Deep Learning-Enhanced\nPaper-Based Vertical Flow Assay\nfor High-Sensitivity Troponin Detection Using Nanoparticle Amplification

Successful integration\nof point-of-care testing (POCT)\ninto clinical\nsettings requires improved assay sensitivity and precision to match\nlaboratory standards. Here, we show how innovations in amplified biosensing,\nimaging, and data processing, coupled with deep learning, can help\nimprove POCT. To demonstrate the performance of our approach, we present\na rapid and cost-effective paper-based high-sensitivity vertical flow\nassay (hs-VFA) for quantitative measurement of cardiac troponin I\n(cTnI), a biomarker widely used for measuring acute cardiac damage\nand assessing cardiovascular risk. The hs-VFA includes a colorimetric\npaper-based sensor, a portable reader with time-lapse imaging, and\ncomputational algorithms for digital assay validation and outlier\ndetection. Operating at the level of a rapid at-home test, the hs-VFA\nenabled the accurate quantification of cTnI using 50 μL of serum\nwithin 15 min per test and achieved a detection limit of 0.2 pg/mL,\nenabled by gold ion amplification chemistry and time-lapse imaging.\nIt also achieved high precision with a coefficient of variation of <7%\nand a very large dynamic range, covering cTnI concentrations over\n6 orders of magnitude, up to 100 ng/mL, satisfying clinical requirements.\nIn blinded testing, this computational hs-VFA platform accurately\nquantified cTnI levels in patient samples and showed a strong correlation\nwith the ground truth values obtained by a benchtop clinical analyzer.\nThis nanoparticle amplification-based computational hs-VFA platform\ncan democratize access to high-sensitivity point-of-care diagnostics\nand provide a cost-effective alternative to laboratory-based biomarker\ntesting.