Machine learning to support visual auditing of home-based lateral flow immunoassay self-test results for SARS-CoV-2 antibodies.

Databases Public health

Journal

Communications medicine
ISSN: 2730-664X
Titre abrégé: Commun Med (Lond)
Pays: England
ID NLM: 9918250414506676

Informations de publication

Date de publication:
2022
Historique:
received: 09 02 2022
accepted: 15 06 2022
entrez: 11 7 2022
pubmed: 12 7 2022
medline: 12 7 2022
Statut: epublish

Résumé

Lateral flow immunoassays (LFIAs) are being used worldwide for COVID-19 mass testing and antibody prevalence studies. Relatively simple to use and low cost, these tests can be self-administered at home, but rely on subjective interpretation of a test line by eye, risking false positives and false negatives. Here, we report on the development of ALFA (Automated Lateral Flow Analysis) to improve reported sensitivity and specificity. Our computational pipeline uses machine learning, computer vision techniques and signal processing algorithms to analyse images of the Fortress LFIA SARS-CoV-2 antibody self-test, and subsequently classify results as invalid, IgG negative and IgG positive. A large image library of 595,339 participant-submitted test photographs was created as part of the REACT-2 community SARS-CoV-2 antibody prevalence study in England, UK. Alongside ALFA, we developed an analysis toolkit which could also detect device blood leakage issues. Automated analysis showed substantial agreement with human experts (Cohen's kappa 0.90-0.97) and performed consistently better than study participants, particularly for weak positive IgG results. Specificity (98.7-99.4%) and sensitivity (90.1-97.1%) were high compared with visual interpretation by human experts (ranges due to the varying prevalence of weak positive IgG tests in datasets). Given the potential for LFIAs to be used at scale in the COVID-19 response (for both antibody and antigen testing), even a small improvement in the accuracy of the algorithms could impact the lives of millions of people by reducing the risk of false-positive and false-negative result read-outs by members of the public. Our findings support the use of machine learning-enabled automated reading of at-home antibody lateral flow tests as a tool for improved accuracy for population-level community surveillance.

Sections du résumé

Background UNASSIGNED
Lateral flow immunoassays (LFIAs) are being used worldwide for COVID-19 mass testing and antibody prevalence studies. Relatively simple to use and low cost, these tests can be self-administered at home, but rely on subjective interpretation of a test line by eye, risking false positives and false negatives. Here, we report on the development of ALFA (Automated Lateral Flow Analysis) to improve reported sensitivity and specificity.
Methods UNASSIGNED
Our computational pipeline uses machine learning, computer vision techniques and signal processing algorithms to analyse images of the Fortress LFIA SARS-CoV-2 antibody self-test, and subsequently classify results as invalid, IgG negative and IgG positive. A large image library of 595,339 participant-submitted test photographs was created as part of the REACT-2 community SARS-CoV-2 antibody prevalence study in England, UK. Alongside ALFA, we developed an analysis toolkit which could also detect device blood leakage issues.
Results UNASSIGNED
Automated analysis showed substantial agreement with human experts (Cohen's kappa 0.90-0.97) and performed consistently better than study participants, particularly for weak positive IgG results. Specificity (98.7-99.4%) and sensitivity (90.1-97.1%) were high compared with visual interpretation by human experts (ranges due to the varying prevalence of weak positive IgG tests in datasets).
Conclusions UNASSIGNED
Given the potential for LFIAs to be used at scale in the COVID-19 response (for both antibody and antigen testing), even a small improvement in the accuracy of the algorithms could impact the lives of millions of people by reducing the risk of false-positive and false-negative result read-outs by members of the public. Our findings support the use of machine learning-enabled automated reading of at-home antibody lateral flow tests as a tool for improved accuracy for population-level community surveillance.

Identifiants

pubmed: 35814295
doi: 10.1038/s43856-022-00146-z
pii: 146
pmc: PMC9259560
doi:

Types de publication

Journal Article

Langues

eng

Pagination

78

Informations de copyright

© The Author(s) 2022.

Déclaration de conflit d'intérêts

Competing interestsThe authors declare no competing interests.

