Postoperative Risk Assessment System, Device, Storage Medium, and Program Product

Abstract

The present invention provides a postoperative risk assessment system, device, storage medium and program product, relating to the field of computer technology. The system includes: an extraction module for extracting independent risk factors for early saphenous vein graft failure from the postoperative data of a patient; a score determination module for determining the scores of each factor in the independent risk factors based on the scoring scale in the nomogram model, where the nomogram model is constructed based on the independent risk factors; and a prediction module for predicting the risk information of early saphenous vein graft failure after coronary artery bypass grafting based on the risk coefficient corresponding to the total score of each factor in the independent risk factors. Based on the results of Lasso regression analysis and Logistic regression analysis, the present invention constructs a prediction model for early saphenous vein graft failure with risk factors such as QFR > 0.80, PI > 3.0, complications, and non-left main stenosis as the main body, thereby improving the accuracy of the risk assessment of early saphenous vein graft failure in postoperative patients.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and particularly to a postoperative risk assessment system, device, storage medium and program product. Background Art

[0002] Coronary artery bypass grafting (CABG) is a common surgical method for treating patients with complex coronary heart disease. Among them, arterial grafts are widely used in CABG because their long-term patency rate is significantly higher than that of venous grafts and can improve the long-term survival rate of patients. However, in a small number of patients, early graft failure (stenosis degree ≥ 50% or complete occlusion) occurs within 1 year after surgery, which may significantly affect the quality of life and medium- and long-term prognosis of patients. Therefore, determining the risk factors for early graft failure after surgery is of great clinical significance for evaluating the benefits of bypass surgery, optimizing postoperative management, reducing the restenosis rate and improving the long-term survival rate of patients. Therefore, how to achieve the postoperative risk assessment of CABG has become an urgent problem to be solved. Summary of the Invention

[0003] The present invention provides a postoperative risk assessment system, device, storage medium and program product to solve the defects of the postoperative risk assessment of CABG in the prior art, realize the risk assessment of early graft failure of postoperative patients, and improve the accuracy of the assessment.

[0004] The present invention provides a postoperative risk assessment system, including the following:

[0005] An extraction module, configured to extract independent risk factors for early graft failure from the postoperative data of patients; the independent risk factors include non-left main stenosis, complications, fractional flow reserve and pulsatility index;

[0006] A score determination module, configured to determine the scores of each factor in the independent risk factors based on the scoring scale in the nomogram model; the nomogram model is constructed based on the independent risk factors;

[0007] A prediction module, configured to predict the risk information of early graft failure after coronary artery bypass grafting based on the risk coefficient corresponding to the total score of each factor in the independent risk factors;

[0008] Wherein, the independent risk factors are determined by respectively performing least absolute shrinkage and selection operator Lasso regression analysis and multivariate logistic Logistic regression analysis on the data set of potential risk factors for cardiovascular diseases, and combining the results of Lasso regression analysis and Logistic regression analysis.

[0009] A postoperative risk assessment system provided by the present invention, wherein the independent risk factors are determined by the following method:

[0010] Perform Lasso regression analysis on the data set to obtain the first type of risk factors for early graft failure;

[0011] Perform Logistic regression analysis on the data set to obtain the second type of risk factors for early graft failure;

[0012] Rank the importance of the first type of risk factors and the second type of risk factors;

[0013] Select the common risk factors from the ranking results of the first type of risk factors and the ranking results of the second type of risk factors as the independent risk factors.

[0014] A postoperative risk assessment system provided by the present invention, wherein the prediction module is further configured to:

[0015] Based on the ranking results of the weight coefficients of the first type of risk factors and the second type of risk factors, select target risk factors; wherein, the target risk factors include the independent risk factors;

[0016] Based on the scoring scale in the nomogram model constructed based on the target risk factors, determine the scores of each factor in the target risk factors;

[0017] Based on the risk coefficients corresponding to the total scores of each factor in the target risk factors, predict the risk information of early graft failure after coronary artery bypass grafting.

[0018] A postoperative risk assessment system provided by the present invention, wherein the first type of risk factors are determined by the following method:

[0019] Construct the objective function of the Lasso regression model;

[0020] Based on the cross-validation method, determine the optimal regularization parameter of the objective function;

[0021] Use the data set to train the objective function after determining the optimal regularization parameter to determine the regression coefficient of each variable; the variable is the risk factor in the data set;

[0022] Based on the regression coefficient of each variable, determine the first type of risk factors.

