A Method and System for Assessing Coronary Artery Occlusion Based on Magnetocardiography

Abstract

This invention discloses a method for assessing the degree of coronary artery occlusion based on magnetocardiography (MCG), comprising: acquiring a MCG dataset, divided into a training set and a test set; assigning labels to each MCG data point; training different classification models using the training set and labels, with different degrees of coronary angiography occlusion as boundaries, to obtain trained classification models; testing the test set using the trained classification models to obtain the positive probability of each MCG data point under different classification models and degrees of coronary angiography occlusion; weighting the positive probability of each classification model under different degrees of coronary angiography occlusion to obtain a score for each classification model; weighting the scores of each MCG data point under different classification models to obtain a final score for each MCG data point; assessing the percentage of coronary artery occlusion based on the final score; and a system for assessing the degree of coronary artery occlusion based on magnetocardiography. This invention can provide a more accurate, objective, and detailed assessment of coronary artery occlusion.

Description

Technical Field

[0001] This invention relates to the field of magnetocardiogram (MCC) image processing. More specifically, this invention relates to a method and system for assessing coronary artery occlusion based on magnetocardiography. Background Technology

[0002] Magnetocardiography (MCG) is a functional test that is more sensitive and reliable than traditional electrocardiography (ECG) because it directly measures magnetic signals unaffected by the musculoskeletal system. However, interpreting MCG reports typically requires experienced professionals, thus limiting its clinical application.

[0003] Currently, magnetocardiography can only provide a rough assessment, such as mild, moderate, and severe, which is difficult to reflect the subtle differences in the severity of the disease. For some diseases that require more detailed assessment, the classification is not accurate enough. In addition, this method is highly subjective and may be affected by the doctor's subjective judgment. Different doctors may classify the severity of the same patient's disease in different ways.

[0004] Artificial intelligence (AI)-based magnetocardiogram (MCC) models for assessing the severity of coronary artery disease (CAD) are currently a hot research topic in this field. They could provide clinicians with an automated and accurate diagnostic tool, potentially promoting clinical acceptance. However, researchers have been working to improve the predictive accuracy of binary classification models, which use a specific degree of blockage in coronary angiography as a threshold to categorize negative and positive MCCs. This approach is not accurate enough for assessing the severity of coronary artery disease. Summary of the Invention

[0005] One object of the present invention is to provide a method and system for assessing the degree of coronary artery blockage based on magnetocardiography, so as to at least solve the above-mentioned problems.

[0006] To achieve the objectives and other advantages of this invention, the present invention provides a method for assessing the degree of coronary artery occlusion based on magnetocardiography, comprising:

[0007] Obtain the magnetocardiogram dataset and randomly divide it into training and test sets;

[0008] Based on the coronary angiography results corresponding to each case of magnetocardiogram data in the training set, a label is given to each case of magnetocardiogram data according to the degree of coronary angiography blockage.

[0009] Using the training set and its corresponding labels, the models are divided into positive and negative based on different degrees of coronary angiography blockage. One or more classification models are trained for binary classification to obtain well-trained classification models under different degrees of coronary angiography blockage.

[0010] The trained classification model was used to test the test set to obtain the positive and negative probabilities of each magnetocardiogram data under different classification models and different degrees of coronary angiography blockage.

[0011] For each case of magnetocardiogram data, the positive probability of each classification model under different degrees of coronary angiography blockage is assigned a weight, and a weighted average is calculated to obtain the score of each case of magnetocardiogram data under each classification model;

[0012] Weights were assigned to the scores of each magnetocardiogram (MCC) data point under different classification models, and a weighted average was calculated to obtain the final score for each MCC data point.

[0013] The percentage of coronary artery blockage is assessed based on the final score.

[0014] Preferably, the method further includes preprocessing each magnetocardiogram (MCC) data point in the MCC dataset to extract MCC parameters, and using the MCC parameters as input to the classification model.

[0015] Preferably, the magnetocardiogram parameters include the T-peak magnetic field angle, T-peak current angle, T-peak maximum-minimum ratio, TT maximum current angle, TT maximum magnetic field angle, TT minimum current angle, TT minimum magnetic field angle, positive electrode area variation standard deviation, and negative electrode area variation standard deviation.

[0016] Preferably, the different degrees of coronary angiography occlusion include 50% occlusion, 70% occlusion, 80% occlusion, and 90% occlusion.

