A method for constructing a diagnostic model of septic cardiomyopathy and a diagnostic system

Abstract

The present invention discloses a method for constructing a diagnostic model for septic cardiomyopathy and a diagnostic system, which relates to the technical field of biomedicine. The method includes: dynamically collecting multimodal ultrasound data and clinical data of septic patients, evaluating the overall and local systolic and diastolic functions of the left and right ventricles of the heart of septic cardiomyopathy patients according to the multimodal ultrasound data, and making a preliminary diagnosis of septic cardiomyopathy in combination with clinical data; using the Rcaret package to divide the multimodal ultrasound data of septic patients into a training set and a validation set, performing least absolute shrinkage and LASSO regression analysis on the training set to determine the preliminary predictors; performing logistics regression analysis on the preliminary predictors to determine the independent predictors; translating the independent predictors into nomoscore through the rms package to obtain a nomogram, visually presenting the independent predictors according to their weights, and quantifying the probability of the occurrence of septic cardiomyopathy. The present invention realizes a comprehensive and dynamic quantitative evaluation of the cardiac function of patients based on multimodal ultrasound parameters.

Description

Technical Field

[0001] The present invention relates to the field of biomedical technologies, and particularly to a method for constructing a sepsis-induced cardiomyopathy diagnosis model and a diagnosis system. Background Art

[0002] Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection and is the leading cause of death in critically ill hospitalized patients. According to statistics, approximately 48.9 million people developed sepsis globally in 2017, of which 11 million died, accounting for 20% of the global death toll. The heart is the main organ affected by sepsis, and the cardiac dysfunction induced by sepsis is called sepsis-induced cardiomyopathy (SIC), which is characterized by global but reversible cardiac dysfunction. 10%-70% of sepsis patients will develop SIC, with a mortality rate as high as 70%-90%, severely deteriorating the prognosis of sepsis. SIC has obvious heterogeneity, with diverse clinical manifestations and unclear pathogenesis, making early diagnosis difficult and lacking specific treatment methods, which is one of the research hotspots in critical care medicine.

[0003] Currently, there is no unified diagnostic standard for SIC, but current research has confirmed that SIC has the following basic characteristics: (1) Acute and reversible, with cardiac function gradually returning to normal 7-10 days after onset; (2) Global and biventricular systolic and / or diastolic dysfunction, accompanied by a decrease in contractility; (3) Left ventricular dilation; (4) Poor response to fluid resuscitation and catecholamine drugs; (5) Excluding myocardial ischemia caused by acute coronary artery stenosis.

[0004] In the 1980s, Parker et al. used ventriculography to study cardiac function and initially recognized sepsis-induced cardiomyopathy, that is, left ventricular dilation accompanied by systolic dysfunction, which recovered in 7-10 days. This study has aroused the attention of many scholars to sepsis-induced cardiomyopathy, and various means such as echocardiography have been used to observe the changes in left ventricular systolic, left ventricular diastolic, and right ventricular functions of sepsis patients. With the in-depth study, people have gradually realized that sepsis-induced cardiomyopathy has a variety of clinical phenotypes, which can be manifested as left ventricular systolic dysfunction (LVSDF), left ventricular diastolic dysfunction (LVDDF), right ventricular dysfunction (RVDF), diffuse ventricular dysfunction, and / or mixed ventricular dysfunction (Mixed). The heterogeneity of the clinical manifestations of sepsis-induced cardiomyopathy and the resulting different types of hemodynamic changes and different clinical outcomes have enabled researchers to recognize the importance of correctly identifying different subtypes of sepsis-induced cardiomyopathy at an early stage for precise treatment of sepsis-induced cardiomyopathy.

