Identification of COPD from primary care spirometry with artificial intelligence (AI) interpretation software: a retrospective, diagnostic accuracy study

<bold>Background:</bold> Primary care clinicians have previously reported low levels of confidence with interpreting spirometry, with significant numbers of undiagnosed or late diagnosed with COPD. <bold>Aim:</bold> Assess the discriminative accuracy of AI software to identify COPD and other respiratory conditions from primary care spirometry. <bold>Methods:</bold> In this retrospective, blinded analysis, we analysed hand-held spirometry data from consecutive patients attending primary care respiratory clinics. The index diagnosis was the preferred diagnosis generated by AI (random forest machine learning) interpretation of raw spirometry data and basic demographics. The reference diagnosis was by consensus of expert pulmonologists with access to medical notes and investigation results. The primary outcome was sensitivity of AI in identifying COPD. <bold>Results:</bold> From the 1113 patients, 543 (48.8%) had a reference diagnosis of COPD. AI preferred diagnosis achieved a sensitivity of 84% (95%CI 80.6-87.0) with an area under curve (AUC) of 0.909 (Fig 1). AI preferred diagnosis also identified patients with interstitial lung disease and asthma with an AUC of 0.899 and 0.813 respectively. <bold>Conclusion:</bold> AI spirometry interpretation software achieved excellent performance in identifying COPD from primary care spirometry and basic demographic data and may support the accurate diagnosis of COPD and other respiratory conditions. <fig><object-id>erj;64/suppl_68/OA1047/F1</object-id><object-id>F1</object-id><object-id>F1</object-id><graphic></graphic></fig>

Authors

Institutions

Publication Details

Published
2024-09-14
DOI
https://doi.org/10.1183/13993003.congress-2024.oa1047
Citations
2
Primary Topic
Chronic Obstructive Pulmonary Disease (COPD) Research
Type
article
Field-Weighted Citation Impact
0.84
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Identification of COPD from primary care spirometry with artificial intelligence (AI) interpretation software: a retrospective, diagnostic accuracy study

Suhani Patel, Ethaar El-Emir, Samantha Kon, Matthew Rutter et al.
2 citations
Chronic Obstructive Pulmonary Disease (COPD) Research
0.84
article

Identification of COPD from primary care spirometry with artificial intelligence (AI) interpretation software: a retrospective, diagnostic accuracy study

Suhani Patel, Ethaar El-Emir, Samantha Kon, Matthew Rutter, Vicky Moore, Nicholas S Hopkinson, Joanna Purvis, Stephanie Taylor, Paul Desbordes, David A. Scott, Karl Sylvester, Michael D. Morgan, William D‐C Man, Maarten De Vos, Gillian Doe, Rachael A Evans, Anthony Paulo Sunjaya, Benoit Cuyvers, Joanna Shakespeare, Richard Russell, A Toby Prevost, Ahmed Elmahy, Ian Jarrold, Karolien Van Orshoven, George Edwards, Marko Topalovic, Jennifer Harvey, Nannette Spain
article en
2 citations

Abstract

<bold>Background:</bold> Primary care clinicians have previously reported low levels of confidence with interpreting spirometry, with significant numbers of undiagnosed or late diagnosed with COPD. <bold>Aim:</bold> Assess the discriminative accuracy of AI software to identify COPD and other respiratory conditions from primary care spirometry. <bold>Methods:</bold> In this retrospective, blinded analysis, we analysed hand-held spirometry data from consecutive patients attending primary care respiratory clinics. The index diagnosis was the preferred diagnosis generated by AI (random forest machine learning) interpretation of raw spirometry data and basic demographics. The reference diagnosis was by consensus of expert pulmonologists with access to medical notes and investigation results. The primary outcome was sensitivity of AI in identifying COPD. <bold>Results:</bold> From the 1113 patients, 543 (48.8%) had a reference diagnosis of COPD. AI preferred diagnosis achieved a sensitivity of 84% (95%CI 80.6-87.0) with an area under curve (AUC) of 0.909 (Fig 1). AI preferred diagnosis also identified patients with interstitial lung disease and asthma with an AUC of 0.899 and 0.813 respectively. <bold>Conclusion:</bold> AI spirometry interpretation software achieved excellent performance in identifying COPD from primary care spirometry and basic demographic data and may support the accurate diagnosis of COPD and other respiratory conditions. <fig><object-id>erj;64/suppl_68/OA1047/F1</object-id><object-id>F1</object-id><object-id>F1</object-id><graphic></graphic></fig>

University Hospitals Birmingham NHS Foundation Trust (GB), University of Leicester (GB), Queen Mary University of London (GB), King's College London (GB), Guy's and St Thomas' NHS Foundation Trust (GB), George Eliot Hospital (GB), House of Representatives (NL), British Lung Foundation (GB), Cambridge University Hospitals NHS Foundation Trust (GB), University Hospitals Coventry and Warwickshire NHS Trust (GB), Lung Institute (US), The George Institute for Global Health (AU), Anheuser-Busch InBev (Belgium) (BE), The London College (GB), Science Oxford (GB), Southampton City Council (GB), University of Southampton (GB), Hull York Medical School (GB)
Openalex Percentile: Top 23%
Chronic Obstructive Pulmonary Disease (COPD) Research
0.84
DOISourceOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.