CN119324064B - Method, device, electronic device and storage medium for determining severity of coronary artery stenosis - Google Patents
Method, device, electronic device and storage medium for determining severity of coronary artery stenosis Download PDFInfo
- Publication number
- CN119324064B CN119324064B CN202411876580.0A CN202411876580A CN119324064B CN 119324064 B CN119324064 B CN 119324064B CN 202411876580 A CN202411876580 A CN 202411876580A CN 119324064 B CN119324064 B CN 119324064B
- Authority
- CN
- China
- Prior art keywords
- actual
- coronary
- model
- score
- stenosis
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
本公开提供了一种冠脉狭窄严重程度确定方法、装置、电子设备及存储介质，该方法包括：获取目标患者的包含低密度脂蛋白胆固醇在内的实际血脂参数；获取目标患者的包含身份信息和既往病史在内的临床特征；将实际血脂参数和临床特征作为输入信息，输入经预先训练得到的冠脉狭窄严重程度预测模型，冠脉狭窄严重程度预测模型采用了可解释增强机的模型框架；接收冠脉狭窄严重程度预测模型输出的实际冠脉狭窄评分和评分解释信息；基于实际冠脉狭窄评分和评分解释信息，确定目标患者的实际冠脉狭窄严重程度。应用该方法可以更准确的确定低密度脂蛋白胆固醇与冠脉狭窄严重程度之间的关联，且结果具有更强的可解释性。
The present disclosure provides a method, device, electronic device and storage medium for determining the severity of coronary stenosis, the method comprising: obtaining actual blood lipid parameters including low-density lipoprotein cholesterol of a target patient; obtaining clinical characteristics including identity information and past medical history of the target patient; using the actual blood lipid parameters and clinical characteristics as input information, inputting a coronary stenosis severity prediction model obtained through pre-training, the coronary stenosis severity prediction model adopts a model framework of an interpretable enhancement machine; receiving the actual coronary stenosis score and score explanation information output by the coronary stenosis severity prediction model; determining the actual coronary stenosis severity of the target patient based on the actual coronary stenosis score and score explanation information. The application of this method can more accurately determine the association between low-density lipoprotein cholesterol and the severity of coronary stenosis, and the result has stronger interpretability.
Description
Technical Field
The present disclosure relates to the field of data processing technologies, and in particular, to the field of artificial intelligence technologies such as machine learning, association prediction, auxiliary discrimination, and the like, and more particularly, to a method and apparatus for determining coronary artery stenosis severity, an electronic device, a computer readable storage medium, and a computer program product.
Background
In the field of cardiovascular disease, atherosclerosis (Atherosclerotic Cardiovascular Disease, ASCVD) is a major pathological mechanism of coronary heart disease and acute myocardial infarction, and Low-density lipoprotein cholesterol (Low-Density Lipoprotein Cholesterol, LDL-C) is considered to be one of the central factors responsible for atherosclerosis. Thus, LDL-C level control is considered a key element in the prevention and treatment of cardiovascular disease. The control targets for LDL-C in the european cardiology guidelines are divided into multiple risk classes, and strict LDL-C control targets are proposed for high-risk patients. For example, for very high risk ASCVD patients, the LDL-C level should be controlled below 1.4 mmol/L.
While numerous studies have shown that LDL-C levels are closely related to atherosclerosis and coronary stenosis, and the concept of "lower LDL-C is better" has been widely accepted, existing clinical assessment methods suffer from deficiencies in identifying the risk of refinement from different LDL-C levels, particularly the inability to capture complex nonlinear relationships between LDL-C and coronary stenosis, making it difficult to formulate more personalized treatment regimens for different patients.
Disclosure of Invention
Embodiments of the present disclosure provide a coronary stenosis severity determination method, apparatus, electronic device, computer readable storage medium, and computer program product.
In a first aspect, embodiments of the present disclosure provide a method for determining severity of coronary stenosis, comprising:
Obtaining actual blood lipid parameters including low density lipoprotein cholesterol of a target patient;
acquiring clinical characteristics of a target patient, including identity information and prior medical history;
Inputting the actual blood fat parameters and clinical characteristics as input information into a coronary artery stenosis severity prediction model obtained through pre-training, wherein the coronary artery stenosis severity prediction model is obtained through training by using a training sample formed by sample input acted by the actual blood fat parameters and clinical characteristics of a historical patient and sample output acted by corresponding real coronary artery stenosis scores, and the coronary artery stenosis severity prediction model adopts a model frame of an interpretable enhancer;
Receiving actual coronary artery stenosis score and score interpretation information output by a coronary artery stenosis severity prediction model;
Based on the actual coronary stenosis score and the score interpretation information, an actual coronary stenosis severity of the target patient is determined.
Optionally, the actual blood lipid parameter at least comprises low density lipoprotein cholesterol and at least one of high density lipoprotein cholesterol, total cholesterol and triglyceride.
Optionally, the identity information comprises sex information and age information, and the prior medical history comprises whether the history of any of hypertension, diabetes, kidney diseases and apoplexy is illness or morbidity.
Optionally, before the training samples are formed using the sample input and the sample output, further comprising:
performing interpolation processing on missing values in the sample input and the sample output;
and performing rejection or replacement processing on the abnormal values in the sample input and the sample output so as to form training samples from the data subjected to the interpolation processing and the rejection or replacement processing.
Optionally, performing interpolation processing on the missing values in the sample input and the sample output includes:
performing interpolation processing on continuous variables with missing values in sample input and sample output by adopting a median interpolation method;
And performing interpolation processing on the classified variables with the missing values in the sample input and the sample output by adopting a mode interpolation method.
Optionally, the scoring interpretation information includes:
at least one of variable importance analysis results, global interpretation information, interaction analysis information and local interpretation information;
The method comprises the steps of obtaining a coronary artery stenosis score, obtaining a global interpretation information, wherein variable importance analysis results are used for presenting independent influence degrees of different types of input variables on obtaining the actual coronary artery stenosis score respectively, the global interpretation information is used for presenting actual corresponding relations between variable values of each input variable and the coronary artery stenosis score respectively, interaction analysis is used for presenting combined influence degrees of the different types of input variables on obtaining the actual coronary artery stenosis score when the input variables are combined mutually, and local interpretation information is used for presenting contribution polarities of variable values of each input variable on obtaining the actual coronary artery stenosis score respectively for a target patient, wherein the contribution polarities comprise positive contribution of positive correlation and negative contribution of negative correlation.
