Jun Hua
Jun Hua has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
Hong-Mei Chen
Department of Emergency Medicine, Children's Hospital of Soochow University
1 shared publication- Jian Pan
Institute of Pediatric Research, Children’s Hospital of Soochow University , Suzhou
1 shared publication - Jian Wang
Department of General Surgery, Children’s Hospital of Soochow University, Suzhou, China
1 shared publication Li Shen
Department of Pharmacy, Suzhou Research Center of Medical School, Suzhou Hospital, Affiliated Hospital of Medical School, Nanjing University, Suzhou, China
1 shared publicationLi-Nong Wang
Department of Emergency Medicine, Children's Hospital of Soochow University
1 shared publication- Lin-Lin Li
Department of Respiratory Medicine, Children's Hospital of Nanjing Medical University, Nanjing, China
1 shared publication
Latest Publications
- Open Access
- 31 August 2026
- Published by Frontiers Media SA in Frontiers in Pediatrics
- Vol. 14, 1933395
Development and validation of an explainable XGBoost model for early mortality prediction in pediatric sepsis
Objective: To develop and internally validate an explainable machine learning model for early mortality prediction in pediatric sepsis using routinely available clinical variables, to support clinical risk stratification and decision-making. Methods: A total of 752 children with sepsis admitted to the Pediatric Intensive Care Unit (PICU) at the Children's Hospital of Soochow University between January 1, 2019, and June 30, 2023, were retrospectively enrolled. Twenty-six routinely available clinical variables obtained within 24 h after PICU admission were included. After data preprocessing, variables with excessive missing values, significant multicollinearity, or limited univariate predictive value were excluded. Clinically meaningful ratio-derived variables were constructed, and logarithmic transformation was applied to selected right-skewed continuous variables. Twelve supervised machine learning algorithms were developed and compared for mortality prediction. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), precision–recall (PR) curve, decision curve analysis (DCA), and other performance metrics. Model interpretability was assessed using SHapley Additive exPlanations (SHAP). Results: Among the 12 machine learning algorithms, extreme gradient boosting (XGBoost) achieved the best overall predictive performance. Following recursive feature elimination based on permutation importance, a simplified XGBoost model incorporating seven variables was established and interpreted using SHAP. In the internal validation cohort, the final model achieved an AUC of 0.803, with a sensitivity of 0.846, specificity of 0.696, accuracy of 0.722, and an F1 score of 0.512. The mean AUCs obtained from 5-fold and 10-fold cross-validation were 0.780 ± 0.044 and 0.771 ± 0.070, respectively, indicating good model stability. SHAP analysis identified pH, PaO₂, glucose, blood urea nitrogen (BUN), lactate dehydrogenase (LDH), international normalized ratio (INR), and the BUN-to-creatinine ratio (BUN/Cr) as the features that made the greatest contributions to the final model's mortality predictions. Conclusion: An explainable XGBoost-based prediction model was developed for early mortality prediction in pediatric sepsis using seven routinely available clinical variables collected within 24 h after PICU admission. The model demonstrated good discrimination, stability, and model-based interpretability, suggesting its potential as a practical tool for early risk stratification and individualized clinical management. Further multicenter external validation is warranted before routine clinical implementation.
