Prediction of eGFR and uACR Using Evolutionary Learning for Identifying High‐Risk Individuals in Renal Disease Prevention

Objectives Many individuals at high risk for renal disease remain undetected in clinical practice due to insufficient testing of estimated glomerular filtration rate (eGFR) and urine albumin‐to‐creatinine ratio (uACR). This study aims to develop models to predict reduced eGFR and elevated uACR, thereby facilitating early identification of individuals at risk. Methods From 72,346 analyzed samples, we derived a dataset containing 15 features (U15). We developed three evolutionary learning‐based models: PeGFR, PuACR1, and PuACR2, to predict eGFR < 60 mL/min/1.73 m 2 , uACR ≥ 30 mg/g, and uACR > 300 mg/g, respectively. These models were built using an intelligent bi‐objective combinatorial genetic algorithm (IBCGA) and SVM parameter settings optimized simultaneously. External dataset without eGFR and uACR measurements to evaluate clinical applicability. Results In the U15 training and test datasets for uACR prediction, gender, pH, and bilirubin were not statistically significant; whereas all features were significant in eGFR prediction ( p < 0.05). The areas under the curve (AUC) for PeGFR, PuACR1, and PuACR2 were 0.846, 0.827, and 0.949, respectively. In the external test, 589 individuals (11.36%) were identified as high‐ or very‐high‐risk, defined as having eGFR < 60 mL/min/1.73 m 2 and uACR ≥ 30 mg/g. Conclusions The evolutionary learning framework proposed accurately predicts renal risk using clinical data as an early screening tool. It also demonstrates scalability in assisting healthcare providers in determining whether further renal‐related examinations are warranted, thereby facilitating the identification of high‐risk individuals for secondary prevention. An online tool has also been developed for clinical and research use: https://nycuiclab.ngrok.io .

Authors

Institutions

Publication Details

Journal
Journal of Clinical Laboratory Analysis
Published
2026-11-19
DOI
https://doi.org/10.1002/jcla.70371
Primary Topic
Artificial Intelligence in Healthcare
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Prediction of eGFR and uACR Using Evolutionary Learning for Identifying High‐Risk Individuals in Renal Disease Prevention

Chia-Heng Yen, Shinn‐Ying Ho, Ting‐Yu Chen, Yen‐Cheng Lin et al.
Journal of Clinical Laboratory Analysis
Artificial Intelligence in Healthcare
article

Prediction of eGFR and uACR Using Evolutionary Learning for Identifying High‐Risk Individuals in Renal Disease Prevention

Chia-Heng Yen, Shinn‐Ying Ho, Ting‐Yu Chen, Yen‐Cheng Lin, Yao‐Feng Hu
article en

Abstract

Objectives Many individuals at high risk for renal disease remain undetected in clinical practice due to insufficient testing of estimated glomerular filtration rate (eGFR) and urine albumin‐to‐creatinine ratio (uACR). This study aims to develop models to predict reduced eGFR and elevated uACR, thereby facilitating early identification of individuals at risk. Methods From 72,346 analyzed samples, we derived a dataset containing 15 features (U15). We developed three evolutionary learning‐based models: PeGFR, PuACR1, and PuACR2, to predict eGFR < 60 mL/min/1.73 m 2 , uACR ≥ 30 mg/g, and uACR > 300 mg/g, respectively. These models were built using an intelligent bi‐objective combinatorial genetic algorithm (IBCGA) and SVM parameter settings optimized simultaneously. External dataset without eGFR and uACR measurements to evaluate clinical applicability. Results In the U15 training and test datasets for uACR prediction, gender, pH, and bilirubin were not statistically significant; whereas all features were significant in eGFR prediction ( p < 0.05). The areas under the curve (AUC) for PeGFR, PuACR1, and PuACR2 were 0.846, 0.827, and 0.949, respectively. In the external test, 589 individuals (11.36%) were identified as high‐ or very‐high‐risk, defined as having eGFR < 60 mL/min/1.73 m 2 and uACR ≥ 30 mg/g. Conclusions The evolutionary learning framework proposed accurately predicts renal risk using clinical data as an early screening tool. It also demonstrates scalability in assisting healthcare providers in determining whether further renal‐related examinations are warranted, thereby facilitating the identification of high‐risk individuals for secondary prevention. An online tool has also been developed for clinical and research use: https://nycuiclab.ngrok.io .

Journal of Clinical Laboratory Analysis
National Yang Ming Chiao Tung University (TW), Kaohsiung Medical University (TW), National Taiwan Ocean University (TW), Mackay Memorial Hospital (TW)
Artificial Intelligence in Healthcare
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon