monitoring related to coronary artery disease. It is particularly suitable for preliminary analysis of ultrasound data of suspected coronary heart disease patients, risk screening of coronary artery-related data in high-risk groups of coronary heart disease (such as patients with hypertension, diabetes, hyperlipidemia and long-term smokers), and early detection of coronary artery-related data in asymptomatic individuals, providing reliable data references for clinical practice.

[0087] In other embodiments of the present invention, the parameters such as the ultrasonic data sampling volume and regression equation coefficients can be reasonably selected within the range defined in the claims, all of which can achieve accurate calculation of coronary artery stenosis-related parameters without affecting the technical effect of the present invention.

[0088] Experimental Example: Validation Experiment of Method and System

[0089] (1) Validation of the method (corresponding to) Figures 1-5 ):

[0090] Experimental Methods: Echocardiographic data and structured clinical data from 64 patients diagnosed with coronary artery stenosis by coronary angiography (gold standard) were collected. The method described in this embodiment was used for automated computer processing to extract the LAD-PDV parameter. Using the degree of stenosis corresponding to the angiographic results as a reference standard, a scatter plot of the correlation between LAD-PDV and the degree of stenosis was constructed, and linear regression analysis was performed. Experimental Results: Computer analysis showed a significant positive linear correlation between LAD-PDV and the degree of coronary artery stenosis (R²=0.73, P<0.0001). The degree of stenosis can be quantitatively calculated from the flow velocity parameter using a regression equation, verifying the scientific validity of the "quantitative correlation of flow velocity parameter and degree of stenosis parameter" technique of this invention. The equation will be continuously iterated and optimized with new clinical data to further improve the calculation accuracy.

[0091] Figure 1 Color Doppler flow imaging (CDFI) results of the proximal-mid segment of the left anterior descending coronary artery (LAD). This image clearly shows the blood flow signal distribution in the proximal-mid segment of the left anterior descending coronary artery. Color Doppler imaging technology enables preliminary localization of the target coronary artery segment, providing a basic vascular anatomy reference for subsequent stenosis screening.

[0092] Figure 2 Color Doppler flow imaging (CDFI) results of the mid-distal segment of the left anterior descending coronary artery (LAD). The arrows in the image clearly indicate abnormal areas with color aliasing of blood flow signals. This abnormal sign is related to local blood flow acceleration caused by coronary artery stenosis and can directly assist in the preliminary localization of suspected stenosis lesions, helping to pinpoint the target area for subsequent flow velocity parameter acquisition.

[0093] Figure 3 : A schematic diagram of transthoracic echocardiography-color Doppler flow imaging (TTE-CDFI) for detecting flow velocity in the LAD stenosis segment. This figure visually demonstrates the precise location of the pulsed Doppler (PW) sampling volume in the LAD stenosis area, clearly indicating that the peak diastolic velocity (LAD-PDV) at the lesion site is 194 cm / s, verifying the practical feasibility and positioning accuracy of the flow velocity parameter acquisition method of this invention;

[0094] Figure 4 Another schematic diagram of transthoracic echocardiography-color Doppler flow imaging (TTE-CDFI) for detecting flow velocity in the LAD stenosis segment. The peak diastolic velocity (LAD-PDV) at the lesion site is marked as 72.1 cm / s in the figure. Figure 3 By comparing flow velocity data under different degrees of stenosis, the correlation trend between the degree of coronary artery stenosis and PDV parameters is intuitively reflected, providing concrete data support for the construction of linear regression equations.

[0095] (2) Model fitting effect verification experiment (corresponding to) Figures 6-8 ):

[0096] Experimental Methods: Residual analysis and a comparison of actual and predicted values ​​were used to verify the fitting stability of the linear regression equation constructed by the computer. The error distribution characteristics were assessed by plotting residual plots, and the fitting curves between the actual narrowing degree and the predicted values ​​were plotted to evaluate the degree of fit. Experimental Results: The residual distribution was concentrated and showed no obvious trend (indicating stable model error). The actual narrowing degree showed a high degree of fit with the computer predictions (R² ≥ 0.7), confirming that the statistical analysis model constructed in this invention has a good fitting effect and providing data support for the reliability of quantitative calculation of the narrowing correlation parameter.

