A model for predicting mortality risk in patients with severe burns complicated by hypoxic hepatitis, its construction method, and its storage medium.

Abstract

This invention relates to a mortality risk prediction model, its construction method, and a storage medium for patients with severe burns complicated by hypoxic hepatitis. The method for constructing the mortality risk prediction model includes: collecting clinical characteristic data of patients with severe burns complicated by hypoxic hepatitis; using a regression model to analyze and screen the training set clinical characteristic data with p-values ​​less than 0.05 against mortality, thereby identifying clinical characteristic data related to mortality in patients with severe burns complicated by hypoxic hepatitis; and constructing a mortality risk prediction model based on the screened clinical characteristic data. This invention also provides a mortality risk prediction model constructed using the aforementioned method. This invention also provides a storage medium. Based on the patient's basic clinical data and disease-related indicators, this invention identifies risk factors leading to death in patients with severe burns and hypoxic hepatitis, enabling convenient, rapid, and accurate prediction of the patient's mortality probability in advance, thereby helping doctors to take appropriate preventive and treatment measures for patients.

Description

Technical Field

[0001] This invention relates to the field of medical diagnostic technology, specifically to a mortality risk prediction model, construction method, and storage medium for severe burns complicated with hypoxic hepatitis. Background Technology

[0002] Burns are an extremely common type of trauma in daily life, posing a significant threat to human life and health. Burns can lead to long-term physical and mental health problems, and globally, a large number of patients lose their lives to burns each year. Burns are classified primarily based on their depth and area. Burn depth ranges from first to fourth degree, while burn area is categorized as minor or severe. Minor burns are characterized by a burn area of ​​less than 10% of the total surface area (TBSA), mainly manifesting as superficial burns. Generally, patients with a burn TBSA >20% or full-thickness burns >10% are defined as having severe burns.

[0003] Although advancements in medical technology have greatly improved the prognosis for patients with severe burns, more than 300,000 people worldwide still die from them. Severe burn patients often develop sepsis with multiple organ dysfunction syndrome (MODS). In fact, MODS is the leading cause of death within 24 hours of severe burn injury. The liver is the most important organ for biochemical synthesis and metabolism. Liver damage following burns has been confirmed. Liver necrosis and liver dysfunction following liver burns were reported in the late 1930s. Subsequently, many studies have shown that liver dysfunction, hepatomegaly, and fatty infiltration are common in burn patients and are associated with the total burn surface area (TBSA).

[0004] Hypoxic hepatitis (HH) is a complication of underlying diseases and an acute liver injury that occurs in patients with heart failure, respiratory failure, septic or toxic shock, and other underlying conditions, characterized by a rapid and transient increase in AST or ALT. HH has a high mortality rate and poor prognosis in the ICU. Hepatic ischemia (venous congestion due to reduced hepatic blood flow and right heart failure) and arterial hypoxemia (due to decreased blood oxygen content) are the two main mechanisms leading to HH. Previous studies have shown that the mortality rate of HH patients during or shortly after hospitalization exceeds 50%. However, the morbidity, prognosis, and risk factors for occurrence and death in burn HH patients remain unclear.

[0005] Currently, treatment for severe burn patients is often based on their existing clinical symptoms, which can easily lead to missing the optimal treatment window for preventing death after the development of hypoxic hepatitis. Although the treatment of burn patients has improved significantly in recent decades, severe burns remain fatal. Among these, the mortality rate from hypoxic hepatitis in severe burn patients remains high. Current technology cannot predict whether a patient with severe burns complicated by hypoxic hepatitis will die, thus hindering timely prevention and treatment. Therefore, timely and accurate identification of mortality risk factors is crucial for assessing patient clinical prognosis and guiding individualized treatment. Summary of the Invention

[0006] In view of this, the purpose of this invention is to provide a mortality risk prediction model, construction method and storage medium for patients with severe burns complicated with hypoxic hepatitis, which identifies risk factors leading to death in patients with severe burns and hypoxic hepatitis based on the patient's basic clinical data and disease-related indicators, and uses these factors to establish a fully validated mortality risk prediction model to predict the patient's probability of death in advance conveniently, quickly and accurately, thereby helping doctors to take appropriate preventive and treatment measures for patients.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A method for constructing a mortality risk prediction model for severe burns complicated with hypoxic hepatitis includes the following steps:

[0009] S1. Data Collection and Preprocessing: Collect clinical characteristic data of deceased and surviving patients with severe burns complicated by hypoxic hepatitis, and preprocess the clinical characteristic data.

[0010] S2. Feature data screening: The preprocessed clinical feature data is divided into training set and validation set. The clinical feature data of the training set with p value less than 0.05 are analyzed and screened for whether or not death occurs. Clinical feature data related to death in severe burns complicated with hypoxic hepatitis are screened out.

[0011] S3. Model Construction: Based on the selected clinical characteristic data related to death, a mortality risk prediction model for severe burns complicated with hypoxic hepatitis was constructed.

[0012] Based on the aforementioned technical methods and grounded in actual clinical data, the practicality and relevance of the model are ensured. Data collection and preprocessing improve data quality and reduce the impact of noise and outliers on the model's predictive accuracy. Utilizing the LASSO regression model to screen for clinical features significantly associated with mortality helps improve the model's predictive power and interpretability. Dividing the data into training and validation sets helps assess the model's generalization ability and reduces the risk of overfitting. The model built based on relevant clinical features can provide a scientific basis for clinical decision-making, contributing to improved treatment outcomes and survival rates for patients with severe burns complicated by hypoxic-ischemic hepatitis.

[0013] Hypoxic hepatitis (HH) is defined as a syndrome in which alanine aminotransferase (ALT) or aspartate aminotransferase (AST) rapidly and transiently rises to more than 10 times the upper limit of normal in the presence of cardiac, circulatory, or respiratory failure. It is characterized by a pattern of primary hepatocellular injury without evidence of cholestasis.

[0014] Preferably, the clinical characteristic data include gender, age, cause of burn, past medical history, smoking history, alcohol consumption history, total burn area, third-degree burn area, BMI, burn location, number of liver injuries before the onset of HH, length of hospital stay, length of ICU stay, mortality rate, AST admission value, ALT admission value, GGT admission value, maximum AST value, maximum ALT value, maximum GGT value, ALP, direct bilirubin, indirect bilirubin, LDH, albumin, prealbumin, fibrinogen, serum creatinine admission value, and maximum serum creatinine value.

[0015] Preferably, the clinical characteristic data related to death from severe burns complicated with hypoxic hepatitis include the patient's age, total burn area, third-degree burn area, serum creatinine level upon admission, and maximum serum creatinine level.

[0016] Preferably, the mortality risk prediction model is a nomogram.

[0017] Preferably, in the column diagram:

[0018] The first column is the score scale, ranging from 0 to 100; the second column is the patient's serum creatinine value upon admission, ranging from 50 to 600; the third column is the patient's maximum serum creatinine value, ranging from 50 to 650; the fourth column is the patient's age, ranging from 20 to 80; the fifth column is the patient's total burn area, ranging from 20 to 100; the sixth column is the area of ​​third-degree burns, ranging from 0 to 70; the seventh column is the total score, ranging from 0 to 130; and the eighth column is the probability of death, ranging from 0.01 to 0.99.

[0019] Preferably, the mortality risk prediction model further includes a mortality risk value formula:

[0020]

[0021] In formula (I), X1 represents the patient's serum creatinine value upon admission; X2 represents the patient's maximum serum creatinine value; X3 represents the patient's age; X4 represents the patient's total burn area; X5 represents the patient's third-degree burn area; and P represents the mortality rate of severe burns complicated with hypoxic hepatitis.

[0022] Preferably, the maximum AUROC value of the nodal plot model is 0.945.

[0023] Preferably, the p-value is obtained by t-test or chi-square test.

[0024] Preferably, in step S1, the clinical feature data is randomly divided into a 70% training set and a 30% test set; the model is trained using the training dataset to obtain a mortality risk prediction model.

[0025] The mortality risk prediction model was validated using the test set.

[0026] The present invention also provides a mortality risk prediction model for severe burns complicated with hypoxic hepatitis constructed by the construction method described in the present invention.