[0023] A postoperative risk assessment system provided by the present invention, wherein the second type of risk factors are determined by the following method:

[0024] Construct a Logistic regression model;

[0025] Based on the Logistic regression model, perform univariate analysis on the data set to screen out candidate risk factors;

[0026] Perform multivariate analysis on the candidate risk factors to obtain the second type of risk factors.

[0027] According to a postoperative risk assessment system provided by the present invention, the prediction module is further configured to:

[0028] Based on the independent risk factors and their weight coefficients, determine the risk scores of the risk factors;

[0029] Based on the risk scores of the risk factors, divide the risk groups;

[0030] Based on the risk groups, predict major adverse cardiovascular and cerebrovascular events.

[0031] According to a postoperative risk assessment system provided by the present invention, the weight rankings of the factors in the independent risk factors are, in sequence, fractional flow reserve, complications, pulsatility index, and non-left main stenosis.

[0032] According to a postoperative risk assessment system provided by the present invention, the fractional flow reserve is greater than 0.8, and the pulsatility index is greater than 3.0.

[0033] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, the method in any one of the above-mentioned postoperative risk assessment systems is implemented.

[0034] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method in any one of the above-mentioned postoperative risk assessment systems is implemented.

[0035] The present invention also provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method in any one of the above-mentioned postoperative risk assessment systems is implemented.

[0036] The postoperative risk assessment system, device, storage medium and program product provided by the present invention. The system includes: an extraction module for extracting independent risk factors for early saphenous vein graft failure from the postoperative data of patients; a score determination module for determining the scores of each factor among the independent risk factors based on the scoring scale in the nomogram model, where the nomogram model is constructed based on the independent risk factors; and a prediction module for predicting the risk information of early saphenous vein graft failure after coronary artery bypass grafting based on the risk coefficient corresponding to the total score of each factor among the independent risk factors. Based on the results of Lasso regression analysis and Logistic regression analysis, the present invention constructs a prediction model for early saphenous vein graft failure with risk factors such as QFR > 0.80, PI > 3.0, complications, and non-left main stenosis as the main body, thereby improving the accuracy of risk assessment for early saphenous vein graft failure in postoperative patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0038] Figure 1 It is a schematic diagram of the modules of the postoperative risk assessment system provided by the present invention.

[0039] Figure 2 It is one of the schematic diagrams of the Lasso regression analysis of the risk factors for early saphenous vein graft failure provided by the present invention.

[0040] Figure 3 It is another schematic diagram of the Lasso regression analysis of the risk factors for early saphenous vein graft failure provided by the present invention.

[0041] Figure 4 It is yet another schematic diagram of the Lasso regression analysis of the risk factors for early saphenous vein graft failure provided by the present invention.

[0042] Figure 5 It is one of the schematic diagrams of the construction of the new model for the risk factors of early saphenous vein graft failure provided by the present invention.

[0043] Figure 6 It is another schematic diagram of the construction of the new model for the risk factors of early saphenous vein graft failure provided by the present invention.

[0044] Figure 7 It is yet another schematic diagram of the construction of the new model for the risk factors of early saphenous vein graft failure provided by the present invention.

[0045] Figure 8It is one of the schematic diagrams for verifying the efficacy of the early graft failure prediction model provided by the present invention.

[0046] Figure 9 It is the second schematic diagram for verifying the efficacy of the early graft failure prediction model provided by the present invention.

[0047] Figure 10 It is the third schematic diagram for verifying the efficacy of the early graft failure prediction model provided by the present invention.

[0048] Figure 11 It is the schematic diagram for comparing the cumulative MACCE-free event rates after CABG provided by the present invention.

[0049] Figure 12 It is the structural schematic diagram of the electronic device provided by the present invention. Detailed implementation manners

[0050] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0051] The influencing factors of early graft failure are complex and diverse, including traditional cardiovascular risk factors such as hypertension, diabetes, smoking history, etc., and also involve intraoperative hemodynamic parameters, complications and other factors. In the related art, a nomogram prediction model constructed with traditional cardiovascular risk factors such as smoking and hypertension has a certain ability to predict the risk of graft stenosis. However, with the in-depth study of hemodynamic and wall compliance evaluation tools such as quantitative flow ratio (QFR) and pulsatility index (PI), more and more evidence shows that blood flow parameters play an important role in evaluating the long-term prognosis of CABG patients. However, in the related art, there is no research on the association between these influencing factors and early graft failure.

[0052] Based on this, the present invention uses the Least Absolute Shrinkage and Selection Operator (Lasso) regression model and the multi-factor Logistic regression model to explore the potential risk factors of early graft failure after CABG, construct and evaluate the early graft failure risk prediction model, providing a new basis for the risk assessment and individualized treatment of postoperative patients.