[0097] Figure 6 : Residual plot. This plot shows the residual distribution characteristics of the linear regression model. It can be seen that the residuals are concentrated near the zero line, with no obvious trend deviation, indicating that the linear regression model constructed in this invention has stable error and reliable fitting effect, providing statistical support for the accuracy of quantitative assessment of narrowing degree;

[0098] Figure 7 Actual vs. Predicted Values ​​(I). This chart presents the correspondence between the actual degree of coronary artery stenosis and the model's predicted values ​​in scatter plot form. It can be seen that most data points closely follow the fitted line, and the deviation between the actual and predicted values ​​is small, further verifying the reliability of the linear regression equation in quantitatively predicting the degree of stenosis.

[0099] Figure 8Actual vs. Predicted Value Comparison Chart (II). This chart shows the fit between the actual narrowness and the predicted value from another perspective. The horizontal and vertical axes correspond to the predicted and actual values, respectively. The data points are evenly distributed on both sides of the equivalence line, further confirming the fitting accuracy of the regression equation of this invention and ensuring the credibility of the quantitative evaluation results.

[0100] (3) Clinical application efficacy verification experiment (corresponding) Figures 9-10 )

[0101] Experimental methods: Echocardiographic data of 64 patients suspected of having coronary heart disease were selected and compared with coronary angiography using the method of this invention. The angiography results were used as the gold standard to calculate the diagnostic sensitivity and specificity of this method. At the same time, the computer data processing time, equipment compatibility and data processing security were recorded.

[0102] Experimental results:

[0103] (3.1) Diagnostic efficacy: The specificity of this method for moderate and above stenosis (≥50%) is 97.8%, the sensitivity is 85.2%, and the AUC is 0.979; the specificity for the presence of stenosis is 95.5%, the sensitivity is 92.0%, and the AUC is 0.992, indicating that the computer model has excellent predictive efficacy.

[0104] (3.2) Data processing security: The entire process involves only computer processing of the collected data, without any invasive procedures, contrast agent use or radiation exposure, and there is no risk of complications such as trauma or infection. The data processing process complies with medical data security standards.

[0105] (3.3) Practicality: The processing time for a single case is ≤15 minutes. It can be implemented using the computer terminal of the conventional ultrasound equipment in primary healthcare institutions without additional hardware investment. The software is easy to install and operate, and it is suitable for large-scale clinical data processing.

[0106] Figure 9 The ROC curve for determining the presence of moderate to severe coronary artery stenosis using LAD-PDV is shown in the figure. The area under the curve (AUC) is 0.979. According to the ROC curve evaluation criteria, an AUC ≥ 0.9 indicates excellent predictive efficacy, confirming the clinical practical value of this invention in determining moderate to severe coronary artery stenosis using PDV parameters, and providing a reliable basis for clinical risk stratification.

[0107] Figure 10 The ROC curve for determining the presence of LAD stenosis using LAD-PDV is shown in the figure. The area under the curve (AUC) is 0.992, close to the ideal value of 1.0, indicating that the core parameter PDV has a very strong ability to identify coronary artery stenosis lesions and can effectively distinguish between normal and stenotic coronary arteries, providing high-sensitivity and high-specificity technical support for early screening of coronary artery stenosis.

[0108] Experimental results show that the method and system of this invention, through the technical path of "CDFI data positioning and processing + PW parameter extraction + multi-factor computer correction + statistical modeling", solves the technical pain points of insufficient data processing accuracy, low efficiency and large interference from individual differences in traditional methods. It has the advantages of being non-invasive and safe, easy to operate and accurate in calculation, and is suitable for widespread clinical application. It provides a brand-new computer technology solution for non-invasive and accurate analysis of coronary artery disease-related data.

