A method and device for constructing a refractory mycoplasma pneumoniae pneumonia prediction model

Abstract

The application belongs to the field of intelligent medical treatment, and particularly relates to a method for constructing a refractory mycoplasma pneumoniae pneumonia prediction model and equipment. The method comprises the following steps: acquiring a baseline time data set and a follow-up time label of a mycoplasma pneumoniae patient; inputting the image data set into a lung blood vessel segmentation model to obtain a lung blood vessel network, obtaining a total lung blood vessel volume based on the lung blood vessel network, and obtaining a blood vessel volume of different cross-sectional areas based on a cross-sectional area of the lung blood vessel network; inputting a ratio of the blood vessel volume of different cross-sectional areas to the total lung blood vessel volume into a machine learning model to obtain a prediction label, and obtaining a refractory mycoplasma pneumoniae pneumonia prediction model after iterative optimization of the machine learning model based on a difference between the prediction label and the follow-up time label. The application segments a lung CT image to obtain a blood vessel network diagram, quantitatively obtains a percentage of a blood vessel volume of different thicknesses in a total lung blood vessel volume, and constructs a refractory mycoplasma pneumoniae pneumonia prediction model based on an innovative feature extraction method.

Description

Technical Field

[0001] This invention relates to the field of intelligent healthcare, and more specifically, to a method, device, medium, and program product for constructing a predictive model for refractory mycoplasma pneumonia. Background Technology

[0002] Mycoplasma pneumoniae (MP) is the smallest pathogenic microorganism between bacteria and viruses. It can survive independently on cell-free culture media, has no cell wall, and takes various forms, such as rod-shaped, spherical or filamentous. It is Gram-negative, facultatively anaerobic, and is mainly transmitted through droplets. It can occur all year round, but is more common in winter and spring. Mycoplasma pneumoniae pneumonia (MPP) is a common pediatric disease. In non-epidemic years, mycoplasma pneumoniae pneumonia accounts for 10% to 20% of community-acquired pneumonia in children, while in epidemic years, this proportion can increase to 30%. Regional epidemics occur every 3 to 7 years [1]. In recent years, there have been numerous reports of refractory mycoplasma pneumoniae pneumonia (RMPP). Conventional macrolide antibiotics are ineffective in treating this disease, and may even worsen the condition. It can involve multiple systems, is chronic, and is difficult to identify and treat in its early stages. If treatment is not timely, it can leave pulmonary sequelae such as bronchiectasis, atelectasis, and obliterative bronchitis. In severe cases, it can lead to pulmonary embolism or even death.

[0003] The standard CT imaging examination for Mycoplasma pneumoniae (MPP) is as follows: Early X-ray findings of MPP may include interstitial pneumonia, increased and blurred lung markings, and reticular shadows. As the condition worsens, lung damage is mainly concentrated in the lower lobes near the hilum, with extensive pulmonary infiltration and consolidation. A small percentage of children may have pleural effusion. Lesions can involve unilateral or bilateral lung tissue; studies show that lesions are more common in the lower lung fields than the upper lung fields, and more common in the right lung than the left. Chest X-rays for MPP are often nonspecific. If a child has severe symptoms but mild physical signs, MPP infection can be considered in conjunction with chest X-ray findings. High-resolution CT can reveal interstitial lesions, pulmonary gas retention, and enlarged hilar lymph nodes in MPP, and increases the likelihood of detecting small amounts of pleural effusion. However, many patients are still hesitant to directly diagnose pneumonia using CT, and this perception needs to be gradually changed. The pulmonary imaging findings of RMPP are more severe than those of MPP. Chest X-rays may show large focal, segmental pneumonia, pulmonary consolidation with atelectasis, and varying degrees of pleural effusion (most commonly moderate to large effusions). Chest CT findings are diverse, showing interstitial infiltration (ground-glass opacities, reticular and irregular linear changes), mostly parenchymal infiltration, large areas of increased density (>2 / 3 of the lung lobes), accompanied by air bronchograms, and moderate to large pleural effusions. Severe cases may present with atelectasis and pulmonary embolism. Therefore, if the pulmonary imaging findings of a child with MPP are severe, the possibility of progression to refractory mycoplasma pneumonia should be considered, and timely administration of hormones should be added to prevent further deterioration leading to atelectasis, bronchiectasis, etc. However, the aforementioned CT examination methods are difficult to quantify and require highly experienced clinicians. Summary of the Invention

[0004] In view of the above problems, the present invention provides a method for constructing a predictive model for refractory mycoplasma pneumonia, which extracts quantitative features from CT images for the prediction of RMPP.

[0005] This application (first aspect) discloses a method for constructing a predictive model for refractory mycoplasma pneumoniae pneumonia, comprising:

[0006] A dataset of baseline time and follow-up time labels for patients with Mycoplasma pneumoniae was obtained. The dataset included an image dataset, and the labels included those indicating progression to refractory Mycoplasma pneumoniae pneumonia and those indicating non-progression to refractory Mycoplasma pneumoniae pneumonia.

[0007] The image dataset is input into the lung vessel segmentation model to obtain the lung vessel network. The total lung vessel volume is obtained based on the lung vessel network. The vessel volume of different cross-sectional areas is obtained based on the cross-sectional area of ​​the vessels in the lung vessel network. The percentage of blood volume in the vessels of different cross-sectional areas to the total lung blood volume is obtained based on the ratio of the vessel volume of different cross-sectional areas to the total lung vessel volume.

[0008] The percentage of intravascular blood volume in different cross-sectional areas relative to total pulmonary blood volume is input into a machine learning model to obtain a prediction label. Based on the difference between the prediction label and the follow-up time label, the machine learning model is iteratively optimized to obtain a prediction model for refractory mycoplasma pneumonia.