[0053] The present invention provides a postoperative risk assessment system, specifically an early graft failure assessment system for coronary artery bypass grafting.

[0054] The following will describe the postoperative risk assessment system, device, storage medium and program product of the present invention in conjunction with Figures 1 - 12 Describe the postoperative risk assessment system, device, storage medium and program product of the present invention.

[0055] Figure 1 is a schematic diagram of the modules of the postoperative risk assessment system provided by the present invention. As Figure 1 shown, the system may include the following modules:

[0056] An extraction module for extracting independent risk factors for early graft failure from the postoperative data of a patient;

[0057] A score determination module for determining the scores of each factor in the independent risk factors based on the scoring scale in the nomogram model; the nomogram model is constructed based on the independent risk factors;

[0058] A prediction module for predicting the risk information of early graft failure after coronary artery bypass grafting based on the risk coefficient corresponding to the total score of each factor in the independent risk factors;

[0059] Among them, the independent risk factors are determined by performing least absolute shrinkage and selection operator Lasso regression analysis and multi-factor logistic Logistic regression analysis on the dataset of potential risk factors for cardiovascular diseases respectively, and combining the results of Lasso regression analysis and Logistic regression analysis.

[0060] In an embodiment of the present invention, the graft may be an internal mammary artery graft; early graft failure after coronary artery bypass grafting refers to the degree of graft stenosis ≥ 50% (including complete occlusion) evaluated by CCTA (Coronary Computed Tomography Angiography).

[0061] The independent risk factors include non-left main stenosis, complications (also known as in-hospital complications), quantitative flow ratio and pulsatility index; the weight ranking of each factor in the independent risk factors is in turn quantitative flow ratio, complications, pulsatility index and non-left main stenosis. Among them, the quantitative flow ratio is greater than 0.8 (i.e., QFR > 0.80), and the pulsatility index is greater than 3.0 (i.e., PI > 3.0).

[0062] A dataset of potential risk factors for cardiovascular diseases is pre-constructed. For example, it can be achieved through the following methods:

[0063] Collect the baseline clinical data of all patients, including demographic characteristics, intraoperative data of CABG, and imaging data (including coronary angiography and postoperative follow-up CCTA images), etc. Among them, coronary angiography was performed using the standard percutaneous femoral or radial artery approach with a 6F or 7F guiding catheter.

[0064] All patients underwent CCTA follow-up at 1 year (12 ± 3 months) after surgery to evaluate the patency of the grafts, and the patients were divided into a non-failing group and a failing group accordingly. Among them, the clinical baseline data of the patients in the non-failing group and the failing group showed that the patients in the early graft failure group had a preoperative QFR > 0.80, a higher proportion of non-left main lesions or in-hospital complications (P < 0.01), refer to Tables 1 to 2.

[0065] Table 1 Comparison of clinical baseline characteristics between the non-failing group and the failing group

[0066]

[0067] Table 2 Comparison of clinical baseline characteristics between the non-failing group and the failing group

[0068]

[0069] In Tables 1 and 2, the failing group was defined as graft vessel stenosis ≥ 50% in diameter (including complete occlusion) at 1 year after surgery; BMI: body mass index; PCI: percutaneous coronary intervention; LVEF: left ventricular ejection fraction; non-left main lesion was defined as a left main stenosis degree < 50%; other concomitant surgeries: including left atrial appendectomy, ventricular aneurysm resection, carotid endarterectomy, and arteriovenous fistula surgery, etc.; the number of anastomotic sites using arterial grafts: including the number of anastomotic sites of the left internal mammary artery, right internal mammary artery, or radial artery graft with the diseased coronary artery; in-hospital complications included infectious symptoms such as postoperative respiratory tract, surgical site, catheter-related bloodstream infection, or infective endocarditis; postoperative bleeding (continuous chest tube drainage ≥ 7 mL / Kg / h for 2 consecutive hours within the first 12 hours after surgery or a total of ≥ 84 mL / kg within the first 24 hours); newly diagnosed heart failure by echocardiogram and laboratory tests after surgery; reoperation after surgery; : t-test; #: chi-square test; : Mann-Whitney U test; ▲: Fisher's exact probability method.

