Atherosclerosis marker detection data integration analysis method and system

Abstract

The present invention relates to the field of medical data analysis, and in particular to a method and system for integrated analysis of atherosclerosis marker detection data. The method comprises the following steps: obtaining multi-dimensional biometric information and medical images of atherosclerosis markers from a patient; adaptively normalizing the multi-dimensional biometric information to obtain standardized biometric information; performing multi-time point visual recognition and plaque geometry analysis on the medical images of atherosclerosis markers to generate geometric features of the plaque at each time point; and tracking multi-time point morphological changes and performing nonlinear morphological trend trajectory evolution based on the geometric features of the plaque at each time point to construct a temporal morphological evolution trajectory of the sclerotic plaque. The present invention achieves personalized and efficient detection and analysis of atherosclerosis markers, thereby effectively improving the accuracy of atherosclerosis prediction.

Description

Technical Field

[0001] The present invention relates to the field of medical data analysis, and in particular to a method and system for integrated analysis of atherosclerosis marker detection data. Background Art

[0002] Atherosclerosis is a common chronic disease characterized by the formation of atherosclerotic plaques within blood vessel walls, leading to narrowing and hardening of the vessels. In severe cases, this can lead to cardiovascular and cerebrovascular diseases, and can even be life-threatening. With an aging population and changing lifestyles, the incidence of atherosclerosis is increasing year by year, becoming one of the leading causes of death and disability worldwide. Therefore, timely and accurate detection and assessment of the development and progression of atherosclerosis are crucial.

[0003] Traditional methods for detecting atherosclerosis usually rely on imaging examinations, blood biomarker tests and other means. Imaging examinations such as ultrasound, CT, and MRI can provide intuitive images of vascular structures, helping doctors identify and evaluate the location and size of plaques and their impact on blood flow. However, imaging examinations require a high level of professional skills and experience from doctors, and cannot directly provide dynamic information on intravascular biomarkers. Although blood biomarker testing can provide a quantitative assessment of the early risk of atherosclerosis, traditional single markers are often difficult to fully and accurately reflect the complex pathological process of atherosclerosis, and their diagnostic effect and predictive accuracy are low.

[0004] Therefore, the demand for integrated analysis methods for atherosclerosis marker detection data is becoming increasingly prominent. How to effectively integrate and analyze data from different data sources through innovative technical solutions to provide more accurate and comprehensive disease prediction and assessment has become a key issue in the current scientific research field. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention proposes a method and system for integrating and analyzing atherosclerosis marker detection data to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a method for integrated analysis of atherosclerosis marker detection data, comprising the following steps:

[0007] Step S1: obtaining multi-dimensional biological detection index information of a patient's atherosclerosis markers and medical images of atherosclerosis markers; performing adaptive standardization processing on the multi-dimensional biological detection index information to obtain standardized biological detection information;

[0008] Step S2: Perform multi-time point visual recognition on the medical images of sclerosis markers and perform plaque geometry analysis to generate plaque geometry features at each time point;

[0009] Step S3: tracking the morphological changes at multiple time points and performing nonlinear morphological trend trajectory evolution based on the geometric morphological characteristics of the plaque at each time point to construct a temporal morphological evolution trajectory of the sclerotic plaque;

[0010] Step S4: mining potential associations among multiple indicators of standardized biological detection information, and then fitting the evolution of the situation over time to construct a multi-dimensional detection indicator situation evolution diagram;

[0011] Step S5: Based on the multi-dimensional detection index trend evolution diagram, a personalized prediction of the temporal morphological evolution trajectory of the sclerotic plaque is performed for multiple time periods in the future, thereby obtaining the prediction data of the atherosclerosis status in multiple time periods;

[0012] Step S6: Quantitatively analyze the atherosclerosis status prediction data for multiple time periods and perform personalized comprehensive diagnosis to generate personalized diagnosis results.

[0013] This invention integrates different types of data (biological test indicators and medical images) to provide a comprehensive understanding of a patient's atherosclerosis status. The multidimensionality of this information provides additional perspectives for subsequent analysis, facilitating a comprehensive assessment of the severity and progression of the disease. By comprehensively considering both biological test and imaging data, the limitations of a single data source can be overcome, errors can be reduced, diagnostic accuracy can be improved, and the comprehensiveness and accuracy of the analysis process can be ensured. Different biomarkers vary in their detection methods, ranges, and units. Adaptive normalization eliminates these differences, ensuring data consistency and comparability across indicators. After normalization, data can be used in multiple analytical methods without bias due to varying data scales, thus ensuring the accuracy of subsequent analysis results. Through normalization, biomarker data from different sources can be analyzed within the same framework, ensuring cross-domain compatibility and further improving analytical accuracy. By analyzing imaging data at different time points, the evolution of atherosclerosis can be tracked in real time, capturing changes in plaque morphology and location at each moment. Multi-time point image recognition helps doctors better understand the dynamics of the disease, enabling timely intervention and treatment. Precise geometric morphological analysis extracts key plaque features (such as size, shape, and boundary smoothness). These geometric features are key indicators for assessing atherosclerotic plaque stability and rupture potential, aiding disease prediction and risk assessment. Multi-point morphological tracking clearly demonstrates the temporal trends of atherosclerotic plaques. This dynamic tracking reveals whether plaques are growing, deforming, or rupturing, helping to accurately predict disease progression. Modeling plaque evolution using nonlinear morphological change models captures the complex patterns underlying plaque morphological changes. Traditional linear analysis overlooks this complexity, while nonlinear analysis provides more accurate and comprehensive trend predictions, making disease assessment more reliable. Multi-indicator latent association analysis identifies complex interactions between different biomarkers. This is crucial for understanding the systemic characteristics of the disease and key factors in disease progression. This approach enables the discovery of new biomarker combinations and improves the sensitivity of early diagnosis. By fitting the changing trends of individual biomarkers at different time points based on the progression of a time series, this helps predict future health status, identify potential risks in advance, and provide more targeted patient monitoring. By constructing a multidimensional trend evolution diagram, the evolving trends of various biomarkers at different time points and their relationship with plaque morphological changes are intuitively displayed. This graphical display allows clinicians to intuitively understand a patient's disease progression and potential future risks. By combining multidimensional data, a personalized disease prediction model is tailored for each patient. Taking into account the changing patterns of markers and disease characteristics of each patient, personalized predictions can provide more precise health management recommendations.By quantitatively analyzing predicted data across multiple time periods, we can accurately assess the risk and development trends of atherosclerosis. This quantitative analysis provides more reliable information than a single result, helping clinicians develop more appropriate treatment plans. Combining a patient's personal data, historical health records, and disease development trends, we provide a comprehensive diagnosis and personalized health recommendations. This personalized diagnosis maximizes treatment effectiveness and quality of life based on the patient's specific circumstances.

[0014] In this specification, a system for integrated analysis of atherosclerosis marker detection data is provided, which is used to execute the method for integrated analysis of atherosclerosis marker detection data as described above, comprising:

[0015] A data processing module is used to obtain multi-dimensional biological detection indicator information of the patient's atherosclerosis markers and medical images of atherosclerosis markers; and adaptively standardize the multi-dimensional biological detection indicator information to obtain standardized biological detection information;

[0016] The geometric morphology analysis module is used to perform multi-time point visual recognition of medical images of sclerosis markers and perform plaque geometric morphology analysis to generate the geometric morphology characteristics of plaques at each time point;

[0017] The morphological trajectory evolution module is used to track the morphological changes at multiple time points and perform nonlinear morphological trend trajectory evolution based on the geometric morphological characteristics of the plaque at each time point, so as to construct the temporal morphological evolution trajectory of the sclerotic plaque;

[0018] The situation evolution module is used to mine the potential correlation of multiple indicators in standardized biological detection information, and then perform time-series situation evolution fitting to construct a multi-dimensional detection indicator situation evolution diagram;

[0019] The multi-period prediction module is used to make personalized predictions of the temporal evolution trajectory of sclerotic plaques based on the multi-dimensional detection indicator trend evolution diagram, thereby obtaining prediction data on the state of atherosclerosis in multiple time periods;

[0020] The personalized diagnosis and analysis module is used to quantitatively analyze the atherosclerosis status prediction data of multiple time periods and perform personalized comprehensive diagnosis to generate personalized diagnosis results.