[0009] Furthermore, the different cross-sectional areas include one or more of the following: 0mm 2 -5mm 2 5 mm 2 ~10 mm 2 Greater than 10 mm 2 ;

[0010] Optionally, the different cross-sectional areas include: 1.25 mm². 2 -5mm 2 5 mm 2 ~10 mm 2 ;

[0011] Furthermore, the method for obtaining the lung vessel segmentation model is as follows:

[0012] Obtain a training set of lung images, the image set including annotations of lung lobe boundaries;

[0013] Based on the lung image set and lung lobe boundary annotations, the UV-Net model is iteratively trained to obtain a trained lung vessel segmentation model;

[0014] Furthermore, the machine learning model includes any one or more of the following: XGBoost, SVM, LightGBM, logistic regression, K-nearest neighbors algorithm, random forest, and multilayer perceptron;

[0015] Optionally, the machine learning model is XGBoost;

[0016] Optionally, the hyperparameters of the machine learning model can be optimized using k-fold cross-validation to obtain the optimal hyperparameters;

[0017] Optionally, the dataset also includes a clinical feature set. The percentage of intravascular blood volume in different cross-sectional areas relative to total lung blood volume and the clinical feature set are input into a machine learning model to obtain prediction labels. The machine learning model is iteratively optimized based on the difference between the prediction labels and the follow-up time labels to obtain a prediction model for refractory mycoplasma pneumonia. The clinical features include one or more of the following: duration of azithromycin treatment before admission, body temperature, presence of severe pneumonia, CRP level, BV10%, albumin, gamma-glutamyl transferase, platelet distribution width, serum prealbumin, heart rate, creatine kinase, platelet count, hemoglobin, BV5%, BV5-10%, adenosine deaminase, monocyte count, whether glucocorticoids were used before admission, and mixed infection.

[0018] Optionally, SHAP analysis is performed on the refractory mycoplasma pneumoniae pneumonia prediction model to obtain the K most influential features, and the model is reconstructed based on the K most influential features to obtain an optimized refractory mycoplasma pneumoniae pneumonia prediction model, where K is a natural number greater than 1.

[0019] Optionally, K is 5, and the K most influential features are: duration of azithromycin treatment before admission, body temperature, presence of severe pneumonia, CRP level, and BV10%.

[0020] Furthermore, the baseline time datasets and follow-up time labels of Mycoplasma pneumoniae patients in different age groups were obtained. After replacing the baseline time datasets of Mycoplasma pneumoniae patients in different age groups with the baseline time datasets of Mycoplasma pneumoniae patients in different age groups, prediction models for refractory Mycoplasma pneumoniae pneumonia in different age groups were trained.

[0021] The second aspect of this application discloses a method for predicting refractory mycoplasma pneumoniae pneumonia using a predictive model, comprising:

[0022] Obtain images of patients with Mycoplasma pneumoniae pneumonia.

[0023] Based on the images, the percentage of intravascular blood volume to total pulmonary blood volume for different cross-sectional areas was obtained;

[0024] The percentage of intravascular blood volume in different cross-sectional areas relative to total pulmonary blood volume is input into the refractory mycoplasma pneumonia prediction model constructed according to any of the above-described methods to obtain the prediction result.

[0025] Optionally, the age of patients with Mycoplasma pneumoniae pneumonia can be obtained simultaneously, and the data can be input into the prediction model for refractory Mycoplasma pneumoniae pneumonia of the corresponding age group constructed according to the above method to obtain the prediction result.

[0026] The third aspect of this application discloses a method for predicting refractory mycoplasma pneumoniae pneumonia using a predictive model, the method comprising:

[0027] Obtain images of patients with Mycoplasma pneumoniae pneumonia.

[0028] Based on the images, the percentage of intravascular blood volume in vessels with a cross-sectional area of ​​less than 5 mm² relative to the total pulmonary blood volume is obtained.

[0029] If the percentage of intravascular blood volume with a cross-sectional area less than 5 mm² to the total lung blood volume is lower than the BV5 threshold, the result is that the risk of developing refractory mycoplasma pneumoniae pneumonia is high; otherwise, the result is that the risk of developing refractory mycoplasma pneumoniae pneumonia is low.

[0030] Furthermore, the percentage of blood volume in blood vessels with a cross-sectional area of ​​less than 5 mm² relative to the total lung blood volume is input into the classifier to obtain the prediction results.

[0031] The fourth aspect of this application discloses a system for constructing a predictive model for refractory mycoplasma pneumonia, comprising:

[0032] First acquisition module 201: used to acquire the baseline time dataset and follow-up time labels of patients with Mycoplasma pneumoniae. The dataset includes an image dataset, and the labels include those indicating development of refractory Mycoplasma pneumoniae pneumonia and those indicating no development of refractory Mycoplasma pneumoniae pneumonia.

[0033] Feature extraction module 202: is used to input the image dataset into the lung vessel segmentation model to obtain the lung vessel network, obtain the total lung vessel volume based on the lung vessel network, obtain the vessel volume of different cross-sectional areas based on the cross-sectional area of ​​the vessels in the lung vessel network, and obtain the percentage of blood volume in the vessels of different cross-sectional areas to the total lung blood volume based on the ratio of the vessel volume of different cross-sectional areas to the total lung vessel volume.

[0034] Iterative training module 203: is used to input the percentage of intravascular blood volume of different cross-sectional areas to the total lung blood volume into the machine learning model to obtain the prediction label, and to obtain the prediction model of refractory mycoplasma pneumonia by iteratively optimizing the machine learning model based on the difference between the prediction label and the follow-up time label.

[0035] The fifth aspect of this application discloses a system for constructing a predictive model for refractory mycoplasma pneumonia, comprising:

[0036] The second acquisition module is used to acquire images of patients with Mycoplasma pneumoniae pneumonia.

[0037] The second feature extraction module is used to obtain the percentage of intravascular blood volume to total pulmonary blood volume for different cross-sectional areas based on the image.

[0038] The second prediction module inputs the percentage of intravascular blood volume in different cross-sectional areas to the total pulmonary blood volume into the refractory mycoplasma pneumonia prediction model constructed according to the method described above, and obtains the prediction result.

[0039] The sixth aspect of this application discloses a system for constructing a predictive model for refractory mycoplasma pneumonia, comprising:

[0040] The third acquisition module is used to acquire images of patients with Mycoplasma pneumoniae pneumonia.

