Empowering an Acute Kidney Injury 3D Graphene-Based Sensor Using Extreme Learning Machine

Area of Science:

• Biomedical Engineering
• Artificial Intelligence
• Nanomaterials

Background:

• Early detection of acute kidney injury (AKI) is crucial for patient outcomes.
• Current methods for monitoring kidney function biomarkers like neutrophil gelatinase-associated lipocalin (NGAL) can be improved for speed and accuracy.
• Graphene-based electrochemical sensors offer potential for sensitive biomarker detection.

Purpose of the Study:

• To develop and evaluate an AI-enhanced electrochemical sensor system for near-real-time NGAL detection.
• To investigate the application of an extreme learning machine (ELM) to improve NGAL detection accuracy using a 3D graphene electrode.
• To assess the performance of the integrated system for predicting AKI.

Main Methods:

• Fabrication of a 3D graphene electrode functionalized with a lipocalin-2 antibody for NGAL capture.
• Application of an extreme learning machine (ELM) algorithm to analyze urine data and predict NGAL levels.
• Comparative analysis of ELM against other machine learning algorithms (SVM, MLP, Random Forest).

Main Results:

• The ELM integration resulted in a 15% increase in the area under the curve (AUC) for NGAL determination.
• Detection limit for NGAL was significantly reduced from 14.8 to 0.89 ng/mL.
• Accuracy, precision, sensitivity, specificity, and F1 score for AKI prediction showed substantial improvements (e.g., 30.69% increase in precision).

Conclusions:

• The combination of ELM and graphene-based sensors provides a highly effective approach for accurate NGAL detection and AKI prediction.
• The developed system demonstrates the potential for miniaturized, AI-enhanced biosensors for practical clinical applications.
• ELM offers an optimal balance of performance and resource utilization for enhancing electrochemical sensor capabilities.