[0070] Tables 1 and 2 include potential risk factors for cardiovascular diseases, such as traditional cardiovascular risk factors, including age, gender, BMI, ejection fraction, diabetes, hypertension, hyperlipidemia, peripheral vascular disease, smoking, history of previous myocardial infarction and PCI, and the time from preoperative angiography to surgery; potential intraoperative risk factors, including non-left main stenosis, QFR>0.80, graft material, and PI; postoperative risk factors, including postoperative complications (such as infection, bleeding, heart failure, and reoperation) and length of hospital stay, etc.

[0071] In one embodiment, the independent risk factors are determined in the following manner: performing Lasso regression analysis on the data set to obtain the first type of risk factors for early graft failure; performing Logistic regression analysis on the data set to obtain the second type of risk factors for early graft failure; ranking the importance of the first type of risk factors and the second type of risk factors; and selecting the common risk factors in the ranking results of the first type of risk factors and the ranking results of the second type of risk factors as the independent risk factors.

[0072] In order to screen out the factors that have the most significant impact on early graft failure and improve the prediction ability of the model, the embodiment of the present invention uses Lasso regression to introduce a regularization parameter ( λ ) to reduce overfitting, as Figures 2 to 3 shown. The included factors include all potential risk factors for cardiovascular diseases in Tables 1 and 2. The model adopts a binomial distribution, and the optimal parameters are determined through cross-validation (such as ten-fold cross-validation, k-fold cross-validation, stratified k-fold cross-validation) (for example, when λ = 0.018, the prediction error of the model is the lowest). At this time, the variables included in the model are: non-left main stenosis, hypertension, hyperlipidemia, complications (i.e., in-hospital complications), QFR>0.80, and PI>3.0. An importance ranking analysis is carried out on the relevant factors, and the results show that QFR>0.80 (weighting coefficient 0.152) and in-hospital complications (weighting coefficient 0.114) are the two most important risk factors, followed by PI>3.0 (weighting coefficient 0.063) and non-left main stenosis (weighting coefficient 0.038), indicating that these factors may have an important impact on the prediction of early graft failure, as Figure 4 shown.

[0073] In one embodiment, the first type of risk factors is determined in the following manner: constructing the objective function of the Lasso regression model; determining the optimal regularization parameter of the objective function based on the cross-validation method; training the objective function after determining the optimal regularization parameter using the data set to determine the regression coefficient of each variable; the variable is the risk factor in the data set; and determining the first type of risk factors based on the regression coefficient of each variable.

[0074] For example, the objective function of the Lasso regression model is set, and its general form is:

[0075] ;

[0076] where, y i is the early decline situation (target variable) of the blood vessel bypass graft of the i th sample, x ij is the i th potential risk factor of the j th sample, β j is the corresponding regression coefficient, n is the number of samples, p is the number of potential risk factors, λ is the regularization parameter.

[0077] Since the model adopts the binomial distribution, the output of the model is a probability value, representing the probability of early decline of the blood vessel bypass graft, which matches the binary classification nature of the target variable.

[0078] Furthermore, through cross-validation, the optimal regularization parameter of the objective function is determined. For example, taking the ten-fold cross-validation as an example for analysis and explanation, the data set is divided into ten parts. For each λ value: Nine of them are used as the training set. Based on the principle of minimizing the objective function on the training set, an optimization algorithm (such as the coordinate descent method) is used to update the regression coefficients, so that the model can fit the training data and be constrained by regularization. The remaining one is used as the test set, and the prediction performance indicators of the model on the test set are calculated, such as accuracy, recall rate, F1-score or log-loss, etc. Different λ values are traversed, and the above process is repeated. The one that makes the average prediction performance of the model on the test set of the ten-fold cross-validation the best

[0079] is selected as the optimal parameter, that is, the optimal regularization parameter. λ After determining the optimal

[0080] Optionally, the variables can be sorted according to the absolute value of the regression coefficients of the variables finally retained in the Lasso regression model. The larger the absolute value of the regression coefficient, the more important the role of the variable in the model.

[0081] To further control confounding factors, a Logistic regression model was constructed to control for: 1) traditional cardiovascular risk factors, including age, sex, BMI, ejection fraction, diabetes, hypertension, hyperlipidemia, peripheral vascular disease, smoking, history of previous myocardial infarction and PCI, and the time from preoperative angiography to surgery; 2) potential intraoperative risk factors, including non-left main stenosis, QFR>0.80, graft material, and PI; 3) postoperative risk factors, including postoperative complications (such as infection, bleeding, heart failure, and reoperation) and length of hospital stay, as shown in Table 3 and Figure 5 as follows.