[0041] The third feature extraction module is used to obtain the percentage of intravascular blood volume with a cross-sectional area of ​​less than 5 mm2 relative to the total lung blood volume based on the image.

[0042] The third prediction module is used to obtain a prediction result based on the percentage of blood volume in blood vessels with a cross-sectional area of ​​less than 5 mm2 to the total lung blood volume. If the percentage of blood volume in blood vessels with a cross-sectional area of ​​less than 5 mm2 to the total lung blood volume is lower than the BV5 threshold, the result is that the risk of mycoplasma pneumoniae pneumonia patients developing refractory mycoplasma pneumoniae pneumonia is high; otherwise, the result is that the risk of developing refractory mycoplasma pneumoniae pneumonia is low.

[0043] A seventh aspect of this application discloses a computer device, the device comprising: a memory and a processor; the memory being used to store program instructions; the processor being used to invoke the program instructions, which, when executed, are used to perform the steps of the method described above.

[0044] The eighth aspect of this application discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0045] The fifth aspect of this application discloses a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0046] This application has the following beneficial effects:

[0047] (1) This application obtains a vascular network map by segmenting lung CT images using a lung vascular segmentation model, and quantifies the percentage of different thicknesses of blood vessel volume in the total lung vascular volume based on the vascular network map, and constructs a prediction model for refractory pneumonia based on an innovative feature extraction method.

[0048] (2) Based on the feature extraction method of this application, important quantitative feature parameters were selected and determined in the process of constructing a prediction model for refractory pneumonia;

[0049] (3) When the model was constructed, the quantitative feature parameters extracted in this application and traditional clinical features were combined to improve the performance of the prediction model. Furthermore, the model was optimized by SHAP analysis of the constructed model. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a schematic diagram of the method flow provided in the first aspect of the present invention;

[0052] Figure 2 This is a schematic diagram of a program product provided in the fourth aspect of the present invention;

[0053] Figure 3 This is a schematic diagram of a computer device provided in an embodiment of the present invention;

[0054] Figure 4 This is a schematic diagram of the architecture of an exemplary computing device provided in an embodiment of the present invention;

[0055] Figure 5 This is a schematic diagram of the storage medium provided in an embodiment of the present invention;

[0056] Figure 6 This is a distribution map of blood vessels with different cross-sectional areas obtained based on a lung vessel segmentation model, provided by an embodiment of the present invention.

[0057] Figure 7 This is a schematic diagram of the predicted AUC on the cross-validation set and external test set of a clinical-BV fusion model, a clinical feature-only model, and an embodiment of the present invention. Detailed Implementation

[0058] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0059] In some of the processes described in the specification, claims, and accompanying drawings of this invention, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as S101, S102, etc., are merely used to distinguish different operations and do not represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to different types.

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Figure 1This is a schematic flowchart of a method for predicting RMPP based on assessment of pulmonary microvascular changes provided by an embodiment of the present invention. Specifically, the method includes the following steps:

[0062] S101: Obtain the baseline time dataset and follow-up time labels of patients with Mycoplasma pneumoniae. The dataset includes an image dataset, and the labels include those indicating progression to refractory Mycoplasma pneumoniae pneumonia and those indicating non-progression to refractory Mycoplasma pneumoniae pneumonia.

[0063] S102: Input the image dataset into the lung vessel segmentation model to obtain the lung vessel network, obtain the total lung vessel volume based on the lung vessel network, obtain the vessel volume of different cross-sectional areas based on the cross-sectional area of ​​the vessels in the lung vessel network, and obtain the percentage of blood volume in the vessels of different cross-sectional areas to the total lung blood volume based on the ratio of the vessel volume of different cross-sectional areas to the total lung vessel volume.

[0064] S103: Input the percentage of intravascular blood volume of different cross-sectional areas to the total pulmonary blood volume into the machine learning model to obtain the prediction label. Based on the difference between the prediction label and the follow-up time label, iteratively optimize the machine learning model to obtain the prediction model for refractory mycoplasma pneumonia.

[0065] The scheme in this application is based on the following research.

[0066] 1. Methods and Materials

[0067] The study has been approved by the institutional ethics committees of both hospitals, and no individual consent was required for the retrospective analysis.

[0068] 1.1. Study Design and Patient Selection

[0069] This retrospective study included consecutively diagnosed pediatric patients with MPP from two medical institutions, forming a cross-validation cohort and an external validation cohort.

[0070] The cross-validation cohort included patients admitted to the general inpatient ward of a tertiary pediatric specialist hospital between July 2019 and December 2023. The external validation cohort included patients hospitalized in the pediatric ward of a tertiary general hospital between January 2023 and April 2024.

[0071] Inclusion criteria were: (1) pediatric patients aged > 28 days to ≤ 18 years; (2) clinically diagnosed with MPP; and (3) all cohorts had a baseline chest CT scan during hospitalization, except for the cross-validation cohort which required a follow-up CT scan within 2 months of discharge. Exclusion criteria were: (1) baseline or follow-up CT image quality was insufficient for quantitative analysis; (2) no laboratory test data available; (3) immunodeficiency or documented prior immunosuppressive therapy; and (4) need for mechanical ventilation during hospitalization. Ultimately, the cross-validation and external validation cohorts included 512 and 124 patients, respectively. The diagnostic criteria for MPP can be found in Supplementary Material 1.1.

[0072] 1.2. Data Collection and Chest CT Imaging

[0073] All patients included in this study were diagnosed with mycoplasmal paroxysmal positional pulmonary embolism (MPP). Clinical, etiological, and laboratory data were retrospectively collected from the digital hospital information system. Clinical variables included demographic characteristics, past comorbidities, duration of fever and cough prior to admission, pre-admission treatment, initial vital signs, and oxygen support requirements at admission. Disease severity was classified according to the 2023 Chinese Pediatric MPP Guidelines. Pathogen detection results (including viruses, bacteria, and atypical pathogens) were recorded. Laboratory variables obtained at admission included complete blood count, liver and kidney function tests, inflammatory markers, coagulation tests, and electrolytes.