Optionally, the score interpretation information further comprises a risk level to which the actual coronary stenosis score belongs, wherein a risk threshold for classifying the risk level is determined by using a training sample based on the coronary stenosis severity prediction model.
In a second aspect, embodiments of the present disclosure provide a coronary stenosis severity determination apparatus, comprising:
An actual blood lipid parameter acquisition unit configured to acquire an actual blood lipid parameter including low density lipoprotein cholesterol of a target patient;
A clinical feature acquisition unit configured to acquire clinical features including identity information and past medical history of a target patient;
An actual parameter input unit configured to input an actual blood lipid parameter and clinical characteristics as input information into a coronary artery stenosis severity prediction model obtained by training in advance, wherein the coronary artery stenosis severity prediction model is obtained by training a training sample formed by a sample input acted by the actual blood lipid parameter and clinical characteristics of a historical patient and a sample output acted by a corresponding real coronary artery stenosis score, and the coronary artery stenosis severity prediction model adopts a model framework of an interpretable enhancer;
A model output result receiving unit configured to receive the actual coronary artery stenosis score and score interpretation information output by the coronary artery stenosis severity prediction model;
an actual coronary artery stenosis severity determination unit configured to determine an actual coronary artery stenosis severity of the target patient based on the actual coronary artery stenosis score and the score interpretation information.
Optionally, the actual blood lipid parameter at least comprises low density lipoprotein cholesterol and at least one of high density lipoprotein cholesterol, total cholesterol and triglyceride.
Optionally, the identity information comprises sex information and age information, and the prior medical history comprises whether the history of any of hypertension, diabetes, kidney diseases and apoplexy is illness or morbidity.
Optionally, before the training samples are formed using the sample input and the sample output, further comprising:
a missing value interpolation unit configured to perform interpolation processing on missing values in the sample input and the sample output;
an outlier processing unit configured to perform a culling or replacing process on outliers in the sample input and the sample output to construct the interpolation-processed and culled or replaced data into training samples.
Optionally, the missing value interpolation unit is further configured to:
performing interpolation processing on continuous variables with missing values in sample input and sample output by adopting a median interpolation method;
And performing interpolation processing on the classified variables with the missing values in the sample input and the sample output by adopting a mode interpolation method.
Optionally, the scoring interpretation information includes:
at least one of variable importance analysis results, global interpretation information, interaction analysis information and local interpretation information;
The method comprises the steps of obtaining a coronary artery stenosis score, obtaining a global interpretation information, wherein variable importance analysis results are used for presenting independent influence degrees of different types of input variables on obtaining the actual coronary artery stenosis score respectively, the global interpretation information is used for presenting actual corresponding relations between variable values of each input variable and the coronary artery stenosis score respectively, interaction analysis is used for presenting combined influence degrees of the different types of input variables on obtaining the actual coronary artery stenosis score when the input variables are combined mutually, and local interpretation information is used for presenting contribution polarities of variable values of each input variable on obtaining the actual coronary artery stenosis score respectively for a target patient, wherein the contribution polarities comprise positive contribution of positive correlation and negative contribution of negative correlation.
Optionally, the score interpretation information further comprises a risk level to which the actual coronary stenosis score belongs, wherein a risk threshold for classifying the risk level is determined by using a training sample based on the coronary stenosis severity prediction model.
In a third aspect, an embodiment of the present disclosure provides an electronic device, including at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to implement a method of determining coronary stenosis severity as described in any of the implementations of the first aspect.
In a fourth aspect, embodiments of the present disclosure provide a non-transitory computer-readable storage medium storing computer instructions for enabling a computer to implement a method of coronary stenosis severity determination as described in any of the implementations of the first aspect when executed.
In a fifth aspect, the presently disclosed embodiments provide a computer program product comprising a computer program which, when executed by a processor, is capable of implementing the steps of the coronary stenosis severity determination method as described in any of the implementations of the first aspect.
According to the coronary artery stenosis severity determination scheme provided by the disclosure, by adopting the interpretable enhancer model which expands the enhancement method based on the regression tree on the generalized addition model as a model framework and adopting blood lipid parameters including low density lipoprotein cholesterol and clinical characteristics including identity information and past medical history as input variables, the accurate relationship between the coronary artery stenosis severity can be predicted jointly by using a plurality of types of influence characteristics, and due to the fact that the multivariate history parameters are used as training samples, clear characteristic interpretation and interaction effects between variables can be provided while the prediction accuracy is ensured by means of the characteristics of the interpretable enhancer model, so that the model can also output interpretation information of how to obtain results, and therefore the acceptance of a doctor on output results can be enhanced, and a better auxiliary doctor can make more accurate diagnosis.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.
Drawings
Other features, objects and advantages of the present disclosure will become more apparent upon reading of the detailed description of non-limiting embodiments, made with reference to the following drawings:
FIG. 1 is a flow chart of a method for determining severity of coronary stenosis provided by an embodiment of the present disclosure;
FIG. 2 is a flow chart of a method of preprocessing sample input and sample output provided by an embodiment of the present disclosure;
fig. 3 is a schematic diagram of a composition structure of score interpretation information according to an embodiment of the present disclosure;
FIG. 4-1 is a schematic diagram showing one presentation of a global interpretation, shown in an embodiment of the present disclosure;
FIG. 4-2 is a schematic diagram showing an overall explanation of the predictive relationship between LDL-C and Gensini scores in accordance with an embodiment of the present disclosure;
FIGS. 4-3 are schematic diagrams illustrating an overall explanation of the predictive relationship between one interaction term (age and LDL-C) and Gensini score, as shown in embodiments of the present disclosure;
FIGS. 4-4 are schematic diagrams showing a presentation of partial explanations according to embodiments of the present disclosure;
FIG. 5 is a block diagram of a coronary stenosis severity determination apparatus according to an embodiment of the present disclosure;
Fig. 6 is a schematic structural diagram of an electronic device adapted to perform a coronary artery stenosis severity determination method according to an embodiment of the present disclosure.
Detailed Description
Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness. It should be noted that, without conflict, the embodiments of the present disclosure and features of the embodiments may be combined with each other.
In the technical scheme of the disclosure, the related processes of collecting, storing, using, processing, transmitting, providing, disclosing and the like of the personal information of the user accord with the regulations of related laws and regulations, and the public order colloquial is not violated.
