Detection of Triple Respiratory Viruses in Saliva Using Multiplexed RT-LAMP on a Machine Learning-Empowered Portable Device

Accurately diagnosing respiratory infections is paramount, particularly during public health crises such as the COVID-19 pandemic, which has stressed the necessity for rapid and reliable point-of-care testing (POCT) for nucleic acid detection. Our study introduces a lens-free optical system integrated with machine learning to streamline multiplexed nucleic acid tests in real-time, overcoming the limitations of spatial parallelization and optical filters found in conventional POCT devices. Through a neural network, our scalable approach efficiently differentiates between fluorescent signals from a mixture of fluorophores, improving detection and quantification capabilities. Moreover, it showcases adaptability in predicting the concentrations of different fluorophores and the concurrent detection of multiple pathogens, such as RSV, Influenza A, and SARS-CoV-2. Notably, our system has been validated with mock saliva samples, affirming its potential for accurate diagnostics in scenarios with limited sample volume and the need for flexible fluorophore deployment, contributing a robust solution for POCT applications.