Automated detection of erythema migrans and other confounding skin lesions via deep learning

Philippe Burlina, Neil J. Joshi, Elise Ng, Seth D. Billings, Alison W. Rebman, John Aucott

Research output: Contribution to journalArticle

Abstract

Lyme disease can lead to neurological, cardiac, and rheumatologic complications when untreated. Timely recognition of the erythema migrans rash of acute Lyme disease by patients and clinicians is crucial to early diagnosis and treatment. Our objective in this study was to develop deep learning approaches using deep convolutional neural networks for detecting acute Lyme disease from erythema migrans images of varying quality and acquisition conditions. This study used a cross-sectional dataset of images to train a model employing a deep convolutional neural network to perform classification of erythema migrans versus other skin conditions including tinea corporis and herpes zoster, and normal, non-pathogenic skin. Evaluation of the machine's ability to classify skin types was also performed on a validation set of images. Machine performance for detecting erythema migrans was further tested against a panel of non-medical humans. Online, publicly available images of both erythema migrans and non-Lyme confounding skin lesions were mined, and combined with erythema migrans images from an ongoing, longitudinal study of participants with acute Lyme disease enrolled in 2016 and 2017 who were recruited from primary and urgent care centers. The final dataset had 1834 images, including 1718 expert clinician-curated online images from unknown individuals with erythema migrans, tinea corporis, herpes zoster, and normal skin. It also included 116 images taken of 63 research participants from the Mid-Atlantic region. Two clinicians carefully annotated all lesion images. A convenience sample of 7 non-medically-trained humans were used as a panel to compare against machine performance. We calculated several performance metrics, including accuracy and Kappa (characterizing agreement with gold standard), as well as a receiver operating characteristic curve and associated area under the curve. For detecting erythema migrans, the machine had an accuracy (95% confidence interval error margin) of 86.53% (2.70), ROCAUC of 0.9510 (0.0171) and Kappa of 0.7143. Our results suggested substantial agreement between machine and clinician criterion standard. Comparison of machine with non-medical expert human performance indicated that the machine almost always exceeded acceptable specificity, and could operate with higher sensitivity. This could have benefits for prescreening prior to physician referral, earlier treatment, and reductions in morbidity.

Original languageEnglish (US)
Pages (from-to)151-156
Number of pages6
JournalComputers in Biology and Medicine
Volume105
DOIs
StatePublished - Feb 1 2019

Fingerprint

Erythema
Skin
Learning
Lyme Disease
Acute Disease
Tinea
Herpes Zoster
Neural networks
Mid-Atlantic Region
Aptitude
Deep learning
Ambulatory Care Facilities
Exanthema
ROC Curve
Area Under Curve
Longitudinal Studies
Early Diagnosis
Primary Health Care
Referral and Consultation
Confidence Intervals

Keywords

  • Automated pre-screening
  • Convolutional neural networks
  • Deep learning
  • Erythema migrans
  • Lyme disease

ASJC Scopus subject areas

  • Computer Science Applications
  • Health Informatics

Cite this

Automated detection of erythema migrans and other confounding skin lesions via deep learning. / Burlina, Philippe; Joshi, Neil J.; Ng, Elise; Billings, Seth D.; Rebman, Alison W.; Aucott, John.

In: Computers in Biology and Medicine, Vol. 105, 01.02.2019, p. 151-156.

Research output: Contribution to journalArticle

Burlina, Philippe ; Joshi, Neil J. ; Ng, Elise ; Billings, Seth D. ; Rebman, Alison W. ; Aucott, John. / Automated detection of erythema migrans and other confounding skin lesions via deep learning. In: Computers in Biology and Medicine. 2019 ; Vol. 105. pp. 151-156.
@article{ee9359827b104b2490e8bd01f8b7f807,
title = "Automated detection of erythema migrans and other confounding skin lesions via deep learning",
abstract = "Lyme disease can lead to neurological, cardiac, and rheumatologic complications when untreated. Timely recognition of the erythema migrans rash of acute Lyme disease by patients and clinicians is crucial to early diagnosis and treatment. Our objective in this study was to develop deep learning approaches using deep convolutional neural networks for detecting acute Lyme disease from erythema migrans images of varying quality and acquisition conditions. This study used a cross-sectional dataset of images to train a model employing a deep convolutional neural network to perform classification of erythema migrans versus other skin conditions including tinea corporis and herpes zoster, and normal, non-pathogenic skin. Evaluation of the machine's ability to classify skin types was also performed on a validation set of images. Machine performance for detecting erythema migrans was further tested against a panel of non-medical humans. Online, publicly available images of both erythema migrans and non-Lyme confounding skin lesions were mined, and combined with erythema migrans images from an ongoing, longitudinal study of participants with acute Lyme disease enrolled in 2016 and 2017 who were recruited from primary and urgent care centers. The final dataset had 1834 images, including 1718 expert clinician-curated online images from unknown individuals with erythema migrans, tinea corporis, herpes zoster, and normal skin. It also included 116 images taken of 63 research participants from the Mid-Atlantic region. Two clinicians carefully annotated all lesion images. A convenience sample of 7 non-medically-trained humans were used as a panel to compare against machine performance. We calculated several performance metrics, including accuracy and Kappa (characterizing agreement with gold standard), as well as a receiver operating characteristic curve and associated area under the curve. For detecting erythema migrans, the machine had an accuracy (95{\%} confidence interval error margin) of 86.53{\%} (2.70), ROCAUC of 0.9510 (0.0171) and Kappa of 0.7143. Our results suggested substantial agreement between machine and clinician criterion standard. Comparison of machine with non-medical expert human performance indicated that the machine almost always exceeded acceptable specificity, and could operate with higher sensitivity. This could have benefits for prescreening prior to physician referral, earlier treatment, and reductions in morbidity.",
keywords = "Automated pre-screening, Convolutional neural networks, Deep learning, Erythema migrans, Lyme disease",
author = "Philippe Burlina and Joshi, {Neil J.} and Elise Ng and Billings, {Seth D.} and Rebman, {Alison W.} and John Aucott",
year = "2019",
month = "2",
day = "1",
doi = "10.1016/j.compbiomed.2018.12.007",
language = "English (US)",
volume = "105",
pages = "151--156",
journal = "Computers in Biology and Medicine",
issn = "0010-4825",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Automated detection of erythema migrans and other confounding skin lesions via deep learning