[0017] Preferably, the classification model includes a machine learning model and / or a deep learning model.

[0018] Preferably, the machine learning model includes a random forest model and an extreme gradient boosting model.

[0019] Preferably, the deep learning model includes a long short-term memory network model and a deep residual network model.

[0020] The present invention also provides a system based on the above method, comprising:

[0021] The module is used to acquire the magnetocardiogram dataset and randomly divides it into training and test sets.

[0022] The labeling module is used to assign a label to each magnetocardiogram data based on the degree of coronary angiography blockage according to the coronary angiography result corresponding to each magnetocardiogram data in the training set.

[0023] The training module is used to divide the coronary angiography into positive and negative categories based on different degrees of coronary angiography blockage, using the training set and its corresponding labels, and to perform binary classification training on one or more classification models to obtain well-trained classification models under different degrees of coronary angiography blockage.

[0024] The testing module is used to test the test set using the trained classification model to obtain the positive and negative probabilities of each magnetocardiogram data under different classification models and different degrees of coronary angiography blockage.

[0025] The first data processing module is used to assign weights to the positive probability of each classification model of each case of magnetocardiogram data under different degrees of coronary angiography blockage, and to calculate the weighted average to obtain the score of each case of magnetocardiogram data under each classification model.

[0026] The second data processing module is used to assign weights to the scores of each magnetocardiogram data under different classification models, and to calculate the weighted average to obtain the final score of each magnetocardiogram data.

[0027] An assessment module is used to evaluate the percentage of coronary artery blockage based on the final score.

[0028] The present invention also provides an electronic device comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including methods described above.

[0029] The present invention also provides a computer-readable storage medium storing a computer program used in conjunction with an electronic device, the computer program being executable by a processor to implement the above-described method.

[0030] The present invention has at least the following beneficial effects:

[0031] This invention constructs a scoring mechanism to quantify the results of magnetocardiogram data, providing a more accurate, objective, and detailed assessment of coronary artery blockage. This avoids the influence of doctors' subjective judgment, making patients' understanding of the degree of coronary artery blockage more intuitive and easier to accept in clinical practice.

[0032] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Attached Figure Description

[0033] Figure 1 This is a schematic flowchart of a coronary artery occlusion assessment method based on magnetocardiography according to an embodiment of the present invention. Detailed Implementation

[0034] The present invention will now be described in further detail with reference to the embodiments and accompanying drawings, so that those skilled in the art can implement it based on the description.

[0035] It should be understood that terms such as “having,” “comprising,” and “including” as used herein do not exclude the presence or addition of one or more other elements or combinations thereof.

[0036] It should be noted that, unless otherwise specified, the experimental methods described in the following implementation plan are all conventional methods, and the reagents and materials described are all commercially available unless otherwise specified.

[0037] In one aspect of the present invention, such as Figure 1 As shown, a method for assessing the degree of coronary artery occlusion based on magnetocardiography is provided, including:

[0038] S1. Obtain the magnetocardiogram dataset and randomly divide it into training and test sets.

[0039] Specifically, the magnetocardiogram (MCG) of patients is collected using a magnetocardiogram (MCG) instrument, and the corresponding coronary angiography results are obtained to form a MCG dataset.

[0040] S2. Based on the coronary angiography results corresponding to each case of magnetocardiogram data in the training set, assign a label to each case of magnetocardiogram data according to the degree of coronary angiography blockage.

[0041] For example, if the coronary angiography result corresponding to the magnetocardiogram data is 55% blockage, then when classifying positive and negative cases based on a 50% blockage level, the magnetocardiogram data is marked as positive; when classifying positive and negative cases based on 70%, 80%, and 90% blockage levels, the magnetocardiogram data is marked as negative. If the coronary angiography result corresponding to the magnetocardiogram data is 75% blockage, then when classifying positive and negative cases based on 50% and 70% blockage levels... The magnetocardiogram (MCC) data in this case is marked as positive. When dividing positive and negative cases based on the degree of coronary angiography blockage of 80% and 90%, the MCC data in this case is marked as negative. If the coronary angiography result corresponding to the MCC data is 85% blockage, then when dividing positive and negative cases based on the degree of coronary angiography blockage of 50%, 70%, and 80%, the MCC data in this case is marked as positive. When dividing positive and negative cases based on the degree of coronary angiography blockage of 90%, the MCC data in this case is marked as negative.