[0082] Table 3 Logistic regression analysis of influencing factors for vascular graft failure 1 year after CABG

[0083]

[0084] In Table 3, PI: pulsatility index; QFR: quantitative flow ratio; PCI: percutaneous coronary intervention; P values of potential cardiovascular risk factors and intraoperative related factors not listed were all >0.05, and those with P<0.2 in univariate Logistic regression analysis were included in the multivariate analysis.

[0085] Multivariate logistic regression analysis was performed on the dataset to obtain the second type of risk factors for early graft failure. Among them, the second type of risk factors can include non-left main stenosis, complications, QFR>0.80, and PI>3.0. Optionally, the second type of risk factors can also include risk factors such as hypertension and hyperlipidemia.

[0086] In one embodiment, the second type of risk factors is determined by the following method: constructing a Logistic regression model; based on the Logistic regression model, performing univariate analysis on the dataset to screen out candidate risk factors; performing multivariate analysis on the candidate risk factors to obtain the second type of risk factors.

[0087] For example, the Logistic regression model is used for binary classification problems. In the context of studying early graft failure, it is assumed that the dependent variable Y (whether the graft fails) follows a Bernoulli distribution, and its probability function is:

[0088] ;

[0089] where X =( x1 , x 2 , ..., x p ) is a vector containing all the predictive variables that may affect the early decline of the vascular graft (such as the patient's age, gender, disease indicators, etc.). β = ( β 0 , β 1 , ..., β p ) is the vector of regression coefficients.

[0090] During the univariate analysis process, each risk factor was separately subjected to a simple Logistic regression analysis with the early decline of the vascular graft, and the odds ratio (OddsRatio, OR) of each factor, its confidence interval, p - value ( P value), β value, S.E, and Z value and other statistics were calculated. Among them, the odds ratio represents the relative probability change of the occurrence of the event (early decline of the vascular graft) when this factor increases by one unit (for continuous variables) or changes from one category to another (for categorical variables) while other factors remain unchanged. Then, based on the univariate analysis results, candidate risk factors were screened out. For example, in Table 3, factors with P value < 0.2 (i.e., candidate risk factors) were selected for subsequent multivariate analysis.

[0091] During the multivariate analysis process, the candidate risk factors screened out in the univariate analysis were simultaneously incorporated into the multivariate Logistic regression model. In the multivariate model, the model takes into account the interactions and combined effects among these factors, thereby more accurately evaluating the independent contribution of each factor to the early decline of the vascular graft. For example, the Wald statistic or likelihood ratio statistic of each factor was calculated, and by comparing with the corresponding distribution (such as the chi-square distribution), p - value ( P value) was obtained. p - value ( P value) is used to test the significance of this factor in the model. If p - value ( P value) < 0.05, then this factor is considered to significantly affect the early decline of the vascular graft statistically. For example, in Table 3, p - value ( PValue) < 0.05 included non-left main stem stenosis, complications, QFR > 0.80, and PI > 3.0. The results of the multivariate analysis in Table 3 showed that QFR > 0.80 (OR = 5.57, 95% CI: 2.98 - 10.41) and in-hospital complications (OR = 4.02, 95% CI: 1.59 - 10.19) were strong predictors of the risk of decline. The results of the Logistic regression analysis were consistent with those of the Lasso regression analysis, indicating the importance of hemodynamics and complication management in reducing the risk of graft decline. In addition, the significance of non-left main stem stenosis in both models suggested that the risk of graft stenosis in such patients might increase, which needed to attract clinical attention.

[0092] Based on the results of the Logistic regression analysis, the importance ranking of the first type of risk factors was QFR > 0.80, complications, PI > 3.0, non-left main stem stenosis, hypertension, and hyperlipidemia in turn. Based on the results of the Lasso regression analysis, the importance ranking of the second type of risk factors was QFR > 0.80, complications, PI > 3.0, and non-left main stem stenosis in turn. Then, multiple risk factors with higher rankings were selected, such as selecting the top 4 risk factors; further, the common risk factors were extracted from the selected first type of risk factors and the second type of risk factors as independent risk factors, that is, QFR > 0.80, complications, PI > 3.0, and non-left main stem stenosis were used as independent risk factors.

[0093] It can be understood that the purpose of selecting the common risk factors of the two regression analyses is to find the factors that are considered to have a significant impact on the early decline of the graft under different analysis models. These factors have higher reliability and importance because they have been screened by two different analysis ideas and model mechanisms, and can provide more powerful evidence support for clinical decision-making, risk prediction, etc. At the same time, these common risk factors have shown a significant impact on the early decline of the graft in both different statistical analysis models, indicating that they may play a core role in the occurrence process of the early decline of the graft. Compared with the factors screened only in one model, the common risk factors are more likely to be the real and stable risk factors, and their relationship with the early decline of the graft is closer and more reliable. In subsequent research and practice, in-depth research, monitoring, or intervention on these common risk factors can more effectively reduce the risk of early decline of the graft or more accurately predict the occurrence of early decline of the graft.

