Performance of a Chest Radiograph AI Diagnostic Tool for COVID-19: A Prospective Observational Study

Ju Sun; Le Peng; Taihui Li; Dyah Adila; Zach Zaiman; Genevieve B Melton-Meaux; Nicholas E Ingraham; Eric Murray; Daniel Boley; Sean Switzer; John L Burns; Kun Huang; Tadashi Allen; Scott D Steenburg; Judy Wawira Gichoya; Erich Kummerfeld; Christopher J Tignanelli

doi:10.1148/ryai.210217

Performance of a Chest Radiograph AI Diagnostic Tool for COVID-19: A Prospective Observational Study

Radiol Artif Intell. 2022 Jun 1;4(4):e210217. doi: 10.1148/ryai.210217. eCollection 2022 Jul.

Authors

Ju Sun¹, Le Peng¹, Taihui Li¹, Dyah Adila¹, Zach Zaiman¹, Genevieve B Melton-Meaux¹, Nicholas E Ingraham¹, Eric Murray¹, Daniel Boley¹, Sean Switzer¹, John L Burns¹, Kun Huang¹, Tadashi Allen¹, Scott D Steenburg¹, Judy Wawira Gichoya¹, Erich Kummerfeld^#¹, Christopher J Tignanelli^#¹

Affiliation

¹ Department of Computer Science and Engineering (J.S., L.P., T.L., D.A., D.B.), Institute for Health Informatics (G.B.M.M., E.K., C.J.T.), Department of Surgery (G.B.M.M., C.J.T.), Department of Medicine, Division of Pulmonary and Critical Care (N.E.I.), Department of Medicine (S.S.), and Department of Radiology (T.A.), University of Minnesota, 420 Delaware St SE, Minneapolis, MN 55455; Departments of Computer Science (Z.Z.) and Radiology (J.W.G.), Emory University, Atlanta, Ga; M Health Fairview Informatics, Minneapolis, Minn (E.M.); The School of Medicine (J.L.B., K.H.) and Department of Radiology (S.D.S.), Indiana University, Indianapolis, Ind; and Department of Surgery, North Memorial Health Hospital, Robbinsdale, Minn (C.J.T.).

^# Contributed equally.

Abstract

Purpose: To conduct a prospective observational study across 12 U.S. hospitals to evaluate real-time performance of an interpretable artificial intelligence (AI) model to detect COVID-19 on chest radiographs.

Materials and methods: A total of 95 363 chest radiographs were included in model training, external validation, and real-time validation. The model was deployed as a clinical decision support system, and performance was prospectively evaluated. There were 5335 total real-time predictions and a COVID-19 prevalence of 4.8% (258 of 5335). Model performance was assessed with use of receiver operating characteristic analysis, precision-recall curves, and F1 score. Logistic regression was used to evaluate the association of race and sex with AI model diagnostic accuracy. To compare model accuracy with the performance of board-certified radiologists, a third dataset of 1638 images was read independently by two radiologists.

Results: Participants positive for COVID-19 had higher COVID-19 diagnostic scores than participants negative for COVID-19 (median, 0.1 [IQR, 0.0-0.8] vs 0.0 [IQR, 0.0-0.1], respectively; P < .001). Real-time model performance was unchanged over 19 weeks of implementation (area under the receiver operating characteristic curve, 0.70; 95% CI: 0.66, 0.73). Model sensitivity was higher in men than women (P = .01), whereas model specificity was higher in women (P = .001). Sensitivity was higher for Asian (P = .002) and Black (P = .046) participants compared with White participants. The COVID-19 AI diagnostic system had worse accuracy (63.5% correct) compared with radiologist predictions (radiologist 1 = 67.8% correct, radiologist 2 = 68.6% correct; McNemar P < .001 for both).

Conclusion: AI-based tools have not yet reached full diagnostic potential for COVID-19 and underperform compared with radiologist prediction.Keywords: Diagnosis, Classification, Application Domain, Infection, Lung Supplemental material is available for this article.. © RSNA, 2022.

Keywords: Application Domain; Classification; Diagnosis; Infection; Lung.