A Risk Prediction Model (CMC-AKIX) for Postoperative Acute Kidney Injury Using Machine Learning: Algorithm Development and Validation

Abstract

Background: Postoperative acute kidney injury (AKI) is a significant risk associated with surgeries under general anesthesia, often leading to increased mortality and morbidity. Existing predictive models for postoperative AKI are usually limited to specific surgical areas or require external validation.

Objective: We proposed to build a prediction model for postoperative AKI using several machine learning methods.

Methods: We conducted a retrospective cohort analysis of noncardiac surgeries from 2009 to 2019 at seven university hospitals in South Korea. We evaluated six machine learning models: deep neural network, logistic regression, decision tree, random forest, light gradient boosting machine, and naïve Bayes for predicting postoperative AKI, defined as a significant increase in serum creatinine or the initiation of renal replacement therapy within 30 days after surgery. The performance of the models was analyzed using the area under the curve (AUC) of the receiver operating characteristic curve, accuracy, precision, sensitivity (recall), specificity, and F₁-score.

Results: Among the 239,267 surgeries analyzed, 7935 cases of postoperative AKI were identified. The models, using 38 preoperative predictors, showed that deep neural network (AUC=0.832), light gradient boosting machine (AUC=0.836), and logistic regression (AUC=0.825) demonstrated superior performance in predicting AKI risk. The deep neural network model was then developed into a user-friendly website for clinical use.

Conclusions: Our study introduces a robust, high-performance AKI risk prediction system that is applicable in clinical settings using preoperative data. This model's integration into a user-friendly website enhances its clinical utility, offering a significant step forward in personalized patient care and risk management.

Keywords: South Korea; acute kidney injury; anesthesia; artificial intelligence; cohort analysis; deep neural networks; digital health; general surgery; hospital; logistic regression; machine learning; morbidity; mortality; patient care; postoperative care; prediction model; retrospective study; risk management; surgery; user-friendly.

Figure 1

Flowchart of the study population. AKI: acute kidney injury; CMC: Catholic Medical Center; Cr: serum creatinine; eGFR: estimated glomerular filtration rate; op: operation; RRT: renal replacement therapy.

Figure 2

Loss function graph and AUC graph of the training and validation sets. (A) Loss functions of the validation and training sets converge at about epoch 86 and stabilize thereafter. (B) AUCs also overlap around epoch 86, and this suggests the model is starting to overfit on the training data. AUC: area under the curve; Val: validation.

Figure 3

ROC-AUC of the AKI prediction models. AKI: acute kidney injury; AUC: area under the curve; DNN: deep neural network; Light GBM: light gradient boosting machine; LR + L1: logistic regression with Least Absolute Shrinkage and Selection Operator penalty; ROC: receiver operating characteristic.

Figure 4

A nomogram based on a simplified logistic regression model. eGFR: estimated glomerular filtration rate; Hb: hemoglobin; Na: sodium; OpDuration: operation duration; Prot: urine protein.

Figure 5

A sample page of the website application. ACE-I: angiotensin-converting enzyme inhibitor; ALT: alanine aminotransferase; ARB: angiotensin II type 1 receptor blocker; AST: aspartate aminotransferase; BP: blood pressure; CMC: Catholic Medical Center; COPD: chronic obstructive pulmonary disease; eGFR: estimated glomerular filtration rate; NSAID: nonsteroidal anti-inflammatory drugs.