Currently, many coronary artery disease risk assessment models are mainly based on statistical methods, such as linear regression models, logistic regression models or Cox (Cox regression) scale risk models. These models analyze the relationship between LDL-C and coronary stenosis in a linear or near linear manner. Although simple and feasible, the method has obvious limitation that complex nonlinear relations cannot be accurately captured. Furthermore, the linear regression-based approach assumes that the relationship between variables is linear, and does not adequately reflect the nonlinear effects between LDL-C and atherosclerotic plaque formation, coronary stenosis. There are also some complex machine learning models such as deep neural networks (Deep Neural Networks, DNN) applied in the field of coronary heart disease risk prediction, but the "black box" nature of these models makes them difficult to apply clinically, as doctors cannot clearly interpret the prediction basis of the model.
For clinical assessment of coronary stenosis severity, arterial stenosis is typically quantitatively analyzed by coronary angiography and calculating Gensini scores. Gensini score is a commonly used tool to assess the severity of coronary lesions, with higher scores indicating more severe coronary stenosis. Although Gensini scoring can well quantify the degree of coronary stenosis, it relies on complex and invasive clinical detection means and cannot be directly correlated with blood lipid indicators such as LDL-C.
Through analysis of the above related art, and by applicant's combination with long-term clinical experience, it is believed that it has the following technical drawbacks:
1) Limitations of the linear hypothesis the existing linear regression or Logistic regression models assume that the relationship between LDL-C and coronary stenosis is linear. However, practical clinical data indicate that LDL-C is often non-linear in relation to coronary stenosis, especially in different patient populations, and that risk assessment of LDL-C requires a more flexible non-linear model, which is affected by factors such as age, sex, past medical history, etc.;
2) The lack of processing power for complex interactions existing models often ignore complex interactions between LDL-C and other variables (e.g., age, gender, blood pressure, diabetes, etc.). These interactions play an important role in the pathological course of coronary heart disease, but it is difficult for traditional linear models to capture these details;
3) Model interpretability is poor, while deep learning and other complex machine learning models have high prediction accuracy, their "black box" nature makes predictions of these models difficult to interpret, resulting in a physician not being able to clinically trust the output of these models. Physicians often need clear knowledge of the basis of risk prediction to be able to accurately apply it in clinical practice in patient management and decision making.
The main body of execution of the coronary artery stenosis severity determination method provided by the present disclosure may be a terminal device, or a computer, or a server, or may be other devices having data processing capabilities. The subject of execution of the method is not limited herein.
Optionally, the terminal device may be a mobile phone, a tablet computer, a wearable device, a vehicle-mounted device, an augmented Reality (Augmented Reality, AR)/Virtual Reality (VR) device, a notebook computer, an Ultra-Mobile Personal Computer (UMPC), a netbook, a personal digital assistant (PersonalDigital Assistant, PDA), or the like, and the specific type of the terminal device is not limited in the embodiments of the present disclosure.
In some embodiments, the server may be a single server, or may be a server cluster formed by a plurality of servers. In some implementations, the server cluster may also be a distributed cluster. The present disclosure is not limited to a specific implementation of the server.
Referring to fig. 1, fig. 1 is a flowchart of a coronary artery stenosis severity determination method according to an embodiment of the present disclosure, wherein a flow 100 includes the following steps:
step 101, obtaining actual blood fat parameters including low density lipoprotein cholesterol of a target patient;
this step aims at acquiring actual blood lipid parameters including low density lipoprotein cholesterol of a target patient by an execution subject (e.g., a terminal device or a server having data analysis processing capability, etc.) of the coronary artery stenosis severity determination method. That is, the actual blood lipid parameter should include at least one of low density lipoprotein cholesterol, high density lipoprotein cholesterol, total cholesterol, and triglyceride.
The actual blood lipid parameters of a patient are important parameters for assessing cardiovascular health and metabolic status. In particular, actual lipid parameters include low density lipoprotein cholesterol (LDL-C) and other key lipid indicators. The following are specific descriptions of these indicators:
Low density lipoprotein cholesterol (LDL-C), a type of lipoprotein in the blood, is primarily responsible for the transport of cholesterol from the liver to various tissues of the body, and high levels of LDL-C are associated with increased risk of cardiovascular diseases such as atherosclerosis, heart disease, etc., typically obtained by blood testing, and are commonly expressed in milligrams per deciliter (mg/dL).
High density lipoprotein cholesterol (HDL-Density Lipoprotein Cholesterol, HDL-C) is also known as "good cholesterol" because it helps to recover cholesterol from blood and tissues, transport it back to the liver, reduce the risk of heart disease, and higher HDL-C levels are typically associated with lower risk of cardiovascular disease.
Total cholesterol (Total Cholesterol, TC), which is the sum of all cholesterol in the blood, including LDL, HDL, and other types of cholesterol, is typically checked to assess the overall risk of cardiovascular disease.
Triglyceride (TG), a fat in the blood mainly used for energy storage, high levels of Triglyceride are also associated with increased risk of cardiovascular disease, while maintaining normal Triglyceride levels is critical for the prevention of metabolic syndrome and heart disease.
In clinical practice, obtaining the blood lipid parameters can help doctors to better perform the following operations of 1) risk assessment, namely understanding cardiovascular disease risk of patients, guiding life style changes and treatment schemes, 2) treatment monitoring, namely adjusting drug treatment (such as statin drugs) according to blood lipid levels, and 3) personalized medical treatment, namely making a personalized health management plan according to specific blood lipid conditions of the patients.
The actual blood lipid parameters including LDL-C are obtained, the LDL-C is taken as a core index, and a plurality of indexes are comprehensively considered to enrich input variables, so that the blood lipid parameters of a patient can be comprehensively and clearly defined, and the model can learn more and deeper knowledge from the parameters.
Step 102, acquiring clinical characteristics of a target patient, including identity information and prior medical history;
On the basis of step 101, this step aims at acquiring clinical characteristics of the target patient including identity information and prior medical history by the execution subject. The identity information comprises sex information and age information, and the past medical history comprises whether the history of any of hypertension, diabetes, kidney diseases and apoplexy is illness or illness history.