AU - Burlina, Philippe

AU - Joshi, Neil J.

AU - Ng, Elise

AU - Billings, Seth D.

AU - Rebman, Alison W.

AU - Aucott, John

PY - 2019/2/1

Y1 - 2019/2/1

N2 - Lyme disease can lead to neurological, cardiac, and rheumatologic complications when untreated. Timely recognition of the erythema migrans rash of acute Lyme disease by patients and clinicians is crucial to early diagnosis and treatment. Our objective in this study was to develop deep learning approaches using deep convolutional neural networks for detecting acute Lyme disease from erythema migrans images of varying quality and acquisition conditions. This study used a cross-sectional dataset of images to train a model employing a deep convolutional neural network to perform classification of erythema migrans versus other skin conditions including tinea corporis and herpes zoster, and normal, non-pathogenic skin. Evaluation of the machine's ability to classify skin types was also performed on a validation set of images. Machine performance for detecting erythema migrans was further tested against a panel of non-medical humans. Online, publicly available images of both erythema migrans and non-Lyme confounding skin lesions were mined, and combined with erythema migrans images from an ongoing, longitudinal study of participants with acute Lyme disease enrolled in 2016 and 2017 who were recruited from primary and urgent care centers. The final dataset had 1834 images, including 1718 expert clinician-curated online images from unknown individuals with erythema migrans, tinea corporis, herpes zoster, and normal skin. It also included 116 images taken of 63 research participants from the Mid-Atlantic region. Two clinicians carefully annotated all lesion images. A convenience sample of 7 non-medically-trained humans were used as a panel to compare against machine performance. We calculated several performance metrics, including accuracy and Kappa (characterizing agreement with gold standard), as well as a receiver operating characteristic curve and associated area under the curve. For detecting erythema migrans, the machine had an accuracy (95% confidence interval error margin) of 86.53% (2.70), ROCAUC of 0.9510 (0.0171) and Kappa of 0.7143. Our results suggested substantial agreement between machine and clinician criterion standard. Comparison of machine with non-medical expert human performance indicated that the machine almost always exceeded acceptable specificity, and could operate with higher sensitivity. This could have benefits for prescreening prior to physician referral, earlier treatment, and reductions in morbidity.

AB - Lyme disease can lead to neurological, cardiac, and rheumatologic complications when untreated. Timely recognition of the erythema migrans rash of acute Lyme disease by patients and clinicians is crucial to early diagnosis and treatment. Our objective in this study was to develop deep learning approaches using deep convolutional neural networks for detecting acute Lyme disease from erythema migrans images of varying quality and acquisition conditions. This study used a cross-sectional dataset of images to train a model employing a deep convolutional neural network to perform classification of erythema migrans versus other skin conditions including tinea corporis and herpes zoster, and normal, non-pathogenic skin. Evaluation of the machine's ability to classify skin types was also performed on a validation set of images. Machine performance for detecting erythema migrans was further tested against a panel of non-medical humans. Online, publicly available images of both erythema migrans and non-Lyme confounding skin lesions were mined, and combined with erythema migrans images from an ongoing, longitudinal study of participants with acute Lyme disease enrolled in 2016 and 2017 who were recruited from primary and urgent care centers. The final dataset had 1834 images, including 1718 expert clinician-curated online images from unknown individuals with erythema migrans, tinea corporis, herpes zoster, and normal skin. It also included 116 images taken of 63 research participants from the Mid-Atlantic region. Two clinicians carefully annotated all lesion images. A convenience sample of 7 non-medically-trained humans were used as a panel to compare against machine performance. We calculated several performance metrics, including accuracy and Kappa (characterizing agreement with gold standard), as well as a receiver operating characteristic curve and associated area under the curve. For detecting erythema migrans, the machine had an accuracy (95% confidence interval error margin) of 86.53% (2.70), ROCAUC of 0.9510 (0.0171) and Kappa of 0.7143. Our results suggested substantial agreement between machine and clinician criterion standard. Comparison of machine with non-medical expert human performance indicated that the machine almost always exceeded acceptable specificity, and could operate with higher sensitivity. This could have benefits for prescreening prior to physician referral, earlier treatment, and reductions in morbidity.

KW - Automated pre-screening

KW - Convolutional neural networks

KW - Deep learning

KW - Erythema migrans

KW - Lyme disease

UR - http://www.scopus.com/inward/record.url?scp=85059841114&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85059841114&partnerID=8YFLogxK

U2 - 10.1016/j.compbiomed.2018.12.007

DO - 10.1016/j.compbiomed.2018.12.007

M3 - Article

C2 - 30654165

AN - SCOPUS:85059841114

VL - 105

SP - 151

EP - 156

JO - Computers in Biology and Medicine

JF - Computers in Biology and Medicine

SN - 0010-4825

ER -