Determination of disease severity in COVID-19 patients using deep learning in chest X-ray images
PURPOSEChest X-ray plays a key role in diagnosis and management of COVID-19 patients and imaging features associated with clinical elements may assist with the development or validation of automated image analysis tools. We aimed to identify associations between clinical and radiographic features as...
Main Authors: | , , , , , , , , , , , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Galenos Publishing House
2021-01-01
|
Series: | Diagnostic and Interventional Radiology |
Online Access: |
http://www.dirjournal.org/archives/archive-detail/article-preview/determination-of-disease-severity-in-covd-19-patie/54534
|
_version_ | 1797691372591906816 |
---|---|
author | Maxime Blain Michael T. Kassin Nicole Varble Xiaosong Wang Ziyue Xu Daguang Xu Gianpaolo Carrafiello Valentina Vespro Elvira Stellato Anna Maria Ierardi Letizia Di Meglio Robert D. Suh Stephanie A. Walker Sheng Xu Thomas H. Sanford Evrim B. Turkbey Stephanie Harmon Baris Turkbey Bradford J. Wood |
author_facet | Maxime Blain Michael T. Kassin Nicole Varble Xiaosong Wang Ziyue Xu Daguang Xu Gianpaolo Carrafiello Valentina Vespro Elvira Stellato Anna Maria Ierardi Letizia Di Meglio Robert D. Suh Stephanie A. Walker Sheng Xu Thomas H. Sanford Evrim B. Turkbey Stephanie Harmon Baris Turkbey Bradford J. Wood |
author_sort | Maxime Blain |
collection | DOAJ |
description | PURPOSEChest X-ray plays a key role in diagnosis and management of COVID-19 patients and imaging features associated with clinical elements may assist with the development or validation of automated image analysis tools. We aimed to identify associations between clinical and radiographic features as well as to assess the feasibility of deep learning applied to chest X-rays in the setting of an acute COVID-19 outbreak.METHODSA retrospective study of X-rays, clinical, and laboratory data was performed from 48 SARS-CoV-2 RT-PCR positive patients (age 60±17 years, 15 women) between February 22 and March 6, 2020 from a tertiary care hospital in Milan, Italy. Sixty-five chest X-rays were reviewed by two radiologists for alveolar and interstitial opacities and classified by severity on a scale from 0 to 3. Clinical factors (age, symptoms, comorbidities) were investigated for association with opacity severity and also with placement of central line or endotracheal tube. Deep learning models were then trained for two tasks: lung segmentation and opacity detection. Imaging characteristics were compared to clinical datapoints using the unpaired student’s t-test or Mann-Whitney U test. Cohen’s kappa analysis was used to evaluate the concordance of deep learning to conventional radiologist interpretation.RESULTSFifty-six percent of patients presented with alveolar opacities, 73% had interstitial opacities, and 23% had normal X-rays. The presence of alveolar or interstitial opacities was statistically correlated with age (P = 0.008) and comorbidities (P = 0.005). The extent of alveolar or interstitial opacities on baseline X-ray was significantly associated with the presence of endotracheal tube (P = 0.0008 and P = 0.049) or central line (P = 0.003 and P = 0.007). In comparison to human interpretation, the deep learning model achieved a kappa concordance of 0.51 for alveolar opacities and 0.71 for interstitial opacities.CONCLUSIONChest X-ray analysis in an acute COVID-19 outbreak showed that the severity of opacities was associated with advanced age, comorbidities, as well as acuity of care. Artificial intelligence tools based upon deep learning of COVID-19 chest X-rays are feasible in the acute outbreak setting. |
first_indexed | 2024-03-12T02:12:23Z |
format | Article |
id | doaj.art-a50fc269bfff4b88b1e67a02b77dd973 |
institution | Directory Open Access Journal |
issn | 1305-3825 1305-3612 |
language | English |
last_indexed | 2024-03-12T02:12:23Z |
publishDate | 2021-01-01 |
publisher | Galenos Publishing House |
record_format | Article |
series | Diagnostic and Interventional Radiology |
spelling | doaj.art-a50fc269bfff4b88b1e67a02b77dd9732023-09-06T12:25:34ZengGalenos Publishing HouseDiagnostic and Interventional Radiology1305-38251305-36122021-01-01271202710.5152/dir.2020.2020513049054Determination of disease severity in COVID-19 patients using deep learning in chest X-ray imagesMaxime Blain0Michael T. Kassin1Nicole Varble2Xiaosong Wang3Ziyue Xu4Daguang Xu5Gianpaolo Carrafiello6Valentina Vespro7Elvira Stellato8Anna Maria Ierardi9Letizia Di Meglio10Robert D. Suh11Stephanie A. Walker12Sheng Xu13Thomas H. Sanford14Evrim B. Turkbey15Stephanie Harmon16Baris Turkbey17Bradford J. Wood18 Center for Interventional Oncology, National Institutes of Health Clinical Center and National Cancer Institute, Bethesda, Maryland, USA Center for Interventional Oncology, National Institutes of Health Clinical Center and National Cancer Institute, Bethesda, Maryland, USA Philips Research North America, Cambridge, Massachusetts, USA NVIDIA Corporation, Bethesda, Maryland, USA NVIDIA Corporation, Bethesda, Maryland, USA Center for Interventional Oncology, National Institutes of Health Clinical Center and National Cancer Institute, Bethesda, Maryland, USA Department of Radiology,Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, University of Milan, Italy Department of Radiology,Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, University of Milan, Italy Department of Radiology,Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, University of Milan, Italy Department of Radiology,Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, University of Milan, Italy Department of Radiology,Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, University of Milan, Italy Department of Radiology, University of California Los Angeles, Los Angeles, California, USA National Cancer Institute, National Institutes of Health, Bethesda, Maryland, USA Center for Interventional Oncology, National Institutes of Health Clinical Center and National Cancer Institute, Bethesda, Maryland, USA Center for Interventional Oncology, National Institutes of Health Clinical Center and National Cancer Institute, Bethesda, Maryland, USA Molecular Imaging Program, National Institutes of Health, Bethesda, Maryland, USA National Cancer Institute, National Institutes of Health, Bethesda, Maryland, USA Department of Radiology and Imaging Sciences, National Institutes of Health, Bethesda, Mayland, USA Center for Interventional Oncology, National Institutes of Health Clinical Center and National Cancer Institute, Bethesda, Maryland, USA PURPOSEChest X-ray plays a key role in diagnosis and management of COVID-19 patients and imaging features associated with clinical elements may assist with the development or validation of automated image analysis tools. We aimed to identify associations between clinical and radiographic features as well as to assess the feasibility of deep learning applied to chest X-rays in the setting of an acute COVID-19 outbreak.METHODSA retrospective study of X-rays, clinical, and laboratory data was performed from 48 SARS-CoV-2 RT-PCR positive patients (age 60±17 years, 15 women) between February 22 and March 6, 2020 from a tertiary care hospital in Milan, Italy. Sixty-five chest X-rays were reviewed by two radiologists for alveolar and interstitial opacities and classified by severity on a scale from 0 to 3. Clinical factors (age, symptoms, comorbidities) were investigated for association with opacity severity and also with placement of central line or endotracheal tube. Deep learning models were then trained for two tasks: lung segmentation and opacity detection. Imaging characteristics were compared to clinical datapoints using the unpaired student’s t-test or Mann-Whitney U test. Cohen’s kappa analysis was used to evaluate the concordance of deep learning to conventional radiologist interpretation.RESULTSFifty-six percent of patients presented with alveolar opacities, 73% had interstitial opacities, and 23% had normal X-rays. The presence of alveolar or interstitial opacities was statistically correlated with age (P = 0.008) and comorbidities (P = 0.005). The extent of alveolar or interstitial opacities on baseline X-ray was significantly associated with the presence of endotracheal tube (P = 0.0008 and P = 0.049) or central line (P = 0.003 and P = 0.007). In comparison to human interpretation, the deep learning model achieved a kappa concordance of 0.51 for alveolar opacities and 0.71 for interstitial opacities.CONCLUSIONChest X-ray analysis in an acute COVID-19 outbreak showed that the severity of opacities was associated with advanced age, comorbidities, as well as acuity of care. Artificial intelligence tools based upon deep learning of COVID-19 chest X-rays are feasible in the acute outbreak setting. http://www.dirjournal.org/archives/archive-detail/article-preview/determination-of-disease-severity-in-covd-19-patie/54534 |
spellingShingle | Maxime Blain Michael T. Kassin Nicole Varble Xiaosong Wang Ziyue Xu Daguang Xu Gianpaolo Carrafiello Valentina Vespro Elvira Stellato Anna Maria Ierardi Letizia Di Meglio Robert D. Suh Stephanie A. Walker Sheng Xu Thomas H. Sanford Evrim B. Turkbey Stephanie Harmon Baris Turkbey Bradford J. Wood Determination of disease severity in COVID-19 patients using deep learning in chest X-ray images Diagnostic and Interventional Radiology |
title | Determination of disease severity in COVID-19 patients using deep learning in chest X-ray images |
title_full | Determination of disease severity in COVID-19 patients using deep learning in chest X-ray images |
title_fullStr | Determination of disease severity in COVID-19 patients using deep learning in chest X-ray images |
title_full_unstemmed | Determination of disease severity in COVID-19 patients using deep learning in chest X-ray images |
title_short | Determination of disease severity in COVID-19 patients using deep learning in chest X-ray images |
title_sort | determination of disease severity in covid 19 patients using deep learning in chest x ray images |
url |
http://www.dirjournal.org/archives/archive-detail/article-preview/determination-of-disease-severity-in-covd-19-patie/54534
|
work_keys_str_mv | AT maximeblain determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT michaeltkassin determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT nicolevarble determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT xiaosongwang determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT ziyuexu determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT daguangxu determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT gianpaolocarrafiello determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT valentinavespro determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT elvirastellato determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT annamariaierardi determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT letiziadimeglio determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT robertdsuh determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT stephanieawalker determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT shengxu determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT thomashsanford determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT evrimbturkbey determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT stephanieharmon determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT baristurkbey determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages AT bradfordjwood determinationofdiseaseseverityincovid19patientsusingdeeplearninginchestxrayimages |