Dataset Information

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

ABSTRACT: Rapid antigen tests in the form of lateral flow devices (LFDs) allow testing of a large population for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). To reduce the variability in device interpretation, we show the design and testing of an artifical intelligence (AI) algorithm based on machine learning. The machine learning (ML) algorithm is trained on a combination of artificially hybridized LFDs and LFD data linked to quantitative real-time PCR results. Participants are recruited from assisted test sites (ATSs) and health care workers undertaking self-testing, and images are analyzed using the ML algorithm. A panel of trained clinicians is used to resolve discrepancies. In total, 115,316 images are returned. In the ATS substudy, sensitivity increased from 92.08% to 97.6% and specificity from 99.85% to 99.99%. In the self-read substudy, 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 assisted testing sites and self-reading.

SUBMITTER: LFD AI Consortium. Electronic address: a.beggs@bham.ac.uk

PROVIDER: S-EPMC9513327 | biostudies-literature | 2022 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

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

Cell reports. Medicine 20220927 10

Rapid antigen tests in the form of lateral flow devices (LFDs) allow testing of a large population for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). To reduce the variability in device interpretation, we show the design and testing of an artifical intelligence (AI) algorithm based on machine learning. The machine learning (ML) algorithm is trained on a combination of artificially hybridized LFDs and LFD data linked to quantitative real-time PCR results. Participants are recruited ...[more]

PMID: 36240756

Similar Datasets

Project description:ObjectiveTo assess the performance of the SARS-CoV-2 antigen rapid lateral flow test (LFT) versus polymerase chain reaction testing in the asymptomatic general population attending testing centres.DesignObservational cohort study.SettingCommunity LFT pilot at covid-19 testing sites in Liverpool, UK.Participants5869 asymptomatic adults (≥18 years) voluntarily attending one of 48 testing sites during 6-29 November 2020.InterventionsParticipants were tested using both an Innova LFT and a quantitative reverse-transcriptase polymerase chain reaction (RT-qPCR) test based on supervised self-administered swabbing at testing sites.Main outcome measuresSensitivity, specificity, and predictive values of LFT compared with RT-qPCR in an epidemic steady state of covid-19 among adults with no classic symptoms of the disease.ResultsOf 5869 test results, 22 (0.4%) LFT results and 343 (5.8%) RT-qPCR results were void (that is, when the control line fails to appear within 30 minutes). Excluding the void results, the LFT versus RT-qPCR showed a sensitivity of 40.0% (95% confidence interval 28.5% to 52.4%; 28/70), specificity of 99.9% (99.8% to 99.99%; 5431/5434), positive predictive value of 90.3% (74.2% to 98.0%; 28/31), and negative predictive value of 99.2% (99.0% to 99.4%; 5431/5473). When the void samples were assumed to be negative, a sensitivity was observed for LFT of 37.8% (26.8% to 49.9%; 28/74), specificity of 99.6% (99.4% to 99.8%; 5431/5452), positive predictive value of 84.8% (68.1% to 94.9%; 28/33), and negative predictive value of 93.4% (92.7% to 94.0%; 5431/5814). The sensitivity in participants with an RT-qPCR cycle threshold (Ct) of <18.3 (approximate viral loads >106 RNA copies/mL) was 90.9% (58.7% to 99.8%; 10/11), a Ct of <24.4 (>104 RNA copies/mL) was 69.4% (51.9% to 83.7%; 25/36), and a Ct of >24.4 (<104 RNA copies/mL) was 9.7% (1.9% to 23.7%; 3/34). LFT is likely to detect at least three fifths and at most 998 in every 1000 people with a positive RT-qPCR test result with high viral load.ConclusionsThe Innova LFT can be useful for identifying infections among adults who report no symptoms of covid-19, particularly those with high viral load who are more likely to infect others. The number of asymptomatic adults with lower Ct (indicating higher viral load) missed by LFT, although small, should be considered when using single LFT in high consequence settings. Clear and accurate communication with the public about how to interpret test results is important, given the chance of missing some cases, even at high viral loads. Further research is needed to understand how infectiousness is reflected in the viral antigen shedding detected by LFT versus the viral loads approximated by RT-qPCR.

