Machine learning for determining lateral flow device results for testing of SARS-CoV-2 infection in asymptomatic populations

Beggs, Andrew D.; Caiado, Camila C.S.; Branigan, Mark; Lewis-Borman, Paul; Patel, Nishali; Fowler, Tom; Dijkstra, Anna; Chudzik, Piotr; Yousefi, Paria; Javer, Avelino; Van Meurs, Bram; Tarassenko, Lionel; Irving, Benjamin; Whalley, Celina; Lal, Neeraj; Robbins, Helen; Leung, Elaine; Lee, Lennard; Banathy, Robert

Abstract

Rapid antigen tests, in the form of lateral flow devices (LFD) allow testing of a large population for SARS-CoV-2. To reduce the variability seen in device interpretation, we show the design and testing of an AI algorithm based on machine learning. The machine learning (ML) algorithm is trained on a combination of artificially hybridised LFDs and LFD data linked to RT-qPCR result. Participants are recruited from assisted test sites (ATS) and health care workers undertaking self-testing and images analysed using the ML algorithm. A panel of trained clinicians are used to resolve discrepancies. In total, 115,316 images are returned. In the ATS sub study, sensitivity increased from 92.08% to 97.6% and specificity from 99.85% to 99.99%. In the self-read sub-study, sensitivity increased from 16.00% to 100%, and specificity from 99.15% to 99.40%. An ML-based classifier of LFD results outperforms human reads in asymptomatic testing sites and self-reading.

Citation

Beggs, A. D., Caiado, C. C., Branigan, M., Lewis-Borman, P., Patel, N., Fowler, T., …Banathy, R. (2022). Machine learning for determining lateral flow device results for testing of SARS-CoV-2 infection in asymptomatic populations. Cell Reports Medicine, 3(10), Article 100784. https://doi.org/10.1016/j.xcrm.2022.100784

• Journal Article Type: Article
• Acceptance Date: Aug 1, 2022
• Online Publication Date: Sep 26, 2022
• Publication Date: Oct 18, 2022
• Deposit Date: May 15, 2024
• Publicly Available Date: May 17, 2024
• Journal: Cell Reports Medicine
• Print ISSN: 2666-3791
• Publisher: Cell Press
• Peer Reviewed: Peer Reviewed
• Volume: 3
• Issue: 10
• Article Number: 100784
• DOI: https://doi.org/10.1016/j.xcrm.2022.100784
• Public URL: https://durham-repository.worktribe.com/output/2441210