[0094] Furthermore, based on the independent risk factors, a nomogram model was constructed, as Figure 6 shown. Among them, the nomogram model was used to predict the risk information of early decline of the graft after coronary artery bypass grafting.

[0095] After determining the independent risk factors and the nomogram model, the independent risk factors for early saphenous vein graft failure are extracted from the postoperative data of the patients. Based on the scoring scale in the nomogram model, the scores of each factor in the independent risk factors are determined. Finally, based on the risk coefficient corresponding to the total score of each factor in the independent risk factors, the risk information of early saphenous vein graft failure after coronary artery bypass grafting is predicted.

[0096] For example, in Figure 6 , based on the scoring scale, the scores of each factor in the independent risk factors can be determined. For example, the score for PI > 3.0 is 55, the score for QFR > 0.80 is 98, the score for complications is 85, and the score for non-left main stenosis is 69. Assuming that PI > 3.0 does not meet the condition, the score is 0, and the total score of all risk factors is 0 + 98 + 85 + 69 = 252. The risk coefficient corresponding to this total score is 252 / 350 = 0.72. At this time, it can be predicted that the risk of early saphenous vein graft failure after coronary artery bypass grafting is relatively high.

[0097] Compared with the old model that included factors such as smoking, hypertension, LAD stenosis < 75%, and coronary flow, the nomogram model provided in the embodiment of the present invention includes four new key variables: non-left main stenosis, QFR > 0.80, PI > 3.0, and complications. The AUC of the new model is 0.758 (95% CI: 0.694~0.820), which is significantly higher than that of the old model (AUC = 0.632, 95% CI: 0.561~0.694) and the new-old hybrid model (AUC = 0.672, 95% CI: 0.577~0.758). The results show that the new model shows higher predictive ability for early saphenous vein graft failure, especially in the identification and early intervention of high-risk patients, and has important application value, such as Figure 7 shown.

[0098] Optionally, a binary Logistic regression is used to evaluate the model performance. The data is randomly split into a training set and a test set, and the test ratio is 0.70. The ROC (Receiver Operating Characteristic Curve) curves and calibration curves of the training set and the test set both show that the actual performance of the prediction model is highly consistent with the ideal performance, indicating that the model has high clinical application value and can provide effective support for clinical decision-making, such as Figures 8 to 10As shown below. Further calculation of the net reclassification index (NRI) and integrated discrimination improvement (IDI) results showed that the classification accuracy of the new model increased by 27.2% (NRI = 0.272, 95% CI: 0.18 - 0.37; IDI = 0.109, 95% CI: 0.059 - 0.158, P < 0.05).

[0099] The postoperative risk assessment system provided by the embodiments of the present invention includes an extraction module for extracting independent risk factors for early saphenous vein graft failure from the postoperative data of patients; the independent risk factors include non-left main stem stenosis, complications, fractional flow reserve, and pulsatility index; a score determination module for determining the scores of the factors in the independent risk factors based on the scoring scale in the nomogram model; the nomogram model is constructed based on the independent risk factors; a prediction module for predicting the risk information of early saphenous vein graft failure after coronary artery bypass grafting based on the risk coefficients corresponding to the total scores of the factors in the independent risk factors; wherein the independent risk factors are determined by performing least absolute shrinkage and selection operator (Lasso) regression analysis and multivariate logistic regression analysis on the dataset of potential risk factors for cardiovascular diseases respectively, and combining the results of the Lasso regression analysis and the logistic regression analysis. The present invention constructs a prediction model for early saphenous vein graft failure with risk factors such as QFR > 0.80, PI > 3.0, complications, and non-left main stem stenosis as the main body based on the results of the Lasso regression analysis and the multivariate logistic regression analysis, thereby improving the accuracy of the risk assessment of early saphenous vein graft failure in postoperative patients.

[0100] Based on the above embodiments, the prediction module is further configured to: select target risk factors based on the ranking results of the weight coefficients of the first type of risk factors and the second type of risk factors; wherein the target risk factors include independent risk factors; determine the scores of the factors in the target risk factors based on the scoring scale in the nomogram model constructed based on the target risk factors; and predict the risk information of early saphenous vein graft failure after coronary artery bypass grafting based on the risk coefficients corresponding to the total scores of the factors in the target risk factors.