[0042] S3. Using the training set and its corresponding labels, divide the coronary angiography into positive and negative categories based on different degrees of occlusion, and perform binary classification training on one or more classification models to obtain well-trained classification models under different degrees of coronary angiography occlusion.

[0043] Specifically, the blockage is measured by coronary angiography at m1%, m2%, ..., m n Using % (where n is an integer) as a boundary to divide Yin and Yang, one or more machine learning models or deep learning models (denoted as models M1, M2, ..., M) are employed. N (where N is an integer) is used for binary classification training. The model trained with the coronary angiography occlusion level m1% as the boundary can be denoted as P. 11 P 21 ... P N1 Let P be the model trained with m2% as the boundary. 12 P 22 ... P N2 ..., the degree of blockage in coronary angiography m n The model trained with % as the boundary is denoted as P. 1n P 2n ... P Nn Using the training set and its corresponding labels, model P is respectively... 11 P 21 ... P N1 P 12 P 22 ... P N2 , ..., P 1n P 2n ... P Nn Training was performed to obtain well-trained classification models under different degrees of coronary angiography occlusion.

[0044] S4. The trained classification model is used to test the test set to obtain the positive and negative probabilities of each magnetocardiogram data under different classification models and different degrees of coronary angiography blockage.

[0045] For example, the positive probability obtained by the classification model is denoted as p The probability of a negative result is denoted as q , p+q=1 .

[0046] S5. Assign weights to the positive probability of each classification model for each case of magnetocardiogram data under different degrees of coronary angiography blockage, and then calculate a weighted average to obtain the score of each case of magnetocardiogram data under each classification model.

[0047] For example, based on the analytic hierarchy process (AHP), a judgment matrix is ​​constructed to assign weights to the positive probability of each classification model corresponding to each case of magnetocardiogram data under different degrees of coronary angiography occlusion. The positive probability is multiplied by the assigned weights to obtain an additive score, thereby obtaining the score of each case of magnetocardiogram data under each classification model, denoted as:

[0048] S i = w 1q 1 +w 2 q 2+…+ w n q n (i≤N, where i is an integer)

[0049] in, S i For model M i The corresponding model score, q 1. q 2、…、 q n For model M i In coronary angiography, the blockage is m1%, m2%, ..., m n The corresponding positive probability at % w 1. w 2、…、 w n This represents the weight corresponding to the positive probability.

[0050] S6. Assign weights to the scores of each magnetocardiogram (MCC) data under different classification models, and then calculate a weighted average to obtain the final score for each MCC data.

[0051] For example, based on the analytic hierarchy process (AHP), a judgment matrix is ​​constructed to assign weights to the scores of each magnetocardiogram (MCC) data point under different classification models. The scores are then multiplied by the assigned weights to obtain additive scores, thus yielding the final score for each MCC data point. This result is within the range of 0-100 points and is denoted as:

[0052] S = ( W 1 S 1 +W 2 S 2+…+ W n S N )×100

[0053] in, S The final score for each magnetocardiogram (MCC) data point. S 1. S 2、…、 S N For models M1, M2, ..., M N The corresponding model score, W 1. W 2、…、 W n These are the weights for the corresponding model scores.

[0054] S7. Assess the percentage of coronary artery blockage based on the final score.

[0055] For example, if the final score of the magnetocardiogram data is 83, then the percentage of coronary artery blockage for the patient corresponding to this magnetocardiogram data is 83%.

[0056] This embodiment first uses a training set and its corresponding labels to pre-train different classification models under different degrees of coronary artery occlusion, obtaining trained classification models. Then, the trained classification models are tested on the test set to obtain the positive and negative probabilities of each magnetocardiogram (MCC) data point under different classification models and different degrees of coronary angiography occlusion. Next, weights are assigned to the positive probabilities of each classification model for each MCC data point under different degrees of coronary angiography occlusion, and a weighted average is performed to obtain a score for each classification model. Finally, weights are assigned to the scores of each MCC data point under different classification models, and a weighted average is performed to obtain the final score for each MCC data point. The final score can be used to assess the percentage of coronary artery occlusion. In other words, this embodiment provides a coronary artery occlusion scoring mechanism based on MCC data. Compared with existing classification models that can only determine positive or negative under different degrees of coronary artery occlusion, the method of this embodiment can directly assess the percentage of coronary artery occlusion through the final score, providing a more accurate, objective, and detailed assessment of coronary artery occlusion, avoiding the influence of doctors' subjective judgment, making patients' understanding of the degree of coronary artery occlusion more intuitive, and promoting clinical acceptance.

[0057] In another embodiment, the method further includes preprocessing each magnetocardiogram (MCC) data point in the MCC dataset to extract MCC parameters, and using the MCC parameters as input to the classification model.

[0058] Specifically, the preprocessing method and parameter extraction method are existing technologies and will not be elaborated here. For example, firstly, the magnetocardiogram (MCC) data is used to obtain the magnetocardiogram (MCC) field map, current map, and butterfly diagram. Then, parameters are extracted from the MCC, current map, and butterfly diagram. These parameters include, but are not limited to, magnetic field or current angle parameters and their variations within the TT segment, magnetic pole distance variations within the TT segment, positive and negative pole position variations within the TT segment, current position variations within the TT segment, positive and negative pole boundary and area variations within the TT segment, positive and negative pole area variations within the TT segment, positive and negative pole area ratio parameters, and R-peak / T-peak amplitude ratio parameters. The number of parameters does not affect the implementation of this method.

[0059] Furthermore, the magnetocardiogram parameters include the T-peak magnetic field angle, T-peak current angle, T-peak maximum-minimum ratio, TT maximum current angle, TT maximum magnetic field angle, TT minimum current angle, TT minimum magnetic field angle, positive electrode area change standard deviation, and negative electrode area change standard deviation.

[0060] In another embodiment, the different degrees of coronary angiography occlusion include 50% occlusion, 70% occlusion, 80% occlusion, and 90% occlusion.

[0061] In another embodiment, the classification model includes a machine learning model and / or a deep learning model.

[0062] Furthermore, the machine learning models include Random Forest (RF) and Extreme Gradient Boosting (XGB).

[0063] Furthermore, the deep learning models include Long Short-Term Memory (LSTM) network models and ResNet (Residual Network) models.

[0064] In another aspect, the present invention provides a system based on the above method, comprising:

[0065] The module is used to acquire the magnetocardiogram dataset and randomly divides it into training and test sets.

[0066] The labeling module is used to assign a label to each magnetocardiogram data based on the degree of coronary angiography blockage according to the coronary angiography result corresponding to each magnetocardiogram data in the training set.

[0067] The training module is used to divide the coronary angiography into positive and negative categories based on different degrees of coronary angiography blockage, using the training set and its corresponding labels, and to perform binary classification training on one or more classification models to obtain well-trained classification models under different degrees of coronary angiography blockage.

[0068] The testing module is used to test the test set using the trained classification model to obtain the positive and negative probabilities of each magnetocardiogram data under different classification models and different degrees of coronary angiography blockage.

[0069] The first data processing module is used to assign weights to the positive probability of each classification model of each case of magnetocardiogram data under different degrees of coronary angiography blockage, and to calculate the weighted average to obtain the score of each case of magnetocardiogram data under each classification model.

[0070] The second data processing module is used to assign weights to the scores of each magnetocardiogram data under different classification models, and to calculate the weighted average to obtain the final score of each magnetocardiogram data.

[0071] An assessment module is used to evaluate the percentage of coronary artery blockage based on the final score.

[0072] This embodiment constructs an acquisition module, a labeling module, a training module, a testing module, a first data processing module, a second data processing module, and an evaluation module through a software program. The training module uses a training set and its corresponding labels to pre-train different classification models under different degrees of coronary artery occlusion, resulting in trained classification models. The testing module uses the trained classification models to test the test set, obtaining the positive and negative probabilities of each magnetocardiogram (MCC) data point under different classification models and different degrees of coronary angiography occlusion. The first data processing module assigns weights to the positive probabilities of each classification model for each MCC data point under different degrees of coronary angiography occlusion, and performs a weighted average to obtain a score for each classification model. The second data processing module assigns weights to the scores of each MCC data point under different classification models, performs a weighted average, and obtains a final score for each MCC data point. The evaluation module assesses the percentage of coronary artery occlusion based on the final score, providing a more accurate, objective, and detailed assessment of coronary artery occlusion, avoiding the influence of doctors' subjective judgment, making patients' understanding of the degree of coronary artery occlusion more intuitive, and promoting clinical acceptance.

[0073] In another aspect, the present invention provides an electronic device comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including methods described above.

[0074] For example, the electronic device is a device that includes a processor (CPU / MCU / SOC) and a memory (ROM / RAM), such as a desktop computer, a portable computer, a smartphone, etc.

[0075] In another aspect, the present invention provides a computer-readable storage medium storing a computer program used in conjunction with an electronic device, the computer program being executable by a processor to implement the above-described method.

[0076] For example, the computer-readable storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0077] The present invention will be further described below with specific embodiments:

[0078] Example 1:

[0079] A method for assessing the degree of coronary artery occlusion based on magnetocardiography, comprising:

[0080] (1) Obtain the magnetocardiogram (MCG) results of a certain number of patients and use existing technology to extract the parameters of the magnetocardiogram, current graph, and butterfly graph for later use. The parameters include the T-peak magnetic field angle, T-peak current angle, T-peak maximum-minimum ratio, TT maximum current angle, TT maximum magnetic field angle, TT minimum current angle, TT minimum magnetic field angle, positive electrode area change standard deviation, and negative electrode area change standard deviation.

[0081] (2) Obtain the patient's coronary angiography results for future use;

[0082] (3) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 50% angiographic blockage. The RF model and XGB model were used to train a prediction model P based on 50% blockage. 11 Model P 21 ;

[0083] (4) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 70% angiographic blockage. The RF model and XGB model were used to train a prediction model P based on 70% blockage. 12 Model P 22 ;

[0084] (5) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 80% angiographic blockage. The RF model and XGB model were used to train a prediction model P based on 80% blockage. 13 Model P 23 ;

[0085] (6) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 90% angiographic blockage. The RF model and XGB model were used to train a prediction model P based on 90% blockage. 14 Model P 24 ;

[0086] (7) Obtain the magnetocardiogram data to be tested and extract the magnetocardiogram parameters. Use the four prediction models P of the RF model. 11 P 12 P 13 P 14 Predictions were performed separately to obtain the prediction results of the RF model under different degrees of coronary angiography occlusion, including positive and negative probabilities; four prediction models P were used from the XGB model. 21 P 22 P 23 P 24 Predictions were made separately to obtain the prediction results of the XGB model under different degrees of coronary angiography occlusion, including positive and negative probabilities;

[0087] (8) Construct the judgment matrix according to the analytic hierarchy process, which is the model P.11 P 12 P 13 P 14 The positive probability is assigned a weight, and the positive probability is multiplied by the assigned weight to obtain the additive score, thus obtaining the score of the RF model for the single-case data.

[0088] (9) Construct the judgment matrix according to the analytic hierarchy process, which is the model P. 21 P 22 P 23 P 24 The positive probability is assigned a weight, and the positive probability is multiplied by the assigned weight to obtain the additive score, thus obtaining the score of the XGB model for the single case data.

[0089] (10) Construct a judgment matrix based on the analytic hierarchy process to assign weights to the RF model and the XGB model, thereby obtaining the final score of the single-case data.

[0090] The above assessment method was used to score 600 cases of magnetocardiogram data and perform correlation analysis with the corresponding coronary angiography results. The Pearson correlation coefficient was 0.81, which is greater than 0.8, indicating a strong correlation. Furthermore, based on the scoring results obtained by this assessment method, the accuracy rate reached 95.7% when the 50% division was positive or negative; similarly, the accuracy rates were 89.3% for 70%, 83.2% for 80%, and 82.5% for 90%. In contrast, the RF model achieved accuracies of approximately 88%, 87%, 79%, and 80% when the 50%, 70%, 80%, and 90% divisions were used, while the XGB model achieved accuracies of approximately 90%, 85%, 80%, and 79%, respectively, demonstrating the effectiveness of the assessment method.

[0091] Example 2:

[0092] A method for assessing the degree of coronary artery occlusion based on magnetocardiography, comprising:

[0093] (1) Obtain the magnetocardiogram (MCG) results of a certain number of patients and use existing technology to extract the parameters of the magnetocardiogram, current graph, and butterfly graph for later use. The parameters include T-peak magnetic field angle, T-peak current angle, T-peak maximum-minimum ratio, TT maximum current angle, TT maximum magnetic field angle, TT minimum current angle, TT minimum magnetic field angle, positive electrode area change standard deviation, negative electrode area change standard deviation, and RT angle.