[0005] Echocardiography is a fundamental method for evaluating cardiac structure and function. In the early stage, a decrease in left ventricular ejection fraction (LVEF) was used as the standard for diagnosing SIC. However, LVEF reflects the result of the coupling and superposition of left ventricular contraction and left ventricular afterload, and is significantly affected by afterload. In patients with septic shock, there is often a decrease in peripheral vascular resistance. Even if the intrinsic contractility of the left ventricular myocardium has decreased, due to the concurrent decrease in afterload, a normal LVEF may be presented. At the same time, the distribution of LVEF in patients with sepsis shows a "U"-shaped non-linear distribution. Even in patients with a hyperdynamic state, that is, LVEF > 70%, they also have a relatively high short-term mortality rate. Therefore, using LVEF as a single diagnostic criterion for SIC may lead to a delay or missed diagnosis of SIC. This indicates that LVEF is not the best echocardiographic indicator for diagnosing septic cardiomyopathy. Therefore, an index that is not affected by cardiac preload and afterload and reflects the intrinsic contractility of the myocardium is more likely to reflect the myocardial function state in sepsis. Tissue Doppler measures the peak systolic velocity (S') of the mitral annulus at the interventricular septum or the lateral wall of the mitral annulus, which can be used to evaluate the overall function of the left ventricle. This index is less dependent on load conditions than LVEF. However, a meta-analysis confirmed that there was no difference in the S' wave value at the mitral annulus between survivors and non-survivors of sepsis. The reason is that the measurement of the S' wave is limited by angle dependence, and the measurement is a one-way (longitudinal) and one-dimensional (one section) evaluation. It is difficult for critically ill patients with limited body positions to obtain the best examination section image. By measuring global longitudinal strain (GLS) of the ventricle, speckle-tracking echocardiography (STE) provides important insights for the early detection of cardiac dysfunction. The application of speckle-tracking echocardiography in evaluating myocardial function in sepsis is still in the early stage of exploration.

[0006] In summary, currently, specific echocardiographic parameters and their significance for SIC have not been clearly defined. Moreover, due to the extremely complex clinical phenotypes and unclear pathogenesis of septic cardiomyopathy, these situations have led to difficulties in the diagnosis of septic cardiomyopathy. As can be seen from the above, relying solely on a single echocardiographic parameter such as global longitudinal strain (GLS) of the ventricle or left ventricular ejection fraction (LVEF) cannot comprehensively reflect cardiac function, and the combined parameters of multiple indicators are not clear and lack a quantification standard. Therefore, there is an urgent need to construct a protocol that can quantitatively diagnose septic cardiomyopathy by combining multiple echocardiographic parameters. Summary of the Invention

[0007] The purpose of the present invention is to provide a method for constructing a diagnostic model for septic cardiomyopathy and a diagnostic system. By combining multimodal cardiac echocardiography data, it comprehensively and dynamically quantifies and evaluates the cardiac function of septic patients, determines independent predictors for ultrasonic diagnosis of septic myocardium through regression analysis and visualizes them, achieving quantitative diagnosis of septic cardiomyopathy.

[0008] To achieve the above object, the present invention provides the following solution:

[0009] A method for constructing a diagnostic model of septic cardiomyopathy, comprising the following steps:

[0010] S1. Multimodal data collection: Dynamically collect multimodal ultrasound data and clinical data of septic patients;

[0011] S2. Evaluate the global and regional systolic and diastolic functions of the left and right ventricles of the heart in septic cardiomyopathy patients based on the multimodal ultrasound data, and conduct a preliminary diagnosis of septic cardiomyopathy in combination with the clinical data;

[0012] S3. Use the Rcaret package to divide the multimodal ultrasound data of septic patients into a training set and a validation set, and perform least absolute shrinkage and LASSO regression analysis on the training set using the glmnet R software package to determine the preliminary predictors of ultrasound diagnosis;

[0013] S4. Perform logistic regression analysis on the preliminary predictors to determine independent predictors, and the independent predictors are used to reflect the cardiac function of septic patients;

[0014] S5. Reverse translate the independent predictors into the nomoscore through the rms package to obtain a nomogram, visually present the independent predictors according to their weights, and quantify the probability of the occurrence of septic cardiomyopathy.

[0015] Further, the clinical data includes demographic and clinical characteristics, clinical and laboratory data, and prognostic indicators;

[0016] The demographic and clinical characteristics include age, gender, height, weight, primary diagnosis, infection site, known etiology, and underlying diseases;

[0017] The clinical and laboratory data includes heart rate, mean arterial pressure, vasoactive drug dosage, pH, oxygenation index, acute physiology and chronic health evaluation score, infection-related acute organ function score, white blood cell count, neutrophil count, lymphocyte count, hematocrit level, platelet count, calcitonin level, C-reactive protein, interleukin 6, total bilirubin level, serum creatinine level, cardiac troponin I level, and N-terminal pro-brain natriuretic peptide;

[0018] The prognostic indicators include primary prognostic indicators and secondary prognostic indicators. The primary prognostic indicator is the 28-day mortality rate, and the secondary prognostic indicators are the ventilator use time and the length of stay in the ICU.