Where the patient's sex information generally includes male and female, this is critical to medical decision-making, since the incidence and symptoms of certain diseases may vary from gender to gender, and the sex differences may affect drug metabolism, disease risk and therapeutic response (e.g., cardiovascular disease may vary in female and male manifestations and risk factors), while the patient's age refers to the actual age of the patient, usually in years, which is an important factor in assessing disease risk, and the incidence of many diseases (e.g., cardiovascular disease, cancer, etc.) is closely related to age. Age may also affect treatment selection and prognosis evaluation.
The hypertension is a state that the blood pressure is continuously increased, and usually more than 140/90 mmHg is taken as a diagnosis standard, the hypertension is an important risk factor of cardiovascular diseases, and knowing whether a patient has history of the hypertension can help doctors to evaluate risks of diseases such as heart diseases, cerebral apoplexy and the like;
Diabetes is a metabolic disease affecting how the body uses glucose, and is largely classified into type 1 and type 2 diabetes, and the risk of cardiovascular disease, kidney disease and neuropathy in diabetics is significantly increased. Acquiring a patient's history of diabetes aids in the formulation of diets, exercise and medication regimens.
Kidney disease refers to a disease in which kidney function gradually decreases, which is usually evaluated by glomerular filtration rate (Glomerular Filtration Rate, GFR), and is closely related to cardiovascular disease, and knowledge of whether a patient has such a history can guide kidney function monitoring and corresponding therapeutic measures.
Stroke is an acute disorder caused by the influence of brain blood supply, and is generally classified into ischemic and hemorrhagic stroke, and the past history of stroke can significantly increase the risk of the patient to re-stroke, and knowledge of this history can help doctors to make comprehensive risk assessment and prevention strategies.
This step aims to more fully understand the actual health condition of the patient, i.e. the influence of the historical health state on the current and future health conditions, by acquiring clinical features including identity information and past medical history.
Step 103, taking actual blood fat parameters and clinical characteristics as input information, and inputting a coronary artery stenosis severity prediction model obtained through pre-training;
based on step 102, the present step aims to input the coronary stenosis severity prediction model obtained by training in advance by using the actual blood lipid parameter and the clinical feature together as input information by the execution subject.
Wherein the coronary stenosis severity prediction model is trained using training samples (i.e., each training sample is a sample pair formed by a sample input-a sample output) formed by a sample input served by actual blood lipid parameters and clinical characteristics of the historic patient and a sample output served by a corresponding real coronary stenosis score, the coronary stenosis severity prediction model employing a model framework of an interpretable enhancer (Explainable Boosting Machine, EBM).
Among them, EBM is essentially an interpretable machine learning model, a variant of the generalized additive model, aimed at providing interpretability of the model while maintaining high prediction accuracy. The EBM combines the advantages of ensemble learning and interpretability modeling, and is suitable for deep analysis of the influence of features.
The method has the following main characteristics:
1) Explanation EBM allows a user to understand the impact of each feature on the final prediction by decomposing the prediction of the model into its contributions to the individual features. This interpretability is particularly important in areas where transparent decision making processes are required, such as medical, financial and legal.
2) The generalized addition model (Generalized Additive Models, GAMs) has a prediction result that is a weighted sum of features. The contribution of each feature is modeled by a non-parametric function, allowing a nonlinear relationship to be captured.
3) Ensemble learning-using enhancement techniques, similar to gradient-lifted trees (Gradient Boosting Trees), models are gradually added to optimize the loss function. Each step is improved on the basis of the previous model, so that the overall performance is improved.
4) Processing feature interactions-interactions between different features can be handled naturally. By modeling the feature interactions, the user can more clearly understand which feature combinations have the greatest impact on the predicted outcome.
5) High efficiency the training process is typically faster than traditional deep learning models and performs well on small to medium scale data sets.
The method and the device specifically use the model obtained through EBM training to help doctors understand how the model obtains diagnosis results, and can be applied to a financial service scene to explain the decision basis of the credit score and risk assessment model and a user behavior analysis scene to analyze the influence of the user behavior characteristics on marketing strategies besides the scene. Namely, the EBM can provide a high-precision prediction result, ensure the transparency of a decision process and is suitable for various application scenes in which the model prediction back cause needs to be understood. By using EBM, data scientists and business analysts can more effectively trench-pass model results, improving the user's sense of trust.
Before data analysis, all collected patient data (i.e., the sample inputs and sample outputs mentioned above) can also be subjected to comprehensive data preprocessing to ensure data integrity and consistency. The method mainly comprises the following two processes of carrying out interpolation processing on missing values in sample input and sample output and carrying out rejection or replacement processing on the missing values in the sample input and the sample output, so that finally, the data subjected to the interpolation processing and the rejection or replacement processing form a training sample for subsequent training, and the same is true if a test set is required to be formed.
Specifically, for the interpolation processing of the missing values, a median interpolation method may be used for continuous variables (e.g., LDL-C, TC, etc.), and a mode interpolation method may be used for classified variables (e.g., gender, medical history, etc.). By the method, the influence of data loss on model training is reduced to the greatest extent;
For the outlier rejection or replacement process, statistical methods such as box-plot (Boxplot) may be used to identify outliers in the data, and then reject or replace outliers in the blood lipid index, gensini score that deviate extremely from the normal range, to prevent these outliers from negatively affecting the stability of the model (the above process corresponds to the schematic diagram shown in fig. 2).
104, Receiving actual coronary artery stenosis score and score interpretation information output by a coronary artery stenosis severity prediction model;
on the basis of step 103, this step aims at receiving the actual coronary stenosis score and score interpretation information outputted by the coronary stenosis severity prediction model by the execution subject.
Specifically, the score may still be obtained based on Gensini scoring algorithm, and the score interpretation information is used to explain how to obtain the actual coronary stenosis score, and may include at least one of variable importance analysis result, global interpretation information, interaction analysis information, and local interpretation information (corresponding to the schematic diagram shown in fig. 3).
The variable importance analysis result is used for presenting independent influence degrees of different types of input variables on obtaining an actual coronary artery stenosis score respectively, the global interpretation information is used for presenting actual corresponding relations between variable values of each input variable and the coronary artery stenosis score respectively, the interaction analysis is used for presenting combined influence degrees of the different types of input variables on obtaining the actual coronary artery stenosis score when the input variables are combined mutually, the local interpretation information is used for presenting contribution polarities of variable values of each input variable on obtaining the actual coronary artery stenosis score respectively for a target patient, and the contribution polarities comprise positive contribution of positive correlation and negative contribution of negative correlation.