[0021] This invention systematically integrates multiple types of data to obtain patient health information, thereby improving the accuracy of disease analysis. This multi-dimensional data fusion provides a more comprehensive basis for subsequent analysis. Adaptive normalization eliminates scale differences between data, ensuring comparability between different biomarkers and imaging data. This not only improves data processing efficiency but also reduces errors caused by data inconsistencies during subsequent modeling, thereby enhancing the reliability of analysis. Visual recognition of images at multiple time points enables tracking the evolution of plaques and capturing the course of atherosclerosis in real time. Multi-time point analysis provides temporal data on disease progression, helping to understand the dynamics of plaque development. Geometric morphological analysis enables detailed quantification of plaque characteristics such as size and shape. This information is crucial for predicting plaque stability, rupture risk, and disease progression, providing clinicians with more decision-making support. Multi-time point morphological tracking accurately captures temporal trends in plaque morphology. This dynamic change provides a reliable basis for early warning of disease. Nonlinear evolutionary models capture complex trends in plaque morphological changes, rather than simple linear relationships. This provides deeper insights into disease progression, particularly for predicting sudden events such as plaque rupture. By exploring potential correlations between biomarkers, new early disease predictors can be discovered, which is crucial for accurately identifying high-risk patients. Time-series analysis reveals patterns in disease marker changes over time, providing quantifiable insights for prognosis and early intervention, and enabling clinicians to deliver more accurate predictions. Based on multidimensional patient data and individual health status, personalized disease predictions are generated. Future predictions allow patients to understand disease progression in advance, allowing them to adjust treatment plans promptly and prevent further progression. Accurately predicting the disease's status at a specific point in the future helps patients and physicians develop more informed treatment strategies and reduce the risk of complications. Quantitative analysis quantifies a patient's health status, providing a more refined risk assessment. This quantitative assessment facilitates the development of targeted intervention strategies. By integrating multiple facets of information, a personalized comprehensive diagnostic report is generated, providing more accurate and personalized treatment recommendations than traditional diagnostic methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a schematic flow chart of the steps of a method for integrating and analyzing atherosclerosis marker detection data according to the present invention;

[0023] Figure 2 Detailed implementation flow chart of step S1;

[0024] Figure 3 Detailed implementation flow chart of step S2;

[0025] Figure 4 Schematic diagram of the detailed implementation steps of step S3. DETAILED DESCRIPTION

[0026] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0027] This application provides a method and system for integrated analysis of atherosclerosis marker detection data. The execution entities of this method and system include, but are not limited to, mechanical equipment, data processing platforms, cloud server nodes, network upload devices, etc., which can be considered general computing nodes of this application. The data processing platform includes, but is not limited to, at least one of an audio and image management system, an information management system, and a cloud data management system.

[0028] See also Figures 1 to 4 The present invention provides an integrated analysis method for atherosclerosis marker detection data, which comprises the following steps:

[0029] Step S1: obtaining multi-dimensional biological detection index information of a patient's atherosclerosis markers and medical images of atherosclerosis markers; performing adaptive standardization processing on the multi-dimensional biological detection index information to obtain standardized biological detection information;

[0030] Step S2: Perform multi-time point visual recognition on the medical images of sclerosis markers and perform plaque geometry analysis to generate plaque geometry features at each time point;

[0031] Step S3: tracking the morphological changes at multiple time points and performing nonlinear morphological trend trajectory evolution based on the geometric morphological characteristics of the plaque at each time point to construct a temporal morphological evolution trajectory of the sclerotic plaque;

[0032] Step S4: mining potential associations among multiple indicators of standardized biological detection information, and then fitting the evolution of the situation over time to construct a multi-dimensional detection indicator situation evolution diagram;

[0033] Step S5: Based on the multi-dimensional detection index trend evolution diagram, a personalized prediction of the temporal morphological evolution trajectory of the sclerotic plaque is performed for multiple time periods in the future, thereby obtaining the prediction data of the atherosclerosis status in multiple time periods;

[0034] Step S6: Quantitatively analyze the atherosclerosis status prediction data for multiple time periods and perform personalized comprehensive diagnosis to generate personalized diagnosis results.

[0035] This invention integrates different types of data (biological test indicators and medical images) to provide a comprehensive understanding of a patient's atherosclerosis status. The multidimensionality of this information provides additional perspectives for subsequent analysis, facilitating a comprehensive assessment of the severity and progression of the disease. By comprehensively considering both biological test and imaging data, the limitations of a single data source can be overcome, errors can be reduced, diagnostic accuracy can be improved, and the comprehensiveness and accuracy of the analysis process can be ensured. Different biomarkers vary in their detection methods, ranges, and units. Adaptive normalization eliminates these differences, ensuring data consistency and comparability across indicators. After normalization, data can be used in multiple analytical methods without bias due to varying data scales, thus ensuring the accuracy of subsequent analysis results. Through normalization, biomarker data from different sources can be analyzed within the same framework, ensuring cross-domain compatibility and further improving analytical accuracy. By analyzing imaging data at different time points, the evolution of atherosclerosis can be tracked in real time, capturing changes in plaque morphology and location at each moment. Multi-time point image recognition helps doctors better understand the dynamics of the disease, enabling timely intervention and treatment. Precise geometric morphological analysis extracts key plaque features (such as size, shape, and boundary smoothness). These geometric features are key indicators for assessing atherosclerotic plaque stability and rupture potential, aiding disease prediction and risk assessment. Multi-point morphological tracking clearly demonstrates the temporal trends of atherosclerotic plaques. This dynamic tracking reveals whether plaques are growing, deforming, or rupturing, helping to accurately predict disease progression. Modeling plaque evolution using nonlinear morphological change models captures the complex patterns underlying plaque morphological changes. Traditional linear analysis overlooks this complexity, while nonlinear analysis provides more accurate and comprehensive trend predictions, making disease assessment more reliable. Multi-indicator latent association analysis identifies complex interactions between different biomarkers. This is crucial for understanding the systemic characteristics of the disease and key factors in disease progression. This approach enables the discovery of new biomarker combinations and improves the sensitivity of early diagnosis. By fitting the changing trends of individual biomarkers at different time points based on the progression of a time series, this helps predict future health status, identify potential risks in advance, and provide more targeted patient monitoring. By constructing a multidimensional trend evolution diagram, the evolving trends of various biomarkers at different time points and their relationship with plaque morphological changes are intuitively displayed. This graphical display allows clinicians to intuitively understand a patient's disease progression and potential future risks. By combining multidimensional data, a personalized disease prediction model is tailored for each patient. Taking into account the changing patterns of markers and disease characteristics of each patient, personalized predictions can provide more precise health management recommendations.By quantitatively analyzing predicted data across multiple time periods, we can accurately assess the risk and development trends of atherosclerosis. This quantitative analysis provides more reliable information than a single result, helping clinicians develop more appropriate treatment plans. Combining a patient's personal data, historical health records, and disease development trends, we provide a comprehensive diagnosis and personalized health recommendations. This personalized diagnosis maximizes treatment effectiveness and quality of life based on the patient's specific circumstances.

[0036] In the embodiment of the present invention, see Figure 1 , is a schematic flow chart of the steps of a method for integrated analysis of atherosclerosis marker detection data according to the present invention. In this example, the steps of the method for integrated analysis of atherosclerosis marker detection data include:

[0037] Step S1: obtaining multi-dimensional biological detection index information of a patient's atherosclerosis markers and medical images of atherosclerosis markers; performing adaptive standardization processing on the multi-dimensional biological detection index information to obtain standardized biological detection information;

[0038] In this example, after obtaining authorization from the hospital and the patient, multi-dimensional biometric testing is performed. Imaging techniques such as ultrasound, CT, or MRI are used to examine the patient and obtain medical images of sclerosis markers. Images should include arterial cross-sections and lengths, as well as the specific location and morphological characteristics of plaques. Standardization of equipment settings (e.g., resolution and contrast agent usage) during image acquisition is ensured to improve image quality and comparability. Biometric indicators and medical imaging data are integrated into a unified database. Each patient should have a unique identifier to facilitate subsequent analysis and tracking. Biometric indicator data should include the test date, indicator name, measurement value, and unit to ensure data integrity. Collected data is cleaned, and missing values, outliers, and duplicate records are identified and addressed. Missing values ​​are addressed using mean imputation or interpolation, while outliers are detected and addressed using box plots or Z-score methods. An appropriate standardization method is selected, typically Z-score or Min-Max standardization. Each biometric indicator is standardized. After calculating the mean and standard deviation for each indicator, the Z-score formula is applied for transformation. Ensure that the data types remain consistent during the standardization process and record the standardized parameters of each indicator for future de-standardization. After the standardization process is completed, check the distribution of the standardized results and draw a histogram or box plot to ensure that the mean of the data is 0 and the standard deviation is 1 to verify the effectiveness of the standardization.