- 31 August 2026
- Published by Frontiers Media SA in Frontiers in Pediatrics
- Vol. 14, 1933395
Open AccessDevelopment and validation of an explainable XGBoost model for early mortality prediction in pediatric sepsis
Objective: To develop and internally validate an explainable machine learning model for early mortality prediction in pediatric sepsis using routinely available clinical variables, to support clinical risk stratification and decision-making. Methods: A total of 752 children with sepsis admitted to the Pediatric Intensive Care Unit (PICU) at the Children's Hospital of Soochow University between January 1, 2019, and June 30, 2023, were retrospectively enrolled. Twenty-six routinely available clinical variables obtained within 24 h after PICU admission were included. After data preprocessing, variables with excessive missing values, significant multicollinearity, or limited univariate predictive value were excluded. Clinically meaningful ratio-derived variables were constructed, and logarithmic transformation was applied to selected right-skewed continuous variables. Twelve supervised machine learning algorithms were developed and compared for mortality prediction. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), precision–recall (PR) curve, decision curve analysis (DCA), and other performance metrics. Model interpretability was assessed using SHapley Additive exPlanations (SHAP). Results: Among the 12 machine learning algorithms, extreme gradient boosting (XGBoost) achieved the best overall predictive performance. Following recursive feature elimination based on permutation importance, a simplified XGBoost model incorporating seven variables was established and interpreted using SHAP. In the internal validation cohort, the final model achieved an AUC of 0.803, with a sensitivity of 0.846, specificity of 0.696, accuracy of 0.722, and an F1 score of 0.512. The mean AUCs obtained from 5-fold and 10-fold cross-validation were 0.780 ± 0.044 and 0.771 ± 0.070, respectively, indicating good model stability. SHAP analysis identified pH, PaO₂, glucose, blood urea nitrogen (BUN), lactate dehydrogenase (LDH), international normalized ratio (INR), and the BUN-to-creatinine ratio (BUN/Cr) as the features that made the greatest contributions to the final model's mortality predictions. Conclusion: An explainable XGBoost-based prediction model was developed for early mortality prediction in pediatric sepsis using seven routinely available clinical variables collected within 24 h after PICU admission. The model demonstrated good discrimination, stability, and model-based interpretability, suggesting its potential as a practical tool for early risk stratification and individualized clinical management. Further multicenter external validation is warranted before routine clinical implementation.
- Open Access
- 10 July 2024
- Published by Frontiers Media SA in Frontiers in Pediatrics
- Vol. 12, 1397750
Establishment and validation of a prediction model for apnea on bronchiolitis
Objective: The objective of this study is to examine the risk factors associated with apnea in hospitalized patients diagnosed with bronchiolitis and to develop a nomogram prediction model for the early identification of patients who are at risk of developing apnea.Methods: The clinical data of patients diagnosed with acute bronchiolitis and hospitalized at the Children's Hospital of Nanjing Medical University between February 2018 and May 2021 were retrospectively analyzed. LASSO regression and logistic regression analysis were used to determine the risk factors for apnea in these patients. A nomogram was constructed based on variables selected through multivariable logistic regression analysis. Receiver operating characteristic (ROC) curve and calibration curve were used to assess the accuracy and discriminative ability of the nomogram model, and decision curve analysis (DCA) was performed to evaluate the model's performance and clinical effectiveness.Results: A retrospective analysis was conducted on 613 children hospitalized with bronchiolitis, among whom 53 (8.6%) experienced apnea. The results of Lasso regression and Logistic regression analyses showed that underlying diseases, feeding difficulties, tachypnea, WBC count, and lung consolidation were independent risk factors for apnea. A nomogram prediction model was constructed based on the five predictors mentioned above. After internal validation, the nomogram model demonstrated an AUC of 0.969 (95% CI 0.951–0.987), indicating strong predictive performance for apnea in bronchiolitis. Calibration curve analysis confirmed that the nomogram prediction model had good calibration, and the clinical decision curve analysis (DCA) indicated that the nomogram was clinically useful in estimating the net benefit to patients.Conclusion: In this study, a nomogram model was developed to predict the risk of apnea in hospitalized children with bronchiolitis. The model showed good predictive performance and clinical applicability, allowing for timely identification and intensified monitoring and treatment of high-risk patients to improve overall clinical prognosis.
- 10 July 2024
- Published by Frontiers Media SA in Frontiers in Pediatrics
- Vol. 12, 1397750
Open AccessEstablishment and validation of a prediction model for apnea on bronchiolitis
Objective: The objective of this study is to examine the risk factors associated with apnea in hospitalized patients diagnosed with bronchiolitis and to develop a nomogram prediction model for the early identification of patients who are at risk of developing apnea.Methods: The clinical data of patients diagnosed with acute bronchiolitis and hospitalized at the Children's Hospital of Nanjing Medical University between February 2018 and May 2021 were retrospectively analyzed. LASSO regression and logistic regression analysis were used to determine the risk factors for apnea in these patients. A nomogram was constructed based on variables selected through multivariable logistic regression analysis. Receiver operating characteristic (ROC) curve and calibration curve were used to assess the accuracy and discriminative ability of the nomogram model, and decision curve analysis (DCA) was performed to evaluate the model's performance and clinical effectiveness.Results: A retrospective analysis was conducted on 613 children hospitalized with bronchiolitis, among whom 53 (8.6%) experienced apnea. The results of Lasso regression and Logistic regression analyses showed that underlying diseases, feeding difficulties, tachypnea, WBC count, and lung consolidation were independent risk factors for apnea. A nomogram prediction model was constructed based on the five predictors mentioned above. After internal validation, the nomogram model demonstrated an AUC of 0.969 (95% CI 0.951–0.987), indicating strong predictive performance for apnea in bronchiolitis. Calibration curve analysis confirmed that the nomogram prediction model had good calibration, and the clinical decision curve analysis (DCA) indicated that the nomogram was clinically useful in estimating the net benefit to patients.Conclusion: In this study, a nomogram model was developed to predict the risk of apnea in hospitalized children with bronchiolitis. The model showed good predictive performance and clinical applicability, allowing for timely identification and intensified monitoring and treatment of high-risk patients to improve overall clinical prognosis.
