Dynamic and interpretable deep learning model for predicting respiratory failure following cardiac surgery
Abstract Background Postoperative respiratory failure following cardiac surgery (CS-PRF) remains a critical complication with substantial morbidity and mortality. Current risk prediction models are limited by static assessments and suboptimal accuracy. This study aimed to develop and validate a dyna...
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-08-01
|
| Series: | BMC Anesthesiology |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12871-025-03239-z |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849234887318437888 |
|---|---|
| author | Man Xu Hao Liu Anran Dai Qilian Tan Xinlong Zhang Rui Ding Chen Chen Jianjun Zou Yongjun Li Yanna Si |
| author_facet | Man Xu Hao Liu Anran Dai Qilian Tan Xinlong Zhang Rui Ding Chen Chen Jianjun Zou Yongjun Li Yanna Si |
| author_sort | Man Xu |
| collection | DOAJ |
| description | Abstract Background Postoperative respiratory failure following cardiac surgery (CS-PRF) remains a critical complication with substantial morbidity and mortality. Current risk prediction models are limited by static assessments and suboptimal accuracy. This study aimed to develop and validate a dynamic, machine learning–based model to enhance perioperative risk stratification for CS-PRF. Methods We retrospectively analyzed 1,016 adult patients who underwent cardiac surgery. Feature selection was conducted via the Least Absolute Shrinkage and Selection Operator (LASSO) and Boruta algorithms. Five machine learning models, including logistic regression, multilayer perceptron, extreme gradient boosting, categorical boosting, and deep neural network (DNN), were trained using preoperative and intraoperative variables. Model performance was evaluated by the area under the receiver operating characteristic curve (AUROC), area under the precision–recall curve (AUPRC), and calibration metrics. Model interpretability was evaluated via SHapley additive exPlanation (SHAP), and restricted cubic spline (RCS) analyses were used to explore nonlinear associations. Results The incidence of CS-PRF was 16.3%. In the validation cohort, the DNN model achieved superior performance, with an AUROC of 0.782 (95% CI 0.703–0.852) and an AUPRC of 0.496 based on preoperative variables, which improved to an AUROC of 0.855 (95% CI 0.796–0.906) and an AUPRC of 0.549 with the addition of intraoperative data. Calibration analysis demonstrated good agreement between predicted and observed risk. SHAP analysis of the preoperative model identified pulmonary artery pressure, age, and preoperative creatinine as key contributors. In the combined model, intraoperative features such as cardiopulmonary bypass duration and autologous blood transfusion volume emerged as additional important predictors. RCS analysis revealed a nonlinear association between age and CS-PRF. A web-based risk calculator integrating DNN predictions and individualized SHAP interpretation was deployed to support clinical decision-making. Conclusions We developed a deep learning model that integrates perioperative variables to predict postoperative respiratory failure following cardiac surgery. Demonstrating high accuracy and interpretability, the model has been deployed as an accessible web-based calculator, offering a practical tool for personalized perioperative risk assessment. Trial registration Not applicable. This study is a retrospective observational study and was not registered as a clinical trial. |
| format | Article |
| id | doaj-art-edca67f6068b44a094283bcfe747331d |
| institution | Kabale University |
| issn | 1471-2253 |
| language | English |
| publishDate | 2025-08-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Anesthesiology |
| spelling | doaj-art-edca67f6068b44a094283bcfe747331d2025-08-20T04:03:00ZengBMCBMC Anesthesiology1471-22532025-08-0125111810.1186/s12871-025-03239-zDynamic and interpretable deep learning model for predicting respiratory failure following cardiac surgeryMan Xu0Hao Liu1Anran Dai2Qilian Tan3Xinlong Zhang4Rui Ding5Chen Chen6Jianjun Zou7Yongjun Li8Yanna Si9Department of Anesthesiology, Perioperative and Pain Medicine, Nanjing First Hospital, Nanjing Medical UniversityState Key Laboratory of Natural Medicines, Key Laboratory of Drug Metabolism, China Pharmaceutical UniversityDepartment of Pharmacy, Sir Run Run Shaw Hospital, Zhejiang University School of MedicineDepartment of Anesthesiology, Perioperative and Pain Medicine, Nanjing First Hospital, Nanjing Medical UniversityDepartment of Anesthesiology, Perioperative and Pain Medicine, Nanjing First Hospital, Nanjing Medical UniversityDepartment of Anesthesiology, Perioperative and Pain Medicine, Nanjing First Hospital, Nanjing Medical UniversityDepartment of Pharmacy, Nanjing First Hospital, Nanjing Medical UniversityDepartment of Pharmacy, Nanjing First Hospital, Nanjing Medical UniversityDepartment