[0074] Clinical outcomes, complications, and hospitalizations (e.g., glucocorticoids, intravenous immunoglobulin, bronchoscopy) were extracted from discharge records. For outcomes, RMPP was defined as persistent fever, worsening clinical symptoms, progression of lung imaging, or extrapulmonary complications after ≥ 7 days of macrolide therapy. Patients were divided into RMPP and non-RMPP groups.

[0075] Thin-slice (≤1.5 mm), non-contrast chest CT images were acquired and analyzed from image archives and communication systems. The earliest chest CT acquired during hospitalization was selected as the baseline CT. For the cross-validation cohort, baseline and follow-up chest CT scans within 2 months after discharge were retrospectively collected. For the external validation cohort, only the baseline chest CT scan was acquired.

[0076] 1.3. Quantitative pulmonary vascular analysis

[0077] Lung vessels were automatically segmented using a previously validated UV-Net-based segmentation model. The detailed method steps include:

[0078] This study uses the deep learning framework UV-Net for three-dimensional pulmonary vessel segmentation.

[0079] This model incorporates a U-shaped network architecture, including a 2D encoder module and a 3D decoder module, which effectively adapts to data features. To expand the receptive field and incorporate global contextual information, the network includes an Atrous Spatial Pyramid Pooling (ASPP) module, which fuses multi-scale feature maps to preserve spatial and semantic details. For upsampling, the model employs the PixelShuffle technique to restore spatial resolution while maintaining detailed anatomical structures, ensuring continuity and topological consistency in the segmentation of pulmonary vessels and airways.

[0080] The segmentation process first involves the preliminary delineation of lung lobes with well-defined boundaries. Based on the delineated lung lobe boundaries, boundary representation B-rep data is constructed. The B-rep data structure contains multiple topological entities—faces, edges, half-edges, and vertices—as well as the connections between them, and interview parameter surfaces (including vascular surfaces in our study).

[0081] Topology UV-Net uses a face adjacency graph derived from B-rep to model the topology, where vertices V represent faces in B-rep and edges E encode the connectivity between faces.

[0082] Airway segmentation is then performed to preserve the tree-like connectivity of the bronchial structure.

[0083] In addition, the vessel segmentation module generates a complete and connected vessel network, which can accurately quantify vessel volume based on cross-sectional area.

[0084] During internal validation, the algorithm demonstrated high segmentation accuracy, with a Dice coefficient of 92.68% for arteries and 89.96% for veins. These results highlight the robustness and clinical applicability of UV-Net in pulmonary vascular analysis.

[0085] The segmented pulmonary vessels were then divided into three categories based on their cross-sectional area: <5 mm² (BV5), 5–10 mm² (BV5–10), and >10 mm² (BV10). Figure 6 Calculate the blood volume in each category and express it as a percentage of total lung blood volume, denoted as BV5%, BV5–10%, and BV10%, respectively.

[0086] 1.4. Development and Validation of the RMPP Prediction Model

[0087] A predictive model for RMPP was developed using clinical variables and PBV parameters.

[0088] First, candidate predictors collected at admission (including demographics, clinical symptoms, previous treatments, vital signs, laboratory tests, and BV parameters) were screened, and predictors with missing data >40% in the cross-validation cohort or external validation cohort were excluded.

[0089] To develop a predictive model for refractory mycoplasma pneumonia (RMPP), we evaluated five machine learning algorithms: Extreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), Lightweight Gradient Boosting Machine (LightGBM), Logistic Regression (LR), and K-Nearest Neighbors (KNN). XGBoost, officially released in 2016, is an advanced ensemble learning algorithm that constructs multiple classification and regression trees within an optimized distributed gradient boosting framework. Compared to traditional machine learning algorithms, it offers higher complexity and superior performance. LightGBM is a high-performance distributed gradient boosting framework based on decision trees, widely used for ranking, classification, and other machine learning tasks. LR is a widely used linear classification model for evaluating the relationship between a dependent variable and multiple independent variables, particularly suitable for binary classification problems. SVM is a supervised learning algorithm designed to construct an optimal hyperplane to separate positive and negative samples, performing well in binary classification and high-dimensional datasets. KNN is an instance-based nonparametric learning algorithm that classifies samples based on their proximity to labeled instances. It is simple yet effective in handling multi-class problems and problems involving nonlinear decision boundaries.

[0090] For these models, we included 19 variables selected through minimum absolute shrinkage and operator-selective regression, and optimized the hyperparameters using 5-fold cross-validation based on average performance metrics. Subsequently, the final models were retrained on the entire cross-validation cohort using the determined optimal hyperparameters. Receiver Operating Characteristic (ROC) curve analysis was used to comprehensively evaluate model performance, with key metrics including area under the ROC curve (AUC), accuracy, precision, sensitivity, specificity, and F1 score. Evaluations were conducted in both the cross-validation and external validation cohorts. XGBoost performed best on both datasets and was therefore selected as the final analysis algorithm.

[0091] All models were developed in R software (version 4.3.1) using packages such as “caret”, “glmnet”, “knn”, “e1071”, “lightgbm”, and “xgboost”.

[0092] Based on the selected predictors, two models were constructed: (1) a clinical model containing 16 selected clinical variables and (2) a clinical BV model incorporating all 19 variables. For model development, the model hyperparameters were optimized through 5x cross-validation and retrained on the full cross-validation cohort with the determined optimal hyperparameters.

[0093] The 16 selected clinical variables were: duration of azithromycin treatment before admission, body temperature, presence of severe pneumonia, CRP level, BV10%, albumin, gamma-glutamyl transferase, platelet distribution width, serum prealbumin, heart rate, creatine kinase, platelet count, hemoglobin, BV5%, BV5-10%, adenosine deaminase, monocyte count, whether glucocorticoids were used before admission, and mixed infection.

[0094] The 19-variable clinical-BV model includes 16 selected clinical variables and 3 PBV parameters.

[0095] Model performance was assessed internally and externally using receiver operating characteristic (ROC) curve analysis, calculating the area under the ROC curve (AUC), accuracy, precision, sensitivity, specificity, and F1 score for both cohorts. The DeLong test was used to compare the AUC differences between the clinical and clinical BV models. Shapley additive interpretation (SHAP) values ​​were calculated to quantify the feature importance in the clinical BV model.

