Fast, accurate, and scalable testing has been recognized unanimously as crucial for mitigating the impact of COVID-19 and future pandemics. We propose a technology that allows rapid (~2 minutes) testing for SARS CoV-2. Our technology combines novel label-free imaging and dedicated deep-learning algorithms to detect and classify viral populations in exhaled air. If successful, this project will result in a device based on quantitative phase imaging and integrated AI tools, which will detect the unlabeled virus acquired by the patient?s breath condensed on a microscope slide. Toward this goal, we will advance Spatial Light Interference Microscopy (SLIM), an ultrasensitive label-free imaging technique, proven to measure structures down to the sub-nanometer scale. SLIM was developed in the PI?s Lab at UIUC, its original publication received 490 citations to date, and has been commercialized by Phi Optics (Research Park, UIUC), with sales across the world in both academia and industry. Applying the computed fluorescence maps back to the QPI data, we propose to measure nanoscale features of viral particles, with high specificity, minimal preparation time, and independent of clinical infrastructure. As a result, the new technology will eventually be ideal for point-of-care settings, surveillance screening and as a home monitoring device. We anticipate that our approach will be scalable to other viruses, with new imaging and training data.
We propose a breath test using label-free imaging and AI: an individual exhales on a microscope slide, which is fed into a SLIM microscope equipped with a computer that runs deep-learning pre-trained algorithms for SARS CoV-2 identification. The result is displayed in real time, with the entire procedure requiring < 2min.