[0101] For example, based on the results of Logistic regression analysis, the weight coefficient ranking of the first type of risk factors is QFR > 0.80, complications, PI > 3.0, non-left main stem stenosis, hypertension, and hyperlipidemia in sequence. Based on the results of Lasso regression analysis, the weight coefficient ranking of the second type of risk factors is QFR > 0.80, complications, PI > 3.0, and non-left main stem stenosis in sequence. Based on this, QFR > 0.80, complications, PI > 3.0, non-left main stem stenosis, hypertension, and hyperlipidemia can be selected as target risk factors. Then, based on the scoring scale in the nomogram model constructed based on the target risk factors, the scores of each factor in the target risk factors are determined. Finally, based on the risk coefficients corresponding to the total scores of each factor in the target risk factors, the risk information of early graft failure after coronary artery bypass grafting is predicted.

[0102] Based on the results of Lasso regression analysis and multi-factor Logistic regression analysis in the embodiments of the present invention, a prediction model for early graft failure with risk factors such as QFR > 0.80, PI > 3.0, complications, non-left main stem stenosis, hypertension, and hyperlipidemia as the main body is constructed, thereby improving the accuracy of risk assessment of early graft failure in postoperative patients.

[0103] Based on the above embodiments, the prediction module is further configured to: determine the risk score of risk factors based on the independent risk factors and their weight coefficients; divide risk groups based on the risk scores of risk factors; and predict major adverse cardiac and cerebrovascular events based on the risk groups.

[0104] In the embodiments of the present invention, major adverse cardiac and cerebrovascular events refer to MACCE (major adverse cardiac and cerebrovascular events), where MACCE events include all-cause death, myocardial infarction, stroke, and re-revascularization due to angina or other reasons.

[0105] Based on each risk factor and its weight coefficient, the risk score of risk factors is determined. Among them, the risk score of risk factors is to comprehensively evaluate the risk degree of an individual patient having early graft failure. By combining multiple risk factors and their weight coefficients, a quantitative risk score can be obtained. For example, the weight coefficient of QFR > 0.80 is 0.152, the weight coefficient of in-hospital complications is 0.114, the weight coefficient of PI > 3.0 is 0.063, and the weight coefficient of non-left main stem stenosis is 0.038. Then, the risk score of risk factors is:

[0106] Risk score of risk factors = (QFR > 0.80 × 0.152) + (in-hospital complications × 0.114) + (PI > 3.0 × 0.063) + (non-left main stem stenosis × 0.038).

[0107] Optionally, the risk scores of risk factors can also be calculated based on the binary classification results (0 or 1) of each risk factor and their weight coefficients. For example, assume that the binary classification result of QFR>0.80 is 1, the weight coefficient is 0.152, the binary classification result of in-hospital complications is 1, the weight coefficient is 0.114, the binary classification result of PI>3.0 is 0, the weight coefficient is 0.063, and the binary classification result of non-left main stem stenosis is 1, the weight coefficient is 0.038. Then the risk score of risk factors is:

[0108] Risk score of risk factors = 1×0.152 + 1×0.114 + 0×0.063 + 1×0.038.

[0109] Furthermore, based on the risk scores of risk factors, risk groups are divided, such as high-risk group, medium-risk group, and low-risk group; finally, based on the risk groups, major adverse cardiovascular and cerebrovascular events are predicted.

[0110] Specifically, the coronary artery risk score of the follow-up patients is calculated according to the importance weighting coefficient of risk factors. Through the enumeration method, a grouped statistical test is performed on the total risk scores of four risk factors, such as QFR, PI, non-left main stem lesion, and in-hospital complications. For example, the grouped statistical test can be implemented using X-tile software; based on the results of the grouped statistical test, the optimal cut-off value related to MACCE is obtained, so as to maximize the difference in survival patterns. According to the risk scores of risk factors, patients are divided into a low-risk group (score < 0.16), a medium-risk group (score ≥ 0.16 and < 0.25), and a high-risk group (score ≥ 0.25). The Kaplan-Meier survival curve shows that there are significant differences in the cumulative MACCE-free survival rates among different risk groups (Log rank P = 0.001), as Figure 11 shown. In the low-risk group, the incidence of MACCE events is 5.85% (10 / 171), in the medium-risk group is 13.30% (31 / 233), while the incidence in the high-risk group is significantly increased to 17.06% (43 / 209), and the difference between groups is statistically significant (P = 0.003), which further verifies the risk stratification ability of the risk factor scores such as QFR, PI, non-left main stem lesion, and in-hospital complications in clinical practice, and helps to strengthen the follow-up and management of high-risk patients after surgery.