[0096] 1.5. Statistical Analysis

[0097] Statistical analysis was performed using R software (version 4.3.1). Chi-square tests were used for comparisons of categorical variables, and Student's t-test (for normally distributed data) or Wilcoxon rank-sum test (for non-normally distributed data) were used for comparisons of continuous variables. Holm's method was used for correction of multiple comparisons. Missing data were imputed using the multiple imputation method in R's "mice" package.

[0098] The association between PBV parameters and RMPP was investigated in a cross-validation cohort. Numerical variables were standardized using z-score standardization, and univariate logistic regression was performed to initially examine the association. Clinically relevant variables with <40% missing data were selected using LASSO regression (“glmnet” package) and included in a multivariate model to independently assess BV parameters. Furthermore, subgroup analyses were performed for the 5–10 year age group based on age distribution, age-related immune differences, and the higher prevalence of MPP in children over 5 years of age in the cross-validation cohort. Pearson correlation analysis further assessed the relationship between significant BV parameters and key laboratory indicators (D-dimer, CRP, LDH) and duration of fever.

[0099] Develop XGBoost models using the "caret" and "xgboost" packages. Perform SHAP analysis using the "SHAPforxgboost" package. Perform ROC analysis using the "pROC" package. A p-value less than 0.05 is considered statistically significant.

[0100] 2. Results

[0101] 2.1. Comparison of clinical and laboratory characteristics between RMPP and non-RMPP

[0102] As shown in Table 1, the cross-validation cohort included 512 patients (256 men and 256 women; median age 6.42 years [IQR, 4.33–8.25]), of whom 265 were classified as non-RMPP and 247 as RMPP. Patients in the RMPP group had significantly longer durations of fever and cough prior to admission, and a higher incidence of severe / critical pneumonia. Furthermore, these patients were more likely to require oxygen support and had elevated temperature, pulse, and respiratory rate (all p ≤ 0.01). In terms of clinical outcomes, compared to the non-RMPP group, the RMPP group had significantly longer hospital stays, prolonged total fever duration, and a higher incidence of pulmonary complications (all p ≤ 0.03).

[0103] Regarding etiology, there were no statistically significant differences among the groups (all p > 0.05). Laboratory results showed that patients with RMPP had significantly higher levels of CRP, erythrocyte sedimentation rate (ESR), neutrophil count, LDH, and D-dimer (all p ≤ 0.03).

[0104] 2.2. PBV parameter analysis at baseline and follow-up

[0105] At baseline, the BV5% in the RMPP group was significantly lower than that in the non-RMPP group (58.50% vs. 60.63%, p = 0.007), while the BV5-10% and BV10% in the RMPP group were both significantly higher than those in the non-RMPP group (median BV5-10%: 15.74% vs. 15.13%, p = 0.03; median BV10%: 20.90% vs. 19.39%, p = 0.004).

[0106] The median interval for follow-up CT scans was 16.00 days [13.00, 23.00]. Follow-up CT scans showed a significant increase in BV5% relative to baseline (p < 0.001). Conversely, both BV5-10% and BV10% values ​​on follow-up CT scans were significantly lower than baseline (all p < 0.01), suggesting that lung blood distribution may have recovered over time. Notably, during the follow-up period, BV5% in the RMPP group remained lower than that in the non-RMPP group (median: 62.42% vs. 63.55%, p = 0.03), while BV10% was higher in the RMPP group (Table 1).

[0107] Table 1. Comparison of quantitative parameters of pulmonary blood volume during follow-up in patients with refractory and non-refractory mycoplasma pneumoniae in cross-validation cohorts.

[0108]

[0109] Table Note: Unless otherwise stated, data are presented in terms of patient numbers, with percentages in parentheses. BV5%, BV5-10%, and BV10% represent the percentage of total lung blood volume contained in vessels with cross-sectional areas less than 5 mm², 5-10 mm², and greater than 10 mm², respectively. * indicates p < 0.05.

[0110] 2.3. Correlation analysis of RMPP and quantitative characteristics of pulmonary blood volume and its association with RMPP-related risk factors.

[0111] As shown in Table 2, in the cross-validation cohort, univariate analysis showed that each PBV parameter was significantly associated with the occurrence of RMPP (all p ≤ 0.03). In the subsequent multivariate logistic regression analysis, after adjusting for clinical covariates, BV5% remained an independent protective factor against RMPP (odds ratio [OR] = 0.70, 95% confidence interval [CI]: 0.54–0.89, p = 0.005), while BV10% became an independent risk factor (OR = 1.49, 95% CI: 1.17–1.92, p = 0.002). Similarly, in patients aged 5–10 years, after adjusting for relevant covariates, BV5% continued to show a protective association with RMPP (OR = 0.68; 95% CI: 0.49–0.83; p = 0.018), while BV10% remained a significant risk factor (OR = 1.58; 95% CI: 1.15–2.19; p = 0.005).

[0112] Table 2. Univariate and multivariate logistic regression analyses of quantitative pulmonary blood volume parameters associated with refractory mycoplasma pneumonia in the cross-validation cohort.

[0113]

[0114] Table Notes: The adjusted model for the entire cohort included the following covariates: duration of azithromycin treatment before admission, severe pneumonia, body temperature, platelet distribution width, albumin, serum prealbumin, prothrombin time, thrombin time, serum phosphorus, sodium, uric acid, and cystatin C. For patients aged >5 years and <10 years, the adjusted model included the following covariates: duration of azithromycin treatment before admission, duration of fever before admission, severe pneumonia, body temperature, platelet distribution width, albumin, adenosine deaminase, prothrombin time, serum phosphorus, uric acid, and cystatin C. Abbreviation: CI, confidence interval. BV5%, BV5–10%, and BV10% represent the percentage of pulmonary blood volume in vessels with cross-sectional areas of 1.25–5 mm², 5–10 mm², and >10 mm², respectively. * indicates p < 0.05.