[0039] Step S2: Perform multi-time point visual recognition on the medical images of sclerosis markers and perform plaque geometry analysis to generate plaque geometry features at each time point;

[0040] In this embodiment, relevant image data are extracted from the patient's medical imaging database. These image data are from ultrasound, CT, or MRI, ensuring good image quality and sufficient resolution (for example, the resolution of CT images should reach 1 mm³, and the resolution of ultrasound images should reach greater than 500 μm). The images at each time point should be clearly labeled to facilitate subsequent comparison and analysis. The extracted images are preprocessed, including denoising, contrast enhancement, and normalization. Techniques such as high-pass filtering and histogram equalization are used to improve image clarity and contrast. This is achieved using open source image processing libraries (such as OpenCV):

[0041] import cv2

[0042] # Read the image

[0043] image = cv2.imread('image_path')

[0044] # Denoise

[0045] denoised_image = cv2.fastNlMeansDenoising(image, None, 30, 7, 21)

[0046] # Histogram equalization

[0047] equalized_image = cv2.equalizeHist(denoised_image)

[0048] Select an appropriate image recognition algorithm for plaque detection and segmentation. Use deep learning models such as convolutional neural networks (CNNs) or U-Net, which have demonstrated excellent performance in medical image analysis and are particularly well-suited for handling complex plaque morphologies. Train the deep learning model using a well-labeled image dataset. The dataset should include a variety of plaque types and morphologies to improve the model's generalization ability. During model training, use cross-validation to monitor model performance and ensure accuracy and robustness. During training, set hyperparameters (such as the learning rate and batch size) and optimize using an appropriate loss function (such as the cross-entropy loss function). Use the trained model to detect plaques on images at each time point. Calculate the output through forward propagation to obtain the plaque boundary and area. This process should record the pixel coordinates of the plaque and their corresponding shape features. Perform geometric shape feature analysis on the segmented plaques, extracting the following features: Area (the total number of pixels in the plaque image); Perimeter (the length of the plaque boundary, calculated from the boundary points); Shape Factor (the circularity of the plaque); Aspect Ratio (the ratio of the length to the width of the plaque, reflecting the degree of plaque stretching). Organize the extracted geometric feature information (such as area, perimeter, shape factor, etc.) into a data frame (DataFrame), ensuring that the feature information at each time point is clearly visible. Each row should contain a timestamp, patch ID, and corresponding geometric feature.

[0049] Step S3: tracking the morphological changes at multiple time points and performing nonlinear morphological trend trajectory evolution based on the geometric morphological characteristics of the plaque at each time point to construct a temporal morphological evolution trajectory of the sclerotic plaque;

[0050] In this embodiment, an appropriate algorithm is selected to track plaque morphological changes, typically using dynamic time warping (DTW) or morphological analysis methods. DTW is suitable for processing data with time series characteristics and can capture morphological change trends between different time points. The selected algorithm is used to analyze the geometric features of each plaque at different time points. First, the geometric feature vector is calculated for each time point. The DTW algorithm is then applied to calculate the similarity between time series and identify changes between adjacent time points. The changes in the plaque at different time points are recorded in a data frame, and a change trajectory graph is generated. The results are visualized using a line graph or scatter plot to display the changes in the plaque along the time axis, facilitating observation of the plaque's evolutionary trends. An appropriate nonlinear model is selected to analyze the evolution of the morphological trend trajectory. Commonly used methods include polynomial regression, spline regression, or Gaussian process regression. These models can handle complex nonlinear relationships and are suitable for describing changes in plaque morphology over time. Using the organized plaque geometric feature data, a nonlinear regression model is constructed. For each plaque, appropriate features are selected as independent variables (such as time, area, etc.) and fitted as dependent variables. For example, a polynomial regression model can be used with different polynomial orders and the optimal model selected through cross-validation. During training, the model's goodness of fit (such as the R² value) is monitored to ensure accuracy. After model fitting is complete, a temporal morphological evolution trajectory for each plaque is generated. The model results are analyzed to observe trends in plaque morphology (such as growth, shrinkage, and shape variation). The evolution trajectory is visualized, generating dynamic charts that illustrate the dynamic changes in plaque morphology over time, facilitating analysis and decision-making for clinicians.

[0051] Step S4: mining potential associations among multiple indicators of standardized biological detection information, and then fitting the evolution of the situation over time to construct a multi-dimensional detection indicator situation evolution diagram;

[0052] In this embodiment, all standardized biological detection indicator data are collected, including blood lipids, blood sugar, inflammatory markers, etc. These data should cover all relevant time points to ensure their integrity and consistency. The biological indicators at each time point should be organized in the form of a data frame (DataFrame) to ensure that the data is clear and easy to read. Clean the data, identify and process missing values ​​and outliers. Missing values ​​are processed by mean filling or interpolation, and outliers can be detected by the Z-score method. The cleaning process ensures the accuracy and stability of subsequent analysis. Select appropriate statistical analysis methods to mine potential associations among multiple indicators. Commonly used methods include Pearson correlation coefficient analysis or Spearman rank correlation coefficient, which can effectively identify linear or nonlinear associations between indicators. Correlation is calculated for each pair of indicators in the standardized data. The correlation results are recorded in a new data frame, and a correlation heat map is generated to visualize the association between indicators. Use the Seaborn library to draw a heat map:

[0053] import seaborn as sns

[0054] import matplotlib.pyplot as plt

[0055] plt.figure(figsize=(10, 8))

[0056] sns.heatmap(correlation_matrix, annot=1, fmt='.2f', cmap='coolwarm')

[0057] plt.title('Correlation Heatmap of Biological Indicators')

[0058] plt.show()

[0059] Select an appropriate time series model to fit the evolving situation. Commonly used models include the Autoregressive Integrated Moving Average (ARIMA) model or the Long Short-Term Memory (LSTM) network, which effectively capture trends and seasonality in time series data. Split the integrated dataset into a training set and a test set, typically using 70% for training and 30% for testing. Ensure data randomness to improve the model's generalization ability. Use the training set to train the selected time series model. For the ARIMA model, determine the optimal p, d, and q parameters, using the AIC or BIC criteria for selection. For the LSTM model, set appropriate hyperparameters (such as the learning rate and batch size) and use cross-validation to monitor model performance.

[0060] After training is completed, the model effect is evaluated on the test set, using the mean square error (MSE) or root mean square error (RMSE) as the evaluation indicator:

[0061] from sklearn.metrics import mean_squared_error

[0062] predictions = model.predict(X_test)

[0063] mse = mean_squared_error(y_test, predictions)

[0064] Based on the output of the fitted model, a trend chart of the multi-dimensional detection indicators is generated. The visualization results include time series graphs, showing the changing trends of each indicator at different time points, facilitating clinical analysis and decision-making.

