Researchers from the University of Oxford's Department of Physics have created a rapid screening test for SARS-CoV-2 using artificial intelligence. With promising early results, the team aims to bring this innovative method to market next year.
Mass testing remains a cornerstone in combating SARS-CoV-2. For context, Chinese firm Coyote recently launched the Flash 20 PCR machine, which delivers results in 30 minutes and processes four samples at once—touted as the world's fastest at the time (details below).
In a preprint published on MedRxiv on October 17, 2020, Oxford physicists detailed a screening test that yields results in under five minutes. Leveraging machine learning for high accuracy, this approach could enable swift screening at public venues like businesses, airports, and concert halls.
“Our method quickly detects intact virus particles; which means the test is simple, extremely fast, and cost-effective," said Achillefs Kapanidis, a key study contributor.
The process starts with a throat swab, analyzed via microscopy to spot SARS-CoV-2 infection markers. Machine learning software then evaluates the sample, capitalizing on the virus's unique size and shape for speed and precision.
Unlike traditional PCR tests, which amplify RNA to isolate it, this method skips that step—explaining its remarkable speed. Researchers report a 98% success rate, though achieved in controlled lab settings.
As a preprint, the study awaits peer review from a scientific journal. Commercialization promises for 2021 should be viewed cautiously. The team is now refining it into a fully integrated, real-time diagnostic device.
