MedFusion-gP-AKI: development and multicenter validation of a machine learning fusion model for early prediction of KDIGO stage 3 acute kidney injury in critically ill traumatic cervicothoracic spinal
12 hours ago
- #acute kidney injury
- #spinal cord injury
- #machine learning
- Development of MedFusion-gP-AKI, a machine learning fusion model for early prediction of KDIGO stage 3 acute kidney injury (AKI) in critically ill traumatic cervicothoracic spinal cord injury (TCTSCI) patients.
- Model trained on MIMIC-IV/eICU cohort and externally validated in 188 patients from four Chinese centers, achieving high AUCs (0.938, 0.909, 0.969, etc.) and reliable discrimination.
- Key predictors identified include lactate, mean arterial pressure, temperature, potassium, and TCTSCI level.
- Utilized advanced data imputation (GAN-based) and balancing techniques (SMOTified-GAN) for robust model performance.
- SHAP analysis confirmed model attributions aligned with clinical patterns, and a web-based calculator was developed for practical use.
- Aims to improve early identification and preventive strategies for AKI in TCTSCI patients, potentially reducing mortality.