Machine Learning-Assisted Molecularly Imprinted Polymer Sensor for Point-of-Care Vancomycin Monitoring in Serum.

Area of Science:

• Analytical Chemistry
• Biomedical Engineering
• Materials Science

Background:

• Optimizing vancomycin dosage is crucial for treating severe infections and preventing antimicrobial resistance.
• Current laboratory testing methods for vancomycin are slow and centralized, creating a clinical gap.
• Rapid, point-of-care vancomycin quantification is needed for personalized therapeutic drug monitoring.

Purpose of the Study:

• To develop a low-cost, disposable electrochemical sensor for rapid vancomycin quantification in undiluted human serum.
• To integrate a selective molecularly imprinted polymer and signal-amplifying gold nanostructures for enhanced sensor performance.
• To utilize a Random Forest machine learning model to process sensor data and address matrix interference.

Main Methods:

• Fabrication of a printed circuit board-based electrochemical sensor incorporating a molecularly imprinted polymer with phenol red.
• Integration of highly porous gold nanostructures for signal amplification.
• Application of a Random Forest machine learning regression model for data analysis and quantification.

Main Results:

• The sensor achieved a limit of detection of 0.848 μg mL^-1 within a dynamic range of 0-100 μg mL^-1.
• Testing with patient serum samples showed excellent correlation (R² = 0.98) and agreement with LC-MS/MS.
• The developed sensor system demonstrates a robust alternative to conventional vancomycin testing methods.

Conclusions:

• The data-driven sensor system provides rapid, accurate vancomycin quantification at the point of care.
• This technology facilitates real-time, personalized vancomycin therapy, optimizing treatment outcomes.
• The developed sensor platform addresses the need for faster diagnostics in managing severe infections and antimicrobial resistance.