Method for constructing prediction model of invasive fungal infection after immunosuppression treatment
By constructing a prediction model based on Nomogram model, using indicators such as age, IgG level and CD4+ T cell count, the problem of lack of effective prediction models in the existing technology is solved, and efficiently the risk of invasive fungal infection in immunosuppressed patients is achieved. Predictive and preventive strategies guidance.
Patent Information
- Application Number
- CN202411022576.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-15
AI Technical Summary
The lack of effective predictive models and indicators to guide the risk of invasive fungal infection in immunosuppressed patients, resulting in the inability to identify the appropriate target population for antifungal prevention treatment, increasing the risk of infection and death.
By collecting clinical data, using single-factor chi-square test and multi-factor Logistic regression analysis, Nomogram model was constructed, and age, IgG level below 6.5g/L and CD4+ T cell count were determined as independent risk factors, and a prediction model was established. Predict the risk of invasive fungal infection after immunosuppressive treatment.
The constructed model has high predictive value. The C-index and AUC evaluation shows that the distinction and calibration degree are high in the training set and verification set, which can effectively guide the clinical prevention of high-risk patients and identify IFIs in patients with different degrees of immunosuppression. risk.
Smart Images

Figure CN118969302A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of biomedicine technology, and relates to a method for constructing a prediction model for invasive fungal infection after immunosuppressive therapy, and specifically to an application of a model for predicting clinical fungal infection risk in patients receiving immunotherapy. Background Art
[0002] Invasive fungal infections (IFIs) are systemic diseases characterized by the invasion of yeasts or molds into deep tissues, which are in stark contrast to superficial fungal infections in terms of severity and prognosis. Unlike superficial fungal infections, IFIs have high morbidity and mortality, leading to considerable health challenges worldwide (Pagano L, Mayor S. Invasive fungal infections in high-risk patients: report from TIMM-8 2017. Future Sci OA. 2018; 4(6): FSO307. doi: 10.4155 / fsoa-2018-0019.). Recent epidemiological data indicate that approximately 1.9 million people are affected by acute fungal infections each year, and an estimated 3 million people worldwide suffer from chronic severe fungal infections. The grim reality is that more than 1.6 million people die from fungal diseases each year, highlighting the urgency of addressing this public health threat (Bongomin F, Gago S, Oladele RO, Denning DW. Global and multi-national prevalence of fungal diseases-estimate precision. J Fungi (Basel) 2017; 3(4): 57. doi: 10.3390 / jof3040057.).
[0003] Most IFIs (nearly 70%) are invasive candidiasis, while cryptococcosis and aspergillosis account for 20% and 10%, respectively (Delaloye J, Calandra T. Invasive candidiasis as a cause of sepsis in the critically ill patient. Virulence. 2014; 5(1): 161-169. doi: 10.4161 / viru.26187.). High-risk groups for infection include patients receiving organ transplants, patients receiving blood stem cell transplants, AIDS patients, diabetes patients, burn patients, patients with neoplastic diseases, and patients receiving long-term immunosuppressive therapy (Badiee P, Hashemizadeh Z. Opportunistic invasive fungal infections: diagnosis & clinical management. Indian J Med Res. 2014; 139(2): 195-204.).
[0004] In order to reduce the morbidity and mortality of IFIs, there is a clinical consensus on the importance of preventing fungal infections. It is crucial to identify the appropriate target population for such prevention to avoid unnecessary side effects and economic pressure.
