Deep learning-enhanced dual-mode multiplexed optical sensor for point-of-care diagnostics of cardiovascular diseases

Area of Science:

• Biomedical Engineering
• Cardiovascular Diagnostics
• Point-of-Care Testing

Background:

• Timely diagnosis of myocardial infarction (MI) and heart failure (HF) is critical due to their high mortality and frequent co-occurrence.
• Current cardiac biomarker tests suffer from long turnaround times, limited dynamic ranges, and single-analyte detection, hindering comprehensive cardiovascular assessment.
• Existing point-of-care testing (POCT) methods often lack the sensitivity and multiplexing capabilities required for complex cardiovascular disease diagnosis.

Purpose of the Study:

• To develop a deep learning-enhanced, dual-mode multiplexed vertical flow assay (xVFA) for rapid and quantitative cardiovascular biomarker detection.
• To integrate colorimetric and chemiluminescent detection for a broad dynamic range and high sensitivity in a single paper-based cartridge.
• To create a portable optical reader and neural network-based quantification pipeline for efficient point-of-care cardiovascular diagnostics.

Main Methods:

• Development of a dual-mode multiplexed vertical flow assay (xVFA) integrating colorimetric and chemiluminescent detection on a paper-based cartridge.
• Utilization of a portable optical reader coupled with a neural network-based quantification pipeline for data analysis.
• Simultaneous quantification of cardiac troponin I (cTnI), creatine kinase-MB (CK-MB), and N-terminal pro-B-type natriuretic peptide (NT-proBNP) using 50 µL of serum.

Main Results:

• The xVFA achieved a broad dynamic range of approximately 6 orders of magnitude, accommodating both low- and high-abundance biomarkers.
• High sensitivity was demonstrated, with sub-pg/mL for cTnI and sub-ng/mL for CK-MB and NT-proBNP, covering clinically relevant ranges.
• The system provided rapid results within 23 minutes, and neural network models showed robust quantification performance (Pearson's r > 0.96) compared to reference assays using 92 patient samples.

Conclusions:

• The deep learning-enhanced dual-mode xVFA offers a sensitive, rapid, and multiplexed solution for cardiovascular biomarker quantification.
• This compact and cost-effective optical sensor technology enables accurate point-of-care diagnostics for conditions like MI and HF.
• The integrated system addresses limitations of current testing methods, paving the way for improved cardiovascular disease management.