[0027] Preferably, the mortality risk prediction model for severe burns complicated by hypoxic hepatitis includes:

[0028] The data collection and preprocessing module is used to collect clinical characteristic data of deceased and surviving patients with severe burns complicated by hypoxic hepatitis, and to preprocess the clinical characteristic data.

[0029] The feature data filtering module is used to divide the preprocessed clinical feature data into training set and validation set. The LASSO regression model is used to analyze and filter the clinical feature data of the training set with p value less than 0.05 and whether or not death occurred, and to filter out the clinical feature data related to death in severe burns complicated with hypoxic hepatitis.

[0030] The prediction model building module constructs a prediction model for the mortality risk of severe burns complicated with hypoxic hepatitis based on the selected clinical characteristic data related to mortality.

[0031] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the construction method as described in the present invention.

[0032] The beneficial effects of this invention are:

[0033] The present invention discloses a method for constructing a mortality risk prediction model for patients with severe burns complicated by hypoxic hepatitis. This method ensures the comprehensiveness and representativeness of the foundational data for the prediction model by collecting clinical characteristic data from these patients. During data preprocessing, the collected clinical characteristic data is cleaned and standardized, improving data quality and providing accurate and reliable data support for subsequent analysis. The use of a LASSO regression model for feature selection effectively handles high-dimensional data and identifies clinical characteristics highly correlated with mortality, thus improving the model's predictive accuracy. The model, based on clinical characteristic data, provides a more comprehensive and integrated reflection of the severity of the patient's condition and mortality risk. The mortality risk prediction model constructed based on the selected indicators provides clinicians with an accurate and practical tool, enabling result visualization and facilitating the assessment of patient mortality risk, thereby allowing for the development of more personalized treatment plans. This method has significant application value in the field of medical diagnostic technology. Attached Figure Description

[0034] Figure 1 A flowchart illustrating the method for constructing a mortality risk prediction model for patients with severe burns complicated by hypoxic hepatitis;

[0035] Figure 2 A nomogram model for predicting the risk of death in patients with severe burns complicated by hypoxic hepatitis;

[0036] Figure 3 The working characteristic curve of the training set subject cohort for the logistic regression model;

[0037] Figure 4 A plot of operating characteristic curves for test set cohort subjects in a logistic regression model;

[0038] Figure 5 The graph shows the results of the queue decision curve analysis for the training set of the nomogram model.

[0039] Figure 6 The result of the test set queue decision curve analysis for the nodal plot model;

[0040] Figure 7 The graph shows the results of the calibration curve analysis for the training set of the nomogram model.

[0041] Figure 8 The result of the test set queue calibration curve analysis for the nomogram model;

[0042] Figure 9 This is a frequency distribution histogram. Detailed Implementation

[0043] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0044] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0045] This invention aims to disclose a mortality risk prediction model, construction method, and storage medium for patients with severe burns complicated by hypoxic hepatitis. Based on the patient's clinical baseline data and disease-related indicators, it identifies risk factors leading to death in patients with severe burns and hypoxic hepatitis, and uses these factors to establish a fully validated mortality risk prediction model. This model allows for the convenient, rapid, and accurate prediction of the patient's probability of death in advance, thereby helping doctors to take appropriate preventive and treatment measures for the patient.

[0046] Among them, such as Figure 1 As shown, the method for constructing a mortality risk prediction model for severe burns complicated with hypoxic hepatitis includes the following steps:

[0047] S1. Data Collection: Collect clinical characteristic data of deceased and surviving patients with severe burns complicated by hypoxic hepatitis, and preprocess the clinical characteristic data to remove patients with incomplete clinical characteristic data.

[0048] S2. Feature data screening: The preprocessed clinical feature data is divided into training set and validation set. The clinical feature data of the training set with p value less than 0.05 of t test or chi-square test are analyzed and screened with whether or not death is found through LASSO regression model to screen out the clinical feature data related to death in severe burns complicated with hypoxic hepatitis.

[0049] S3. Model Construction: Based on the selected clinical characteristic data related to death, a mortality risk prediction model for severe burns complicated with hypoxic hepatitis was constructed.

[0050] By basing the model on real clinical data, its practicality and relevance are ensured. Data collection and preprocessing improve data quality and reduce the impact of noise and outliers on model prediction accuracy. Utilizing the LASSO regression model to screen for clinical features significantly associated with mortality helps improve the model's predictive power and interpretability. Dividing the data into training and validation sets helps assess the model's generalization ability and reduces the risk of overfitting. The model built based on relevant clinical features can provide a scientific basis for clinical decision-making, helping to improve the treatment outcomes and survival rate of patients with severe burns complicated by hypoxic-ischemic hepatitis.

[0051] Hypoxic hepatitis (HH) is defined as a syndrome in which alanine aminotransferase (ALT) or aspartate aminotransferase (AST) rapidly and transiently rises to more than 10 times the upper limit of normal in the presence of cardiac, circulatory, or respiratory failure. It is characterized by a pattern of primary hepatocellular injury without evidence of cholestasis.

[0052] In some embodiments, clinical characteristic data include sex, age, cause of burn, past medical history, smoking history, alcohol consumption history, burn area, third-degree burn area, BMI, burn location, number of liver injuries prior to HH, length of hospital stay, length of ICU stay, mortality rate, AST admission value, ALT admission value, GGT admission value, serum creatinine admission value, maximum AST value, maximum ALT value, maximum GGT value, maximum serum creatinine value, ALP, direct bilirubin, indirect bilirubin, LDH, albumin, prealbumin, and fibrinogen.

[0053] In some embodiments, the clinical characteristics associated with death in severe burns complicated by hypoxic hepatitis include the patient's age, total burn area, third-degree burn area, admission serum creatinine level, and maximum serum creatinine level.

[0054] In some embodiments, the mortality risk prediction model is a nomogram model.

[0055] For example, in S1, data collection and preprocessing include:

[0056] Data from 742 patients consecutively admitted to the Burn Unit of Southwest Hospital between January 2017 and December 2022 were collected. After preprocessing, data from 88 patients with missing information and 81 patients with TBSA ≤20% were removed, resulting in 573 patients with severe burns. Among them, 51 patients developed hypoxic-hepatitis (HH), and 19 of these HH patients died after surgery for HH, resulting in a mortality rate of 37.25%. Multivariate logistic regression was used to screen for factors influencing mortality. In the mortality model, variables with p-values ​​less than 0.1 were included in univariate logistic regression, and variables with p-values ​​less than 0.05 in univariate logistic regression were included in the multivariate model. LASSO regression analysis, nomogram construction, and evaluation were performed using R Studio 2024.09.0 software.

[0057] During the pretreatment process, the exclusion criteria were: patients with severely incomplete outpatient or medical records; patients in the BICU without liver function tests; patients in the BICU for ≤48 hours; patients admitted more than 7 days after burns; and patients with abnormal ALT / AST levels before admission to the BICU.

[0058] For example, S2, feature data filtering, includes:

[0059] Patients with hypoxic hepatitis (HH) were further divided into two groups: a survival group (N=19, 37.3%) and a non-survival group (N=32). Analysis revealed that the non-survival group had a significantly older age [54(43,70) vs. 44(31.3,49), p=0.007]. The total burn surface area was significantly higher in the non-survival group [55(37,81) vs. 44(31.3,49), p=0.014], and the third-degree burn surface area was also higher [38(22,48) vs. 21(13.3,30), p=0.008]. Compared with the survival group, the non-survival group also had a higher burn index [46(31.5,62.5) vs. 30.3(20.6,42.4), p=0.0100].

[0060] The analysis showed that the survivors experienced more electrical burns after HH (21.1% vs. 53.1%, p = 0.024). The survivors were also more likely to have a history of alcohol consumption (21.1% vs. 53.1%, p = 0.024) (see Table 1).

[0061] Table 1. Clinical characteristics of patients with severe burns complicated with hypoxic hepatitis (* indicates Fisher's test was used; chi-square test was used for other categorical variables; continuous variables: second degree burns, length of hospital stay, and length of ICU stay were tested using nonparametric tests; t-test was used for other continuous variables).