of Anesthesiology, Lianshui County People’s HospitalDepartment of Anesthesiology, Perioperative and Pain Medicine, Nanjing First Hospital, Nanjing Medical UniversityAbstract Background Postoperative respiratory failure following cardiac surgery (CS-PRF) remains a critical complication with substantial morbidity and mortality. Current risk prediction models are limited by static assessments and suboptimal accuracy. This study aimed to develop and validate a dynamic, machine learning–based model to enhance perioperative risk stratification for CS-PRF. Methods We retrospectively analyzed 1,016 adult patients who underwent cardiac surgery. Feature selection was conducted via the Least Absolute Shrinkage and Selection Operator (LASSO) and Boruta algorithms. Five machine learning models, including logistic regression, multilayer perceptron, extreme gradient boosting, categorical boosting, and deep neural network (DNN), were trained using preoperative and intraoperative variables. Model performance was evaluated by the area under the receiver operating characteristic curve (AUROC), area under the precision–recall curve (AUPRC), and calibration metrics. Model interpretability was evaluated via SHapley additive exPlanation (SHAP), and restricted cubic spline (RCS) analyses were used to explore nonlinear associations. Results The incidence of CS-PRF was 16.3%. In the validation cohort, the DNN model achieved superior performance, with an AUROC of 0.782 (95% CI 0.703–0.852) and an AUPRC of 0.496 based on preoperative variables, which improved to an AUROC of 0.855 (95% CI 0.796–0.906) and an AUPRC of 0.549 with the addition of intraoperative data. Calibration analysis demonstrated good agreement between predicted and observed risk. SHAP analysis of the preoperative model identified pulmonary artery pressure, age, and preoperative creatinine as key contributors. In the combined model, intraoperative features such as cardiopulmonary bypass duration and autologous blood transfusion volume emerged as additional important predictors. RCS analysis revealed a nonlinear association between age and CS-PRF. A web-based risk calculator integrating DNN predictions and individualized SHAP interpretation was deployed to support clinical decision-making. Conclusions We developed a deep learning model that integrates perioperative variables to predict postoperative respiratory failure following cardiac surgery. Demonstrating high accuracy and interpretability, the model has been deployed as an accessible web-based calculator, offering a practical tool for personalized perioperative risk assessment. Trial registration Not applicable. This study is a retrospective observational study and was not registered as a clinical trial.https://doi.org/10.1186/s12871-025-03239-zCardiac surgeryPostoperative respiratory failureMachine learningDeep neural networkRisk predictionSHAP |
| spellingShingle | Man Xu Hao Liu Anran Dai Qilian Tan Xinlong Zhang Rui Ding Chen Chen Jianjun Zou Yongjun Li Yanna Si Dynamic and interpretable deep learning model for predicting respiratory failure following cardiac surgery BMC Anesthesiology Cardiac surgery Postoperative respiratory failure Machine learning Deep neural network Risk prediction SHAP |
| title | Dynamic and interpretable deep learning model for predicting respiratory failure following cardiac surgery |
| title_full | Dynamic and interpretable deep learning model for predicting respiratory failure following cardiac surgery |
| title_fullStr | Dynamic and interpretable deep learning model for predicting respiratory failure following cardiac surgery |
| title_full_unstemmed | Dynamic and interpretable deep learning model for predicting respiratory failure following cardiac surgery |
| title_short | Dynamic and interpretable deep learning model for predicting respiratory failure following cardiac surgery |
| title_sort | dynamic and interpretable deep learning model for predicting respiratory failure following cardiac surgery |
| topic | Cardiac surgery Postoperative respiratory failure Machine learning Deep neural network Risk prediction SHAP |
| url | https://doi.org/10.1186/s12871-025-03239-z |
| work_keys_str_mv | AT manxu dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery AT haoliu dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery AT anrandai dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery AT qiliantan dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery AT xinlongzhang dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery AT ruiding dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery AT chenchen dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery AT jianjunzou dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery AT yongjunli dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery AT yannasi dynamicandinterpretabledeeplearningmodelforpredictingrespiratoryfailurefollowingcardiacsurgery |