[0115] Furthermore, as shown in Figure 4, correlation analysis revealed that BV5% was significantly negatively correlated with key inflammatory and coagulation biomarkers, including NR, CRP, and D-dimer (r = –0.12, –0.13, and –0.19, respectively; all p < 0.01). Conversely, BV10% was significantly positively correlated with these biomarkers (r = 0.14, 0.14, and 0.21, respectively; all p < 0.01).

[0116] 2.4. XGBoost-based RMPP prediction enhanced by quantitative PBV parameters

[0117] To improve the predictive performance of RMPP, we incorporated clinical variables and quantitative PBV features into an XGBoost-based model. Feature selection using LASSO regression identified 19 variables, including 16 clinical features and 3 BV-derived features.

[0118] Table 3. Performance metrics of the XGBoost model in the cross-validation queue and external validation queue.

[0119]

[0120] Table Notes: This table presents the performance metrics of the Extreme Gradient Boosting (XGBoost) model in the clinical-blood volume model and the clinical model. The clinical-blood volume model included all 19 variables selected by LASSO regression (16 clinical variables and 3 quantitative features derived from blood volume), while the clinical model used only 16 clinical variables. Metrics include the area under the curve (AUC) and its 95% confidence interval (CI), accuracy, sensitivity, specificity, precision, and F1 score, presented in both the cross-validation and external validation cohorts. p-values ​​were obtained using the DeLong test, which assesses the difference in AUC between the clinical-blood volume model and the clinical model in the cross-validation and external validation cohorts. * indicates p < 0.05.

[0121] As shown in Table 3, Figure 7 As shown, in the cross-validation cohort, the clinical-BV model had a significantly higher AUC than the clinical model (AUC: 0.91 vs. 0.88, p < 0.001), and also had higher accuracy, precision, sensitivity, specificity, and F1 score.

[0122] Table 4 Comparison of Predicted AUC for Different Indicators

[0123]

[0124] SHAP analysis of the clinical-BV model showed that the five most influential features in the cross-validation cohort were the duration of azithromycin treatment before admission, body temperature, presence of severe pneumonia, CRP level, and BV10%, highlighting the importance of inflammatory markers and BV characteristics.

[0125] To assess the generalizability of the model, we evaluated its performance in an independent external validation cohort comprising 124 patients (median age: 7.00 years [IQR: 5.00–9.00], 52% male). In this cohort, the clinical-BV model had a higher AUC than the clinical model, although the difference was not statistically significant (AUC: 0.84 vs. 0.81, p = 0.09), demonstrating superior performance in accuracy, precision, specificity, and F1 score. Furthermore, as shown in Table 4, SHAP analysis of the clinical-BV model revealed that the AUC of the five most influential factors reached 0.90, highlighting that the combination of the quantitative PBV parameters and clinical indicators identified in this application can achieve better predictive results.

[0126] 3. Discussion

[0127] In this study, RMPP patients exhibited a consistent pattern of microvascular remodeling on CT scans, characterized by a 5% decrease in BV (microvessels < 5 mm) at baseline and follow-up. 2 ) and BV 10% increase (large vessels > 10 mm) 2 These changes indicate altered pulmonary vascular volume distribution, a change also observed in patients with acute pulmonary microcirculatory disturbances during COVID-19. Furthermore, multivariate analysis revealed that BV5%, negatively correlated with inflammation and coagulation biomarkers, was an independent protective predictor of RMPP. Conversely, BV10% emerged as an independent risk factor, positively correlated with these same biomarkers. When integrated into an XGBoost-based model, these quantitative PBV parameters significantly improved RMPP prediction in the cross-validation cohort compared to models using only clinical variables. Although the AUC difference was not statistically significant in the external validation cohort, the clinical-BV model still outperformed the clinical model in terms of accuracy, precision, specificity, and F1 score. These findings suggest that incorporating PBV parameters can enhance the predictive performance of RMPP and maintain clinical applicability in independent cohorts.

[0128] In patients with RMPP, we observed a significant decrease in BV5%, while BV5–10% and BV10% increased. Consistent with these CT findings, histopathological analysis in a mouse model of recurrent MP infection revealed thickening of the pulmonary arteriolar walls and narrowing of the lumen, leading to elevated pulmonary artery pressure. Microscopically, pulmonary arterioles were defined as vessels with a diameter less than 500 μm, corresponding to the subset of vessels classified as BV5% on CT. The consistency between microscopic and imaging observations further suggests that patients with RMPP do indeed exhibit a reduction in pulmonary microvascular volume.

[0129] Given that the pathogenesis of MPP involves direct toxicity, immune-mediated damage, and vascular inflammation / thrombosis, we hypothesized that the observed decrease in BV5% (reflecting reduced microvascular volume in RMPP patients) may be related to microvascular endothelial inflammation or thrombosis induced by enhanced coagulation and inflammatory states. The observed increase in BV10% may reflect compensatory dilation of larger vessels due to reduced microvascular volume. This hypothesis is supported by three pieces of indirect evidence. First, RMPP patients in the cross-validation cohort showed significantly elevated D-dimer, CRP, ESR, and neutrophil counts. Second, correlation analysis further confirmed these findings, revealing a negative correlation between BV5% and inflammatory or coagulation markers. Finally, all recorded vascular thrombotic events occurred in the RMPP group. Furthermore, SHAP value analysis using the XGBoost model showed that BV10% was more important than BV5%. This observation suggests that macrovascular dilation is an active compensatory response, not merely a passive consequence.

