Dynamic machine learning architecture integrating time-phased predictions for acute kidney injury following extracorporeal membrane oxygenation
4 affiliations
- 1Department of Intensive Care Unit, The Eighth Affiliated Hospital, Southern Medical University (The First People’s Hospital of Shunde, Foshan), Foshan, Guangdong, China
- 2Department of Critical Care Medicine, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong, China
- 3School of Medicine, Southern Medical University, Foshan, Guangdong, China
- 4Editor Office, iMeta, Shenzhen, Guangdong, China
Abstract
Background: Acute kidney injury (AKI) is a frequent, severe complication of extracorporeal membrane oxygenation (ECMO) therapy, substantially contributing to unfavorable outcomes. No effective machine learning (ML) models exist for early, accurate detection. This study developed a two-phased prediction model using clinical data from pre-ECMO and on-ECMO stages for ECMO-induced AKI (ECMO-AKI), and further integrated two models to formulate a consensus risk-stratification strategy.
Methods: This retrospective study included 344 patients undergoing ECMO from 2015 to 2026. Clinical and laboratory variables from pre-ECMO and on-ECMO stages were respectively used to develop two ECMO-AKI diagnostic models via a analytical framework incorporating 12 ML algorithms. Multiple analytical methods were utilised to evaluate pre-ECMOAKI and on-ECMOAKI models. The Shapley Additive Explanations (SHAP) method was used to interpret two models and to quantify and visualize the features. Latent class growth mixture model (LCGMM) was applied to investigate the association between longitudinal feature trajectories and ECMO-AKI.
Results: Pre-ECMOAKI and on-ECMOAKI models incorporated into 12 ML, yielding 66 and 91 combinations, respectively. Among these, glmBoost + gradient boosting machine (GBM) as a superior algorithm for predicting ECMO-AKI. Pre-ECMOAKI (9 features) and on-ECMOAKI (10 features) models exhibited superior discrimination performance, with area under curve (AUC) of 0.867 (95%CI: 0.819-0.911) and 0.907 (95%CI: 0.870-0.941) in derivation cohort, 0.732 (95%CI: 0.678-0.780) and 0.803 (95%CI: 0.711-0.889) in held-out validation cohort, respectively. A risk-stratification framework integrating pre-ECMOAKI and on-ECMOAKI predictions achieved a class-specific precision of 90.23% for ECMO-AKI events in the consensus high-risk group and 90.79% for non-ECMO-AKI events in the consensus low-risk populations. SHAP analysis reveals that endothelial activation and stress index (EASIX), cardiopulmonary resuscitation time (CPRT) and total bilirubin (TBI) are important feature variables in two models. Restricted cubic spline (RCS) regression, threshold effect analysis, and LCGMM indicated EASIX > 4.7 (with sustained increasing trend), CRPT > 24 min, and TBI > 20.2 μmol/L (with continuous upward trend) were associated with an elevated ECMO-AKI risk.
Conclusion: Two-phased dynamic prediction models (pre-ECMOAKI and on-ECMOAKI) were constructed based on glmBoost +GBM, which act as an accurate and interpretable diagnostic tool for ECMO-AKI. Based on consistency analysis of predictions of two models, this framework allows early risk stratification and thus holds potential for individualized management.
Related
- "White-Out Lung" Following Veno-Venous Extracorporeal Membrane Oxygenation (ECMO) Initiation for Acute Refractory Respiratory Failure From Obstructive Lung Disease: A Case Series.
- Infectious complications following heart transplantation in the era of high-priority allocation and extracorporeal membrane oxygenation
- Mixing point thrombus during veno-arterial extracorporeal membrane oxygenation support
- Veno-arterial extracorporeal membrane oxygenation versus off-pump lung transplantation: a long-standing debate.