[0005] Previous studies have highlighted the susceptibility of patients with prolonged neutropenia, such as patients with acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS) receiving remission-induction chemotherapy and patients with severe aplastic anemia, to IFIs (Hoenigl M, Salmanton-García J, Walsh TJ, Nucci M, Neoh CF, Jenks JD, et al. Global guideline for the diagnosis and management of rare mould infections: an initiative of the European Confederation of Medical Mycology incooperation with the International Society for Human and Animal Mycology and the American Society for Microbiology. Lancet Infect Dis. 2021;21(8):e246-e257. doi:10.1016 / S1473-3099(20)30784-2.). However, for patients without agranulocytosis, the relationship between IFIs and host immune status is complex (Pagano L, Busca A, Candoni A et al. Riskstratification for invasive fungal infections in patients with hematological malignancies: SEIFEM recommendations. Blood Rev 2017; 31: 17-29.10.1016 / j.blre.2016.09.002).Clinical guidelines emphasize the need for prophylactic antifungal treatment in immunosuppressed patients (Ascioglu S, Rex JH, de Pauw B, Bennett JE, Bille J, Crokaert F, Denning DW, Donnelly JP, Edwards JE, Erjavec Z, Fiere D, Lortholary O, Maertens J, Meis JF, Patterson TF, Ritter J, Selleslag D, Shah PM, Stevens DA, Walsh TJ; Invasive Fungal Infections Cooperative Group of the European Organization for Research and Treatment of Cancer; Mycoses Study Group of the National Institute of Allergy and Infectious Diseases. Defining opportunistic invasive fungal infections in immunocompromised patients with cancer and hematopoietic stem cell transplants: an international consensus. Clin Infect Dis 2002; 34: 7-14.). However, there is currently a lack of models and indicators to predict IFI risk and guide prevention strategies.
[0006] Our previous work identified immunoglobulin G levels and natural killer (NK) cell counts as potential risk factors for IFIs (Ma J, Hu Y, Wu M, Wang X, Xie Y. Timing Determination of Invasive Fungal Infection Prophylaxis According to Immune Function in HSCT Patients. Front Microbiol. 2018 Mar 2; 9: 370.), which provided valuable insights for this study but was limited by the small sample size. This study included nearly 700 hospitalized patients from Huashan Hospital and Huadong Hospital. By closely monitoring the patients' cellular and humoral immunity levels, we identified their relationship with the development of IFIs and constructed a clinically relevant risk prediction model. Summary of the invention
[0007] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and provide a method for constructing a predictive model for invasive fungal infection after immunosuppressive therapy. The present invention is based on research findings that existing clinical guidelines only emphasize the necessity of preventive antifungal treatment for immunosuppressed patients, but the specific definition of immunosuppression, monitoring indicators and prevention strategies are still unclear. The significance of this study is to identify immunosuppressed patients who have received different degrees of immunosuppressive therapy, establish models and indicators for predicting IFI risk and guiding prevention strategies, so as to determine the appropriate target population for antifungal preventive treatment.
[0008] The purpose of the present invention can be achieved by the following scheme:
[0009] In a first aspect, the present invention provides a method for constructing a prediction model for invasive fungal infection after immunosuppressive therapy, comprising the following steps:
[0010] S1. Collect clinical information and data to obtain a data set, and divide it into a training set and a validation set;
[0011] S2. Single-factor chi-square test was used to determine the single-factor indicators of invasive fungal infection after immunosuppressive therapy in the training set;
[0012] S3, performing multivariate logistic regression analysis on the univariate indicators obtained in step S2 to determine the independent risk factors for invasive fungal infection as indicators for establishing a prediction model;
[0013] S4, constructing a Nomogram model based on the indicators in step S3, which is a prediction model for invasive fungal infection after immunosuppressive therapy;
[0014] S5. Verify the prediction model through the validation set.
[0015] As an embodiment of the present invention, in step S1, the clinical information and data include: age, gender, disease diagnosis, transplant type, chemotherapy regimen, combined chronic diseases and other clinical complications, IgG level, IgM level, IgA level, CD4 + / CD8 + T cell count, CD4 + T cell count, CD8 + T cell count, CD19 + T cell count, NK cell count.
[0016] As an embodiment of the present invention, in step S1, the data set is randomly divided into a training set and a validation set in a ratio of 4:3.
[0017] The clinical characteristics and laboratory test results of the patients were collected, and the occurrence of fungal infection was used as the research endpoint. Univariate and multivariate logistic regression was used to confirm that age, IgG level below 6.5 g / L and CD4 + T cell count was an independent risk factor.
[0018] As an embodiment of the present invention, in step S2, the single factor indicators include age, IgG level lower than 6.5g / L, CD4 + T cell count, CD8 + T cell count.
[0019] As an embodiment of the present invention, in step S3, the independent risk factors include age, IgG level lower than 6.5g / L, CD4 + T cell count.
[0020] As an embodiment of the present invention, in step S5, the verification uses C-index, AUC and calibration graph to evaluate the performance of the prediction model.
[0021] Furthermore, the C-index and AUC evaluate the discrimination of the prediction model, and the calibration graph evaluates the calibration of the prediction model in the training set and the validation set. Both groups showed high discrimination and calibration.
