Label-free urinary protein detection through machine learning analysis of single droplet evaporation patterns

Area of Science:

• Biomedical Engineering
• Analytical Chemistry
• Machine Learning

Background:

• Chronic kidney disease (CKD) is a growing global health concern.
• Urinary protein is a key biomarker for CKD diagnosis and management.
• Current protein detection methods have limitations for routine CKD screening.

Purpose of the Study:

• To develop and validate a novel, rapid method for urine protein quantification.
• To address the need for a convenient diagnostic approach for kidney function assessment.

Main Methods:

• Developed a method based on dried droplet morphology analysis for urine protein quantification.
• Utilized machine learning models to analyze droplet drying patterns.
• Validated performance against common interfering substances and sample variations.

Main Results:

• The novel method accurately quantifies urine protein across various concentrations.
• Demonstrated robustness against common interfering substances and sample processing variations.
• Showed excellent agreement with colorimetric assays, with reduced sample volume and faster analysis.

Conclusions:

• A new dried droplet morphology analysis method offers a convenient alternative for proteinuria assessment.
• This simple, low-cost, and fast system is suitable for point-of-care CKD surveillance.
• Machine learning enhances protein content identification without staining or antibodies.