[0019] Further, the dynamic collection of multimodal ultrasound data specifically includes:

[0020] On the 1st to 5th day after the patient enters the ICU, multimodal cardiac ultrasound technology including two-dimensional imaging, M-mode, color Doppler, pulsed Doppler, tissue Doppler, and speckle tracking imaging is used to evaluate the cardiac function of septic patients. The evaluation criteria include the evaluation of left ventricular systolic and diastolic functions and the evaluation of right ventricular systolic function.

[0021] Among them, the evaluation of left ventricular systolic and diastolic functions includes: left ventricular longitudinal axis strain measured by speckle tracking, left ventricular ejection fraction, time-velocity integral of the left ventricular outflow tract, early diastolic mitral valve flow velocity, late diastolic mitral valve flow velocity, E / A ratio, early diastolic velocity of mitral annulus tissue Doppler, on the ventricular septal side and ventricular free wall side, E / e' ratio, left atrial volume index calculated from the left atrial diameter measured by two-dimensional ultrasound, and maximum tricuspid regurgitation velocity;

[0022] The evaluation of right ventricular systolic function includes measuring tricuspid annular plane systolic displacement, tricuspid systolic blood flow velocity, and inferior vena cava collapse index.

[0023] Furthermore, the evaluation of the overall and local systolic and diastolic functions of the left and right ventricles of the heart in septic patients specifically includes:

[0024] The evaluation criterion for left ventricular systolic dysfunction is that the left ventricular ejection fraction is less than 50%;

[0025] Defining left ventricular diastolic dysfunction requires at least meeting the following two criteria: tricuspid peak flow velocity exceeds 2.8 m / s, left atrial volume is greater than 34 mL / m², tissue Doppler imaging shows that the e' wave velocity of the septal annulus is lower than 7 cm / s, the lateral annulus is less than 10 cm / s, the e' ratio of the lateral annulus exceeds 13, and the e' ratio of the septal annulus exceeds 15;

[0026] Defining right ventricular systolic dysfunction requires meeting at least the following two criteria; tricuspid systolic annular plane displacement is less than 16 mm, tricuspid systolic lateral annulus velocity is less than 15 cm / S, and right ventricular fractional area change is less than 35%.

[0027] The present invention also provides a septic cardiomyopathy diagnosis model constructed according to the construction method of the septic cardiomyopathy diagnosis model.

[0028] The present invention also provides a diagnosis system for septic cardiomyopathy, which is used to execute the construction method of the septic cardiomyopathy diagnosis model, including:

[0029] A data acquisition module for dynamically acquiring multimodal ultrasound data and clinical data of septic patients;

[0030] An evaluation module, configured to preliminarily evaluate the overall and local systolic and diastolic functions of the left and right ventricles of the heart of a patient with septic cardiomyopathy based on multimodal ultrasound data, and perform a preliminary diagnosis of septic cardiomyopathy in combination with the clinical data;

[0031] A preliminary predictor determination module, configured to use the Rcaret package to divide the multimodal ultrasound data of septic patients into a training set and a validation set, and perform least absolute shrinkage and LASSO regression analysis on the training set using the glmnet R software package to determine the preliminary predictors for ultrasonic diagnosis;

[0032] A Logistics analysis module, configured to perform logistics regression analysis on the preliminary predictors to determine independent predictors, where the independent predictors are used to reflect the cardiac function of septic patients;

[0033] A prediction visualization module, configured to back-translate the independent predictors into the nomoscore through the rms package to obtain a nomogram, visually present the independent predictors according to their weights, and quantify the probability of the occurrence of septic cardiomyopathy.

[0034] According to the specific embodiments provided by the present invention, the following technical effects are disclosed: The method and system for quantitatively diagnosing septic cardiomyopathy provided by the present invention comprehensively evaluate the cardiac function of septic patients based on multimodal echocardiography and establish a reliable ultrasonic diagnosis prediction model for septic cardiomyopathy. Independent predictors are screened out through this model, which comprehensively reflects the cardiac function of septic patients and simplifies the cardiac evaluation process for septic patients. Moreover, through visual presentation by nomoscore, the probability of the occurrence of SIC is quantified, and the weights of ultrasonic diagnosis predictors in the diagnosis of SIC are clarified. The present invention also proves that the performance of this model in predicting SIC is excellent, providing a highly sensitive and specific visual means for early prediction of the occurrence of septic cardiomyopathy, facilitating clinical development and application, and having profound clinical significance.