[0022] In a second aspect, the present invention provides a prediction model construction system for invasive fungal infection after immunosuppressive therapy, comprising:
[0023] Module M1: collect clinical information and data to obtain a data set, and divide it into a training set and a validation set;
[0024] Module M2, using the one-way chi-square test method to determine the univariate indicators of invasive fungal infection after immunosuppressive therapy in the training set;
[0025] Module M3: Perform multivariate logistic regression analysis on the univariate indicators obtained in module M2 to determine the independent risk factors for invasive fungal infection as indicators for establishing a prediction model;
[0026] Module M4: constructing a Nomogram model based on the indicators in module M3, which is a prediction model for invasive fungal infection after immunosuppressive therapy;
[0027] Module M5: Verify the prediction model through the verification set.
[0028] In a third aspect, the present invention further provides a method for predicting invasive fungal infection after immunosuppressive therapy, based on the prediction model construction method described in the first aspect, the clinical information and data of the predicted object are input into the prediction model to obtain a corresponding nomogram;
[0029] According to the nomogram, each model indicator variable was assigned a value to obtain the score of each model indicator variable, and the total score was calculated to obtain the corresponding probability of occurrence of invasive fungal infection.
[0030] To be more specific, the prediction method includes: assigning values according to these factors through a model visualized by a nomogram, drawing a vertical line from the coordinate axis of each risk indicator to obtain a corresponding score, adding these scores to obtain a total score, and the total score can be combined with the predicted value of the last coordinate axis, thereby obtaining the risk percentage of each individual.
[0031] In a fourth aspect, the present invention further provides an electronic device comprising a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the method in any possible implementation of the first aspect is completed.
[0032] Compared with the prior art, the present invention has the following beneficial effects:
[0033] 1. Identify immunosuppressed patients who have received different degrees of immunosuppressive therapy and determine the independent prognostic factors of IFI from cellular and humoral immune indicators. The nomogram prediction constructed based on these indicators can guide clinical prevention in high-risk patients.
[0034] 2. The C-index of the model constructed by the present invention in the training set and the validation set were 0.72 and 0.73 respectively, and the area under the curve was 0.718 and 0.723 respectively, which had good predictive value.
[0035] 3. The indicators required by the model constructed by the present invention are easy to obtain clinically, which is convenient for clinicians to evaluate patients. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:
[0037] Figure 1 A flowchart for patient screening and establishing a nomogram prediction in Example 2;
[0038] Figure 2 Visually assign values to risk indicators based on prediction of the nomogram in Example 2;
[0039] Figure 3 is the area under the curve of the training set and the validation set in Example 2;
[0040] Figure 4 is a calibration diagram predicted by the nomogram in Example 2;
[0041] Figure 5 The discrimination and calibration of the prediction model 1 in Example 2;
[0042] Figure 6 The discrimination and calibration of the prediction model 2 in Example 2;
[0043] Figure 7 The discrimination and calibration of the prediction model 3 in Example 2 are shown. DETAILED DESCRIPTION
[0044] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following examples are implemented on the premise of the technical solution of the present invention, provide detailed implementation methods and specific operation processes, and will help those skilled in the art to further understand the present invention. It should be pointed out that the protection scope of the present invention is not limited to the following embodiments, and several adjustments and improvements made under the premise of the concept of the present invention all belong to the protection scope of the present invention.
[0045] The present invention provides a method for constructing a prediction model for invasive fungal infection after immunosuppressive therapy, comprising the following steps:
[0046] S1. Collect clinical information and data to obtain a data set, and divide it into a training set and a validation set;
[0047] S2. Single-factor chi-square test was used to determine the single-factor indicators of invasive fungal infection after immunosuppressive therapy in the training set;
[0048] S3, performing multivariate logistic regression analysis on the univariate indicators obtained in step S2 to determine the independent risk factors for invasive fungal infection as indicators for establishing a prediction model;
[0049] S4, constructing a Nomogram model based on the indicators in step S3, which is a prediction model for invasive fungal infection after immunosuppressive therapy;
[0050] S5. Verify the prediction model through the validation set.