[0065] Step S5: Based on the multi-dimensional detection index trend evolution diagram, a personalized prediction of the temporal morphological evolution trajectory of the sclerotic plaque is performed for multiple time periods in the future, thereby obtaining the prediction data of the atherosclerosis status in multiple time periods;

[0066] In this example, data on the temporal morphological evolution of sclerotic plaques is collected, including geometric features (such as area, perimeter, and shape factor) at each time point. This data should be combined with biological indicators to form a dataset containing multidimensional features. Data cleaning is performed to address missing values ​​and outliers to ensure data integrity and accuracy. Missing values ​​can be addressed using mean imputation or interpolation, while outliers can be detected and removed using methods such as the Z-score. An appropriate prediction model is selected for personalized predictions for multiple future time periods. Commonly used models include long short-term memory (LSTM), support vector machines (SVM), or random forests. These models are capable of processing time series data and are suitable for capturing the complex relationship between plaque morphology and biological indicators. The integrated dataset is divided into training and test sets. Typically, 70% is used for training and 30% for testing. In the training set, ensure that biological indicators and plaque characteristics from multiple time points are included to improve model learning. The selected personalized prediction model is trained using the training set. For the LSTM model, appropriate hyperparameters such as the learning rate and batch size are set, and cross-validation is used to monitor model performance. During the training process, the training and validation loss functions are monitored to prevent overfitting. After the model training is completed, the plaque geometry and biological indicator data at the last known time point are used as input to generate prediction data for multiple time periods in the future. These time periods are set according to clinical needs, such as status predictions for the next 3 months, 6 months, and 12 months. The input data is passed into the trained model to generate prediction data for the atherosclerosis status at future time points. These prediction data should include changes in the geometric characteristics of the plaque, expected values ​​of relevant biological indicators, etc., to form a comprehensive prediction result. The generated prediction results are recorded in a data frame to ensure that the prediction information for each time period is clearly visible. The changing trend of the future atherosclerosis status is displayed through visualization tools (such as charts or heat maps) to facilitate clinicians' analysis and decision-making.

[0067] Step S6: Quantitatively analyze the atherosclerosis status prediction data for multiple time periods and perform personalized comprehensive diagnosis to generate personalized diagnosis results.

[0068] In this example, the collected data was collated to ensure that the predicted data for each patient was centrally located and sorted by time period. Missing values ​​and outliers were checked for data, and missing values ​​were imputed using methods such as mean imputation. Outliers were detected and removed using methods such as the Z-score to ensure the accuracy of subsequent analysis. Descriptive statistical analysis was performed on the predicted atherosclerosis status data for each time period. Statistics such as the mean, standard deviation, maximum, and minimum values ​​were calculated for each time period to assess changes in status over time. For example, the mean plaque area was calculated for different time periods to observe its trend over time. The standard deviation of inflammatory marker levels was calculated to understand the degree of variability within the patient population. Correlation analysis methods (such as the Pearson correlation coefficient or the Spearman rank correlation coefficient) were used to assess potential associations between the various biomarkers in the predicted data. A correlation matrix was calculated to identify indicators that showed significant correlations across different time periods, facilitating subsequent comprehensive diagnosis. The quantitative analysis results were compiled into a report, documenting key statistical indicators and correlation analysis results for each time period. The data was visualized for easier understanding, for example, using charts to display changes in plaque area and trends in biomarkers over different time periods. Select an appropriate comprehensive assessment model and combine biological indicators, plaque geometric characteristics and the patient's clinical background information for personalized comprehensive diagnosis. The model is a score-based assessment system, such as the Framingham risk score or the SCORE model, which assesses the risk of cardiovascular events based on multiple risk factors. The predicted data of each patient's atherosclerosis status at different time periods are summarized to form a personalized input data set. The input data should include various biological indicators, plaque characteristics, patient age, gender and other information. Input the prepared data into the selected comprehensive assessment model to calculate the personalized risk score for each patient at different time periods. Based on the evaluation results of the model, patients are divided into different risk levels (such as low risk, medium risk, high risk), and corresponding diagnostic recommendations are generated.

[0069] In this embodiment, refer to Figure 2 , is a flowchart of the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0070] Step S11: obtaining multi-dimensional biological detection index information of the patient's atherosclerosis markers and medical images of atherosclerosis markers;

[0071] Step S12: performing marker intelligent classification on the multi-dimensional biological detection indicator information to obtain marker intelligent classification data;

[0072] Step S13: Screening key markers based on the marker intelligent classification data, and extracting multi-dimensional biological detection indicator information of the key markers;

[0073] Step S14: Adaptively standardize the multi-dimensional biological detection index information of the key markers to obtain standardized biological detection information.

[0074] In this example, after obtaining authorization from the hospital and the patient, multidimensional biometric testing was performed, including blood biochemical markers (such as cholesterol, triglycerides, and C-reactive protein). These markers can reflect biomarkers of atherosclerosis. Samples were standardized within the laboratory to ensure that the collection process adhered to bioethical and experimental standards. Imaging data of arteriosclerosis were obtained from the patient using imaging modalities (such as ultrasound, CT, or MRI). Image clarity and resolution were ensured to facilitate subsequent image analysis. The imaging data should be annotated with the patient's basic information and clinical background to facilitate data integration. The collected biometric indicators and medical imaging data were formatted to ensure that all data could be processed on the same platform. Data management software (such as Excel or SQL databases) was used for data import and integration. The collected data was checked for completeness and consistency, and missing values, outliers, and duplicate records were removed to ensure data quality. Reasonable ranges were set for the biometric indicators, and obviously erroneous physiological values ​​were excluded. Feature selection was performed on the multidimensional biometric indicators to identify key indicators related to atherosclerosis. Potential biomarkers were identified through literature review and expert consultation. Preprocess the bioassay indicators, including imputing missing values, normalizing and standardizing data, to ensure that different features are on the same scale for subsequent analysis. Select an appropriate machine learning algorithm for marker classification, such as random forest, support vector machine (SVM), or neural network. Random forest is suitable for processing high-dimensional data and can effectively perform feature selection and classification. Divide the dataset into a training set (80%) and a test set (20%). Use the training set to train the selected model and adjust the model's hyperparameters (such as the number of trees and maximum depth). Evaluate the model using metrics such as accuracy, F1 score, and receiver operating characteristic (ROC) curve. Select the best-performing model and apply it to the test set for intelligent marker classification. Store the classification results in a database to generate marker intelligent classification data, including each marker's category label, predicted probability, and other information, for subsequent analysis. Based on the intelligent classification results, set key marker screening criteria, such as selecting markers with a predicted probability greater than 0.75 or assessing the relevance of the markers based on the opinions of clinical experts. A threshold or ranking approach can be used to sort markers based on their importance scores (such as Gini importance or information gain), selecting the top N markers as key markers. Multidimensional bioassay indicator information for the selected key markers should be extracted from the original dataset, including the specific values ​​of each marker and relevant clinical information. This extracted key marker information should be integrated with the patient's basic information to generate a comprehensive data table containing the key markers and their corresponding indicators. Appropriate normalization methods should be used, such as Z-score normalization or Min-Max normalization.Z-score normalization is applicable to features of various distributions, transforming each feature value to a mean of 0 and a standard deviation of 1, while Min-Max normalization scales the data to the interval [0, 1]. Adaptive normalization: Dynamically selects a normalization method based on the distribution characteristics of key markers. Z-score normalization is used for most features, while Min-Max normalization is used for features with significant skewed distributions. Standardization is performed using the Python Scikit-learn library. For each marker, the mean and standard deviation are calculated and transformed according to the formula. The normalized data is stored in a new data frame to facilitate subsequent analysis and evaluation of the normalized data distribution, ensuring that the data for each marker is on the same scale and avoiding bias during model training. Use visualization tools (such as Seaborn) to plot the data distribution before and after normalization to ensure the effectiveness of normalization.

[0075] In this embodiment, the specific steps of step S14 are:

[0076] Perform parameterized data recognition on the multi-dimensional biological detection indicator information of key markers and calculate the average parameter threshold to obtain the parameterized average threshold;

[0077] Detect abnormal outlier data based on parameterized average threshold and mark abnormal outlier data points;

[0078] Perform outlier filtering and cleaning on abnormal outlier data points to obtain cleaning optimization indicator information;

[0079] Perform multi-dimensional difference index analysis on cleaning optimization index information to generate difference data between multi-dimensional data;

[0080] Adaptive standardization processing is performed based on the difference data between multi-dimensional data to obtain standardized biological detection information.

[0081] In this embodiment, a suitable parameterization method is selected, such as normalization or discretization. Normalization can use Z-score normalization to convert the data into a form with a mean of 0 and a standard deviation of 1; discretization can divide continuous data into discrete intervals.

[0082] Implementing parametric identification:

[0083] Use Python's Pandas library to parameterize the selected biological detection indicators. Use the StandardScaler class for standardization:

[0084] from sklearn.preprocessing import StandardScaler

[0085] scaler = StandardScaler()

[0086] standardized_data = scaler.fit_transform(raw_data)

[0087] Calculate the mean and standard deviation of each parameterized metric to generate a parameterized average threshold. Use Pandas’ mean() and std() methods for calculations.