[0035] In summary, the present invention solves the problems in the prior art that a single ultrasonic parameter cannot comprehensively reflect cardiac function, the combined parameters of multiple indicators are unclear and lack a quantitative standard, and specific ultrasonic parameters for SIC cannot be screened. Specific ultrasonic parameters (i.e., 3 independent predictors) for SIC diagnosis are screened out. The present invention provides an intuitive, highly sensitive and specific method for early prediction of the occurrence of SIC, improves the efficiency of SIC diagnosis, promotes clinical development and application, and has important clinical significance. Description of the Drawings

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only 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.

[0037] Figure 1 It is a flowchart of the method for constructing a diagnostic model for sepsis cardiomyopathy according to the present invention;

[0038] Figure 2 It is a schematic diagram of variable selection using the LASSO regression analysis method in the embodiments of the present invention. Among them, Figure A shows variable selection using the LASSO regression model, and Figure B shows the screening results of the LASSO regression model;

[0039] Figure 3 It is a forest plot of logistic regression analysis of the predictive factors for the risk of septic cardiomyopathy in sepsis patients in the embodiments of the present invention. Among them, A represents the logistic regression analysis of the predictive factors for the risk of septic cardiomyopathy in sepsis patients, and B represents the independent predictive factors GLS, TAPSE, and E predicted by the SIC modality map. Detailed implementation manners

[0040] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0041] The purpose of the present invention is to provide a method and a diagnostic system for constructing a diagnostic model for sepsis cardiomyopathy. By combining multimodal cardiac ultrasound, the cardiac function of sepsis patients is comprehensively and dynamically evaluated. The independent predictive factors for ultrasonic diagnosis of septic myocardium are determined by Lasso regression, and the quantitative diagnosis of sepsis cardiomyopathy is realized.

[0042] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the drawings and specific implementation manners.

[0043] A method for constructing a diagnostic model for sepsis cardiomyopathy provided by an embodiment of the present invention includes the following steps:

[0044] S1. Multimodal data acquisition: Dynamically acquire multimodal ultrasound data and clinical data of sepsis patients, and perform a preliminary diagnosis of sepsis cardiomyopathy based on the clinical data;

[0045] S2. Evaluate the global and regional systolic and diastolic functions of the left and right ventricles of the heart in patients with septic cardiomyopathy based on multimodal ultrasound data, and make a preliminary diagnosis of septic cardiomyopathy in combination with clinical data;

[0046] S3. Use the Rcaret package to divide the multimodal ultrasound data of septic patients into a training set and a validation set, and perform least absolute shrinkage and selection operator (LASSO) regression analysis on the training set using the glmnet R software package to determine the preliminary predictors for ultrasound diagnosis;

[0047] S4. Perform logistic regression analysis on the preliminary predictors to determine the independent predictors, and the independent predictors are used to reflect the cardiac function of septic patients;

[0048] S5. Reverse translate the independent predictors into nomoscore through the rms package to obtain a nomogram, visually present the independent predictors according to their weights, and quantify the probability of the occurrence of septic cardiomyopathy.

[0049] The specific implementation process of the above method is as follows:

[0050] 1. Select patients who visited a certain tertiary general hospital from March 2023 to April 2024 as the research objects, and screen them according to the following inclusion and exclusion criteria:

[0051] Inclusion criteria: 1) Meet the diagnostic criteria of sepsis 3.0; 2) Age ≥ 18 years old.

[0052] Exclusion criteria: 1) Patients with combined acute coronary syndrome or those who received relevant surgeries or interventions during hospitalization; 2) Patients with a history of chronic heart failure, chronic renal insufficiency (requiring long-term dialysis), or chronic liver insufficiency (Child-Pugh class C); 3) Patients with a history of hypertrophic cardiomyopathy, dilated cardiomyopathy, restrictive cardiomyopathy, rheumatic heart disease, etc.; 4) Patients with a history of severe arrhythmia; 5) Patients who were unable to undergo echocardiography during their stay in the ICU for various reasons; 6) Patients or their guardians who refused to participate in this study or withdrew midway.

[0053] Among them, 73 patients with acute or chronic cardiac dysfunction unrelated to septic cardiomyopathy or end-stage functional failure caused by chronic organ diseases were excluded from the study to avoid misdiagnosis of septic cardiomyopathy. Subsequently, 221 patients with sepsis were subjected to dynamic multimodal cardiac ultrasound examination. However, 40 patients were excluded from the subsequent analysis due to incomplete dynamic cardiac ultrasound examination. Finally, a total of 181 patients with sepsis were considered eligible for the analysis of this study, including 107 males and 74 females, aged between 19 and 86 years, with a median age of 58 years. The participants were randomly assigned to the training set and the validation set at a ratio of 7:3, with 126 cases in the training set and 55 cases in the validation set. All patients underwent necessary examinations.

[0054] 2. Dynamic acquisition of multimodal ultrasound data

[0055] For patients meeting the above criteria, on the 1st - 5th day after the patients entered the ICU, multimodal cardiac ultrasound including two-dimensional imaging, M-mode, color Doppler, pulsed Doppler, tissue Doppler, and speckle tracking imaging was used to evaluate the cardiac function of patients with sepsis. The evaluation criteria covered the assessment of left ventricular systolic and diastolic functions, including left ventricular longitudinal strain (GLS) measured by speckle tracking, left ventricular ejection fraction (LVEF), left ventricular outflow tract time-velocity integral (VTI), early diastolic mitral valve flow velocity (E), late diastolic mitral valve flow velocity (A), E / A ratio, early diastolic velocity of mitral annulus tissue Doppler (e', at the interventricular septum side (e'-S) and the free wall side of the ventricle (e'-L)), E / e' ratio, left atrial volume index (LAVI) calculated from the left atrial diameter measured by two-dimensional ultrasound, and maximum tricuspid regurgitation velocity (TRV). In addition, the assessment of right ventricular systolic function was also included, including measurement of tricuspid annular plane systolic excursion (TAPSE), tricuspid systolic blood flow velocity (S), and inferior vena cava collapse index (IVC-CI).

[0056] A total of 14 ultrasound factors were obtained in this process. A standardized ultrasound examination procedure was adopted to ensure optimal image acquisition and data collection. The left and right ventricular global and regional systolic and diastolic functions of septic cardiomyopathy were quantitatively and comprehensively evaluated through multimodal ultrasound data.

[0057] 3. Collection of clinical data

[0058] (1) Demographics and clinical characteristics: age, gender, height, weight, primary diagnosis, site of infection, etiology (if known), underlying diseases (hypertension, diabetes, coronary artery disease, chronic kidney disease, chronic obstructive pulmonary disease, tumor diseases, autoimmune diseases, etc.).

[0059] (2)Clinical and laboratory data: heart rate, mean arterial pressure, dosage of vasoactive drugs (such as norepinephrine, dopamine, dobutamine), pH, oxygenation index, Acute Physiology and Chronic Health Evaluation II (APACHE II) score, Sequential Organ Failure Assessment (SOFA) score related to infection, white blood cell count, neutrophil count, lymphocyte count, hematocrit level, platelet count, calcitonin level, C-reactive protein, interleukin 6, total bilirubin level, serum creatinine level, cardiac troponin I level, N-terminal pro-brain natriuretic peptide (NT-proBNP).

[0060] (3)Prognostic indicators: The primary prognostic indicator is the 28-day mortality rate, and the secondary prognostic indicators are the duration of ventilator use (h) and the length of stay in the intensive care unit (ICU) (h).

[0061] 4. Diagnostic criteria for septic cardiomyopathy

[0062] For the enrolled septic patients, whether they have septic cardiomyopathy is comprehensively judged by physicians with more than 10 years of ICU practice and deputy senior or higher professional titles in combination with the medical history according to the following conditions: (1) Acute and reversible, at least 2-3 dynamic multimodal ultrasounds; (2) Global and biventricular dysfunction (systolic and / or diastolic), accompanied by decreased contractility; (3) Elevated myocardial injury markers (cardiac troponin I > 0.012 ng / ml and NT-proBNP > 300 pg / ml); (4) Excluding myocardial ischemia caused by acute coronary artery stenosis.

[0063] The diagnostic criteria for left ventricular systolic dysfunction (LVSDF) in echocardiogram reports are: left ventricular ejection fraction (LVEF) less than 50%.

[0064] The diagnostic criteria for left ventricular diastolic dysfunction (LVDDF) are: at least meeting the following two criteria: tricuspid peak flow velocity exceeding 2.8 m / s, left atrial volume greater than 34 mL / m², tissue Doppler imaging showing septal annulus e' wave velocity lower than 7 cm / s, lateral annulus less than 10 cm / s, lateral annulus e' ratio exceeding 13, and septal annulus e' ratio exceeding 15.