[0051] Among them, the assignment principle of Nomogram variables is:
[0052] When constructing the multi-factor Cox regression, the regression coefficient β (i.e. coef) of each variable is obtained; in the constructed Nomogram, the specific score of each variable can also be obtained.
[0053] The prediction model is essentially a visualization of the multivariate Cox regression.
[0054] 1. Model regression coefficient β (coef)
[0055] The regression coefficient shows the contribution of a variable to the model. The larger the absolute value, the greater the contribution. The positive or negative value indicates whether the impact on the outcome is positive or negative. At the same time, the hazard ratio HR value is also calculated from the regression coefficient. The formula is HR = exp(coef), coef = regression coefficient.
[0056] 2. The risk score of the nomogram is calculated based on the regression coefficient. The regression coefficients of all variables are calculated based on one of their own subvariables as a reference (assuming that the reference variable coef = 0).
[0057] For example: rt: yes = -1.2162, relative to No (whose coef is set to 0), Yes will reduce the risk of recurrence. However, this reference cannot be applied to lymph node metastasis n. The nomogram is a whole, and there must be a value that all variables can refer to.
[0058] The reference of the nomogram in this code is the maximum regression coefficient. That is, the variable with the largest regression coefficient n=≥4, coef=2.4296 has a score of 100 in the nomogram. All other variables are transformed based on it.
[0059] 3. Calculation of risk scores for categorical variables
[0060] Take the variable g as an example: the regression coefficient of histological grade II relative to grade I (coef = 0) = 1.3497. The conversion formula is:
[0061] The histological grade was II and the score on the Nomogram was 55.55≈56, which was consistent with the figure.
[0062] 4. Calculation of Risk Scores for Continuous Variables
[0063] First, the regression coefficients for continuous variables are interpreted as the change in risk per 1-unit change. Therefore, calculating the score for 1 unit can be used to calculate all scores.
[0064] Taking age as an example: coef = -0.0315, which means that the risk decreases by 0.0315 for every 1 year of age. Converted into risk score:
[0065] Results: The risk score for a 1-unit change in age, i.e., 1 year old, was 1.2963 points.
[0066] If the patient's recurrence risk is 0 points at the age of 80, then at the age of 20 it is 1.2963*(80-20)=77.79≈78 points.
[0067] nomo2#After running, the scores of all variables will be obtained
[0068] The calculated result is exactly the same as our manual calculation, and the result also shows that the score of each scale is 0.02429555, which is exactly one percent of the maximum regression coefficient.
[0069] Example 1
[0070] This example provides a clinical risk model for predicting invasive fungal infection in patients receiving immunotherapy, wherein the clinical test contents included in the risk model include age, IgG level, CD4 + T cell count.
[0071] Example 2
[0072] 2.1 Patient information
[0073] The information of 1,000 patients who visited two medical centers (East China Hospital Affiliated to Fudan University and Huashan Hospital Affiliated to Fudan University) from June 2021 to June 2022 was collected retrospectively.
[0074] 2.2 Patient inclusion criteria
[0075] The criteria for patients included in this modeling include: 1) patients who have received chemotherapy or immunosuppressive therapy within the last 3 months, 2) patients with chronic diseases such as diabetes, cancer, organ transplantation, and chronic infection. 3) All participants underwent humoral and cellular immunity assessments, including immunoglobulin levels, peripheral blood lymphocyte counts, and flow cytometry analysis. Excluded were: 1) patients with neutropenia, 2) patients with a history of invasive fungal infection or patients who had received fungal prevention in the past. The number of patients included in this study was 664. Clinical data included age, sex, disease diagnosis, transplant type, chemotherapy regimen, concurrent chronic diseases, and other clinical complications.
[0076] 2.3 Detection of immune function
[0077] The levels of IgG, IgM, and IgA were measured by immunoturbidimetry to evaluate humoral immune function. CD4 + / CD8 + 、CD4 + 、CD8 + 、CD19 + Cellular immune function was assessed by counting NK cells and 50,000 cells per analysis. Cell counts and ratios were determined by peripheral blood testing and flow cytometry.
