System for predicting live delivery probability of uterus mediastinum patient

Abstract

The invention provides a system for predicting the live birth probability of a uterus mediastinum patient, and the system comprises an input module which is used for inputting the information of the uterus mediastinum patient, enabling the information to be related to the pregnancy outcome of the uterus mediastinum patient, and dividing the information into a training set and a verification set; the model construction module is used for screening candidate prediction indexes from the information by using a single-factor binary logistic regression model in the training set, screening a final prediction index from the candidate prediction indexes by using a multi-factor binary logistic regression model, and constructing the multi-factor binary logistic regression model based on the final prediction index; and the prediction module is used for calculating an optimal cut-off value by using an ROC curve, calculating the live delivery probability of the uterine mediastinum patient by using the constructed multi-factor binary Logistic regression model in the verification set, and predicting whether the uterine mediastinum patient can live delivery or not according to the optimal cut-off value. The method is high in prediction accuracy, simple in user operation and easy to accept.

Description

Technical Field

[0001] The present invention relates to the technical field of data evaluation, and in particular to a system for predicting the probability of live birth in patients with uterine septate. Background Art

[0002] Congenital uterine malformation is one of the important causes of female infertility, among which uterine septate is the most common. Uterine septate leads to low fertility and recurrent miscarriage in women. However, the actual incidence of these adverse events is seriously underestimated, because a large proportion of women with uterine septate are missed due to the lack of obvious symptoms in the early stage.

[0003] The main treatment for uterine septum is hysteroscopic septectomy, but there is controversy in current clinical practice as to whether surgery can improve the reproductive outcomes of patients with uterine septum. Some studies have shown that the reproductive outcomes of patients with uterine septum who received surgical treatment were significantly improved compared with expectant treatment, but some studies have found that the pregnancy outcomes of patients after surgery did not improve. At the same time, surgical operations increase the risk of complications such as uterine bleeding, endometrial damage, and intrauterine adhesions. The above evidence shows that not all patients can benefit from surgery, but there is currently no method to predict the risk of pregnancy outcomes.

[0004] Although previous studies have found that factors such as age and primary infertility may be related to the pregnancy outcomes of patients with uterine septate, the effects of the diaphragm and uterus itself on pregnancy have been neglected. In addition, no model has been developed to predict the outcome of pregnancy with uterine septate. Therefore, how to develop a comprehensive, highly accurate, simple to operate, and easy-to-accept system to predict the outcome of pregnancy with uterine septate is an urgent problem to be solved. Summary of the invention

[0005] The present invention provides a system for predicting the probability of live birth in patients with uterine septate, so as to solve the defect that the risk of pregnancy outcome cannot be predicted in the prior art, and to develop a comprehensive, highly accurate, simple to operate and easy to accept system for predicting pregnancy outcome of uterine septate.

[0006] The present invention provides a system for predicting the probability of live birth in patients with uterine septate, comprising:

[0007] An input module, used for inputting information of patients with uterine septate, the information being related to the pregnancy outcome of the patients with uterine septate, and dividing the information into a training set and a validation set;

[0008] A model building module, used to use a univariate binary Logistic regression model in the training set to screen candidate prediction indicators from the information, use a multivariate binary Logistic regression model to screen final prediction indicators from the candidate prediction indicators, and build a multivariate binary Logistic regression model based on the final prediction indicators;

[0009] A prediction module is used to calculate the optimal cutoff value of the constructed multi-factor binary Logistic regression model using the ROC curve, calculate the probability of live birth of the patient with uterine septate using the constructed multi-factor binary Logistic regression model in the validation set, and predict whether the patient with uterine septate can give birth live based on the optimal cutoff value.

[0010] According to a system for predicting the probability of live birth in patients with uterine septate provided by the present invention, the information of the patients with uterine septate includes one or more of the age of consultation, main complaint, whether menstruation is regular, whether dysmenorrhea occurs, number of pregnancies, mode of pregnancy, number of parities, mode of delivery, number of miscarriages, type of miscarriage, whether endometriosis is present, whether adenomyoma is present, whether endometrial polyps are present, type of uterine septate, morphology of the uterus, length of the septum, width of the septum, length of the uterus, width of the uterus, and whether other malformations are present.

[0011] According to a system for predicting the probability of live birth in patients with uterine septate provided by the present invention, the pregnancy mode includes natural pregnancy and assisted pregnancy;

[0012] The delivery methods include natural delivery and cesarean section;

[0013] The types of miscarriage include early spontaneous abortion and late spontaneous abortion;

[0014] The types of uterine septum include complete and incomplete;

[0015] The morphology of the uterus includes normal and abnormal;

[0016] The presence or absence of other malformations includes presence or absence of vaginal septum, vaginal septum, and renal agenesis.

[0017] According to a system for predicting the probability of live birth in patients with uterine septate provided by the present invention, the formula of the univariate binary logistic regression model is:

[0018]

[0019] Among them, x is a single information, p is the correlation between each information x and the pregnancy outcome, β0 and β are model parameters;

[0020] The model parameters are solved by the maximum likelihood method, and the formula is as follows:

[0021]

[0022] Wherein, L is the likelihood function, n is the number of information, p(xi) is the correlation between the i-th information and the pregnancy outcome, and yi is 1 when the patient with uterine septate to which the i-th information belongs has a live birth, otherwise it is 0.

[0023] According to a system for predicting the probability of live birth in patients with uterine septate provided by the present invention, the formula of the multi-factor binary logistic regression model is:

[0024]

[0025] Among them, x j is the jth candidate prediction indicator, β j is the model parameter corresponding to the jth candidate prediction indicator.

[0026] According to a system for predicting the probability of live birth in patients with uterine septate provided by the present invention, the model building module uses a generalized linear model to perform linear fitting on the multi-factor binary Logistic regression model through the following formula to build the multi-factor binary Logistic regression model:

[0027]

[0028] Wherein, m is the number of the final prediction indicators.

[0029] According to a system for predicting the probability of live birth in patients with uterine septum provided by the present invention, the final prediction index includes the age of the patient with uterine septum, the number of pregnancies, the width of the septum in the ultrasound report, and the width of the uterus.

[0030] According to a system for predicting the probability of live birth in patients with uterine septate provided by the present invention, the formula of the multi-factor binary logistic regression model is:

[0031] Logit(P|y=live birth)=β0-β1*age+β2*gravidity-β3*W1 / W2

[0032] Among them, Logit(p|y=live birth) represents the probability of live birth of the patient with uterine septate, β0, β1, β2 and β3 represent model parameters, age, gravidity, W1 and W2 represent the age of the patient with uterine septate, the number of pregnancies, the width of the septum in the ultrasound report and the width of the uterus, respectively.

[0033] A system for predicting the probability of live birth in patients with uterine septate provided by the present invention further includes a visualization module for:

[0034] Nomogram was used to visualize the constructed multi-factor binary logistic regression model.

[0035] A system for predicting the probability of live birth in patients with uterine septate provided by the present invention further includes an evaluation module for:

[0036] Draw the ROC curve, calibration curve and clinical decision curve of the constructed multi-factor binary logistic regression model;

[0037] The performance of the constructed multi-factor binary logistic regression model was evaluated based on the ROC curve, calibration curve and clinical decision curve.

[0038] The system for predicting the probability of live birth of patients with uterine septate provided by the present invention obtains information of patients with uterine septate through an input module, obtains important indicators through two indicator screenings through a model construction module to construct a multi-factor binary Logistic regression model, and calculates the probability of live birth of patients with uterine septate by using the constructed multi-factor binary Logistic regression model through a prediction module. The prediction accuracy is high, and the user operation is simple and easy to accept. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 It is a schematic diagram of the structure of a system for predicting the probability of live birth in patients with uterine septate provided by the present invention;

[0041] Figure 2 It is a schematic diagram of a flow chart of establishing a pregnancy outcome prediction model for patients with uterine septum in the system for predicting the probability of live birth in patients with uterine septum provided by the present invention;

[0042] Figure 3 are the ROC curve and AUC value of the model in the training set and the validation set in the system for predicting the probability of live birth in patients with uterine septum provided by the present invention;

[0043] Figure 4 is a schematic diagram of a method for using Nomogram in a system for predicting the probability of live birth in patients with uterine septate provided by the present invention;

[0044] Figure 5 It is the clinical decision curve of the model in the training set and the validation set in the system for predicting the live birth probability of patients with uterine septum provided by the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] Combine the following Figure 1 A system for predicting the probability of live birth in a patient with uterine septate according to the present invention is described, comprising:

[0047] An input module, used for inputting information of patients with uterine septate, the information being related to the pregnancy outcome of the patients with uterine septate, and dividing the information into a training set and a validation set;

[0048] The input module provides an interactive interface for users, through which users input the medical records of patients with uterine septate of childbearing age to extract information that may be related to pregnancy outcomes and standardize the collected data.

[0049] Assign values ​​to character variables and convert the extracted data into numerical and vectorized data for subsequent analysis. Binary variables can be assigned values ​​of "0" or "1", such as whether menstruation is regular (irregular = 0, regular = 1); measurement data are processed as continuous variables, such as age, number of pregnancies, length and width of the mediastinum, length and width of the uterus, etc.

[0050] Clean the collected data, remove duplicate data, interpolate or delete missing values, remove outliers, etc., to ensure that there are no null values ​​in the data set for analysis. In addition, the data format can also be converted (for example, some numerical variables need to be converted into categorical variables).

[0051] In Rstudio, the sample function is used to randomly divide the data set into a training set and a validation set. The implementation of the sample function is as follows: sample(X, size, replace = FALSE, ro = NULL). The data set can be divided into a training set and a validation set in a ratio of 3:2.

[0052] The training set and validation set can be compared in SPSS software to compare the differences between the two data sets and verify that the grouping is random and uniform. For continuous variables that conform to normal distribution, the t test is used, and for those that do not conform to normal distribution, the non-parametric test is used. The chi-square test is used for count data, and P < 0.05 is considered to be statistically significant.

[0053] A model building module, used to use a univariate binary Logistic regression model in the training set to screen candidate prediction indicators from the information, use a multivariate binary Logistic regression model to screen final prediction indicators from the candidate prediction indicators, and build a multivariate binary Logistic regression model based on the final prediction indicators;

[0054] In the training set, a univariate binary logistic regression model was used to evaluate the correlation between each indicator and pregnancy outcome (live birth). Live birth or not was used as the dependent variable (0 = failure to live birth, 1 = live birth), and other variables were used as covariates, and univariate logistic regression analysis was performed on each variable.

[0055] P<0.05 was considered statistically significant, indicating that this indicator has a certain predictive value for the pregnancy outcome of patients with uterine septate. An OR value greater than 1 indicates that this indicator is a favorable factor for live birth, and an OR value less than 1 indicates that this indicator is a harmful factor for live birth. Candidate predictive indicators that may be related to the pregnancy outcome of patients with uterine septate were preliminarily screened from the information of patients with uterine septate and further included in the multivariate logistic regression analysis.

[0056] The variables with significant difference in univariate logistic regression analysis were input as covariates. The dependent variable was still live birth. Multivariate logistic regression analysis was performed to further eliminate the influence of mixed factors and exclude variables with no statistical significance. The important indicators that can predict the pregnancy outcome of patients with uterine septate were screened out, which were the final predictive indicators. Solve the model parameters β0, β1...β j , j is the number of the final prediction indicators screened out, and a multi-factor binary Logistic regression model was established.

[0057] A prediction module is used to calculate the optimal cutoff value of the constructed multi-factor binary Logistic regression model using a ROC (Receiver Operating Characteristic) curve, calculate the probability of live birth of the patient with uterine septate using the constructed multi-factor binary Logistic regression model in the validation set, and predict whether the patient with uterine septate can give birth live based on the optimal cutoff value.

[0058] Flow chart of the establishment of a prediction model for pregnancy outcomes in patients with uterine septate Figure 2 As shown. Figure 3 As shown in (a), the ROC curve of the training set is drawn, and the AUC (Area Under Curve, the area under the ROC curve) value is calculated to be 0.785, indicating that the accuracy of the model is 78.5%. The point in the upper left corner of the ROC curve can be used as the optimal cutoff value.

