Related Experiment Video
Updated: Jul 8, 2025
5/6 Nephrectomy Using Sharp Bipolectomy Via Midline Laparotomy in Rats
Published on: April 4, 2025
Machine learning models to predict end-stage kidney disease in chronic kidney disease stage 4
Kullaya Takkavatakarn1,2, Wonsuk Oh3, Ella Cheng4
1Division of Nephrology, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA.
Insights
Predicting kidney failure in stage 4 chronic kidney disease (CKD4) is crucial. Machine learning models, including artificial neural networks, accurately forecast progression to end-stage kidney disease (ESKD), aiding patient care.
Area of Science:
- Nephrology
- Data Science
- Biomedical Informatics
Background:
- End-stage kidney disease (ESKD) significantly increases morbidity and mortality.
- Accurate prediction of progression from stage 4 chronic kidney disease (CKD4) to ESKD is challenging but vital for patient management.
- Early risk identification in CKD4 patients facilitates advanced care planning and optimizes healthcare resource allocation.
Purpose of the Study:
- To develop and validate predictive models for identifying CKD4 patients at high risk of progressing to ESKD within three years.
- To compare the performance of machine learning algorithms including LASSO regression, random forest, XGBoost, and artificial neural network (ANN) for ESKD prediction.
- To utilize feature importance analysis to understand the drivers of predicted kidney failure.
Main Methods:
- Utilized electronic health record data from 3,160 CKD4 patients (2006-2016).
- Developed and validated four predictive models: LASSO regression, random forest, XGBoost, and ANN.
- Evaluated model performance using Area Under the Receiver Operating Characteristic Curve (AUROC) and employed SHAP values for feature interpretation.
Main Results:
- Of 3,160 CKD4 patients, 538 (21%) progressed to ESKD.
- All models demonstrated comparable predictive performance, with ANN and LASSO regression achieving the highest AUROC of 0.77.
- ANN (0.77, 95% CI 0.75-0.79) and LASSO regression (0.77, 95% CI 0.75-0.79) slightly outperformed random forest (0.76) and XGBoost (0.76).
Conclusions:
- Developed and validated multiple machine learning models for predicting near-term kidney failure in CKD4 patients.
- ANN, random forest, and XGBoost models showed similar, effective predictive capabilities.
- These models can enable customized interventions based on individual risk, improving population health management and resource allocation.
Introduction:
End-stage kidney disease (ESKD) is associated with increased morbidity and mortality. Identifying patients with stage 4 CKD (CKD4) at risk of rapid progression to ESKD remains challenging. Accurate prediction of CKD4 progression can improve patient outcomes by improving advanced care planning and optimizing healthcare resource allocation.
Methods:
We obtained electronic health record data from patients with CKD4 in a large health system between January 1, 2006, and December 31, 2016. We developed and validated four models, including Least Absolute Shrinkage and Selection Operator (LASSO) regression, random forest, eXtreme Gradient Boosting (XGBoost), and artificial neural network (ANN), to predict ESKD at 3 years. We utilized area under the receiver operating characteristic curve (AUROC) to evaluate model performances and utilized Shapley additive explanation (SHAP) values and plots to define feature dependence of the best performance model.
Results:
We included 3,160 patients with CKD4. ESKD was observed in 538 patients (21%). All approaches had similar AUROCs; ANN yielded the highest AUROC (0.77; 95%CI 0.75 to 0.79) and LASSO regression (0.77; 95%CI 0.75 to 0.79), followed by random forest (0.76; 95% CI 0.74 to 0.79), and XGBoost (0.76; 95% CI 0.74 to 0.78).
Conclusions:
We developed and validated several models for near-term prediction of kidney failure in CKD4. ANN, random forest, and XGBoost demonstrated similar predictive performances. Using this suite of models, interventions can be customized based on risk, and population health and resources appropriately allocated.
Related Concept Videos
01:17Factors Affecting Renal Clearance: Renal Impairment
One condition associated with renal failure is uremia. Uremia is characterized by impaired glomerular filtration and fluid accumulation in the body. This condition hinders the renal clearance of drugs, resulting in drug accumulation and potential...
01:27Dialysis
Acute kidney injury develops suddenly and can be caused by pre-renal causes (e.g., hypovolemia, shock), intrinsic renal causes (e.g., acute tubular necrosis), or post-renal causes (e.g., urinary obstruction). In contrast, chronic renal failure progresses gradually over time and is often...
01:10Nephrons