[0094] (2) Obtain the patient's coronary angiography results for future use;

[0095] (3) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 30% angiographic blockage. The RF model and XGB model were used to train a prediction model P based on 30% blockage. 11 Model P 21;

[0096] (4) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 50% angiographic blockage. The RF model and XGB model were used to train a prediction model P based on 50% blockage. 12 Model P 22 ;

[0097] (5) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 70% angiographic blockage. The RF model and XGB model were used to train a prediction model P based on 70% blockage. 13 Model P 23 ;

[0098] (6) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 80% angiographic blockage. The RF model and XGB model were used to train a prediction model P based on 80% blockage. 14 Model P 24 ;

[0099] (7) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 90% angiographic blockage. The RF model and XGB model were used to train a prediction model P based on 90% blockage. 15 Model P 25 ;

[0100] (8) Obtain the magnetocardiogram data to be tested and extract the magnetocardiogram parameters. Use the five prediction models P of the RF model. 11 P 12 P 13 P 14 P 15 Predictions were performed separately to obtain the prediction results of the RF model under different degrees of coronary angiography occlusion, including positive and negative probabilities; five prediction models P were used from the XGB model. 21 P 22 P 23 P 24 P 25 Predictions were made separately to obtain the prediction results of the XGB model under different degrees of coronary angiography occlusion, including positive and negative probabilities;

[0101] (9) Construct the judgment matrix according to the analytic hierarchy process, which is the model P. 11 P 12 P 13 P 14 P 15 The positive probability is assigned a weight, and the positive probability is multiplied by the assigned weight to obtain the additive score, thus obtaining the score of the RF model for the single-case data.

[0102] (10) Construct a judgment matrix based on the analytic hierarchy process (AHP) for model P. 21 P 22 P 23 P 24 P 25 The positive probability is assigned a weight, and the positive probability is multiplied by the assigned weight to obtain the additive score, thus obtaining the score of the XGB model for the single case data.

[0103] (11) Construct a judgment matrix based on the analytic hierarchy process to assign weights to the RF model and the XGB model, thereby obtaining the final score of the single-case data.

[0104] The above assessment method was used to score the magnetocardiogram data of 600 cases and perform correlation analysis with the corresponding coronary angiography results. The Pearson correlation coefficient was 0.85, which is greater than 0.8, indicating a strong correlation.

[0105] Example 3:

[0106] A method for assessing the degree of coronary artery occlusion based on magnetocardiography, comprising:

[0107] (1) Obtain the magnetocardiogram (MCG) results of a certain number of patients and use existing technology to extract the parameters from the magnetocardiogram, current graph, and butterfly graph for later use. The parameters include the T-peak magnetic field angle, T-peak current angle, T-peak maximum-minimum ratio, TT maximum current angle, TT maximum magnetic field angle, TT minimum current angle, TT minimum magnetic field angle, standard deviation of positive electrode area change, standard deviation of negative electrode area change, RT angle, sum of positive pole position changes, and sum of negative pole position changes.

[0108] (2) Obtain the patient's coronary angiography results for future use;

[0109] (3) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 30% angiographic blockage. RF model, XGB model, and LSTM network were used to train a prediction model P based on 30% blockage. 11 Model P 21 Model P 31 ;

[0110] (4) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 50% angiographic blockage. RF model, XGB model, and LSTM network were used to train a prediction model P based on 50% blockage. 12 Model P 22 Model P 32 ;

[0111] (5) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 70% angiographic blockage. RF model, XGB model, and LSTM network were used to train a prediction model P based on 70% blockage. 13 Model P 23 Model P 33 ;

[0112] (6) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 80% angiographic blockage. RF model, XGB model, and LSTM network were used to train a prediction model P based on 80% blockage. 14 Model P 24 Model P 34 ;

[0113] (7) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 90% angiographic blockage. RF model, XGB model, and LSTM network were used to train a prediction model P based on 90% blockage. 15 Model P 25 Model P 35 ;

[0114] (8) Obtain the magnetocardiogram data to be tested and extract the magnetocardiogram parameters. Use the five prediction models P of the RF model. 11 P 12 P 13 P 14 P 15 Predictions were performed separately to obtain the prediction results of the RF model under different degrees of coronary angiography occlusion, including positive and negative probabilities; five prediction models P were used from the XGB model. 21 P 22 P 23 P 24 P 25 Predictions were performed separately to obtain the XGB model's prediction results under different degrees of coronary angiography occlusion, including positive and negative probabilities; five prediction models P were used with an LSTM network. 31 P 32 P 33 P 34 P 35 Predictions were made separately to obtain the prediction results of the LSTM network under different degrees of coronary angiography occlusion, including positive and negative probabilities.