[0130] Compared to traditional pneumonia volume analysis (such as consolidation volume), quantitative pulmonary vascular analysis can reveal more subtle pathophysiological changes. Traditional volume analysis primarily reflects solid tissue damage but has limited understanding of the underlying pathogenic mechanisms. In contrast, pulmonary vascular analysis, especially by reducing BV by 5%, can effectively capture the reduction in microvessel volume in patients with RMPP. This reduction highlights specific pathophysiological processes in RMPP, where enhanced coagulation and inflammatory states may affect pulmonary microvessels, leading to microvessel volume reduction. Furthermore, incorporating BV parameters into clinical prediction models improves early RMPP identification (including clinical BV models (constructed from 16 clinical features + 3 BV parameters) and the most influential 5-feature model). Clinical BV models demonstrate superior performance compared to clinical models alone (AUC: 0.91 vs. 0.88), enabling timely use of immunomodulators or second-line antibiotics and potentially preventing complications such as necrotizing pneumonia. Moreover, these quantitative vascular measurements can be obtained through routine CT scans, requiring no additional examinations or radiation exposure. Furthermore, compared to non-RMPP patients, RMPP patients continued to show a relative decrease in BV5% during follow-up (median duration: 16 days), similar to the vascular recovery pattern observed in post-COVID-19 studies, where 87.4% of patients still had residual vascular abnormalities at 6 months. This further underscores the necessity of continued vascular monitoring after symptom relief.

[0131] 4. Conclusion

[0132] This study demonstrates that pulmonary vascular changes (RMPP) in children are associated with significant pulmonary microvascular alterations, characterized by decreased BV5% and increased BV10%. These quantitative PBV parameters can serve as independent predictors of RMPP, enhancing the performance of XGBoost-based predictive models and can be easily integrated into routine CT workflows. Assessing pulmonary vascular changes via routine CT imaging is a non-invasive method that can help identify high-risk patients and guide early risk stratification and intervention.

[0133] In some embodiments, the BV5% and BV10% of the training set, and the corresponding labels are obtained, and a first classifier is obtained through iterative training as a prediction model for refractory mycoplasma pneumonia. When used, the BV5% and BV10% of the subject are obtained and input into the first classifier to obtain the predicted label.

[0134] In some embodiments, the duration of azithromycin treatment before admission, body temperature, presence of severe pneumonia, CRP level, and BV10% of the training set, along with the corresponding labels, are obtained. A second classifier is obtained through iterative training as a prediction model for refractory mycoplasma pneumonia. When used, the duration of azithromycin treatment before admission, body temperature, presence of severe pneumonia, CRP level, and BV10% of the subject are obtained and input into the second classifier to obtain the predicted label.

[0135] Figure 3 This is a schematic diagram of a computer device provided in an embodiment of the present invention, such as... Figure 3 As shown, the device 2000 may include: one or more processors 2010 and one or more memories 2020; wherein the memories store computer-readable code that, when run by the one or more processors, can perform the methods described above.

[0136] The processor in this embodiment can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, operations, and logic block diagrams disclosed in this embodiment. The general-purpose processor can be a microprocessor or any conventional processor, and can be based on an x86 or ARM architecture.

[0137] In general, the various exemplary embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, firmware, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device. When aspects of embodiments of this disclosure are illustrated or described as block diagrams, flowcharts, or using some other graphical representation, it will be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as non-limiting examples in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0138] For example, the method or apparatus according to embodiments of this disclosure can also be used by means of Figure 4 The architecture of the computing device 3000 shown is used for implementation. For example... Figure 4 As shown, the computing device 3000 may include a bus 3010, one or more CPUs 3020, a read-only memory (ROM) 3030, a random access memory (RAM) 3040, a communication port 3050 connected to a network, an input / output component 3060, a hard disk 3070, etc. The storage devices in the computing device 3000, such as the ROM 3030 or the hard disk 3070, may store various data or files used for processing and / or communication of the methods provided in this disclosure, as well as program instructions executed by the CPU. The computing device 3000 may also include a user interface 3080. Of course, Figure 4 The architecture shown is merely exemplary and can be omitted as needed when implementing different devices. Figure 4 One or more components in the computing device shown.

[0139] This invention also includes a computer-readable storage medium, such as... Figure 5The diagram illustrates a storage medium 4000 provided in an embodiment of the present invention. The computer storage medium 4020 stores computer-readable instructions 4010. When the computer-readable instructions 4010 are executed by a processor, the method described above according to embodiments of the present disclosure can be performed. The computer-readable storage medium in the embodiments of the present disclosure may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. Non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), Synchronous Link Dynamic Random Access Memory (SLDRAM), and Direct Memory Bus Random Access Memory (DR RAM). It should be noted that the memory used in the methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0140] This disclosure also provides a computer program product or computer program that, when executed by a processor, implements the steps of the above-described method, such as... Figure 2 As shown, the computer program product or computer program includes:

[0141] First acquisition module 201: used to acquire the baseline time dataset and follow-up time labels of patients with Mycoplasma pneumoniae. The dataset includes an image dataset, and the labels include those indicating development of refractory Mycoplasma pneumoniae pneumonia and those indicating no development of refractory Mycoplasma pneumoniae pneumonia.

[0142] Feature extraction module 202: is used to input the image dataset into the lung vessel segmentation model to obtain the lung vessel network, obtain the total lung vessel volume based on the lung vessel network, obtain the vessel volume of different cross-sectional areas based on the cross-sectional area of ​​the vessels in the lung vessel network, and obtain the percentage of blood volume in the vessels of different cross-sectional areas to the total lung blood volume based on the ratio of the vessel volume of different cross-sectional areas to the total lung vessel volume.

[0143] Iterative training module 203: is used to input the percentage of intravascular blood volume of different cross-sectional areas to the total lung blood volume into the machine learning model to obtain the prediction label, and to obtain the prediction model of refractory mycoplasma pneumonia by iteratively optimizing the machine learning model based on the difference between the prediction label and the follow-up time label.

[0144] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0145] In general, the various exemplary embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, firmware, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device. When aspects of embodiments of this disclosure are illustrated or described as block diagrams, flowcharts, or using some other graphical representation, it will be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as non-limiting examples in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0146] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0147] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0148] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0149] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0150] The exemplary embodiments of this disclosure described in detail above are merely illustrative and not restrictive. Those skilled in the art will understand that various modifications and combinations can be made to these embodiments or their features without departing from the principles and spirit of this disclosure, and such modifications should fall within the scope of this disclosure.

