Sample-to-answer automated centrifugal microfluidics with computer-vision readout for point-of-care molecular diagnostics

Abstract The COVID-19 pandemic exposed how unprepared the world is to respond to an infectious disease outbreak at scale. Conventional diagnostic testing relies on centralized laboratories, and when testing demand surges during an outbreak, these facilities become strained for resources and personnel, a problem compounded in low- and middle-income countries with less developed health and transport infrastructure. Decentralized, point-of-care diagnostics offer a path around this bottleneck, but automated platforms capable of true sample-to-answer operation remain limited. Here we report a fully automated centrifugal microfluidic platform that combines loop-mediated isothermal amplification (LAMP) with computer-vision-based readout, demonstrated using SARS-CoV-2 detection as a model application. The integrated platform performs the complete assay protocol, including heating, rotation, and optical detection, at the push of a single button, without user intervention, processing 26 samples in 40 minutes. Using RNA fragments spiked into water and artificial saliva, the platform consistently detected amplification down to approximately 1000 copies per 10 µL reaction. This platform architecture is broadly applicable beyond SARS-CoV-2 and offers a template for rapid, decentralized molecular diagnostics for future infectious disease outbreaks.