[0088] mean_values ​​= standardized_data.mean(axis=0)

[0089] std_values ​​= standardized_data.std(axis=0)

[0090] A parameterized average threshold is set based on the mean and standard deviation. The threshold range is set as: threshold = mean ± 2 × standard deviation. This range is used for subsequent abnormal outlier data detection.

[0091] Choose an appropriate anomaly detection algorithm, such as a Z-score-based method or a machine learning method like Isolation Forest or Local Outlier Factor (LOF). The Z-score method is suitable for small datasets, while Isolation Forest is more suitable for large datasets. Use the Z-score method to detect outliers. Calculate the Z-score for each data point and compare it to a set threshold. If the absolute value of the Z-score exceeds the threshold, the data point is marked as an outlier.

[0092] z_scores = (standardized_data - mean_values) / std_values

[0093] outliers = (np.abs(z_scores)>2).any(axis=1)

[0094] Mark the detected abnormal outlier data points in the data frame for subsequent processing. For example, add a column is_outlier to the original data frame and mark it as 1 or 0.

[0095] Use simple filtering methods to remove outlier data points marked as abnormal, or choose to fill the values ​​of these data points with the median or mean to reduce their impact on the overall data analysis. Use the drop() method of the Pandas library to remove data points marked as abnormal to ensure the integrity of the cleaned data. For indicators marked as abnormal, use the fillna() method to fill them with the median or mean. If you choose to fill with the median,

[0096] cleaned_data.fillna(cleaned_data.median(), inplace=1)

[0097] Select an appropriate statistical analysis method, such as a t-test, analysis of variance (ANOVA), or multivariate linear regression, to compare differences in indicators between groups. To compare biological indicators between different groups (e.g., healthy versus pathological groups), use the ttest_ind() function in the SciPy library to perform a t-test and calculate a p-value to assess the significance of the difference. Record the results of the differential analysis (including t-values, p-values, and effect sizes) in a data frame to generate a table of differential data results.

[0098] Select an adaptive normalization method, such as Z-score-based normalization or Min-Max normalization. Adaptive normalization should take into account the data distribution characteristics and the results of the variance analysis. Use Scikit-learn's StandardScaler or MinMaxScaler for final data normalization.

[0099] from sklearn.preprocessing import MinMaxScaler

[0100] scaler = MinMaxScaler()

[0101] standardized_final_data = scaler.fit_transform(cleaned_data)

[0102] Check the standardized data to ensure that all indicators are on the same scale, and use visualization tools (such as Seaborn) to draw data distribution graphs before and after standardization.

[0103] In this embodiment, refer to Figure 3 , is a flowchart of the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0104] Step S21: performing global contrast enhancement on the medical image of the sclerosis marker, thereby constructing a global contrast optimized image;

[0105] Step S22: performing multi-time point visual recognition on the global contrast optimized image, marking atherosclerotic plaques at multiple time points;

[0106] Step S23: performing plaque edge detection on atherosclerotic plaques at multiple time points and marking the plaque edge contour lines;

[0107] Step S24: performing semantic segmentation of the image boundary according to the edge contour of the patch, thereby extracting the patch area image at each time point;

[0108] Step S25: performing plaque geometric morphology analysis on the plaque region image at each time point, thereby generating plaque geometric morphology features at each time point.

[0109] In this example, medical imaging data for sclerosis markers was extracted from a medical imaging database to ensure that image quality and resolution met analysis requirements. Selected images should have a resolution of at least 512x512 pixels to ensure accuracy in subsequent processing. Images should be saved in a common format (such as DICOM or PNG) for easy access and processing using image processing tools. A global contrast enhancement algorithm, such as histogram equalization or adaptive histogram equalization (CLAHE), was selected. CLAHE effectively enhances local contrast and reduces noise, making it particularly suitable for medical imaging. Each image was processed using the CLAHE algorithm to enhance contrast. After processing, the image histogram distribution was examined to ensure a significant contrast improvement. The enhanced image should more clearly demonstrate the characteristics of the sclerotic plaque. The images before and after contrast enhancement were visualized and the contrast improvement was confirmed by comparing the histograms. Ideally, the enhanced image histogram should be more uniform and cover a wider grayscale range. An appropriate image recognition algorithm was selected, such as a deep learning convolutional neural network (CNN) or traditional image processing methods such as threshold segmentation. For tasks requiring high accuracy, using a pretrained deep learning model (such as U-Net) is an effective option. If using a deep learning model, a well-labeled training set is required. After training, the model is used to process global contrast-optimized images to identify and label atherosclerotic plaques. Based on the model's predictions, a binary mask image is generated at each time point, marking the location of the atherosclerotic plaque for subsequent analysis and processing. Edge detection can be performed using the Canny edge detection algorithm or the Sobel operator; both effectively identify edge features in images. Implementing edge detection: Apply the Canny edge detection algorithm to the marked plaque regions. This algorithm calculates the gradient of grayscale changes in the image to identify distinct edges and label the plaque's edge contours. Visualize the edge detection results to ensure the accuracy and completeness of the edge detection. Compare the original image with the edge-detected image to confirm that the plaque edges are clearly labeled. Select an appropriate semantic segmentation algorithm, such as U-Net or FCN (Fully Convolutional Network). These algorithms effectively extract plaque regions based on edge information. Perform image segmentation based on the edge contours of the plaques and extract the plaque area image at each time point. Overlay the segmentation results with the original image to verify the accuracy of the extracted area. Ensure that the extracted plaque area meets expectations and record the segmentation results at each time point for subsequent analysis. Determine the geometric features to be extracted, such as the area, perimeter, shape factor, aspect ratio, etc. of the plaque. These features help describe the geometric morphology of the plaque and its changes. Perform geometric morphological analysis on each extracted plaque area image and calculate the relevant features of each plaque. Calculate indicators such as area and perimeter using image processing tools. Organize the geometric morphological features of the plaques at each time point into a table to facilitate subsequent data analysis and comparison.These features will be used to analyze plaque evolution and clinical relevance.

[0110] In this embodiment, the specific steps of step S21 are:

[0111] Perform image noise analysis on medical images of sclerosis markers and extract multiple image noise points;

[0112] Mining pixel features of multiple image noise points one by one to generate pixel-level features of each noise point;

[0113] Perform dynamic filtering and noise reduction based on the pixel-level characteristics of each noise point to generate filtered and denoised medical images;

[0114] Perform multi-region decomposition on the filtered and denoised medical image to generate multiple sub-image blocks;

[0115] Performing histogram equalization processing on the multiple sub-image blocks one by one, thereby generating multiple histogram equalized sub-image blocks;

[0116] Global contrast enhancement is performed based on multiple histogram equalized sub-image blocks to construct a global contrast optimized image.

[0117] In this embodiment, an appropriate noise analysis algorithm, such as median filtering or Gaussian filtering, is selected to identify noise points in the image. Median filtering excels at removing salt and pepper noise, while Gaussian filtering is more suitable for Gaussian noise. First, the medical image is preprocessed, including image denoising and enhancement, to improve the accuracy of subsequent noise analysis. The image is processed using the selected algorithm to identify the location and intensity of noise points within the image. Using a threshold segmentation method, pixels with intensities exceeding a certain threshold are marked as noise points. Pixel features to be extracted are determined, such as brightness, contrast, and texture features (such as entropy and variance). These features help assess the nature and impact of the noise points. Pixel-level feature mining is performed on each extracted noise point. The extracted feature values ​​are obtained by calculating the statistical information of pixels within its local neighborhood. The pixel-level features of each noise point are stored in a data table for subsequent analysis and comparison. Dynamic filtering algorithms, such as adaptive filtering or bilateral filtering, are selected. These algorithms can dynamically adjust filtering parameters based on noise point characteristics, effectively removing noise. Based on the characteristics of each noise point, an adaptive filtering algorithm is applied to the original medical image to reduce the impact of noise on image quality. Evaluate the quality of the filtered image to check whether noise is effectively suppressed and that important structures are not blurred. Determine the size and method of decomposition, such as fixed-size or adaptive segmentation, to ensure that each sub-image block contains sufficient structural information. Divide the processed medical image into multiple sub-image blocks according to the specified rules. Each sub-image block should preserve local details. Record the location information of each sub-image block for subsequent processing and analysis. Select a histogram equalization algorithm to enhance the contrast of each sub-image block. Use traditional histogram equalization or CLAHE. Perform histogram equalization on each sub-image block to enhance its local contrast. After equalization, check the histogram distribution of each sub-image block to ensure the equalization effect. Save the equalized sub-image blocks for subsequent global contrast enhancement. Select a global contrast enhancement algorithm, such as global histogram equalization or adaptive histogram equalization (CLAHE), to improve the overall visual quality of the image. Merge all equalized sub-image blocks and apply the global contrast enhancement algorithm to generate the final global contrast-optimized image. Visualize the final image and confirm the contrast improvement by comparing the histograms of the original and optimized images, and evaluate the image clarity and detail.