Claims

1. A method for constructing a predictive model for refractory mycoplasma pneumoniae pneumonia, characterized in that, The method includes: A dataset of baseline time and follow-up time labels for patients with Mycoplasma pneumoniae was obtained. The dataset included an image dataset, and the labels included those indicating progression to refractory Mycoplasma pneumoniae pneumonia and those indicating non-progression to refractory Mycoplasma pneumoniae pneumonia. The image dataset is input into the lung vessel segmentation model to obtain the lung vessel network. The total lung vessel volume is obtained based on the lung vessel network. The vessel volume of different cross-sectional areas is obtained based on the cross-sectional area of ​​the vessels in the lung vessel network. The percentage of blood volume in the vessels of different cross-sectional areas to the total lung blood volume is obtained based on the ratio of the vessel volume of different cross-sectional areas to the total lung vessel volume. The percentage of intravascular blood volume in different cross-sectional areas relative to total pulmonary blood volume is input into a machine learning model constructed using XGBoost to obtain a prediction label. The machine learning model is then iteratively optimized based on the difference between the prediction label and the follow-up time label to obtain a prediction model for refractory mycoplasma pneumonia.

2. The method for constructing a predictive model for refractory mycoplasma pneumonia according to claim 1, characterized in that, The different cross-sectional areas include one or more of the following: 0mm 2 -5mm 2 5 mm 2 ~10mm 2 Greater than 10mm 2 .

3. The method for constructing a predictive model for refractory mycoplasma pneumonia according to claim 2, characterized in that, The different cross-sectional areas include: 1.25mm. 2 -5mm 2 5 mm 2 ~10mm 2 .

4. The method for constructing a predictive model for refractory mycoplasma pneumoniae pneumonia according to claim 1, characterized in that, The method for obtaining the lung vessel segmentation model is as follows: Obtain a training set of lung images, the image set including annotations of lung lobe boundaries; Based on the lung image set and lung lobe boundary annotations, the UV-Net model is iteratively trained to obtain a trained lung vessel segmentation model.

5. The method for constructing a predictive model for refractory mycoplasma pneumoniae pneumonia according to claim 1, characterized in that, The optimal hyperparameters of the machine learning model were obtained by optimizing the hyperparameters using k-fold cross-validation.

6. The method for constructing a predictive model for refractory mycoplasma pneumonia according to claim 1, characterized in that, The dataset also includes a clinical feature set. The percentage of intravascular blood volume in different cross-sectional areas relative to total lung blood volume and the clinical feature set are input into a machine learning model to obtain prediction labels. The machine learning model is iteratively optimized based on the difference between the prediction labels and the follow-up time labels to obtain a prediction model for refractory mycoplasma pneumonia. The clinical features include one or more of the following: duration of azithromycin treatment before admission, body temperature, presence of severe pneumonia, CRP level, BV10%, albumin, gamma-glutamyl transferase, platelet distribution width, serum prealbumin, heart rate, creatine kinase, platelet count, hemoglobin, BV5%, BV5-10%, adenosine deaminase, monocyte count, whether glucocorticoids were used before admission, and mixed infection.

7. The method for constructing a predictive model for refractory mycoplasma pneumonia according to claim 1, characterized in that, The method further includes: performing SHAP analysis on the refractory mycoplasma pneumoniae pneumonia prediction model to obtain the K most influential features, and reconstructing the model based on the K most influential features to obtain an optimized refractory mycoplasma pneumoniae pneumonia prediction model, where K is a natural number greater than 1.

8. The method for constructing a predictive model for refractory mycoplasma pneumonia according to claim 7, characterized in that, K is 5, and the K most influential features are: duration of azithromycin treatment before admission, body temperature, presence of severe pneumonia, CRP level, and BV10%.

9. The method for constructing a predictive model for refractory mycoplasma pneumonia according to claim 1, characterized in that, Data sets of baseline time and follow-up time labels of Mycoplasma pneumoniae patients in different age groups were obtained. After replacing the baseline time data set of Mycoplasma pneumoniae patients in different age groups with the baseline time data set of Mycoplasma pneumoniae patients in different age groups, prediction models of refractory Mycoplasma pneumoniae pneumonia in different age groups were trained.

10. A method for predicting refractory mycoplasma pneumoniae pneumonia, characterized in that, The method includes: Obtain images of patients with Mycoplasma pneumoniae pneumonia. Based on the images, the percentage of intravascular blood volume to total pulmonary blood volume for different cross-sectional areas was obtained; The percentage of intravascular blood volume in the blood vessels with different cross-sectional areas relative to the total pulmonary blood volume is input into the refractory mycoplasma pneumonia prediction model constructed according to any one of claims 1-9 to obtain the prediction result.

11. The method for predicting refractory mycoplasma pneumoniae pneumonia according to claim 10, characterized in that, The method further includes: simultaneously obtaining the age of patients with Mycoplasma pneumoniae pneumonia, and inputting the data into the corresponding age group refractory Mycoplasma pneumoniae pneumonia prediction model constructed according to the method described in claim 9 to obtain the prediction result.

12. A method for predicting refractory mycoplasma pneumoniae pneumonia, characterized in that, The method includes: Obtain images of patients with Mycoplasma pneumoniae pneumonia. Based on the image, the cross-sectional area is less than 5mm. 2 The percentage of intravascular blood volume to total pulmonary blood volume, greater than 10 mm. 2 The percentage of intravascular blood volume in the lungs relative to the total lung blood volume; The image has a cross-sectional area of ​​less than 5 mm². 2 The percentage of intravascular blood volume to total pulmonary blood volume, greater than 10 mm. 2 The percentage of intravascular blood volume to total lung blood volume is input into the refractory mycoplasma pneumoniae pneumonia prediction model constructed according to any one of claims 1-9 to obtain the results of the risk of mycoplasma pneumoniae pneumonia patients developing refractory mycoplasma pneumoniae pneumonia.

13. A computer device, characterized in that, The device includes: a memory and a processor; the memory is used to store a computer program; the processor executes the computer program to implement the steps of the method according to any one of claims 1-12.

14. A computer-readable storage medium, characterized in that, It stores a computer program thereon, which, when executed by a processor, implements the steps of the method as described in any one of claims 1-12.

15. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method described in any one of claims 1-12.

Images