[0059] In the validation set, a multivariate binary logistic regression model was used to calculate the probability of live birth for each patient with uterine septate, and the cutoff value was used to predict whether the patient with uterine septate could have a live birth (greater than the cutoff value = 1, live birth; less than the cutoff value = 0, no live birth).

[0060] like Figure 3 As shown in (b), the ROC curve of the validation cohort was drawn, and the area under the curve was calculated to be 0.813, indicating that the accuracy of the model prediction was 81.3%. This shows that the constructed multi-factor binary logistic regression model has high accuracy, and the area under the curve (AUC) in the validation cohort can reach 0.813, with stable performance.

[0061] In this embodiment, the information of patients with uterine septate is obtained through an input module, and a multi-factor binary Logistic regression model is constructed by obtaining important indicators through two indicator screenings through a model construction module. The probability of live birth of patients with uterine septate is calculated by using the constructed multi-factor binary Logistic regression model through a prediction module. The prediction accuracy is high, and the user operation is simple and easy to accept.

[0062] On the basis of the above embodiment, the information of the patient with uterine septate described in this embodiment includes one or more of the following: age at consultation, main complaint, whether menstruation is regular, whether dysmenorrhea occurs, number of pregnancies, mode of pregnancy, number of parities, mode of delivery, number of miscarriages, type of miscarriage, whether endometriosis is present, whether adenomyoma is present, whether endometrial polyps are present, type of uterine septate, uterine morphology, length of septum, width of septum, length of uterus, width of uterus, and whether other malformations are present;

[0063] Based on the above embodiment, the pregnancy methods described in this embodiment include natural pregnancy and assisted pregnancy;

[0064] The delivery methods include natural delivery and cesarean section;

[0065] The types of miscarriage include early spontaneous abortion and late spontaneous abortion;

[0066] The types of uterine septum include complete and incomplete;

[0067] The morphology of the uterus includes normal and abnormal;

[0068] The presence or absence of other malformations includes presence or absence of vaginal septum, vaginal septum, and renal agenesis.

[0069] Based on the above embodiment, the formula of the univariate binary logistic regression model in this embodiment is:

[0070]

[0071] Among them, x is a single information, p is the correlation between each information x and the pregnancy outcome, β0 and β are model parameters;

[0072] The model parameters are solved by the maximum likelihood method, and the formula is as follows:

[0073]

[0074] Wherein, L is the likelihood function, n is the number of information, p(xi) is the correlation between the i-th information and the pregnancy outcome, and yi is 1 when the patient with uterine septate to which the i-th information belongs has a live birth, otherwise it is 0.

[0075] Based on the above embodiment, the formula of the multi-factor binary logistic regression model in this embodiment is:

[0076]

[0077] Among them, x j is the jth candidate prediction indicator, β j is the model parameter corresponding to the jth candidate prediction indicator.

[0078] On the basis of the above embodiment, the model building module in this embodiment uses the generalized linear model to perform linear fitting on the multi-factor binary Logistic regression model through the following formula to build the multi-factor binary Logistic regression model:

[0079]

[0080] Wherein, m is the number of the final prediction indicators.

[0081] On the basis of the above-mentioned embodiment, the final prediction index in this embodiment includes the age of the patient with uterine septum, the number of pregnancies, the width of the septum in the ultrasound report, and the width of the uterus.

[0082] For the constructed multivariate binary logistic regression model, the probability of live birth in patients with uterine septate can be predicted by simply inputting the age, number of pregnancies, width of the septum in the ultrasound report, and width of the uterus. These indicators are easy to obtain clinically, and there is no need to add additional test items for patients, which has strong clinical practicality.

[0083] On the basis of the above embodiment, the formula of the multi-factor binary logistic regression model constructed in this embodiment is:

[0084] Logit(P|y=live birth)=β0-β1*age+β2*gravidity-β3*W1 / W2

[0085] Among them, Logit(P|y=live birth) represents the probability of live birth of the patient with uterine septum, β0, β1, β2 and β3 represent model parameters, age, gravidity, W1 and W2 represent the age of the patient with uterine septum, the number of pregnancies, the width of the septum in the ultrasound report and the width of the uterus, respectively.