[0111] In the embodiment of the present invention, risk groups are divided through the risk scores of risk factors, and then MACCE events are predicted through the risk groups, which helps to strengthen the follow-up and management of high-risk patients after surgery.

[0112] Figure 12 Illustrates a schematic diagram of the physical structure of an electronic device, such as Figure 12As shown in the figure, the electronic device may include: a processor 1210, a communications interface 1220, a memory 1230, and a communication bus 1240. Among them, the processor 1210, the communications interface 1220, and the memory 1230 communicate with each other through the communication bus 1240. The processor 1210 may call the logical instructions in the memory 1230 to execute the method in the postoperative risk assessment system.

[0113] In addition, when the logical instructions in the above-mentioned memory 1230 are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, may be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs, Read-Only Memories), random access memories (RAMs, Random Access Memories), magnetic disks, or optical discs that can store program codes.

[0114] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method in the above-mentioned postoperative risk assessment system.

[0115] On yet another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it can execute the method in the above-mentioned postoperative risk assessment system.

[0116] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0117] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A postoperative risk assessment system, characterized in that: include: An extraction module is used to extract independent risk factors for early failure of vascular grafts from the postoperative data of patients; the independent risk factors include non-left main stenosis, complications, quantitative blood flow fraction and pulsatility index; the quantitative blood flow fraction is greater than 0.8, and the pulsatility index is greater than 3.0; A score determination module, used to determine the score of each factor in the independent risk factor based on the scoring scale in the nomogram model; The nomogram model is constructed based on the independent risk factors; A prediction module, used to predict the risk information of early failure of vascular graft after coronary artery bypass grafting based on the risk coefficient corresponding to the total score of each factor in the independent risk factors; The independent risk factors are determined by performing the least absolute value shrinkage and selection operator Lasso regression analysis and multivariate logistic regression analysis on the data set of potential risk factors for cardiovascular disease, and combining the results of Lasso regression analysis and logistic regression analysis; The independent risk factors were determined by: Lasso regression analysis was performed on the data set to obtain the first risk factor for early failure of vascular grafts; Logistic regression analysis was performed on the data set to obtain the second type of risk factors for early failure of vascular grafts; rank the first-category risk factors and the second-category risk factors in terms of importance; The common risk factors in the ranking results of the first category of risk factors and the ranking results of the second category of risk factors are selected as the independent risk factors.

2. The postoperative risk assessment system according to claim 1, characterized in that: The prediction module is further used for: Based on the ranking results of the weight coefficients of the first and second risk factors, a target risk factor is selected; wherein the target risk factor includes the independent risk factor; Determine the score of each factor in the target risk factor based on the scoring scale in the nomogram model constructed for the target risk factor; Based on the risk coefficient corresponding to the total score of each factor in the target risk factors, the risk information of early failure of the vascular graft after coronary artery bypass grafting is predicted.

3. The postoperative risk assessment system according to claim 1 or 2, characterized in that: The first category of risk factors is determined by: Construct the objective function of the Lasso regression model; Determining the optimal regularization parameter of the objective function based on a cross-validation method; Using the data set, training the objective function after determining the optimal regularization parameter to determine the regression coefficient of each variable; The variables are risk factors in the data set; The first category of risk factors is determined based on the regression coefficient of each of the variables.

4. The postoperative risk assessment system according to claim 1 or 2, characterized in that: The second category of risk factors is determined by: Build a Logistic regression model; Based on the Logistic regression model, a univariate analysis is performed on the data set to screen out candidate risk factors; Perform a multifactor analysis on the candidate risk factors to obtain the second type of risk factors.

5. The postoperative risk assessment system according to claim 1, characterized in that: The prediction module is further used for: Determining a risk factor risk score based on the independent risk factors and their weight coefficients; Dividing risk groups based on the risk scores of the risk factors; Based on the risk groups, major adverse cardiovascular and cerebrovascular events are predicted.

6. The postoperative risk assessment system according to claim 1, characterized in that: The weight ranking of each factor in the independent risk factors is quantitative blood flow fraction, complications, pulsatility index and non-left main coronary artery stenosis.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the method in the postoperative risk assessment system according to any one of claims 1 to 6 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method in the postoperative risk assessment system according to any one of claims 1 to 6 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method in the postoperative risk assessment system according to any one of claims 1 to 6 is implemented.

Images