[0118] In this embodiment, refer to Figure 4 , is a flowchart of the detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0119] Step S31: performing multi-time point morphological change tracking based on the geometric morphological characteristics of the plaque at each time point to obtain morphological change tracking data of the sclerotic plaque;

[0120] Step S32: performing adjacent time difference calculation on the morphological change tracking data of the sclerotic plaque, thereby obtaining morphological difference data at multiple adjacent time intervals;

[0121] Step S33: performing morphological change trend analysis on the morphological difference data of multiple adjacent time intervals to obtain the temporal morphological change trend of the sclerotic plaque;

[0122] Step S34: performing nonlinear morphological trend trajectory evolution on the temporal morphological change trend of the sclerotic plaque to construct a temporal morphological evolution trajectory of the sclerotic plaque.

[0123] In this example, geometric morphological data of plaques are collected at each time point. These features include the area, perimeter, shape factor, aspect ratio, and other characteristics of the plaque. Ensure that this data is organized chronologically to facilitate subsequent analysis. The morphological feature information for each time point is stored in a data structure, such as a data frame, where each row represents the plaque feature at that time point, and the columns include the timestamp and the corresponding geometric feature. Change calculation: Differences are calculated for the morphological features at each time point to track changes in the plaque from one time point to the next. For example, the percentage change in area between adjacent time points can be calculated as: Percentage change = Area(t+1) − Area t / Area t × 100%. The change from each time point to the next is recorded in a data frame, forming "morphological change tracking data" for subsequent analysis. A specific method for calculating differences between adjacent time points should be determined, for example, using absolute differences or relative rates of change to represent changes. For each geometric feature (such as area, perimeter, etc.), the difference between adjacent time points is calculated. Select an appropriate trend analysis method, such as linear regression analysis, moving average analysis, or time series analysis, to identify trends in morphological change. Apply the selected trend analysis method to the morphological difference data. If linear regression is used, use time as the independent variable and morphological difference as the dependent variable, fit a trend line and calculate the slope. Visualize the analysis results and draw a graph of the relationship between time and morphological difference to intuitively display the morphological change trend of the plaque. Based on the slope and correlation of the trend line, interpret the acceleration or slowdown of morphological changes and evaluate the evolution trend of the plaque. Select a suitable nonlinear model, such as polynomial regression, spline regression or neural network model, to fit complex morphological change trajectories. Construct a nonlinear model based on the time-series morphological change data. Train the model to fit the morphological change data and predict future morphological change trends. Generate a morphological evolution trajectory and visualize the evolution path of the plaque through the model output. Spline curves or polynomial curves are used to describe the evolution trend of the plaque. Verify whether the generated evolution trajectory is reasonable and compare it with the actual observation data. Based on the model output, analyze the evolution law of plaque morphology and explore potential connections with clinical relevance.

[0124] In this embodiment, step S4 includes the following steps:

[0125] Step S41: performing multi-indicator potential association mining on the standardized biological detection information to obtain potential association features between multiple indicators;

[0126] Step S42: Calculating the time series change rate of each indicator of the standardized biological detection information to generate the time series change rate of each detection indicator;

[0127] Step S43: performing a change range analysis based on the standardized biological detection information to obtain a change range of the detection index;

[0128] Step S44: performing a time series change characteristic analysis based on the time series change rate and the change amplitude of each detection indicator to obtain the time series change characteristics of the multi-dimensional detection indicators;

[0129] Step S45: fitting the temporal variation characteristics of the multi-dimensional detection indicators to the temporal evolution trend is performed according to the potential correlation characteristics among the multiple indicators, thereby constructing a multi-dimensional detection indicator trend evolution diagram.

[0130] In this example, relevant indicator data is extracted from standardized bioassay information. Ensure that the data is collected within the same timeframe, typically including various biochemical indicators (such as cholesterol, triglycerides, and inflammatory markers), and that the time points for each indicator are consistent. Organize the data into a format suitable for analysis, such as a data frame, with each row representing a time point and columns representing different indicators. Select an appropriate statistical analysis method, such as correlation analysis (Pearson or Spearman correlation coefficient) or multivariate linear regression analysis, to identify potential correlations between indicators. Calculate the correlation coefficient between each pair of indicators to identify significantly correlated pairs. Set a threshold (such as an absolute value of the correlation coefficient greater than 0.5) to screen for potential correlations. Record the correlation results in a data frame, generate a correlation matrix, and visualize it using a heat map to intuitively understand the degree of association between indicators. For each standardized indicator, calculate the rate of change at each time point to generate rate of change data. Ensure that no time points are omitted during the calculation process, especially when processing long time series data. Record the rate of change for each indicator in a data frame to facilitate subsequent analysis and comparison. Ensure that the data is clearly formatted and linked to the original indicator data. Determine a method for calculating the magnitude of change, using both absolute and relative magnitudes. The formula for calculating absolute magnitude is: Magnitude of change = Indicator max − Indicator min. For each standardized detection indicator, calculate its maximum and minimum values ​​over the entire observation period to determine the magnitude of change. Ensure that all time points are considered in the calculation. Organize the magnitude of change data into a table, recording the magnitude of change for each indicator to facilitate subsequent comparison and analysis. Select appropriate analytical methods, such as cluster analysis or principal component analysis (PCA), to extract the temporal variation characteristics of the multidimensional detection indicators. Based on the rate and magnitude of change, perform clustering or PCA analysis to identify groups of indicators with similar variation characteristics. Use k-means clustering or hierarchical clustering to analyze similarities and differences between indicators. Present the analysis results in a chart to ensure a clear display of the temporal variation characteristics of the multidimensional detection indicators, facilitating subsequent interpretation and application. Select an appropriate fitting model, such as a time series model (such as ARIMA) or a machine learning model (such as random forest), to predict future trends. Use previously discovered potential correlation features to construct and train a model to fit the temporal variation characteristics of the multidimensional detection indicators. Ensure sufficient training data to improve the model's predictive accuracy. Generate a trend chart based on the fitted model's output, showing the changing trends of multiple indicators over future time periods. Use visualization tools to ensure the charts are clear and easy to understand for clinicians and researchers.

[0131] In this embodiment, the specific steps of step S5 are:

[0132] Step S51: performing a deep dynamic change correlation analysis on the temporal morphological evolution trajectory of the sclerotic plaque based on the multi-dimensional detection index trend evolution diagram, thereby obtaining the dynamic evolution law of biological information-plaque evolution;

[0133] Step S52: performing end-to-end iterative learning on the biological information-plaque evolution dynamics law to generate a plaque trajectory dynamics evolution model;

[0134] Step S53: Perform personalized predictions for multiple time periods in the future based on the plaque trajectory dynamic evolution model, thereby obtaining atherosclerosis status prediction data for multiple time periods.