[0086] For example, through linear fitting, β0, β1, β2 and β3 are obtained to be 0.68, 0.083, 2.087 and 1.743 respectively. Therefore, the corresponding multi-factor binary logistic regression model can be Logit (P|y = live birth) = 0.68-0.083*age+2.087*gravidity-1.743*W1 / W2. Among them, W1 / W2 represents the ratio of the width of the uterine septum to the width of the uterus.

[0087] Based on the above embodiment, this embodiment further includes a visualization module for:

[0088] Nomogram was used to visualize the constructed multi-factor binary logistic regression model.

[0089] The model is visualized using Nomogram, which is easy to operate and understand. From the Nomogram, you can intuitively read the scores corresponding to each indicator, add up the scores to get the total score, and then you can get the probability of live birth at the corresponding position below the total score.

[0090] The visualization of the Nomogram is as follows Figure 4 For example, a patient with uterine septum is 26 years old and scored 67.5 points; the ratio of the uterine septum to the width of the uterus is 0.4, which is 77 points; the number of pregnancies is 0, which is 0 points; the total score is 144.5 points. Draw a vertical line downward from the total score and intersect with the line segment representing the probability of live birth. The probability value at the intersection is the patient's probability of live birth, which is about 0.12, indicating that the patient has a low probability of live birth without treatment.

[0091] This embodiment visualizes the complex regression equation with the help of a nomogram, making the prediction model more intuitive and more operable, and facilitating the evaluation of patients.

[0092] Based on the above embodiment, this embodiment further includes an evaluation module for:

[0093] Draw the ROC curve, calibration curve and clinical decision curve of the constructed multi-factor binary logistic regression model;

[0094] The performance of the constructed multi-factor binary logistic regression model was evaluated based on the ROC curve, calibration curve and clinical decision curve.

[0095] The Hosmer-Lemeshow goodness of fit test can be used to evaluate the consistency between the predicted probability and the actual probability of the multi-factor binary logistic regression model, and the calibration curve of the model can be drawn to visualize the results.

[0096] The samples are sorted according to the prediction results, and the samples are predicted one by one in this order as positive examples. The true positive rate (TPR) and false positive rate (FPR) are calculated each time, and they are plotted as the horizontal and vertical coordinates to obtain the ROC curve to evaluate the accuracy of the model.

[0097] The true positive rate and false positive rate are calculated as follows:

[0098]

[0099]

[0100] Among them, TP (True Positive) is the number of true positive cases, FP (False Positive) is the number of false positive cases, TN (True Negative) is the number of true negative cases, and FN (False Negative) is the number of false positive cases.

[0101] Calculate the net benefit of the model, net benefit of the model = (TPR × Treatment Benefit) – (FPR × Treatment Harm). Draw the decision curve with net benefit as the ordinate and the sensitivity and specificity of the model as the abscissa. The clinical decision curve is a curve with the sensitivity and specificity of the prediction model as the abscissa and the benefit as the ordinate. Figure 5 (a) and Figure 5 (b) Clinical decision curves of the model in the training set and validation set, respectively.

[0102] In the clinical decision curve, the curves representing the net benefit of the model are higher than "All" and "None", indicating that the clinical practical value of the model is high.

[0103] The calibration curve and clinical decision curve in the training set were drawn to evaluate the model's fit and clinical application value. In the calibration curve, the better the calibration curve of the model fits the ideal curve, the better the consistency between the model's predicted probability and observed probability, and the higher the accuracy; in the clinical decision curve, the curve representing the model's net benefit is higher than "All" and "None", indicating that the model has high clinical practical value.

[0104] The calibration curve and clinical decision curve were used to evaluate the accuracy of the model in the validation set. The results showed that the model had good fitting performance, high accuracy and high clinical application value in the training set.