[0062]

[0063]

[0064] Compared with the survival group, the death group had a higher TBSA [45(30,61) vs. 35(25,52), p = 0.006], a higher area of ​​third-degree burns [27(14,40) vs. 8(0,22), p < 0.001], and a higher burn index [36(24,46) vs. 18.75(9.5,31.63), p < 0.001]. Liver function tests were normal upon admission, and there were no statistically significant differences between the death and survival groups. However, the death group had higher mean AST [392.86(104.83,1267.25)vs.112.33(82.97,181.45), p=0.01], higher mean direct bilirubin [2.80(2.29,16.90)vs.2.07(1.14,2.96), p=0.01], higher mean indirect bilirubin [14.59(11.70,31.64)vs.10.63(8.02,13.96), p=0.006], and higher mean total bilirubin [23.89(17.91,57.70)vs.14.90(12.21,22.93), p<0.001]. The deceased group also had poorer renal function, including significantly elevated serum creatinine levels upon admission [121.66 (80.02, 174.3) vs. 77.43 (65.34, 97.65), p = 0.003], higher maximum serum creatinine [183.61 (61.32), vs. 55.54 (50.05, 67.30), p < 0.001], and higher mean serum creatinine [85.4 (67.73, 131.21), vs. 66.01 (54.78, 80.96), p < 0.001]. Specifically, the mean GGT [52.04 (38.48, 76.05) vs. 91.85 (61.85, 141.51)] and the maximum [101 (55.7, 205) vs. 243.4 (138.18, 404.65), p = 0.003] in the death group were significantly lower than those in the survival group. The maximum ALP was similarly [242.77 ± 170.17 vs. 264.43 ± 95.09, p = 0.001]. Furthermore, the maximum total bilirubin [58.53 (28.01, 112.67) vs. 35.40 (21.27, 51.97), p = 0.013] was significantly elevated in patients who died from severe burns at the time of death (see Table 2).

[0065] Table 2. Laboratory indicators in clinical characteristics of patients with severe burns complicated with hypoxic hepatitis (admission value, mean, extreme value, albumin / globulin ratio, albumin, extreme value, fibrinogen, and mean). ALP follows a normal distribution (t-value).

[0066] (For the test, all others use non-parametric tests.)

[0067]

[0068]

[0069] Clinical treatment and prognostic data for HH patients (see Table 3). These data include the use of hepatoprotective drugs, hormones, and antibiotics. Patients in the non-survival group received less dexamethasone treatment than the survival group (21.1% vs. 68.8%, p<0.001) (0 vs. 46.9%, p<0.001). To investigate the differences in the types of urinary penicillin used between the two groups, we further categorized and examined urinary penicillin. Piperacillin use showed a significant difference (0 vs. 34.4%, p = 0.04).

[0070] Table 3. Clinical outcomes and treatment of patients with severe burns complicated by hypoxic hepatitis (* indicates Fisher's test was used, and chi-square test was used for other categorical variables).

[0071]

[0072]

[0073] The indicators with significant differences (p<0.05) (age, total burn area, third-degree burn area, Burn Index, electrical burns, history of alcohol consumption, serum creatinine level on admission, mean AST, mean GGT, mean direct bilirubin, mean indirect bilirubin, mean total bilirubin, mean serum creatinine, maximum GGT, maximum ALP, maximum total bilirubin, maximum serum creatinine, whether dexamethasone was used, whether piperacillin was used) were selected for further screening using univariate logistic regression. Logistic regression analysis was used to screen 12 potential risk factors for death due to HH (see Table 4). The results showed that age (OR=1.087, p=0.020) and TBSA (OR=1.057, p=0.037) had a significant impact on postoperative mortality in HH (see Table 5).

[0074] Table 4. Univariate logistic regression analysis of mortality risk factors in HH patients

[0075]

[0076]

[0077] Table 5. Multivariate logistic regression analysis of mortality risk factors in HH patients

[0078]

[0079] LASSO regression analysis revealed that these indicators were positively correlated with the occurrence of mortality.

[0080] For example, S3, model construction includes: constructing a nomogram predicting mortality based on the above-selected clinical characteristic data related to death and clinical characteristic data related to death (all case data were retrieved from the Southwest Hospital Burn Biodatabase), such as... Figure 2 As shown.