[0135] In this example, relevant biological indicators and temporal morphological evolution data of sclerotic plaques are extracted from a multidimensional indicator trend evolution graph. This data should include biological indicators (such as cholesterol and blood glucose) and plaque geometric characteristics (such as area and perimeter) at each time point. The extracted data is integrated into a comprehensive dataset, ensuring that the biological information at each time point matches the corresponding plaque characteristics. The data format should be a data frame (DataFrame), with each row representing a time point and columns representing different biological indicators and plaque characteristics. Multivariate regression analysis or dynamic time warping (DTW) is selected as the correlation analysis method. Multivariate regression can reveal linear relationships between biological indicators and plaque evolution, while DTW is suitable for processing time series data and can identify similarities between different time series. Data analysis is performed using the selected analysis method. For example, if multivariate regression is used, a model is constructed with biological indicators as independent variables and plaque geometric characteristics as dependent variables to analyze the impact of biological information on plaque evolution. Based on the model results, biological indicators that significantly influence plaque evolution are identified, and corresponding regression plots or time series plots are plotted to visualize the dynamic relationship between biological information and plaque morphological evolution. Select a suitable deep learning framework for modeling, such as a recurrent neural network (RNN) or a long short-term memory network (LSTM). These models are well-suited for time series data and can capture dynamic features within the data. Partition the comprehensive dataset into training and test sets, typically using 70% for training and 30% for testing. Ensure the data split is random to avoid model overfitting. Standardize the input features (such as with the Z-score) to improve model convergence and prediction accuracy. Ensure that each feature has a mean of 0 and a standard deviation of 1. Iteratively train the selected deep learning model using the training set. Set an appropriate learning rate (such as 0.001) and batch size (such as 32), and use cross-validation to monitor model performance. Evaluate model performance on the test set, using mean squared error (MSE) or root mean squared error (RMSE) as an evaluation metric to ensure model generalization. Select an appropriate forecasting method, typically based on the trained dynamic evolution model, for future time periods. Use a sliding window approach to predict future values ​​using past observations as input. By using a trained model and inputting biological indicators and plaque characteristics from the last known time period, forecasts for future time points are generated. The forecast period can be set (e.g., three, six, or 12 months in the future). The forecast results are combined with the original data to form a complete time series, enabling physicians and researchers to analyze the progression of atherosclerosis. Visualization tools (such as Matplotlib) are used to generate forecast charts that demonstrate future state trends. The forecast results can be validated by comparing the observed values ​​with the predicted values ​​to analyze the accuracy of the predictions.

[0136] In this embodiment, step S6 includes the following steps:

[0137] Step S61: performing quantitative analysis on the atherosclerosis status prediction data of multiple time periods to generate status prediction analysis data for each time period;

[0138] Step S62: performing risk prediction on the status prediction analysis data of each time period to generate pathology risk prediction data for each time period;

[0139] Step S63: performing personalized comprehensive diagnosis based on the pathological risk prediction data of each time period to generate a personalized diagnosis result;

[0140] Step S64: performing dynamic feedback optimization on the plaque trajectory dynamic evolution model according to the personalized diagnosis result, thereby constructing a dynamic feedback optimization model.

[0141] In this embodiment, state prediction data is collected for each time period, including plaque characteristics such as size, morphology, and density. Next, this data is processed using statistical analysis methods (e.g., mean, standard deviation, confidence interval, etc.). Multivariate regression analysis is used to assess the impact of different factors on the progression of atherosclerosis. For example, biomarkers such as age, gender, and blood lipid levels are used as independent variables, and plaque development is used as the dependent variable. Experimental parameters include sample size, analysis confidence level (e.g., 95%), and the choice of regression model (e.g., linear regression or logistic regression). After quantitative analysis, state prediction analysis data for each time period is obtained, describing the relationship between plaque change trends and relevant biological information, laying the foundation for risk prediction. Risk prediction is performed on the state prediction analysis data for each time period. Based on the state data generated in step S61, a risk assessment model is constructed using a machine learning algorithm (e.g., random forest, support vector machine, etc.). First, the state prediction data is divided into a training set and a test set for model training and validation. During the training process, appropriate feature variables (such as plaque size, morphological changes, and biomarkers) are selected as input, and the corresponding risk status (such as high risk, moderate risk, and low risk) is output. Cross-validation is used to evaluate the model's accuracy and ensure its reliability in risk prediction. Experimental parameters include the training set ratio (e.g., 70% training, 30% testing), model evaluation metrics (such as accuracy, sensitivity, and specificity), and hyperparameter optimization. A personalized comprehensive diagnosis is performed based on the pathological risk prediction data for each time period. First, the pathological risk prediction results for each patient are collected and analyzed in combination with clinical data (such as medical history, family history, and lifestyle). A decision tree or fuzzy logic system is used to integrate the risk prediction results with the patient's specific circumstances to accurately assess each patient's health risk. A diagnostic model is developed to assign weights to different risk factors to produce a personalized comprehensive diagnosis. Experimental parameters include the depth of the decision tree, the number of nodes, and the number of fuzzy logic rules. Based on the generated personalized diagnostic results, dynamic feedback optimization of the plaque trajectory dynamic evolution model is performed. First, specific data and feedback from the personalized diagnosis are collected and analyzed to analyze their impact on the plaque evolution model. By introducing a feedback mechanism, model parameters can be adjusted to improve prediction accuracy. For example, Bayesian optimization methods can be used to update the model's prior knowledge based on new clinical data, gradually improving model performance. Experimental parameters include the adjustment amplitude, feedback cycle, and model evaluation metrics (such as root mean square error and R² value).

[0142] In this embodiment, a system for integrated analysis of atherosclerosis marker detection data is provided, which is used to execute the above-mentioned method for integrated analysis of atherosclerosis marker detection data, including:

[0143] A data processing module is used to obtain multi-dimensional biological detection indicator information of the patient's atherosclerosis markers and medical images of atherosclerosis markers; and adaptively standardize the multi-dimensional biological detection indicator information to obtain standardized biological detection information;

[0144] The geometric morphology analysis module is used to perform multi-time point visual recognition of medical images of sclerosis markers and perform plaque geometric morphology analysis to generate the geometric morphology characteristics of plaques at each time point;

[0145] The morphological trajectory evolution module is used to track the morphological changes at multiple time points and perform nonlinear morphological trend trajectory evolution based on the geometric morphological characteristics of the plaque at each time point, so as to construct the temporal morphological evolution trajectory of the sclerotic plaque;

[0146] The situation evolution module is used to mine the potential correlation of multiple indicators in standardized biological detection information, and then perform time-series situation evolution fitting to construct a multi-dimensional detection indicator situation evolution diagram;

[0147] The multi-period prediction module is used to make personalized predictions of the temporal evolution trajectory of sclerotic plaques based on the multi-dimensional detection indicator trend evolution diagram, thereby obtaining prediction data on the state of atherosclerosis in multiple time periods;

[0148] The personalized diagnosis and analysis module is used to quantitatively analyze the atherosclerosis status prediction data of multiple time periods and perform personalized comprehensive diagnosis to generate personalized diagnosis results.

[0149] This invention systematically integrates multiple types of data to obtain patient health information, thereby improving the accuracy of disease analysis. This multi-dimensional data fusion provides a more comprehensive basis for subsequent analysis. Adaptive normalization eliminates scale differences between data, ensuring comparability between different biomarkers and imaging data. This not only improves data processing efficiency but also reduces errors caused by data inconsistencies during subsequent modeling, thereby enhancing the reliability of analysis. Visual recognition of images at multiple time points enables tracking the evolution of plaques and capturing the course of atherosclerosis in real time. Multi-time point analysis provides temporal data on disease progression, helping to understand the dynamics of plaque development. Geometric morphological analysis enables detailed quantification of plaque characteristics such as size and shape. This information is crucial for predicting plaque stability, rupture risk, and disease progression, providing clinicians with more decision-making support. Multi-time point morphological tracking accurately captures temporal trends in plaque morphology. This dynamic change provides a reliable basis for early warning of disease. Nonlinear evolutionary models capture complex trends in plaque morphological changes, rather than simple linear relationships. This provides deeper insights into disease progression, particularly for predicting sudden events such as plaque rupture. By exploring potential correlations between biomarkers, new early disease predictors can be discovered, which is crucial for accurately identifying high-risk patients. Time-series analysis reveals patterns in disease marker changes over time, providing quantifiable insights for prognosis and early intervention, and enabling clinicians to deliver more accurate predictions. Based on multidimensional patient data and individual health status, personalized disease predictions are generated. Future predictions allow patients to understand disease progression in advance, allowing them to adjust treatment plans promptly and prevent further progression. Accurately predicting the disease's status at a specific point in the future helps patients and physicians develop more informed treatment strategies and reduce the risk of complications. Quantitative analysis quantifies a patient's health status, providing a more refined risk assessment. This quantitative assessment facilitates the development of targeted intervention strategies. By integrating multiple facets of information, a personalized comprehensive diagnostic report is generated, providing more accurate and personalized treatment recommendations than traditional diagnostic methods.