Furthermore, the score interpretation information can further comprise risk levels, such as a low risk level, a medium risk level and a high risk level, to which the actual coronary stenosis score belongs, so as to correspond to the severity and the treatment mode through different risk levels, wherein the risk threshold value of each risk level obtained by dividing can be determined by self through using a training sample based on the coronary stenosis severity prediction model.
Step 105, determining the actual coronary stenosis severity of the target patient based on the actual coronary stenosis score and the score interpretation information.
Based on step 104, the present step aims at pushing the actual coronary stenosis score and the score interpretation information to the corresponding doctor by the execution subject as auxiliary diagnosis information, so that the doctor can make a diagnosis finally, and further, the actual coronary stenosis severity of the target patient can be determined finally.
According to the coronary artery stenosis severity determination method provided by the embodiment of the disclosure, by adopting the interpretable enhancer model which expands the enhancement method based on the regression tree on the generalized addition model as a model framework and adopting the blood lipid parameters including the low density lipoprotein cholesterol and the clinical characteristics including the identity information and the past medical history as input variables, the accurate relationship between the coronary artery stenosis severity can be predicted jointly by using the influence characteristics of a plurality of types, and the characteristics of the interpretable enhancer model can provide clear characteristic interpretation and interaction effect between variables while guaranteeing the prediction accuracy, so that the model can output interpretation information of results, and the acceptance degree of a doctor on output results can be enhanced, and a doctor can be better assisted to make more accurate diagnosis.
For deepening understanding, the disclosure further provides a whole set of specific implementation schemes in combination with a specific application scenario:
Aiming at the technical defects existing in the prior related art, the embodiment provides an LDL-C risk layering scheme based on an EBM model, aims at accurately identifying the influence of different levels of LDL-C on coronary stenosis risk by utilizing the nonlinear regression capability and high interpretability of the model, and provides more personalized and accurate risk assessment. Meanwhile, the embodiment hopes to provide reliable prediction basis for clinicians through transparency of the EBM model, so that the EBM model is effectively applied to actual clinical decisions.
To analyze the complex nonlinear relationship between LDL-C levels and coronary stenosis degree (Gensini score), in combination with other cardiovascular risk factors (e.g., age, gender, past medical history, etc.), a machine learning model with good interpretation needs to be established first. The model has the following characteristics:
1) Multivariate input is not only to analyze LDL-C levels, but also to comprehensively evaluate the coronary stenosis risk of the patient by combining clinical characteristics including age, sex, past medical history (such as diabetes, hypertension, renal insufficiency) and the like.
2) Nonlinear modeling the EBM model captures the nonlinear relationship between LDL-C and coronary stenosis severity by stepwise addition regression, precisely dividing the risk threshold for LDL-C.
3) The method has high interpretation, namely the contribution degree and interaction of each feature output by the EBM model provide clear prediction basis for clinicians, so that the method has high clinical practicability.
The following steps illustrate the main steps:
Step 1 data collection
Based on a single-center cross-sectional study, subjects were acute myocardial infarction (Acute Myocardial Infarction, AMI) patients treated in the cardiovascular department of the beijing XX hospital during the period of 2017, 9, to 2023, 6. All patients were adults aged 18 years and older. Exclusion criteria included individuals lacking coronary angiography recordings and blood lipid measurements. The patient was selected based on diagnosis by a cardiologist and the extent of coronary stenosis was confirmed by coronary angiography. The data source is a hospital electronic medical record system (Electronic Medical Record, EMR) and laboratory detection system.
The specific patient screening procedure was to retrieve 1264 patients from the electronic medical record system, first remove participants (n=56) lacking coronary angiography records, then exclude participants (n=8) lacking data of four blood lipid items [ low density lipoprotein cholesterol (LDL-C), total Cholesterol (TC), high density lipoprotein cholesterol (HDL-C), triglyceride (TG) ], and finally obtain a total of 1200 participants. Table 1 below shows the characteristics of 1200 participants, the values in table 1 being shown as median [ Q1, Q3] or n (%):
TABLE 1 actual blood lipid parameters and clinical profile for 1200 participants
The data in table 1 includes:
1) Coronary angiography data the extent of coronary stenosis was obtained for each patient by coronary angiography (Coronary Angiography) and a Gensini score was used to quantitatively evaluate the extent of lesions. Gensini the score was based on the stenosis degree (in percent) and lesion location (different segment weights) of each coronary segment. The higher the score, the more severe the coronary stenosis.
2) The blood lipid index is that the key blood lipid data such as low density lipoprotein cholesterol (LDL-C), total Cholesterol (TC), high density lipoprotein cholesterol (HDL-C), triglyceride (TG) and the like of each patient are obtained through laboratory detection. Wherein LDL-C is the core index of the invention, and is used for evaluating the main risk factors of coronary artery stenosis.
3) Patient characteristics:
Age (age) the actual age of the patient was included in the model as a continuous variable.
Sex (sex) male or female as binary classification variables.
Medical history including hypertension (hypertension), diabetes (diabetes), kidney disease (KIDNEY DISEASE), past stroke (stroke) as binary classification variables, which affect the progression of coronary stenosis and thus are important covariates in the model.
Step 2, data preprocessing
Before data analysis, the embodiment performs comprehensive data preprocessing on all collected patient data, and ensures the integrity and consistency of the data. The method comprises the following specific steps:
1) And (3) processing missing values, namely adopting a reasonable interpolation strategy for variable values missing in partial patient data. For continuous variables (e.g., LDL-C, TC, etc.), the median interpolation method is used, and for classified variables (e.g., gender, medical history, etc.), the mode interpolation method is used. By the method, the influence of data loss on model training is reduced to the greatest extent.
2) Abnormal value detection and processing, namely identifying the abnormal value in the data by using a statistical method such as a box diagram method. Abnormal points in the blood lipid index and Gensini scores, which deviate from the normal range extremely, are removed or replaced so as to prevent the abnormal values from negatively affecting the stability of the model.
Step 3 EBM model construction
The core of this embodiment is modeling of the relationship of LDL-C with coronary stenosis risk by the EBM model. The EBM model is an extension of a generalized addition model, and provides clear characteristic interpretation and interaction effect between variables while ensuring prediction accuracy by using a regression tree-based enhancement method.
