Perioperative medicine & outcomes
Artificial Intelligence–Driven Prediction of Acute Kidney Injury Following Acute Type A Aortic Dissection Surgery in a Chinese Population
Authors' conclusion
ML models, particularly LightGBM, accurately predict postoperative AKI risk in ATAAD patients, offering a promising tool to enhance perioperative management and patient outcomes.
Study context
- Population
- Patients undergoing surgery for acute type A aortic dissection, Beijing Anzhen Hospital, November 2018 to October 2023
- Participants
- 1,350 patients
- Comparison
- Seven machine learning (ML) algorithms for predicting postoperative acute kidney injury (AKI)
- Outcome
- AUC for postoperative AKI, testing set
Methodological detailssingle-centre, machine learning, random split training (85%) and testing (15%) sets, SHAP analysis; web-based application developed using the optimal model
- Single-centre
- Machine learning
- Random split training (85%) and testing (15%) sets
- SHAP analysis; web-based application developed using the optimal model
Performance
As reported in the abstract
GBM
0.849
AUC · 95% CI 0.798–0.902
LightGBM
0.874
AUC · 95% CI 0.831–0.918
Random Forest (RF)
0.800
AUC · 95% CI 0.737–0.855
K-Nearest Neighbors (KNN)
0.672
AUC · 95% CI 0.598–0.739
Multi-Layer Perceptron Neural Network (MLP-NN)
0.529
AUC · 95% CI 0.486–0.574
Naive Bayes (NB)
0.833
AUC · 95% CI 0.775–0.886
Logistic Regression (LR)
0.866
AUC · 95% CI 0.810–0.912
Further findings
Incidence of postoperative AKI
- Postoperative AKI
- 586 (43.4%)
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Composed by AI from the published abstract and checked against it. How Briefs are made →