[0078] 2.4 Clinical evaluation and definition of IFI
[0079] The diagnosis of IFI follows the revised criteria of the EORTC / MSG consensus group (Fleming, S., Yannakou, CK, Haeusler, GM, Clark, J., Grigg, A., Heath, CH, et al. (2014). Consensus guidelines for antifungal prophylaxis in haematological malignancy and haemopoietic stemcell transplantation.Intern.Med.J.44,1283-1297.doi:10.1111 / imj.12595; DonnellyJP,Chen SC,Kauffman CA.,et al.Revision and Update of the ConsensusDefinitions ofInvasive Fungal Disease From the European Organization forResearch and Treatment of Cancer and the Mycoses Study Group Education andResearch Consortium.Clin Infection Dis. 2020 Sep12; 71(6): 1367-1376. doi: 10.1093 / cid / ciz1008). Possible IFI requires clinical and mycological evidence. Clinical criteria include imaging signs suggestive of pulmonary, sinonasal, or central nervous system infection, supplemented by mycological criteria such as positive cytology, direct microscopy, fungal culture, or detection of galactomannan antigen in various clinical specimens. Most patients with IFI respond to antifungal therapy. Confirmation of IFI includes microscopic identification of fungal components, positive culture results of lesion tissue or blood, or positive fungal detection by polymerase chain reaction or high-throughput sequencing technology in clinical specimens.
[0080] 2.5 Feature selection, model building and evaluation
[0081] The prediction model for invasive fungal infection was developed as follows: the clinical information and data of 664 patients included in this study were used as the data set and randomly assigned to the training set and the validation set in a ratio of 4:3. Univariate logistic regression analysis was used to determine the clinical characteristics significantly associated with IFI. The significant variables in the univariate analysis were included in the logistic regression model and then eliminated step by step. We calculated the significance using a multivariate logistic regression model, which combined the features with the most predictive value and weighted them according to their coefficients. The regression coefficients were converted to 0-100 points and a nomogram was constructed. The receiver operating characteristic (ROC) curve and the area under the curve (AUC) were used to evaluate the accuracy of the prediction. Decision curve analysis (DCA) was used to evaluate the risk of IFI, clinical risk factors, and the clinical utility of the nomogram, and to quantify the net benefit at different threshold probabilities. The research process is as follows Figure 1 shown.
[0082] 2.6 Statistical analysis
[0083] Continuous variables were reported as mean and range, and categorical variables were reported as frequency and percentage. If the data were normally distributed, t-test was used. Continuous data were expressed as mean and range. The rank-sum test was used to compare continuous and categorical data, and the chi-square test was used to compare categorical data. The ROC curve, AUC, and Harrell's C-index predicted by the nomogram were used to evaluate the discrimination of the model, and the calibration plot was used to evaluate the calibration of the model in the training set and validation set. The c-index ranged from 0.5 (random) to 1.0 (complete discrimination). The performance of the model was verified by 10-fold cross-validation. All analyses were performed using R software (version 3.6.3), and P < 0.05 in all tests indicated statistical significance.
[0084] 2.7 Clinical characteristics of patients
[0085] Table 1. Clinical characteristics of the patients
[0086]
[0087] As shown in Table 1, the training set and validation set were well balanced. No significant differences were observed between the groups in terms of gender, age, or clinical and laboratory variables. Univariate logistic regression analysis found that age, IgG level less than 6.5 g / L, CD4 + T cell count and CD8 + T cell count and lung disease were significantly associated with the occurrence of IFI (p < 0.05, Table 2). Multivariate logistic regression analysis confirmed that age, IgG level (less than 6.5 g / L) and CD4 + T cell count was an independent risk factor (p<0.05, Table 2).
[0088] Table 2. Univariate and multivariate logistic regression analysis of patients with invasive fungal infection in the training set
[0089]
[0090]
[0091] 2.8 Establishment and verification of the nomogram prediction model
[0092] The independent prognostic factors (age, IgG level less than 6.5 g / L and CD4 + T cell count) into the prediction model of IFI probability (i.e., omogram model) ( Figure 2 The model was visualized using a nomogram, which assigned scores based on these factors, and example calculations were provided. The performance of the nomogram was evaluated using the C-index, AUC, and calibration plots, which showed high discrimination and calibration in both groups (training set C-index: 0.72, validation set C-index: 0.73; AUC: training set 0.718, validation set 0.723; Figure 3 and Figure 4 ). Based on IgG levels and CD4 + T cell count, age and IgG level (less than 6.5 g / L), age and CD4 + Prediction model 1, prediction model 2, and prediction model 3 were established for the three groups of indicators of T cell count. The performance of the nomograms of the three prediction models was evaluated using C-index, AUC, and calibration graphs. The results are shown in the figure. Figure 5-7 As shown, all showed low discrimination and calibration (prediction model 1 C-index: 0.67, prediction model 2 C-index: 0.69, prediction model 3 C-index: 0.68; AUC: prediction model 1 is 0.664, prediction model 2 is 0.698, prediction model 3 is 0.682). It can be seen that the prediction model established by the present invention based on independent prognostic factors (age, IgG level less than 6.5g / L and CD4+T cell count) has high discrimination and calibration, and can provide more reliable guidance for clinicians to judge whether patients should be given preventive antifungal treatment.
