학술논문
Machine learning application for the prediction of SARS-CoV-2 infection using blood tests and chest radiograph.
Document Type
Academic Journal
Author
Du R; Department of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China.; Artificial Intelligence Lab, Head Office Information Technology and Health Informatics Division, Hospital Authority, Hong Kong, SAR, China.; Tsougenis ED; Artificial Intelligence Lab, Head Office Information Technology and Health Informatics Division, Hospital Authority, Hong Kong, SAR, China.; Ho JWK; The School of Biomedical Sciences, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China.; Chan JKY; Clinical Systems, Information Technology and Health Informatics Division, Hospital Authority, Hong Kong, SAR, China.; Chiu KWH; Department of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China.; Fang BXH; Department of Radiology, Queen Mary Hospital, Hong Kong, SAR, China.; Ng MY; Department of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China.; Department of Medical Imaging, The University of Hong Kong-Shenzhen Hospital, Shenzhen, China.; Leung ST; Department of Radiology, Pamela Youde Nethersole Eastern Hospital, Hong Kong, SAR, China.; Lo CSY; Department of Radiology, Hong Kong Sanatorium & Hospital, Hong Kong, SAR, China.; Wong HF; Department of Radiology, Queen Mary Hospital, Hong Kong, SAR, China.; Lam HS; Department of Radiology, Queen Mary Hospital, Hong Kong, SAR, China.; Chiu LJ; Department of Radiology and Imaging, Queen Elizabeth Hospital, Hong Kong, SAR, China.; So TY; Department of Imaging and Interventional Radiology, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong, China.; Wong KT; Department of Imaging and Interventional Radiology, Prince of Wales Hospital, Hong Kong, SAR, China.; Wong YCI; Department of Radiology, Tuen Muen Hospital, Hong Kong, SAR, China.; Yu K; Department of Radiology, Tuen Muen Hospital, Hong Kong, SAR, China.; Yeung YC; Department of Medicine, Princess Margaret Hospital, Hong Kong, SAR, China.; Chik T; Department of Medicine, Princess Margaret Hospital, Hong Kong, SAR, China.; Pang JWK; Health Informatics, Information Technology and Health Informatics Division, Hospital Authority, Hong Kong, SAR, China.; Wai AK; Emergency Medicine Unit, Li Ka Shing, Faculty of Medicine, The University of Hong Kong, Hong Kong, China.; Kuo MD; Department of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China.; Lam TPW; Department of Radiology, Queen Mary Hospital, Hong Kong, SAR, China.; Khong PL; Department of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China.; Cheung NT; Information Technology and Health Informatics Division, Hospital Authority, Hong Kong, SAR, China.; Vardhanabhuti V; Department of Diagnostic Radiology, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, SAR, China. varv@hku.hk.
Source
Publisher: Nature Publishing Group Country of Publication: England NLM ID: 101563288 Publication Model: Electronic Cited Medium: Internet ISSN: 2045-2322 (Electronic) Linking ISSN: 20452322 NLM ISO Abbreviation: Sci Rep Subsets: MEDLINE
Subject
Language
English
Abstract
Triaging and prioritising patients for RT-PCR test had been essential in the management of COVID-19 in resource-scarce countries. In this study, we applied machine learning (ML) to the task of detection of SARS-CoV-2 infection using basic laboratory markers. We performed the statistical analysis and trained an ML model on a retrospective cohort of 5148 patients from 24 hospitals in Hong Kong to classify COVID-19 and other aetiology of pneumonia. We validated the model on three temporal validation sets from different waves of infection in Hong Kong. For predicting SARS-CoV-2 infection, the ML model achieved high AUCs and specificity but low sensitivity in all three validation sets (AUC: 89.9-95.8%; Sensitivity: 55.5-77.8%; Specificity: 91.5-98.3%). When used in adjunction with radiologist interpretations of chest radiographs, the sensitivity was over 90% while keeping moderate specificity. Our study showed that machine learning model based on readily available laboratory markers could achieve high accuracy in predicting SARS-CoV-2 infection.
(© 2021. The Author(s).)
(© 2021. The Author(s).)