This paper presents a deep learning-based framework for designing and quantifying point-of-care sensors. A low-cost, rapid paper-based vertical flow assay (VFA) for high-sensitivity C-reactive protein (hsCRP) testing is demonstrated. A machine learning framework optimized the immunoreaction spot configuration and accurately inferred analyte concentration. A clinical study with 85 human samples showed a coefficient-of-variation of 11.2% and linearity of R² = 0.95. The system mitigated the hook-effect. This computational VFA expands access to CVD testing, and the framework is broadly applicable to cost-effective, mobile point-of-care sensors.

Zachary S. Ballard, Hyou-Arm Joung, Artem Goncharov, Jesse Liang, Karina Nugroho, Dino Di Carlo, Omai B. Garner, Aydogan Ozcan