[0065] The diagnostic criteria for right ventricular systolic dysfunction (RVSDF) are: meeting at least the following two criteria: tricuspid systolic annular plane displacement less than 16 mm, tricuspid systolic lateral annular velocity less than 15 cm / s, and right ventricular fractional area change less than 35%.

[0066] 5. Statistical methods

[0067] The Shapiro-Wilk normality test was used to evaluate whether the data distribution of continuous variables was normal. The data with normal distribution were described by the mean and standard deviation, and the independent samples t-test was used for comparison. For the data with non-normal distribution, the median and interquartile range were used for generalization, and the Mann-Whitney U test was used for comparison. Descriptive statistics included the frequencies of categorical variables expressed as percentages (%), and the chi-square test or Fisher's exact test was used for comparison.

[0068] Statistical analysis was performed using the Rcaret package (version 4.2.1). First, 181 patients with sepsis were randomly assigned to the training set (126 cases) and the validation set (55 cases) using the R caret package, with an allocation ratio of 7:3. Subsequently, the glmnet package was used for LASSO (Least Absolute Shrinkage and Selection Operator) regression analysis. LASSO regression analysis is a method for scaling and selecting variables in a linear regression model to achieve univariate analysis. By using LASSO regression to minimize the prediction error of the quantitative response variable, the predictive variables were selected. LASSO regression scales the regression coefficients of some variables towards zero. In the LASSO regression analysis with ten-fold cross-validation using R software, the variables with non-zero coefficients were considered to have the strongest correlation with the response variable. This process included standardizing and normalizing the variables to determine the optimal λ value, where Lambda.1se produced a model with good performance and fewer independent variables. As Figure 2 shown, on the training set, according to the specified inclusion criteria, 4 preliminary predictors, namely GLS, e'-L, E, and TAPSE, were determined using the LASSO regression analysis method.

[0069] 6. Establishment and Visualization of the Sepsis Cardiomyopathy Diagnostic Model

[0070] As Figure 3 shown, logistic regression analysis was performed using the glm package in R language, and a predictive diagnostic model was constructed by combining the features selected in the LASSO regression model. By introducing and analyzing the statistically significant predictive factors, an SIC risk prediction model was established. All the features selected in this embodiment were statistically significant and were used to create a nomogram prediction model. The nomogram diagram was developed using the rms package in R language.

[0071] 7. Research Results

[0072] During the period from March 2023 to April 2024, 294 patients with sepsis were initially screened. Among these participants, 110 were diagnosed with septic cardiomyopathy (SIC), while 71 were undiagnosed. The prevalence of septic cardiomyopathy (SIC) in this cohort was 60.77%. Example 2

[0073] An embodiment of the present invention provides a diagnostic model for septic cardiomyopathy constructed according to the diagnostic method of septic cardiomyopathy described in Example 1. The performance of this model is verified as follows:

[0074] The performance of this diagnostic model is evaluated by discrimination ability, calibration ability, and clinical effectiveness. The discrimination ability is evaluated by the area under the receiver operating characteristic curve (AUC); the calibration accuracy is determined by visualizing the calibration plot and the Hosmer-Lemeshow test; the clinical effectiveness is analyzed by decision curve analysis.

[0075] To avoid overfitting, 500 bootstrap internal validations were performed on the prediction efficiency, and a two-tailed p-value less than 0.05 was considered statistically significant. The specific results are as follows:

[0076] First, the ROC curve was used to evaluate the classification ability of the predictive diagnostic model. For this model, the integrated area under the ROC curve in the training set was 0.879 (95% confidence interval (CI) was 0.821 - 0.936), and the integrated area under the ROC curve in the validation set was 0.888 (95% CI was 0.801 - 0.976), indicating that the model has accurate predictive performance.

[0077] The diagnostic model was calibrated using the calibration plot and the Hosmer-Lemeshow test. The results showed that the predicted probability of the model was consistent with the actual probability, indicating good model fitting. The results of the Hosmer-Lemeshow test showed a high degree of consistency between the predicted probability and the observed results in the training set and the validation set (P = 0.386). (p = 0.957).