[0150] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0151] The foregoing description is intended only to provide specific embodiments of the present invention, which are intended to enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for integrated analysis of atherosclerosis marker detection data, characterized in that: The following steps are involved: Step S1: obtaining multi-dimensional biological detection index information of a patient's atherosclerosis markers and medical images of atherosclerosis markers; performing adaptive standardization processing on the multi-dimensional biological detection index information to obtain standardized biological detection information; Step S2: Perform multi-time point visual recognition on the medical images of sclerosis markers and perform plaque geometry analysis to generate plaque geometry features at each time point; Step S3: tracking the morphological changes at multiple time points and performing nonlinear morphological trend trajectory evolution based on the geometric morphological characteristics of the plaque at each time point to construct a temporal morphological evolution trajectory of the sclerotic plaque; Step S4: mining potential associations among multiple indicators of standardized biological detection information, and then fitting the evolution of the situation over time to construct a multi-dimensional detection indicator situation evolution diagram; Step S5: Based on the multi-dimensional detection index trend evolution diagram, a personalized prediction of the temporal morphological evolution trajectory of the sclerotic plaque is performed for multiple time periods in the future, thereby obtaining the prediction data of the atherosclerosis status in multiple time periods; Step S6: quantitatively analyzing the atherosclerosis status prediction data for multiple time periods and performing personalized comprehensive diagnosis to generate personalized diagnosis results; Among them, the specific steps of step S2 are: Step S21: performing global contrast enhancement on the medical image of the sclerosis marker, thereby constructing a global contrast optimized image; Step S22: performing multi-time point visual recognition on the global contrast optimized image, marking atherosclerotic plaques at multiple time points; Step S23: performing plaque edge detection on atherosclerotic plaques at multiple time points and marking the plaque edge contour lines; Step S24: performing semantic segmentation of the image boundary according to the edge contour of the patch, thereby extracting the patch area image at each time point; Step S25: performing plaque geometric morphology analysis on the plaque region image at each time point, thereby generating plaque geometric morphology features at each time point; The specific steps of step S21 are as follows: Perform image noise analysis on medical images of sclerosis markers and extract multiple image noise points; Mining pixel features of multiple image noise points one by one to generate pixel-level features of each noise point; Perform dynamic filtering and noise reduction based on the pixel-level characteristics of each noise point to generate filtered and denoised medical images; Perform multi-region decomposition on the filtered and denoised medical image to generate multiple sub-image blocks; Performing histogram equalization processing on the multiple sub-image blocks one by one, thereby generating multiple histogram equalized sub-image blocks; Global contrast enhancement is performed based on multiple histogram equalized sub-image blocks to construct a global contrast optimized image.

2. The method for integrated analysis of atherosclerosis marker detection data according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: obtaining multi-dimensional biological detection index information of the patient's atherosclerosis markers and medical images of atherosclerosis markers; Step S12: performing marker intelligent classification on the multi-dimensional biological detection indicator information to obtain marker intelligent classification data; Step S13: Screening key markers based on the marker intelligent classification data, and extracting multi-dimensional biological detection indicator information of the key markers; Step S14: Adaptively standardize the multi-dimensional biological detection index information of the key markers to obtain standardized biological detection information.

3. The method for integrated analysis of atherosclerosis marker detection data according to claim 2, characterized in that: The specific steps of step S14 are: Perform parameterized data recognition on the multi-dimensional biological detection indicator information of key markers and calculate the average parameter threshold to obtain the parameterized average threshold; Detect abnormal outlier data based on parameterized average threshold and mark abnormal outlier data points; Perform outlier filtering and cleaning on abnormal outlier data points to obtain cleaning optimization indicator information; Perform multi-dimensional difference index analysis on cleaning optimization index information to generate difference data between multi-dimensional data; Adaptive standardization processing is performed based on the difference data between multi-dimensional data to obtain standardized biological detection information.

4. The method for integrated analysis of atherosclerosis marker detection data according to claim 1, wherein: The specific steps of step S3 are: Step S31: performing multi-time point morphological change tracking based on the geometric morphological characteristics of the plaque at each time point to obtain morphological change tracking data of the sclerotic plaque; Step S32: performing adjacent time difference calculation on the morphological change tracking data of the sclerotic plaque, thereby obtaining morphological difference data at multiple adjacent time intervals; Step S33: performing morphological change trend analysis on the morphological difference data of multiple adjacent time intervals to obtain the temporal morphological change trend of the sclerotic plaque; Step S34: performing nonlinear morphological trend trajectory evolution on the temporal morphological change trend of the sclerotic plaque to construct a temporal morphological evolution trajectory of the sclerotic plaque.

5. The method for integrated analysis of atherosclerosis marker detection data according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: performing multi-indicator potential association mining on the standardized biological detection information to obtain potential association features between multiple indicators; Step S42: Calculating the time series change rate of each indicator of the standardized biological detection information to generate the time series change rate of each detection indicator; Step S43: performing a change range analysis based on the standardized biological detection information to obtain a change range of the detection index; Step S44: performing a time series change characteristic analysis based on the time series change rate and the change amplitude of each detection indicator to obtain the time series change characteristics of the multi-dimensional detection indicators; Step S45: fitting the temporal variation characteristics of the multi-dimensional detection indicators to the temporal evolution trend is performed according to the potential correlation characteristics among the multiple indicators, thereby constructing a multi-dimensional detection indicator trend evolution diagram.

6. The method for integrated analysis of atherosclerosis marker detection data according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: performing a deep dynamic change correlation analysis on the temporal morphological evolution trajectory of the sclerotic plaque based on the multi-dimensional detection index trend evolution diagram, thereby obtaining the dynamic evolution law of biological information-plaque evolution; Step S52: performing end-to-end iterative learning on the biological information-plaque evolution dynamics law to generate a plaque trajectory dynamics evolution model; Step S53: Perform personalized predictions for multiple time periods in the future based on the plaque trajectory dynamic evolution model, thereby obtaining atherosclerosis status prediction data for multiple time periods.

7. The method for integrated analysis of atherosclerosis marker detection data according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: performing quantitative analysis on the atherosclerosis status prediction data of multiple time periods to generate status prediction analysis data for each time period; Step S62: performing risk prediction on the status prediction analysis data of each time period to generate pathology risk prediction data for each time period; Step S63: performing personalized comprehensive diagnosis based on the pathological risk prediction data of each time period to generate a personalized diagnosis result; Step S64: performing dynamic feedback optimization on the plaque trajectory dynamic evolution model according to the personalized diagnosis result, thereby constructing a dynamic feedback optimization model.

8. An integrated analysis system for atherosclerosis marker detection data, characterized in that: The method for performing the integrated analysis of atherosclerosis marker detection data according to claim 1 comprises: A data processing module is used to obtain multi-dimensional biological detection indicator information of the patient's atherosclerosis markers and medical images of atherosclerosis markers; and adaptively standardize the multi-dimensional biological detection indicator information to obtain standardized biological detection information; The geometric morphology analysis module is used to perform multi-time point visual recognition of medical images of sclerosis markers and perform plaque geometric morphology analysis to generate the geometric morphology characteristics of plaques at each time point; The morphological trajectory evolution module is used to track the morphological changes at multiple time points and perform nonlinear morphological trend trajectory evolution based on the geometric morphological characteristics of the plaque at each time point, so as to construct the temporal morphological evolution trajectory of the sclerotic plaque; The situation evolution module is used to mine the potential correlation of multiple indicators in standardized biological detection information, and then perform time-series situation evolution fitting to construct a multi-dimensional detection indicator situation evolution diagram; The multi-period prediction module is used to make personalized predictions of the temporal evolution trajectory of sclerotic plaques based on the multi-dimensional detection indicator trend evolution diagram, thereby obtaining prediction data on the state of atherosclerosis in multiple time periods; The personalized diagnosis and analysis module is used to quantitatively analyze the atherosclerosis status prediction data of multiple time periods and perform personalized comprehensive diagnosis to generate personalized diagnosis results.

Images