- Legionella Septic Cardiomyopathy Requiring Venoarterial Extracorporeal Membrane Oxygenation Followed by Presumed Complement-Mediated Thrombotic Microangiopathy: A Case Report
BibTeX
@article{lu2026dynamic,
title = {{Dynamic machine learning architecture integrating time-phased predictions for acute kidney injury following extracorporeal membrane oxygenation}},
author = {Jianhai Lu and Zhen Chen and Donghua Zheng and Genglong Liu and Yijia Wu and Wanfang Li and Liuer Zuo and Qiang Tai},
journal = {Frontiers in Physiology},
year = {2026},
month = oct,
doi = {10.3389/fphys.2026.1922656},
url = {https://doi.org/10.3389/fphys.2026.1922656},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Dynamic machine learning architecture integrating time-phased predictions for acute kidney injury following extracorporeal membrane oxygenation
AU - Jianhai Lu
AU - Zhen Chen
AU - Donghua Zheng
AU - Genglong Liu
AU - Yijia Wu
AU - Wanfang Li
AU - Liuer Zuo
AU - Qiang Tai
PY - 2026
DA - 2026/10/22
JO - Frontiers in Physiology
DO - 10.3389/fphys.2026.1922656
UR - https://doi.org/10.3389/fphys.2026.1922656
AB - Background: Acute kidney injury (AKI) is a frequent, severe complication of extracorporeal membrane oxygenation (ECMO) therapy, substantially contributing to unfavorable outcomes. No effective machine learning (ML) models exist for early, accurate detection. This study developed a two-phased prediction model using clinical data from pre-ECMO and on-ECMO stages for ECMO-induced AKI (ECMO-AKI), and further integrated two models to formulate a consensus risk-stratification strategy. Methods: This retrospective study included 344 patients undergoing ECMO from 2015 to 2026. Clinical and laboratory variables from pre-ECMO and on-ECMO stages were respectively used to develop two ECMO-AKI diagnostic models via a analytical framework incorporating 12 ML algorithms. Multiple analytical methods were utilised to evaluate pre-ECMOAKI and on-ECMOAKI models. The Shapley Additive Explanations (SHAP) method was used to interpret two models and to quantify and visualize the features. Latent class growth mixture model (LCGMM) was applied to investigate the association between longitudinal feature trajectories and ECMO-AKI. Results: Pre-ECMOAKI and on-ECMOAKI models incorporated into 12 ML, yielding 66 and 91 combinations, respectively. Among these, glmBoost + gradient boosting machine (GBM) as a superior algorithm for predicting ECMO-AKI. Pre-ECMOAKI (9 features) and on-ECMOAKI (10 features) models exhibited superior discrimination performance, with area under curve (AUC) of 0.867 (95%CI: 0.819-0.911) and 0.907 (95%CI: 0.870-0.941) in derivation cohort, 0.732 (95%CI: 0.678-0.780) and 0.803 (95%CI: 0.711-0.889) in held-out validation cohort, respectively. A risk-stratification framework integrating pre-ECMOAKI and on-ECMOAKI predictions achieved a class-specific precision of 90.23% for ECMO-AKI events in the consensus high-risk group and 90.79% for non-ECMO-AKI events in the consensus low-risk populations. SHAP analysis reveals that endothelial activation and stress index (EASIX), cardiopulmonary resuscitation time (CPRT) and total bilirubin (TBI) are important feature variables in two models. Restricted cubic spline (RCS) regression, threshold effect analysis, and LCGMM indicated EASIX > 4.7 (with sustained increasing trend), CRPT > 24 min, and TBI > 20.2 μmol/L (with continuous upward trend) were associated with an elevated ECMO-AKI risk. Conclusion: Two-phased dynamic prediction models (pre-ECMOAKI and on-ECMOAKI) were constructed based on glmBoost +GBM, which act as an accurate and interpretable diagnostic tool for ECMO-AKI. Based on consistency analysis of predictions of two models, this framework allows early risk stratification and thus holds potential for individualized management.
SN - 1664-042X
ER -