[0109] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A computer-based method for assessing the degree of coronary artery stenosis based on data, characterized in that: The method uses acquired echocardiographic data as the core analysis object. It uses a computer to perform target coronary artery localization processing on the color Doppler flow imaging (CDFI) data in the echocardiographic data, extracts the pulsed Doppler (PW) coronary blood flow velocity parameters from the echocardiographic data, and combines the blood flow velocity parameters with a preset computer-readable coronary blood flow velocity influencing factor database. Through standardized statistical analysis performed by the computer, the method realizes the localization correlation and quantitative calculation of parameters related to the degree of coronary artery stenosis.

2. The computer implementation method according to claim 1, characterized in that: The target coronary vessels for computer-based localization processing include the ultrasound image data areas corresponding to the left anterior descending coronary artery, the left circumflex coronary artery, the right main coronary artery, and their branches at all levels.

3. The computer implementation method according to claim 1, characterized in that: The blood flow velocity parameters extracted by the computer include peak systolic velocity, peak diastolic velocity, average velocity, and blood flow velocity integral. The core analytical parameter is the peak diastolic velocity at the lesion site in the ultrasound data corresponding to the suspected coronary artery stenosis area.

4. The computer implementation method according to claim 1, characterized in that: The pre-defined computer-readable database of factors affecting coronary flow velocity covers one or more computer-readable structured data, including age, gender, blood pressure, heart rate, blood lipid level, blood glucose concentration, degree of myocardial hypertrophy, cardiac function classification, coronary artery anatomical variations, number of coronary artery lesions, smoking history, drinking history, duration of diabetes, duration of hypertension, obesity index, left ventricular ejection fraction, valvular disease status, anemia status, electrolyte level, and medication use history.

5. The computer implementation method according to claim 1, characterized in that: The standardized statistical analyses performed by the computer include Pearson product-moment correlation analysis, simple linear regression analysis, and receiver operating characteristic (ROC) curve analysis. A linear regression equation Y=aX+b was constructed using a computer, where Y is a parameter related to the degree of coronary artery stenosis (in %), X is the peak diastolic velocity at the site of coronary artery stenosis (in cm / s), a is the regression coefficient, and b is a constant term. By combining the optimal diagnostic cutoff value of the ROC curve calculated by computer, accurate quantitative calculation of parameters related to the degree of coronary artery stenosis can be achieved.

6. The computer implementation method according to claim 1, characterized in that: The extraction conditions for pulse Doppler related data in the echocardiogram data are as follows: the sampling volume is placed in the center of the coronary artery lumen, the sampling volume size is 2.0-5.0 mm, the angle between the ultrasound beam and the blood flow direction is ≤60°, the blood flow velocity spectrum of at least 3 stable cardiac cycles is continuously recorded during data acquisition, and the computer performs interference signal and artifact removal on the data before using it for parameter analysis.

7. A computer-based system for assessing the degree of coronary artery stenosis based on data, characterized in that: The system includes a data input module, a multi-dimensional clinical data integration module, and an intelligent data analysis and evaluation module. The system uses the computer implementation method of any one of claims 1-6 to process the collected echocardiographic data and realize the localization, correlation, and quantitative analysis of coronary artery stenosis-related parameters.

8. The computer system according to claim 7, characterized in that: The intelligent data analysis and evaluation module includes a computer-readable database of preset linear regression equations, a database of optimal diagnostic cutoff values ​​for ROC curves, and a computer-executed coronary flow velocity influencing factor correction unit. Based on the computer-readable database of coronary flow velocity influencing factors, the coronary flow velocity influencing factor correction unit performs individualized correction on the extracted blood flow velocity parameters, so that the system's calculation accuracy for parameters related to the degree of coronary artery stenosis is ≥92%.

9. The computer system according to claim 7, characterized in that: The system also includes computer-compatible auxiliary components, which are selected from one or more of the following: a surface electrocardiogram synchronous data interface module, a multi-source clinical data structured input interface, a computer-readable data storage and update module, and an analysis report generation module; the application form of the system is selected from one or more of the following: ultrasound equipment matching computer software module, independent data analysis computer terminal, ultrasound equipment integrated computer processing system, and remote data processing and evaluation platform.

Images