The model is constructed as follows:
1) Data set partitioning data from 1200 patients were randomly partitioned into training and testing sets, with 70% of the data being used for training and 30% of the data being used for testing. By this partitioning, it is ensured that the model is able to learn the data patterns on the training set and evaluate its generalization ability on the test set.
2) Variable modeling a regression EBM model was created using the "INTERPRETML" library of Python software. For each input variable (including LDL-C, age, gender, etc.), the EBM model was fitted step by step using a regression tree. In each iteration, the model continually updates the additive structure of each variable by minimizing the prediction error. The nonlinear regression capability of EBM is able to capture these complex patterns due to the significant nonlinear relationship between LDL-C and coronary stenosis.
3) Interaction item detection the EBM model has the ability to automatically detect interactions. During model training, the model automatically recognizes the interactive effects of LDL-C and other variables (e.g., age, gender, hypertension, etc.). For example, models find that LDL-C affects coronary stenosis differently in patients of different ages, and this interaction is difficult to capture in traditional linear models.
4) Model training and parameter setting the EBM model directly uses default parameters for training without feature scaling. Each feature is fitted using a shape function, and interactions between features are automatically detected from the actual data and incorporated into the model. Finally, the prediction capability of the model is gradually optimized through iterative updating of the model.
Step 4 EBM model interpretation
The EBM model provides variable importance analysis, global interpretation, interaction analysis and local interpretation, and can clearly display the contribution of each feature to final prediction. The output of the model includes:
1) Variable importance analysis this embodiment ranks the importance of each variable by calculating the average absolute score of each feature in the training set, see fig. 4-1. Of all variables, age was found to be the most influential variable, followed by LDL-C and diabetes. The interpretability of the model allows the physician to clearly understand the effect of each variable on the risk of coronary stenosis. Wherein the term "importance" refers to the average absolute contribution (score) of each term (feature or interaction) to the prediction. These contributions are averaged over the training dataset taking into account the number of samples in each bin and the sample weights (e.g., correlation)
2) Global interpretation-the relationship between each predicted variable and coronary stenosis (Gensini score) is shown. The score (y-axis) for each feature represents its contribution to the Gensini score. For example, the higher the score for LDL-C, the more severe the predicted coronary stenosis. FIG. 4-2 shows the non-linear relationship between LDL-C and Gensini scores. It should be noted that since EBM is an additive model, FIG. 4-2 can accurately quantify the contribution of LDL-C to Gensini scoring predictions. For example, if a new data point of LDL-c=1 mmol/L is entered, the model will increase the final Gensini score prediction by about-7. The density is a histogram depicting the data distribution of LDL-C. Error bars are indicated by grey marks. These are approximations of the uncertainty of the model to the particular feature region. The impact of slight modification of the training data increases with the width of the bars.
3) Interaction analysis figures 4-3 fully explain the predictive relationship between interaction terms (age and LDL-C) and Gensini scores. When age > 75 years, LDL-C > 4.5mmol/L, the predicted Gensini fraction of interaction term >3, indicating that the risk of coronary stenosis increases significantly when the values of age and LDL-C are both above the corresponding ranges.
4) Local interpretation-for a prediction of a particular patient, the model can interpret the specific contribution of each feature to the patient Gensini's score, see in particular figures 4-4. For example, for a new patient, the patient had an LDL-C of 3.07 mmol/L and a predicted Gensini score contribution of +0.83, indicating that LDL-C increases the patient's risk of coronary stenosis. And adding the scores related to each predicted variable of the patient, and finally predicting that the total score of Gensini of the patient is 58, the total score of the real Gensini of the patient is 47 by using an EBM model, wherein the scores are relatively close to each other, so that the prediction effect of the EBM model on the patient is relatively good. Orange in fig. 4-4 represents a positive contribution and blue represents a negative contribution.
Step 4 LDL-C danger stratification
By training patient data, the EBM model automatically recognizes the nonlinear variation trend of LDL-C on coronary stenosis risk, so that the threshold value can be accurately selected to carry out risk stratification on LDL-C. In this example, the layering of LDL-C levels resulted in the following:
1) Low risk group LDL-C < 1.9 mmol/L. The Gensini scores were lower for this group of patients and the risk of coronary stenosis was small. When LDL-C was at 1.9 mmol/L, the predicted Gensini score reached zero, indicating a change in contribution to coronary artery disease risk from negative to positive. According to model predictions, coronary stenosis risk is relatively low when LDL-C levels are below 1.9 mmol/L, and the model outputs predictions that are highly consistent with clinical observations.
2) 1.9 Mmol/L.ltoreq.LDL-C.ltoreq.2.7. 2.7 mmol/L. Within this range, the nonlinear relationship of LDL-C levels to coronary stenosis appears, with a gradual increase in the Gensini score of the model prediction. This group of patients, while at moderate risk, can significantly reduce the incidence of coronary stenosis by early intervention.
3) High risk group LDL-C > 2.7 mmol/L. At this point LDL-C levels were significantly elevated, with a significant positive correlation with coronary stenosis risk. The model predicts that Gensini scores are higher for this group of patients, indicating that these patients may have severe coronary stenosis, requiring immediate clinical intervention.
Step 5, model performance evaluation
The performance of the EBM model can be comprehensively evaluated by the following indexes:
Root Mean Square Error (RMSE) the prediction error of the model is estimated by the Root Mean Square Error (RMSE). On the test set, RMSE of the EBM model was 47.60, indicating that the model had higher accuracy in predicting coronary stenosis.
The whole set of method provided by the embodiment has the following technical effects:
1) Capturing nonlinear relationship the EBM model is able to accurately capture nonlinear relationship between LDL-C and coronary stenosis compared to traditional linear regression models. While the traditional linear model assumes that the relationship between variables is linear, the present invention can better reflect the actual effect of LDL-C on coronary stenosis at different levels through the application of EBM models.
2) Personalized risk assessment the present invention achieves personalized LDL-C risk stratification by combining other clinical characteristics of the patient (e.g., age, sex, hypertension, diabetes, etc.). The method enables a clinician to provide more accurate treatment advice according to the specific situation of patients, and has important application value in the management of patients with coronary heart disease and acute myocardial infarction.