Références

Nat Med. 2021 Jul;27(7):1165-1170
pubmed: 34140702
JAMA. 2020 Jun 16;323(23):2425-2427
pubmed: 32421144
N Engl J Med. 2020 Jul 16;383(3):201-203
pubmed: 32374952
Thorax. 2020 Dec;75(12):1082-1088
pubmed: 32796119
Clin Cancer Res. 2019 Jun 1;25(11):3266-3275
pubmed: 31010833
JAMA. 2017 Aug 8;318(6):517-518
pubmed: 28727867
Open Forum Infect Dis. 2021 Oct 04;8(11):ofab496
pubmed: 34796246
IEEE Trans Med Imaging. 2016 May;35(5):1273-1284
pubmed: 26886969
Nat Med. 2019 Jan;25(1):24-29
pubmed: 30617335
Clin Infect Dis. 2021 May 4;72(9):e384-e393
pubmed: 32785665
Nat Med. 2018 Sep;24(9):1342-1350
pubmed: 30104768
Lancet. 2020 Aug 22;396(10250):535-544
pubmed: 32645347
Biochem Med (Zagreb). 2012;22(3):276-82
pubmed: 23092060
Proc Natl Acad Sci U S A. 2021 Mar 23;118(12):
pubmed: 33674422
Wellcome Open Res. 2020 Aug 25;5:200
pubmed: 33997297
Nature. 2017 Feb 2;542(7639):115-118
pubmed: 28117445
JAMA Intern Med. 2020 Jul 21;:
pubmed: 32692365
Nat Med. 2020 Jun;26(6):845-848
pubmed: 32350462
Wellcome Open Res. 2020 Jun 11;5:139
pubmed: 33748431
Lancet Reg Health Eur. 2021 May;4:100098
pubmed: 33969335

Auteurs

Nathan C K Wong (NCK)

Department of Bioengineering, Imperial College London, London, UK.

Sepehr Meshkinfamfard (S)

London Centre for Nanotechnology, University College London, London, UK.

Valérian Turbé (V)

London Centre for Nanotechnology, University College London, London, UK.

Matthew Whitaker (M)

School of Public Health, Imperial College London, London, UK.

Maya Moshe (M)

Department of Infectious Disease, Imperial College London, London, UK.

Alessia Bardanzellu (A)

Department of Bioengineering, Imperial College London, London, UK.

Tianhong Dai (T)

Department of Bioengineering, Imperial College London, London, UK.

Eduardo Pignatelli (E)

Department of Bioengineering, Imperial College London, London, UK.

Wendy Barclay (W)

Imperial College Healthcare NHS Trust, London, UK.
Department of Infectious Disease, Imperial College London, London, UK.
National Institute for Health Research Imperial Biomedical Research Centre, London, UK.

Ara Darzi (A)

Imperial College Healthcare NHS Trust, London, UK.
Institute of Global Health Innovation, Imperial College London, London, UK.
National Institute for Health Research Imperial Biomedical Research Centre, London, UK.

Paul Elliott (P)

School of Public Health, Imperial College London, London, UK.
Imperial College Healthcare NHS Trust, London, UK.
National Institute for Health Research Imperial Biomedical Research Centre, London, UK.
MRC Centre for Environment and Health, School of Public Health, Imperial College London, London, UK.

Helen Ward (H)

School of Public Health, Imperial College London, London, UK.
Imperial College Healthcare NHS Trust, London, UK.
National Institute for Health Research Imperial Biomedical Research Centre, London, UK.

Reiko J Tanaka (RJ)

Department of Bioengineering, Imperial College London, London, UK.

Graham S Cooke (GS)

Department of Infectious Disease, Imperial College London, London, UK.
National Institute for Health Research Imperial Biomedical Research Centre, London, UK.

Rachel A McKendry (RA)

London Centre for Nanotechnology, University College London, London, UK.
Division of Medicine, University College London, London, UK.

Christina J Atchison (CJ)

School of Public Health, Imperial College London, London, UK.
Imperial College Healthcare NHS Trust, London, UK.

Anil A Bharath (AA)

Department of Bioengineering, Imperial College London, London, UK.

Classifications MeSH