[0078] Decision curve analysis (DCA) showed that the threshold probability of this diagnostic model varied between 5% - 100% in the training set and between 21% - 100% in the validation set, indicating that the model has considerable net clinical benefits. Example 3

[0079] An embodiment of the present invention provides a diagnostic system for septic cardiomyopathy, which is used to execute the construction method of the diagnostic model for septic cardiomyopathy described in Example 1, including:

[0080] A data acquisition module for dynamically acquiring multimodal ultrasound data and clinical data of septic patients;

[0081] An evaluation module, configured to preliminarily evaluate the overall and local systolic and diastolic functions of the left and right ventricles of the heart of a patient with septic cardiomyopathy based on multimodal ultrasound data, and perform a preliminary diagnosis of septic cardiomyopathy in combination with the clinical data;

[0082] A preliminary predictor determination module, configured to use the Rcaret package to divide the multimodal ultrasound data of septic patients into a training set and a validation set, and perform least absolute shrinkage and LASSO regression analysis on the training set using the glmnet R software package to determine the preliminary predictors for ultrasound diagnosis;

[0083] A Logistics analysis module, configured to perform logistics regression analysis on the preliminary predictors to determine independent predictors, where the independent predictors are used to reflect the cardiac function of septic patients;

[0084] A prediction visualization module, configured to back-translate the independent predictors into the nomoscore through the rms package to obtain a nomogram, visually present the independent predictors according to their weights, and quantify the probability of the occurrence of septic cardiomyopathy.

[0085] In summary, the diagnostic method, system, and diagnostic model for septic cardiomyopathy provided by the present invention comprehensively evaluate the cardiac function of septic patients based on multimodal echocardiography, establish a diagnostic model for septic cardiomyopathy, screen out 3 SIC ultrasound diagnostic predictors through this model, comprehensively reflect the cardiac function of septic patients, and simplify the bedside cardiac assessment process for septic patients. The predictors are visually presented according to their weights through the nomoscore, the probability of the occurrence of SIC is quantified, and the weights of the ultrasound diagnostic predictors in the diagnosis of SIC are clarified. It is confirmed that the performance of the model in predicting SIC is excellent, providing a highly sensitive and specific visualization method for the early prediction of the occurrence of septic cardiomyopathy, facilitating clinical implementation and application, and having profound clinical significance.

[0086] For the remaining technical features in this embodiment, those skilled in the art can flexibly select them according to the actual situation to meet different specific actual needs. However, it is obvious to those of ordinary skill in the art that these specific details do not have to be adopted to implement the present invention.

[0087] Modifications and changes made by those skilled in the art that do not depart from the spirit and scope of the present invention shall fall within the protection scope of the appended claims of the present invention. In the above description, in order to provide a thorough understanding of the present invention, a large number of specific details are elaborated. However, it is obvious to those of ordinary skill in the art that these specific details do not have to be adopted to implement the present invention. In other instances, well-known technologies, such as specific construction details, operating conditions, and other technical conditions, are not specifically described to avoid confusing the present invention.

[0088] In this article, specific examples are used to elaborate on the principles and implementation manners of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. To sum up, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for constructing a diagnostic model for septic cardiomyopathy, characterized in that: The following steps are involved: S1. Multimodal data acquisition: Dynamic multimodal ultrasound data acquisition and clinical data acquisition of patients with sepsis, including: on the 1st to 5th day after the patient enters the ICU, multimodal cardiac ultrasound technology including two-dimensional imaging, M mode, color Doppler, pulsed Doppler, tissue Doppler and speckle tracking imaging is used to evaluate the cardiac function of patients with sepsis. The evaluation criteria include the evaluation of left ventricular systolic and diastolic function, and the evaluation of right ventricular systolic function; S2. Evaluate the overall and local systolic and diastolic function of the left and right ventricles of patients with septic cardiomyopathy based on multimodal ultrasound data, and make a preliminary diagnosis of septic cardiomyopathy in combination with the clinical data; S3. The multimodal ultrasound data of sepsis patients were divided into training and validation sets using the R caret package, and the training set was subjected to least absolute shrinkage and LASSO regression analysis using the glmnet R package to determine preliminary predictors of ultrasound diagnosis. S4. Performing logistic regression analysis on the preliminary predictors to determine the independent predictors GLS, TAPSE and E, which are used to reflect the cardiac function of patients with sepsis; S5. The independent predictors were back-translated into nomoscore using the rms package to obtain a nomogram, which visualized the independent predictors according to their weights and quantified the probability of septic cardiomyopathy.