3) High interpretive-the interpretive nature of the EBM model makes it more clinically useful than the "black box" model of deep learning, etc. The doctor can clearly see the specific contribution of each prediction factor to the coronary artery stenosis risk of the patient through the interpretation output of the model, and the application potential of the model in actual clinical decisions is greatly increased.
With further reference to fig. 5, as an implementation of the method shown in the above figures, the present disclosure provides an embodiment of a coronary stenosis severity determination apparatus, which corresponds to the method embodiment shown in fig. 1, which is particularly applicable in a variety of electronic devices.
As shown in fig. 5, the coronary artery stenosis severity determining apparatus 500 of the present embodiment may include an actual blood lipid parameter acquisition unit 501, a clinical feature acquisition unit 502, an actual parameter input unit 503, a model output result receiving unit 504, and an actual coronary artery stenosis severity determining unit 505. The system comprises an actual blood lipid parameter obtaining unit 501 configured to obtain an actual blood lipid parameter including low density lipoprotein cholesterol of a target patient, a clinical characteristic obtaining unit 502 configured to obtain clinical characteristics including identity information and past medical history of the target patient, an actual parameter input unit 503 configured to input a coronary artery stenosis severity prediction model obtained through pre-training by taking the actual blood lipid parameter and the clinical characteristics as input information, wherein the coronary artery stenosis severity prediction model is obtained through training by using a sample input acted by the actual blood lipid parameter and the clinical characteristics of a historical patient and a training sample output acted by a corresponding real coronary artery stenosis score, the coronary artery stenosis severity prediction model adopts a model frame of an interpretable enhancer, a model output result receiving unit 504 configured to receive an actual coronary artery stenosis score and score interpretation information output by the coronary artery stenosis severity prediction model, and an actual coronary artery stenosis severity determining unit 505 configured to determine the actual coronary artery stenosis severity of the target patient based on the actual coronary artery stenosis score and the severity interpretation information.
In the embodiment, the specific processing and the technical effects of the actual blood lipid parameter obtaining unit 501, the clinical feature obtaining unit 502, the actual parameter input unit 503, the model output result receiving unit 504, and the actual coronary artery stenosis severity determining unit 505 in the coronary artery stenosis severity determining apparatus 500 can refer to the relevant descriptions of steps 101-105 in the corresponding embodiment of fig. 1, and are not repeated here.
In some other alternative implementations of the present embodiment, the actual blood lipid parameter includes at least low density lipoprotein cholesterol and further includes at least one of high density lipoprotein cholesterol, total cholesterol, and triglycerides.
In some other alternative implementations of the present embodiment, the identity information includes gender information, age information, and the prior medical history includes whether there is a history of suffering from or experiencing any of hypertension, diabetes, kidney disease, stroke.
In some other alternative implementations of the present embodiment, before constructing the training samples using the sample inputs and the sample outputs, further comprising:
a missing value interpolation unit configured to perform interpolation processing on missing values in the sample input and the sample output;
an outlier processing unit configured to perform a culling or replacing process on outliers in the sample input and the sample output to construct the interpolation-processed and culled or replaced data into training samples.
In some other alternative implementations of the present embodiment, the missing-value interpolation unit is further configured to:
performing interpolation processing on continuous variables with missing values in sample input and sample output by adopting a median interpolation method;
And performing interpolation processing on the classified variables with the missing values in the sample input and the sample output by adopting a mode interpolation method.
In some other alternative implementations of the present embodiment, the scoring interpretation information includes:
at least one of variable importance analysis results, global interpretation information, interaction analysis information and local interpretation information;
The method comprises the steps of obtaining a coronary artery stenosis score, obtaining a global interpretation information, wherein variable importance analysis results are used for presenting independent influence degrees of different types of input variables on obtaining the actual coronary artery stenosis score respectively, the global interpretation information is used for presenting actual corresponding relations between variable values of each input variable and the coronary artery stenosis score respectively, interaction analysis is used for presenting combined influence degrees of the different types of input variables on obtaining the actual coronary artery stenosis score when the input variables are combined mutually, and local interpretation information is used for presenting contribution polarities of variable values of each input variable on obtaining the actual coronary artery stenosis score respectively for a target patient, wherein the contribution polarities comprise positive contribution of positive correlation and negative contribution of negative correlation.
In some other optional implementations of the present embodiment, the score interpretation information further includes a risk level to which the actual coronary stenosis score belongs, wherein the risk threshold for classifying the risk level is determined by using training samples based on the coronary stenosis severity prediction model.
The present embodiment exists as an embodiment of an apparatus corresponding to the above-described method embodiment, and the coronary stenosis severity determining apparatus provided in the present embodiment, by adopting an interpretable enhancer model, which expands a regression tree-based enhancement method on a generalized addition model, as a model framework, and adopting blood lipid parameters including low-density lipoprotein cholesterol and clinical features including identity information and past medical history as input variables together, an accurate relationship between a coronary stenosis severity and a plurality of types of influence features can be predicted together, and by using the above-described multivariate history parameters as training samples, clear feature interpretation and interaction effects between variables can be provided while ensuring prediction accuracy by means of the characteristics of the interpretable enhancer model, so that the model can also output interpretation information of how to obtain results, thereby also enhancing the acceptance of a doctor on output results, and better assisting a doctor in making a more accurate diagnosis.
According to an embodiment of the present disclosure, there is also provided an electronic device including at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the coronary stenosis severity determination method described in any of the embodiments above.
According to an embodiment of the present disclosure, there is also provided a readable storage medium storing computer instructions for enabling a computer to implement the coronary stenosis severity determination method described in any of the above embodiments when executed.
According to an embodiment of the present disclosure, the present disclosure further provides a computer program product, which, when being executed by a processor, is capable of implementing the steps of the coronary stenosis severity determination method described in any of the embodiments described above.
Fig. 6 illustrates a schematic block diagram of an example electronic device 600 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.
As shown in fig. 6, the apparatus 600 includes a computing unit 601 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 602 or a computer program loaded from a storage unit 608 into a Random Access Memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 may also be stored. The computing unit 601, ROM 602, and RAM 603 are connected to each other by a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.
Various components in the device 600 are connected to the I/O interface 605, including an input unit 606, e.g., keyboard, mouse, etc., an output unit 607, e.g., various types of displays, speakers, etc., a storage unit 608, e.g., magnetic disk, optical disk, etc., and a communication unit 609, e.g., network card, modem, wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.