[0081] In the nomogram model:

[0082] The first column is the score scale, ranging from 0 to 100; the second column is the patient's serum creatinine value upon admission, ranging from 50 to 600; the third column is the patient's maximum serum creatinine value, ranging from 50 to 650; the fourth column is the patient's age, ranging from 20 to 80; the fifth column is the patient's total burn area, ranging from 20 to 100; the sixth column is the patient's third-degree burn area, ranging from 0 to 70; the seventh column is the total score, ranging from 0 to 130; and the eighth column is the probability of death, ranging from 0.01 to 0.99.

[0083] The maximum AUROC value of the nomogram model was 0.945 (95% CI 0.786–1.000), indicating a good fit.

[0084] In some embodiments, the mortality risk prediction model further includes a mortality risk value formula:

[0085]

[0086] In formula (I), X1 represents the patient's serum creatinine value upon admission; X2 represents the patient's maximum serum creatinine value; X3 represents the patient's age; X4 represents the patient's total burn area; X5 represents the patient's third-degree burn area; and P represents the probability of death from severe burns complicated by hypoxic hepatitis.

[0087] Internal validation was employed, randomly dividing all clinical characteristic data of the aforementioned HH patients into a training set (70%) and a test set (30%). The model was then trained using the training set and validated using the test set. The results are as follows: Figures 3 to 8 As shown.

[0088] in, Figure 3 The working characteristic curve of the training set subject cohort for the logistic regression model; Figure 4This is a plot of the subject operating characteristics (AUC) of the test set cohort for the logistic regression model. The model's AUC on the training and test sets are 0.955 and 0.600, respectively, indicating good predictive performance. Figure 5 Analysis results of the calibration curves for the training set of the nomogram model; Figure 6 This section presents the calibration curve analysis results for the test set cohort of the nomogram model. Calibration is an index of model fit, measuring the consistency between the actual probability of an outcome and the model's predicted probability, reflecting the model's accuracy in predicting absolute risk. In this model, the calibration curve in the results graph is close to the ideal line, indicating that the model's predicted probability matches the actual probability. Figure 7 The results of the queue decision curve analysis for the training set of the nomogram model; Figure 8 This is the result of the decision curve analysis for the test set of the nomogram model. In the results graph, the overall benefit of the predicted model in the range of 0 to 1 is generally higher than that of the "all" and "none" lines at the corresponding threshold probabilities, indicating that the model has practical value.

[0089] Frequency distribution histograms were created for the maximum and average ALT and AST values ​​of the deceased population (19 individuals), as shown below. Figure 9 As shown.

[0090] from Figure 9 Analysis revealed that the maximum ALT values ​​were concentrated between 200 and 400, and the maximum AST values ​​were concentrated around 600; the average ALT values ​​were concentrated between 200 and 400, and the average AST values ​​were concentrated around 600. Therefore, it can be concluded that the AST and ALT values ​​in patients with hypoxic hepatitis are generally more than 10 times higher than normal.

[0091] In some embodiments, a mortality risk prediction model for severe burns complicated with hypoxic hepatitis constructed by the construction method in any of the above embodiments is also provided, comprising:

[0092] The data collection and preprocessing module is used to collect clinical characteristic data of deceased and surviving patients with severe burns complicated by hypoxic hepatitis, and to preprocess the clinical characteristic data.

[0093] The feature data filtering module is used to divide the preprocessed clinical feature data into training set and validation set. The LASSO regression model is used to analyze and filter the clinical feature data of the training set with p value less than 0.05 and whether or not death occurred, and to filter out the clinical feature data related to death in severe burns complicated with hypoxic hepatitis.

[0094] The prediction model building module constructs a prediction model for the mortality risk of severe burns complicated with hypoxic hepatitis based on the selected clinical characteristic data related to mortality.

[0095] In some embodiments, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the construction method as described in any of the above embodiments.

[0096] In summary, the nomogram model for predicting mortality in severe burn patients with hypoxic hepatitis, as presented in this invention, can comprehensively evaluate patients' clinical characteristics and laboratory data, reducing judgment errors caused by incomplete information. When a severely burned patient has developed hypoxic hepatitis, the nomogram model based on clinical characteristic data can assist in predicting whether the patient with hypoxic hepatitis will die. The model of this invention is simple, intuitive, and easy to promote and apply. It effectively enables clinicians to conduct accurate individualized assessments of severely burned patients during treatment, bringing better survival benefits to patients. It has significant value for the effective application in the prognosis and monitoring of severe burn patients with hypoxic hepatitis, and provides personalized and professional health management solutions for different stratifications. It has promotional application value in the field of medical diagnostic technology.