- Open Access
- 28 March 2014
- Published by Springer Nature in BMC Pediatrics
- Vol. 14 (1) , 83
Effectiveness of predicting in-hospital mortality in critically ill children by assessing blood lactate levels at admission
Background: Hyperlactatemia upon admission is a documented risk factor for mortality in critically ill adult patients. However, the predictive significance of a single lactate measurement at admission for mortality in the general population of critically ill children remains uncertain. This study evaluated the predictive value of blood lactate levels at admission and determined the cut-off values for predicting in-hospital mortality in the critically ill pediatric population. Methods: We enrolled 1109 critically ill children who were admitted to a pediatric intensive care unit between July 2008 and December 2010. Arterial blood samples were collected in the first 2 hours after admission, and the lactate levels were determined. The Pediatric Risk of Mortality III (PRISM III) scores were calculated during the first 24 hours after admission. Results: Of the 1109 children admitted, 115 (10.4%) died in the hospital. The median (interquartile range) blood lactate level in critically ill children was 3.2 mmol/l (2.2-4.8). Among the children, 859 (77.5%) had a lactate concentration >2.0 mmol/l. The blood lactate level upon admission was significantly associated with mortality (odds ratio [OR] = 1.38; 95% confidence interval [CI], 1.30-1.46; p p p = 0.001), a high PRISM III score (OR = 1.15; 95% CI, 1.11-1.20; p p p p p-value for a comparison of both AUCs was 0.318. Blood lactate displayed a sensitivity of 61% and a specificity of 86% in predicting mortality at the optimal cut-off value of 5.55 mmol/l, and the positive and negative likelihood ratios were 4.5 and 0.45, respectively. Conclusions: A high blood lactate level at admission is independently associated with and predictive of in-hospital mortality in the general population of critically ill children.
- 28 March 2014
- Published by Springer Nature in BMC Pediatrics
- Vol. 14 (1) , 83
Open AccessEffectiveness of predicting in-hospital mortality in critically ill children by assessing blood lactate levels at admission
Background: Hyperlactatemia upon admission is a documented risk factor for mortality in critically ill adult patients. However, the predictive significance of a single lactate measurement at admission for mortality in the general population of critically ill children remains uncertain. This study evaluated the predictive value of blood lactate levels at admission and determined the cut-off values for predicting in-hospital mortality in the critically ill pediatric population. Methods: We enrolled 1109 critically ill children who were admitted to a pediatric intensive care unit between July 2008 and December 2010. Arterial blood samples were collected in the first 2 hours after admission, and the lactate levels were determined. The Pediatric Risk of Mortality III (PRISM III) scores were calculated during the first 24 hours after admission. Results: Of the 1109 children admitted, 115 (10.4%) died in the hospital. The median (interquartile range) blood lactate level in critically ill children was 3.2 mmol/l (2.2-4.8). Among the children, 859 (77.5%) had a lactate concentration >2.0 mmol/l. The blood lactate level upon admission was significantly associated with mortality (odds ratio [OR] = 1.38; 95% confidence interval [CI], 1.30-1.46; p p p = 0.001), a high PRISM III score (OR = 1.15; 95% CI, 1.11-1.20; p p p p p-value for a comparison of both AUCs was 0.318. Blood lactate displayed a sensitivity of 61% and a specificity of 86% in predicting mortality at the optimal cut-off value of 5.55 mmol/l, and the positive and negative likelihood ratios were 4.5 and 0.45, respectively. Conclusions: A high blood lactate level at admission is independently associated with and predictive of in-hospital mortality in the general population of critically ill children.