The computing unit 601 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 601 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the various methods and processes described above, such as the coronary stenosis severity determination method. For example, in some embodiments, the coronary stenosis severity determination method may be implemented as a computer software program, tangibly embodied on a machine-readable medium, such as the storage unit 608. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device 600 via the ROM 602 and/or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the coronary stenosis severity determination method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the coronary stenosis severity determination method in any other suitable way (e.g., by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be a special or general purpose programmable processor, operable to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.
Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user, for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback), and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a Local Area Network (LAN), a Wide Area Network (WAN), and the Internet.
The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so as to solve the defects of large management difficulty and weak service expansibility in the traditional physical host and Virtual Private Server (VPS) PRIVATE SERVER service.
According to the technical scheme of the embodiment of the disclosure, by adopting the interpretable enhancer model which expands the enhancement method based on the regression tree on the generalized addition model as a model framework and adopting blood fat parameters including low density lipoprotein cholesterol and clinical characteristics including identity information and past medical history as input variables together, the accurate relationship between coronary artery stenosis severity can be predicted together by using a plurality of types of influence characteristics, and as the training sample using the multivariate history parameters, clear characteristic interpretation and interaction effect between variables can be provided while ensuring the prediction precision by means of the characteristics of the interpretable enhancer model, so that the model can also output interpretation information of how to obtain results, thereby enhancing the acceptance of doctors on output results and better assisting doctors in making more accurate diagnoses.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel or sequentially or in a different order, provided that the desired results of the technical solutions of the present disclosure are achieved, and are not limited herein.
The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411876580.0A CN119324064B (en) | 2024-12-19 | 2024-12-19 | Method, device, electronic device and storage medium for determining severity of coronary artery stenosis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411876580.0A CN119324064B (en) | 2024-12-19 | 2024-12-19 | Method, device, electronic device and storage medium for determining severity of coronary artery stenosis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119324064A CN119324064A (en) | 2025-01-17 |
| CN119324064B true CN119324064B (en) | 2025-03-21 |
Family
ID=94234584
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411876580.0A Active CN119324064B (en) | 2024-12-19 | 2024-12-19 | Method, device, electronic device and storage medium for determining severity of coronary artery stenosis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119324064B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115132363A (en) * | 2022-08-11 | 2022-09-30 | 中国医科大学附属盛京医院 | Method for constructing coronary disease risk discrimination model based on conventional physical examination indexes |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100030035A1 (en) * | 2008-08-04 | 2010-02-04 | The Hong Kong Polytechnic University | Fuzzy system for cardiovascular disease and stroke risk assessment |
| WO2012166798A2 (en) * | 2011-06-01 | 2012-12-06 | Wake Forest University Health Sciences | Systems and apparatus for indicating risk of coronary stenosis |
| CN109919931B (en) * | 2019-03-08 | 2020-12-25 | 数坤（北京）网络科技有限公司 | Coronary stenosis degree evaluation model training method and evaluation system |
| CN118762826B (en) * | 2024-09-06 | 2024-11-29 | 杭州倍佐健康科技有限公司 | Method for predicting coronary artery stenosis and FFR based on big data of coronary artery diagnosis and treatment |
-
2024
- 2024-12-19 CN CN202411876580.0A patent/CN119324064B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115132363A (en) * | 2022-08-11 | 2022-09-30 | 中国医科大学附属盛京医院 | Method for constructing coronary disease risk discrimination model based on conventional physical examination indexes |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119324064A (en) | 2025-01-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Walsh et al. | Imaging research in fibrotic lung disease; applying deep learning to unsolved problems | |
| Retson et al. | Machine learning and deep neural networks in thoracic and cardiovascular imaging | |
| Angraal et al. | Machine learning prediction of mortality and hospitalization in heart failure with preserved ejection fraction | |
| Ramezankhani et al. | Applying decision tree for identification of a low risk population for type 2 diabetes. Tehran Lipid and Glucose Study | |
| US20160063212A1 (en) | System for Generating and Updating Treatment Guidelines and Estimating Effect Size of Treatment Steps | |
| US20110202486A1 (en) | Healthcare Information Technology System for Predicting Development of Cardiovascular Conditions | |
| US20160358290A1 (en) | Health severity score predictive model | |
| Lüscher et al. | Artificial intelligence in cardiovascular medicine: clinical applications | |
| Yadav et al. | [Retracted] FVC‐NET: An Automated Diagnosis of Pulmonary Fibrosis Progression Prediction Using Honeycombing and Deep Learning | |
| Chinnasamy et al. | Machine learning based cardiovascular disease prediction | |
| Wang et al. | Accurate estimation of biological age and its application in disease prediction using a multimodal image Transformer system | |
| Abdollahi et al. | Artificial intelligence in assessing cardiovascular diseases and risk factors via retinal fundus images: A review of the last decade | |
| Koloi et al. | Predicting early-stage coronary artery disease using machine learning and routine clinical biomarkers improved by augmented virtual data | |
| Lai et al. | Development of a metabolite-based deep learning algorithm for clinical precise diagnosis of the progression of diabetic kidney disease | |
| Panigrahy et al. | Predictive Modelling of Diabetes Complications: Insights from Binary Classifier on Chronic Diabetic Mellitus | |
| Mahmood et al. | Enhancing Coronary Artery Disease Prognosis: A Novel Dual-Class Boosted Decision Trees Strategy for Robust Optimization | |
| Lin et al. | Improving sensitivity and connectivity of retinal vessel segmentation via error discrimination network | |
| CN119324064B (en) | Method, device, electronic device and storage medium for determining severity of coronary artery stenosis | |
| Noori et al. | A comparative analysis for diabetic prediction based on machine learning techniques | |
| US8880457B2 (en) | Method and system for recommending a decision based on combined entity modeling | |
| Jandy et al. | A machine learning approach to classifying New York Heart Association (NYHA) heart failure | |
| CN115274116A (en) | PRO prognosis risk model for chronic heart failure patient and construction method | |
| Al-Sarayrah | RECENT ADVANCES AND APPLICATIONS OF APRIORI ALGORITHM IN EXPLORING INSIGHTS FROM HEALTHCARE DATA PATTERNS | |
| Mittal et al. | Advanced Hybrid Machine Learning Model for Accurate Detection of Cardiovascular Disease | |
| Almarzooqi | Using ML to understand the factors impacting diabetes in diabetic patients |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |