{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,10,10]],"date-time":"2026-10-10T01:00:18Z","timestamp":1791594018598,"version":"4.3.4"},"posted":{"date-parts":[[2024,6,19]]},"group-title":"Public and Global Health","reference-count":25,"publisher":"openRxiv","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2024,6,19]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                <jats:sec>\n                  <jats:title>Background<\/jats:title>\n                  <jats:p>Computer-aided detection (CAD) algorithms for automated chest X-ray (CXR) reading have been endorsed by the World Health Organization for tuberculosis (TB) triage, but independent, multi-country assessment and comparison of current products are needed to guide implementation.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We conducted a head-to-head evaluation of five CAD algorithms for TB triage across seven countries. We included CXRs from adults who presented to outpatient facilities with at least two weeks of cough in India, Madagascar, the Philippines, South Africa, Tanzania, Uganda, and Vietnam. The participants completed a standard evaluation for pulmonary TB, including sputum collection for Xpert MTB\/RIF Ultra and culture. Against a microbiological reference standard, we calculated and compared the accuracy overall, by country and key groups for five CAD algorithms: CAD4TB (Delft Imaging), INSIGHT CXR (Lunit), DrAid (Vinbrain), Genki (Deeptek), and qXR (qure.AI). We determined the area under the ROC curve (AUC) and if any CAD product could achieve the minimum target accuracy for a TB triage test (\u226590% sensitivity and \u226570% specificity). We then applied country- and population-specific thresholds and recalculated accuracy to assess any improvement in performance.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>Of 3,927 individuals included, the median age was 41 years (IQR 29-54), 12.9% were people living with HIV (PLWH), 8.2% living with diabetes, and 21.2% had a prior history of TB. The overall AUC ranged from 0.774-0.819, and specificity ranged from 64.8-73.8% at 90% sensitivity. CAD4TB had the highest overall accuracy (73.8% specific, 95% CI 72.2-75.4, at 90% sensitivity), although qXR and INSIGHT CXR also achieved the target 70% specificity. There was heterogeneity in accuracy by country, and females and PLWH had lower sensitivity while males and people with a history of TB had lower specificity. The performance remained stable regardless of diabetes status. When country- and population-specific thresholds were applied, at least one CAD product could achieve or approach the target accuracy for each country and sub-group, except for PLWH and those with a history of TB.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>Multiple CAD algorithms can achieve or exceed the minimum target accuracy for a TB triage test, with improvement when using setting- or population-specific thresholds. Further efforts are needed to integrate CAD into routine TB case detection programs in high-burden communities.<\/jats:p>\n                <\/jats:sec>","DOI":"10.1101\/2024.06.19.24309061","type":"posted-content","created":{"date-parts":[[2024,6,19]],"date-time":"2024-06-19T15:15:17Z","timestamp":1718810117000},"source":"Crossref","is-referenced-by-count":11,"title":["An independent, multi-country head-to-head accuracy comparison of automated chest x-ray algorithms for the triage of pulmonary tuberculosis"],"prefix":"10.64898","author":[{"given":"William","family":"Worodria","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Robert","family":"Castro","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sandra V.","family":"Kik","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Victoria","family":"Dalay","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Brigitta","family":"Derendinger","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Charles","family":"Festo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Thanh Quoc","family":"Nguyen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mihaja","family":"Raberahona","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Swati","family":"Sudarsan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alfred","family":"Andama","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Balamugesh","family":"Thangakunam","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Issa","family":"Lyimo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Viet Nhung","family":"Nguyen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Rivo","family":"Rakotoarivelo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Grant","family":"Theron","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Charles","family":"Yu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Claudia M.","family":"Denkinger","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Simon Grandjean","family":"Lapierre","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Adithya","family":"Cattamanchi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Devasahayam J.","family":"Christopher","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Devan","family":"Jaganath","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"name":"R2D2 TB Network","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2024062207350490000_2024.06.19.24309061v1.1","unstructured":"World Health Organization. Global Tuberculosis Report 2023. Geneva: WHO, 2023."},{"key":"2024062207350490000_2024.06.19.24309061v1.2","unstructured":"World Health Organization. WHO consolidated guidelines on tuberculosis. Module 2: screening \u2013 systematic screening for tuberculosis disease. Geneva: WHO, 2021. 2021."},{"key":"2024062207350490000_2024.06.19.24309061v1.3","unstructured":"World Health Organization. High priority target product profiles for new tuberculosis diagnostics: report of a consensus meeting, 28-29 April 2014, Geneva, Switzerland. 2014."},{"key":"2024062207350490000_2024.06.19.24309061v1.4","unstructured":"FIND. Digital Chest Radiography and Computer-Aided Detection (CAD) Solutions for Tuberculosis Diagnostics: Technology Landscape Analysis. FIND: Geneva, 2021."},{"key":"2024062207350490000_2024.06.19.24309061v1.5","doi-asserted-by":"crossref","unstructured":"Vo LNQ , Codlin A , Ngo TD , et al. Early Evaluation of an Ultra-Portable X-ray System for Tuberculosis Active Case Finding. Trop Med Infect Dis 2021; 6(3).","DOI":"10.3390\/tropicalmed6030163"},{"key":"2024062207350490000_2024.06.19.24309061v1.6","unstructured":"STOP TB and FIND. AI4Hlth. 2021. https:\/\/www.ai4hlth.org\/."},{"issue":"12","key":"2024062207350490000_2024.06.19.24309061v1.7","doi-asserted-by":"crossref","first-page":"e0277393","DOI":"10.1371\/journal.pone.0277393","article-title":"Economic analysis of different throughput scenarios and implementation strategies of computer-aided detection software as a screening and triage test for pulmonary TB","volume":"17","year":"2022","journal-title":"PLoS One"},{"issue":"9","key":"2024062207350490000_2024.06.19.24309061v1.8","doi-asserted-by":"crossref","first-page":"e543","DOI":"10.1016\/S2589-7500(21)00116-3","article-title":"Tuberculosis detection from chest x-rays for triaging in a high tuberculosis-burden setting: an evaluation of five artificial intelligence algorithms","volume":"3","year":"2021","journal-title":"The Lancet Digital Health"},{"issue":"6","key":"2024062207350490000_2024.06.19.24309061v1.9","doi-asserted-by":"crossref","first-page":"e0000067","DOI":"10.1371\/journal.pdig.0000067","article-title":"Comparing different versions of computer-aided detection products when reading chest X-rays for tuberculosis","volume":"1","year":"2022","journal-title":"PLOS Digit Health"},{"issue":"8","key":"2024062207350490000_2024.06.19.24309061v1.10","doi-asserted-by":"crossref","first-page":"1390","DOI":"10.1093\/cid\/ciab639","article-title":"Chest X-ray Analysis With Deep Learning-Based Software as a Triage Test for Pulmonary Tuberculosis: An Individual Patient Data Meta-Analysis of Diagnostic Accuracy","volume":"74","year":"2022","journal-title":"Clin Infect Dis"},{"issue":"1","key":"2024062207350490000_2024.06.19.24309061v1.11","doi-asserted-by":"crossref","first-page":"23895","DOI":"10.1038\/s41598-021-03265-0","article-title":"Independent evaluation of 12 artificial intelligence solutions for the detection of tuberculosis","volume":"11","year":"2021","journal-title":"Scientific Reports"},{"issue":"7","key":"2024062207350490000_2024.06.19.24309061v1.12","doi-asserted-by":"crossref","first-page":"e0000402","DOI":"10.1371\/journal.pgph.0000402","article-title":"Diagnostic accuracy of three computer-aided detection systems for detecting pulmonary tuberculosis on chest radiography when used for screening: Analysis of an international, multicenter migrants screening study","volume":"3","year":"2023","journal-title":"PLOS Glob Public Health"},{"key":"2024062207350490000_2024.06.19.24309061v1.13","unstructured":"Rapid Research in Diagnostics Development for TB Network (R2D2). https:\/\/www.r2d2tbnetwork.org\/ (accessed 19 June 2024)."},{"key":"2024062207350490000_2024.06.19.24309061v1.14","unstructured":"Cepheid. Xpert MTB\/RIF Ultra Package Insert. Sunnyvale: Cepheid, 2019.."},{"key":"2024062207350490000_2024.06.19.24309061v1.15","unstructured":"Global Laboratory Initiative. Mycobacteriology Laboratory Manual. Geneva: Stop TB Partnership, 2014.."},{"key":"2024062207350490000_2024.06.19.24309061v1.16","doi-asserted-by":"publisher","DOI":"10.1136\/bmjopen-2016-012799"},{"issue":"1","key":"2024062207350490000_2024.06.19.24309061v1.17","doi-asserted-by":"crossref","first-page":"15000","DOI":"10.1038\/s41598-019-51503-3","article-title":"Using artificial intelligence to read chest radiographs for tuberculosis detection: A multi-site evaluation of the diagnostic accuracy of three deep learning systems","volume":"9","year":"2019","journal-title":"Sci Rep"},{"issue":"2","key":"2024062207350490000_2024.06.19.24309061v1.18","doi-asserted-by":"crossref","first-page":"ofae020","DOI":"10.1093\/ofid\/ofae020","article-title":"Diagnostic Accuracy of Computer-Aided Detection During Active Case Finding for Pulmonary Tuberculosis in Africa: A Systematic Review and Meta-analysis","volume":"11","year":"2024","journal-title":"Open forum infectious diseases"},{"issue":"9","key":"2024062207350490000_2024.06.19.24309061v1.19","doi-asserted-by":"crossref","first-page":"1088","DOI":"10.5588\/ijtld.17.0827","article-title":"Computer-assisted chest radiography reading for tuberculosis screening in people living with diabetes mellitus","volume":"22","year":"2018","journal-title":"Int J Tuberc Lung Dis"},{"issue":"1","key":"2024062207350490000_2024.06.19.24309061v1.20","doi-asserted-by":"crossref","first-page":"6276","DOI":"10.1038\/s41598-020-63084-7","article-title":"Evaluation of computer aided detection of tuberculosis on chest radiography among people with diabetes in Karachi Pakistan","volume":"10","year":"2020","journal-title":"Sci Rep"},{"key":"2024062207350490000_2024.06.19.24309061v1.21","first-page":"15","article-title":"Diagnostic accuracy of a commercially available, deep learning-based chest X-ray interpretation software for detecting culture-confirmed pulmonary tuberculosis","volume":"122","year":"2022","journal-title":"International journal of infectious diseases: IJID: official publication of the International Society for Infectious Diseases"},{"key":"2024062207350490000_2024.06.19.24309061v1.22","doi-asserted-by":"publisher","DOI":"10.5588\/ijtld.13.0325"},{"key":"2024062207350490000_2024.06.19.24309061v1.23","unstructured":"Special Programme for TDR SCI. Determining the local calibration of computer-assisted detection (CAD) thresholds and other parameters. Geneva: WHO, 2021."},{"key":"2024062207350490000_2024.06.19.24309061v1.24","unstructured":"FIND. Digital Chest Radiography and Computer-aided Detection (CAD) Solutions for Tuberculosis Diagnostics: Technology Landscape Analysis. Geneva: FIND, 2021. Available at: https:\/\/www.finddx.org\/wp-content\/uploads\/2021\/04\/FIND-CXR-CAD-solutions-for-TB-diagnosis-7Apr2021.pdf."},{"key":"2024062207350490000_2024.06.19.24309061v1.25","unstructured":"World health Organization. Call for expression of interest by manufacturers of software for computer-aided detection of tuberculosis (CAD) to submit products for WHO expert assessment. Available at: https:\/\/cdn.who.int\/media\/docs\/default-source\/hq-tuberculosis\/public-calls\/cad.developer.submission.requirements.for.who.evaluation.pdf?sfvrsn=3425d4a2_3. Last accessed 13 June 2024."}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2024.06.19.24309061","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T20:03:43Z","timestamp":1768507423000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2024.06.19.24309061"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,6,19]]},"references-count":25,"URL":"https:\/\/doi.org\/10.1101\/2024.06.19.24309061","relation":{},"subject":[],"published":{"date-parts":[[2024,6,19]]},"subtype":"preprint"}}