[0105] 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 it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A system for predicting the probability of live birth in patients with uterine septate, characterized in that: include: An input module, used for inputting information of patients with uterine septate, the information being related to the pregnancy outcome of the patients with uterine septate, and dividing the information into a training set and a validation set; A model building module, used to use a univariate binary Logistic regression model in the training set to screen candidate prediction indicators from the information, use a multivariate binary Logistic regression model to screen final prediction indicators from the candidate prediction indicators, and build a multivariate binary Logistic regression model based on the final prediction indicators; A prediction module is used to calculate the optimal cutoff value of the constructed multi-factor binary Logistic regression model using the ROC curve, calculate the probability of live birth of the patient with uterine septate using the constructed multi-factor binary Logistic regression model in the validation set, and predict whether the patient with uterine septate can give birth live based on the optimal cutoff value.

2. The system for predicting the probability of live birth in patients with uterine septate according to claim 1, characterized in that: The information of the patients with uterine septate includes the age of consultation, main complaint, whether menstruation is regular, whether there is dysmenorrhea, number of pregnancies, mode of pregnancy, number of parities, mode of delivery, number of miscarriages, type of miscarriage, whether there is endometriosis, whether there is adenomyosis, whether there is endometrial polyp, type of uterine septate, morphology of the uterus, length of the septum, width of the septum, length of the uterus, width of the uterus and whether there are one or more of other malformations.

3. The system for predicting the probability of live birth in patients with uterine septate according to claim 2, characterized in that: The pregnancy methods include natural pregnancy and assisted pregnancy; The delivery methods include natural delivery and cesarean section; The types of miscarriage include early spontaneous abortion and late spontaneous abortion; The types of uterine septum include complete and incomplete; The morphology of the uterus includes normal and abnormal; The presence or absence of other malformations includes presence or absence of vaginal septum, vaginal septum, and renal agenesis.

4. The system for predicting the probability of live birth in patients with uterine septate according to claim 1, characterized in that: The formula of the univariate binary logistic regression model is: Where x is a single piece of information, p is the correlation between each piece of information x and the pregnancy outcome, and β0 and β are model parameters; The model parameters are solved by the maximum likelihood method, and the formula is as follows: Wherein, L is the likelihood function, n is the number of information, p(xi) is the correlation between the i-th information and the pregnancy outcome, and yi is 1 when the patient with uterine septate to which the i-th information belongs has a live birth, otherwise it is 0.

5. The system for predicting the probability of live birth in patients with uterine septate according to claim 1, characterized in that: The formula of the multi-factor binary logistic regression model is: Among them, x j is the jth candidate prediction indicator, β j is the model parameter corresponding to the jth candidate prediction indicator.

6. The system for predicting the probability of live birth in patients with uterine septate according to claim 5, characterized in that: The model building module uses the generalized linear model to perform linear fitting on the multi-factor binary logistic regression model through the following formula to build the multi-factor binary logistic regression model: Wherein, m is the number of the final prediction indicators.

7. The system for predicting the probability of live birth in patients with uterine septate according to claim 1, characterized in that: The final predictors included the age of the patient with uterine septate, number of pregnancies, width of the septum in the ultrasound report, and width of the uterus.

8. The system for predicting the probability of live birth in patients with uterine septate according to claim 7, characterized in that: The formula of the constructed multi-factor binary logistic regression model is: Logit(P|y=live birth)=β0-β1*age+β2*gravidity-β3*W1 / W2 Among them, Logit(P|y=live birth) represents the probability of live birth of the patient with uterine septum, β0, β1, β2 and β3 represent model parameters, age, gravidity, W1 and W2 represent the age of the patient with uterine septum, the number of pregnancies, the width of the septum in the ultrasound report and the width of the uterus, respectively.

9. The system for predicting the probability of live birth in patients with uterine septate according to claim 1, characterized in that: Also includes visualization modules for: Nomogram was used to visualize the constructed multi-factor binary logistic regression model.

10. The system for predicting the probability of live birth in patients with uterine septate according to claim 1, characterized in that: Also included are evaluation modules for: Draw the ROC curve, calibration curve and clinical decision curve of the constructed multi-factor binary logistic regression model; The performance of the constructed multi-factor binary logistic regression model was evaluated based on the ROC curve, calibration curve and clinical decision curve.

Images