2. The method for constructing a diagnostic model for septic cardiomyopathy according to claim 1, characterized in that: The clinical data included demographic and clinical characteristics, clinical and laboratory data, and prognostic indicators; The demographic and clinical characteristics included age, sex, height, weight, primary diagnosis, site of infection, known etiology, and underlying diseases; The clinical and laboratory data included heart rate, mean arterial pressure, vasoactive drug dosage, pH, oxygenation index, acute physiology and chronic pathology score, infection-related acute organ function score, white blood cell count, neutrophil count, lymphocyte count, hematocrit level, platelet count, calcitonin level, C-reactive protein, interleukin-6, total bilirubin level, serum creatinine level, troponin I level, and pro-brain natriuretic peptide; The prognostic indicators include primary prognostic indicators and secondary prognostic indicators. The primary prognostic indicator is the 28-day mortality rate, and the secondary prognostic indicators are the duration of ventilator use and the duration of ICU stay.

3. The method for constructing a diagnostic model for septic cardiomyopathy according to claim 1, characterized in that: In S1, the assessment of left ventricular systolic and diastolic function included: left ventricular long-axis strain measured by speckle tracking, left ventricular ejection fraction, left ventricular outflow tract time-velocity integral, early diastolic mitral valve velocity E, late diastolic mitral valve velocity A, E / A ratio, mitral valve annular tissue Doppler early diastolic velocity e' on the ventricular septum side and ventricular free wall side, E / e' ratio on the ventricular septum side and ventricular free wall side, left atrial volume index LAVI calculated by left atrial diameter measured by two-dimensional ultrasound, and maximum tricuspid regurgitation velocity TRV; The evaluation of right ventricular systolic function includes measuring tricuspid annular plane systolic displacement TAPSE, tricuspid systolic blood flow velocity S and inferior vena cava collapsibility index IVC-CI.

4. The method for constructing a diagnostic model for septic cardiomyopathy according to claim 3, characterized in that: Assess the overall and regional systolic and diastolic function of the left and right ventricles in patients with sepsis, including: The evaluation criteria for left ventricular systolic dysfunction were left ventricular ejection fraction less than 50%; Left ventricular diastolic dysfunction requires at least two of the following three criteria: Early diastolic mitral flow velocity exceeds 2.8 m / s; Left atrial volume greater than 34 mL / m²; Tissue Doppler imaging showed that the early diastolic velocity of the mitral valve annulus on the ventricular septum side was less than 7 cm / s, the early diastolic velocity of the mitral valve tissue Doppler on the ventricular free wall side was less than 10 cm / s, the E / e' ratio on the ventricular free wall side was more than 13, and the E / e' ratio on the ventricular septum side was more than 15; Right ventricular systolic dysfunction requires at least two of the following three criteria: The plane displacement of the tricuspid valve systolic ring is less than 16 mm; The tricuspid valve systolic annular velocity is less than 15 cm / s; The right ventricular fractional area change was less than 35%.

5. A septic cardiomyopathy diagnostic model constructed according to the method for constructing a septic cardiomyopathy diagnostic model according to any one of claims 1 to 4.

6. A diagnostic system for septic cardiomyopathy, used to execute the method for constructing a diagnostic model for septic cardiomyopathy according to any one of claims 1 to 4, characterized in that: include: The data acquisition module is used to dynamically collect multimodal ultrasound data and clinical information of patients with sepsis, including: on the 1st to 5th day after the patient enters the ICU, the cardiac function of patients with sepsis is evaluated using multimodal cardiac ultrasound technology including two-dimensional imaging, M mode, color Doppler, pulsed Doppler, tissue Doppler and speckle tracking imaging. The evaluation criteria include the evaluation of left ventricular systolic and diastolic function, and the evaluation of right ventricular systolic function; An evaluation module, used to preliminarily evaluate the overall and local systolic and diastolic functions of the left and right ventricles of patients with septic cardiomyopathy based on multimodal ultrasound data, and to make a preliminary diagnosis of septic cardiomyopathy in combination with the clinical data; A preliminary predictor determination module was used to divide the multimodal ultrasound data of sepsis patients into a training set and a validation set using the Rcaret package, and to perform least absolute shrinkage and LASSO regression analysis on the training set using the glmnet R package to determine preliminary predictors of ultrasound diagnosis; Logistics analysis module, used to perform logistics regression analysis on preliminary predictors to determine independent predictors GLS, TAPSE and E, which are used to reflect the cardiac function of patients with sepsis; The prediction visualization module is used to translate the independent predictors into nomoscore through the rms package to obtain a nomogram, visualize the independent predictors according to their weights, and quantify the probability of septic cardiomyopathy.

Images