[0115] (9) Construct the judgment matrix according to the analytic hierarchy process, which is the model P. 11 P 12 P 13 P 14 P 15The positive probability is assigned a weight, and the positive probability is multiplied by the assigned weight to obtain the additive score, thus obtaining the score of the RF model for the single-case data.

[0116] (10) Construct a judgment matrix based on the analytic hierarchy process (AHP) for model P. 21 P 22 P 23 P 24 P 25 The positive probability is assigned a weight, and the positive probability is multiplied by the assigned weight to obtain the additive score, thus obtaining the score of the XGB model for the single case data.

[0117] (11) Construct a judgment matrix based on the analytic hierarchy process (AHP) for model P. 31 P 32 P 33 P 34 P 35 The positive probability is divided into weights, and the positive probability is multiplied by the assigned weights to obtain the additive score, thus obtaining the score of the LSTM network for a single data instance.

[0118] (12) Construct a judgment matrix based on the analytic hierarchy process to assign weights to the RF model, XGB model and LSTM network, thereby obtaining the final score of the single data.

[0119] The above assessment method was used to score the magnetocardiogram data of 600 cases and perform correlation analysis with the corresponding coronary angiography results. The Pearson correlation coefficient was 0.88, which is greater than 0.8, indicating a strong correlation.

[0120] Example 4:

[0121] A method for assessing the degree of coronary artery occlusion based on magnetocardiography, comprising:

[0122] (1) Obtain the magnetocardiogram (MCG) results of a certain number of patients and use existing technology to extract the parameters from the magnetocardiogram, current graph, and butterfly graph for later use. The parameters include the T-peak magnetic field angle, T-peak current angle, T-peak maximum-minimum ratio, TT maximum current angle, TT maximum magnetic field angle, TT minimum current angle, TT minimum magnetic field angle, standard deviation of positive electrode area change, standard deviation of negative electrode area change, RT angle, sum of positive pole position changes, and sum of negative pole position changes.

[0123] (2) Obtain the patient's coronary angiography results for future use;

[0124] (3) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 30% angiographic blockage. The XGB model, KNN model, and ResNet network were used to train a prediction model P based on 30% blockage. 11 Model P 21 Model P31 ;

[0125] (4) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 50% angiographic blockage. The XGB model, KNN model, and ResNet network were used to train a prediction model P based on 50% blockage. 12 Model P 22 Model P 32 ;

[0126] (5) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 70% angiographic blockage. The XGB model, KNN model, and ResNet network were used to train a prediction model P based on 70% blockage. 13 Model P 23 Model P 33 ;

[0127] (6) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 80% angiographic blockage. The XGB model, KNN model, and ResNet network were used to train a prediction model P based on 80% blockage. 14 Model P 24 Model P 34 ;

[0128] (7) Using coronary angiography results as the label for each patient's parameters, patients were classified as positive or negative based on 90% angiographic blockage. The XGB model, KNN model, and ResNet network were used to train a prediction model P based on 90% blockage. 15 Model P 25 Model P 35 ;

[0129] (8) Obtain the magnetocardiogram data to be tested and extract the magnetocardiogram parameters. Use the five prediction models P of the XGB model. 11 P 12 P 13 P 14 P 15 Predictions were performed separately to obtain the XGB model's prediction results under different degrees of coronary angiography occlusion, including positive and negative probabilities; five prediction models P were used from the KNN model. 21 P 22 P 23 P 24 P 25 Predictions were performed separately to obtain the prediction results of the KNN model under different degrees of coronary angiography occlusion, including positive and negative probabilities; five prediction models P were used from the ResNet network. 31 P 32 P 33 P 34 P35 Predictions were made separately to obtain the prediction results of the ResNet network under different degrees of coronary angiography occlusion, including positive and negative probabilities.

[0130] (9) Construct the judgment matrix according to the analytic hierarchy process, which is the model P. 11 P 12 P 13 P 14 P 15 The positive probability is assigned a weight, and the positive probability is multiplied by the assigned weight to obtain the additive score, thus obtaining the score of the XGB model for the single case data.

[0131] (10) Construct a judgment matrix based on the analytic hierarchy process (AHP) for model P. 21 P 22 P 23 P 24 P 25 The positive probability is divided into weights, and the positive probability is multiplied by the assigned weights to obtain the additive score, thus obtaining the score of the KNN model for the single-case data.