[0097] The above embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention.

Claims

1. A method for constructing a mortality risk prediction model for severe burns complicated with hypoxic hepatitis, characterized in that, Includes the following steps: S1. Data Collection and Preprocessing: Collect clinical characteristic data of deceased and surviving patients with severe burns complicated by hypoxic hepatitis, and preprocess the clinical characteristic data. S2. Feature data screening: The preprocessed clinical feature data is divided into training set and validation set. The LASSO regression model is used to analyze and screen the clinical feature data with p value less than 0.05 and whether or not death occurred, and to screen out the clinical feature data related to death in severe burns complicated with hypoxic hepatitis. S3. Model Construction: Based on the selected clinical characteristic data related to death, a mortality risk prediction model for severe burns complicated with hypoxic hepatitis was constructed. The clinical characteristics data include gender, age, cause of burn, past medical history, smoking history, alcohol consumption history, total burn area, third-degree burn area, BMI, burn location, number of liver injuries before the onset of HH, length of hospital stay, length of ICU stay, mortality rate, AST, ALT, GGT, ALP, direct bilirubin, indirect bilirubin, LDH, albumin, prealbumin, fibrinogen, and the admission value and maximum value of serum creatinine. And / or, the clinical characteristic data related to death associated with severe burns complicated by hypoxic hepatitis selected include the patient's age, the patient's total burn area, the patient's third-degree burn area, the patient's serum creatinine level upon admission, and the patient's maximum serum creatinine level. The mortality risk prediction model is a nomogram model; In the nomogram model: The first column is the score scale, ranging from 0 to 100; the second column is the patient's serum creatinine value upon admission, ranging from 50 to 600; the third column is the patient's maximum serum creatinine value, ranging from 50 to 650; the fourth column is the patient's age, ranging from 20 to 80; the fifth column is the patient's total burn area, ranging from 20 to 100; the sixth column is the patient's third-degree burn area, ranging from 0 to 70; the seventh column is the total score, ranging from 0 to 130; and the eighth column is the probability of death, ranging from 0.01 to 0.

99. The mortality risk prediction model also includes a mortality risk value formula: (I) In formula (I), X1 represents the patient's serum creatinine value upon admission; X2 represents the patient's maximum serum creatinine value; X3 represents the patient's age; X4 represents the patient's total burn area; X5 represents the patient's third-degree burn area; and P represents the probability of death from severe burns complicated by hypoxic hepatitis. The maximum AUROC value of the noctilinear graph model is 0.

945.

2. The method for constructing a mortality risk prediction model for severe burns complicated with hypoxic hepatitis according to claim 1, characterized in that, The p-value is obtained through a t-test or a chi-square test; And / or, in S1, the clinical characteristic data is randomly divided into a 70% training set and a 30% test set; the model is trained using the training set to obtain a mortality risk prediction model; The mortality risk prediction model was validated using the test set.

3. A mortality risk prediction model for severe burns complicated with hypoxic hepatitis constructed by the construction method as described in claim 1 or 2.

4. The mortality risk prediction model for severe burns complicated with hypoxic hepatitis according to claim 3, characterized in that, include: The data collection and preprocessing module is used to collect clinical characteristic data of deceased and surviving patients with severe burns complicated by hypoxic hepatitis, and to preprocess the clinical characteristic data. The feature data filtering module is used to divide the preprocessed clinical feature data into training set and validation set. The LASSO regression model is used to analyze and filter the clinical feature data of the training set with p value less than 0.05 and whether or not death occurred, and to filter out the clinical feature data related to death in severe burns complicated with hypoxic hepatitis. The prediction model building module constructs a prediction model for the mortality risk of severe burns complicated with hypoxic hepatitis based on the selected clinical characteristic data related to mortality.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the construction method as described in claim 1 or 2.

Images