[0093] In summary, although there is a consensus on the necessity of fungal prevention in immunosuppressed patients, the specific definition of immunosuppression, monitoring indicators and prevention strategies are still unclear. The significance of this study is to identify immunosuppressed patients who have received different degrees of immunosuppressive treatment and determine the independent prognostic factors of IFI from cellular and humoral immune indicators. The nomogram model constructed based on these indicators can guide clinical prevention in high-risk patients.
[0094] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.
[0095] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art may make various modifications or variations within the scope of the claims, which do not affect the essence of the present invention.
Claims
1. A method for constructing a prediction model for invasive fungal infection after immunosuppressive therapy. It is characterized in that The following steps are involved: S1. Collect clinical information and data to obtain a data set, and divide it into a training set and a validation set; S2. Single-factor chi-square test was used to determine the single-factor indicators of invasive fungal infection after immunosuppressive therapy in the training set; S3, performing multivariate logistic regression analysis on the univariate indicators obtained in step S2 to determine the independent risk factors for invasive fungal infection as indicators for establishing a prediction model; S4, constructing a Nomogram model based on the indicators in step S3, which is a prediction model for invasive fungal infection after immunosuppressive therapy; S5. Verify the prediction model through the validation set.
2. The prediction model construction method according to claim 1, in step S1, the clinical information and data include: Age, sex, disease diagnosis, transplant type, chemotherapy regimen, concurrent chronic diseases and other clinical complications, IgG level, IgM level, IgA level, CD4 + / CD8 + T cell count, CD4 + T cell count, CD8 + T cell count, CD19 + T cell count, NK cell count.
3. According to the prediction model construction method according to claim 1, in step S1, the data set is randomly allocated into a training set and a validation set in a ratio of 4:
3.
4. The prediction model construction method according to claim 1, in step S2, the single factor indicators include age, IgG level lower than 6.5g / L, CD4 + T cell count, CD8 + T cell count.
5. The prediction model construction method according to claim 1, in step S3, the independent risk factors include age, IgG level lower than 6.5g / L, CD4 + T cell count.
6. The prediction model construction method according to claim 1, in step S5, the verification uses C-index, AUC and calibration graph to evaluate the performance of the prediction model.
7. The prediction model construction method according to claim 6, wherein the C-index and AUC evaluate the discrimination of the prediction model, and the calibration graph evaluates the calibration of the prediction model in the training set and the validation set.
8. A prediction model construction system for invasive fungal infection after immunosuppressive therapy, It is characterized in that include: Module M1: collect clinical information and data to obtain a data set, and divide it into a training set and a validation set; Module M2, using the one-way chi-square test method to determine the univariate indicators of invasive fungal infection after immunosuppressive therapy in the training set; Module M3: Perform multivariate logistic regression analysis on the univariate indicators obtained in module M2 to determine the independent risk factors for invasive fungal infection as indicators for establishing a prediction model; Module M4: constructing a Nomogram model based on the indicators in module M3, which is a prediction model for invasive fungal infection after immunosuppressive therapy; Module M5: Verify the prediction model through the verification set.
9. A method for predicting invasive fungal infection after immunosuppressive therapy. It is characterized in that Based on the prediction model construction method described in claim 1, the clinical information and data of the prediction object are input into the prediction model to obtain the corresponding nomogram; According to the nomogram, each model indicator variable was assigned a value to obtain the score of each model indicator variable, and the total score was calculated to obtain the corresponding probability of occurrence of invasive fungal infection.
10. An electronic device comprising a memory and a processor and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the steps of the prediction model construction method according to any one of claims 1 to 7 are completed.