[0132] (11) Construct a judgment matrix based on the analytic hierarchy process (AHP) for model P. 31 P 32 P 33 P 34 P 35 The positive probability is divided into weights, and the positive probability is multiplied by the assigned weight to obtain the additive score, thus obtaining the score of the ResNet network for a single example.

[0133] (12) Construct a judgment matrix based on the analytic hierarchy process to assign weights to the XGB model, KNN model and ResNet network, thereby obtaining the final score of the single data.

[0134] The above assessment method was used to score the magnetocardiogram data of 600 cases and perform correlation analysis with the corresponding coronary angiography results. The Pearson correlation coefficient was 0.84, which is greater than 0.8, indicating a strong correlation.

[0135] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A method for assessing the degree of coronary artery occlusion based on magnetocardiography, characterized in that, include: Obtain the magnetocardiogram dataset and randomly divide it into training and test sets; Based on the coronary angiography results corresponding to each case of magnetocardiogram data in the training set, a label is given to each case of magnetocardiogram data according to the degree of coronary angiography blockage. Using the training set and its corresponding labels, negative and positive cases are divided according to different degrees of coronary angiography blockage. One or more classification models are trained for binary classification to obtain well-trained classification models under different degrees of coronary angiography blockage. The trained classification model was used to test the test set to obtain the positive and negative probabilities of each magnetocardiogram data under different classification models and different degrees of coronary angiography blockage. For each case of magnetocardiogram data, the positive probability of each classification model under different degrees of coronary angiography blockage is assigned a weight, and a weighted average is calculated to obtain the score of each case of magnetocardiogram data under each classification model; Weights were assigned to the scores of each magnetocardiogram (MCC) data point under different classification models, and a weighted average was calculated to obtain the final score for each MCC data point. The percentage of coronary artery blockage is assessed based on the final score; It also includes preprocessing each magnetocardiogram data point in the magnetocardiogram dataset to extract magnetocardiogram parameters, and using the magnetocardiogram parameters as input to the classification model; The magnetocardiogram parameters include T-peak magnetic field angle, T-peak current angle, T-peak maximum-minimum ratio, TT-maximum current angle, TT-maximum magnetic field angle, TT-minimum current angle, TT-minimum magnetic field angle, positive electrode area variation standard deviation, and negative electrode area variation standard deviation. The different degrees of coronary angiography blockage include 50% blockage, 70% blockage, 80% blockage, and 90% blockage.

2. The method for assessing the degree of coronary artery occlusion based on magnetocardiography as described in claim 1, characterized in that, The classification model includes machine learning models and / or deep learning models.

3. The method for assessing the degree of coronary artery occlusion based on magnetocardiography as described in claim 2, characterized in that, The machine learning models include the random forest model and the extreme gradient boosting model.

4. The method for assessing the degree of coronary artery occlusion based on magnetocardiography as described in claim 2, characterized in that, The deep learning models include Long Short-Term Memory Network (LSTM) models and Deep Residual Network (DRN) models.

5. The system according to any one of claims 1 to 4, characterized in that, include: The module is used to acquire the magnetocardiogram dataset and randomly divides it into training and test sets. The labeling module is used to assign a label to each magnetocardiogram data based on the degree of coronary angiography blockage according to the coronary angiography result corresponding to each magnetocardiogram data in the training set. The training module is used to divide the coronary angiography into positive and negative categories based on different degrees of coronary angiography blockage, using the training set and its corresponding labels, and to perform binary classification training on one or more classification models to obtain well-trained classification models under different degrees of coronary angiography blockage. The testing module is used to test the test set using the trained classification model to obtain the positive and negative probabilities of each magnetocardiogram data under different classification models and different degrees of coronary angiography blockage. The first data processing module is used to assign weights to the positive probability of each classification model of each case of magnetocardiogram data under different degrees of coronary angiography blockage, and to calculate the weighted average to obtain the score of each case of magnetocardiogram data under each classification model. The second data processing module is used to assign weights to the scores of each magnetocardiogram data under different classification models, and to calculate the weighted average to obtain the final score of each magnetocardiogram data. An assessment module is used to evaluate the percentage of coronary artery blockage based on the final score.

6. An electronic device, characterized in that, include: One or more processors; Memory; as well as One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs comprising methods for performing any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, A computer program stored in conjunction with an electronic device, the computer program being executable by a processor to implement the method of any one of claims 1 to 4.

Images