Project description:BackgroundThe SARS-CoV-2 pandemic presented a challenge for caregiving relatives in the home care setting. Caregivers can transmit SARS-CoV-2 to their relatives who are often at high risk for a severe course of COVID-19. Regular testing of asymptomatic caregivers for SARS-CoV-2 may reduce the risk of transmission. The optimal method and frequency of regular asymptomatic testing is unknown. We conducted a prospective, randomised trial to assess the feasibility, recruitment and acceptance of different testing frequencies. This serves to inform a future definitive randomised controlled trial.MethodsWe carried out a parallel three-armed feasibility trial, enrolling adult participants who provided home-based care for a relative at least twice a week. Participants were randomly assigned using sealed envelopes to either conduct saliva-based antigen self-testing at a frequency of once a week (group I), twice a week (group II), or every two days (group III). The participants completed questionnaires on a weekly basis. Main outcome measures were feasibility of recruitment, adherence to self-tests and distress caused by self-testing. We further collected data on the use of mouth-nose mask.ResultsFrom 25 March to 7 May 2021 we assessed 27 participants and randomised 26 in the study: 8 participants in group I, 8 in group II and 10 in group III. All participants completed the study. In group I 48/48 (100.0%; 95% CI 92.6% to 100.0%), in group II 93/96 (96.9%; 95% CI 91.2% to 98.9%) and in group III 209/210 (99.5%; 95% CI 97.4% to 99.9%) self-tests were carried out at home. Participants did not perceive regular self-testing as burdensome in any of the study arms. We did not observe any infection with SARS-CoV-2. During the study, mask adherence decreased from 35% to 19% in all groups.ConclusionConducting such a study was feasible. The participants tolerated regular self-testing well, which was reflected in a high level of test adherence. However, regular self-testing may have led to decreased protective behaviour. To demonstrate that regular asymptomatic testing reduces infection transmission, a future definitive trial should be performed at a time of a high prevalence of SARS-CoV-2 and be implemented as a multicentre study.Trial registrationThe trial is registered with the German Clinical Trials Register, DRKS00026234.

Project description:Testing remains a key tool for managing health care and making health policy during the coronavirus pandemic, and it will probably be important in future pandemics. Because of false negative and false positive tests, the observed fraction of positive tests-the surface positivity-is generally different from the fraction of infected individuals (the incidence rate of the disease). In this paper a previous method for translating surface positivity to a point estimate for incidence rate, then to an appropriate range of values for the incidence rate consistent with the model and data (the test range), and finally to the risk (the probability of including one infected individual) associated with groups of different sizes is illustrated. The method is then extended to include asymptomatic infections. To do so, the process of testing is modeled using both analysis and Monte Carlo simulation. Doing so shows that it is possible to determine point estimates for the fraction of infected and symptomatic individuals, the fraction of uninfected and symptomatic individuals, and the ratio of infected asymptomatic individuals to infected symptomatic individuals. Inclusion of symptom status generalizes the test range from an interval to a region in the plane determined by the incidence rate and the ratio of asymptomatic to symptomatic infections; likelihood methods can be used to determine the contour of the rest region. Points on this contour can be used to compute the risk (defined as the probability of including one asymptomatic infected individual) in groups of different sizes. These results have operational implications that include: positivity rate is not incidence rate; symptom status at testing can provide valuable information about asymptomatic infections; collecting information on time since putative virus exposure at testing is valuable for determining point estimates and test ranges; risk is a graded (rather than binary) function of group size; and because the information provided by testing becomes more accurate with more tests but at a decreasing rate, it is possible to over-test fixed spatial regions. The paper concludes with limitations of the method and directions for future work.

Dataset Information

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

Publications

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

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets