ScienceOpen:
research and publishing network
For Publishers
Discovery
Metadata
Peer review
Hosting
Publishing
For Researchers
Join
Publish
Review
Collect
My ScienceOpen
Sign in
Register
Dashboard
Blog
About
Search
Advanced search
My ScienceOpen
Sign in
Register
Dashboard
Search
Search
Advanced search
For Publishers
Discovery
Metadata
Peer review
Hosting
Publishing
For Researchers
Join
Publish
Review
Collect
Blog
About
4
views
0
references
Top references
cited by
62
Cite as...
0 reviews
Review
0
comments
Comment
0
recommends
+1
Recommend
0
collections
Add to
0
shares
Share
Twitter
Sina Weibo
Facebook
Email
2,071
similar
All similar
Record
: found
Abstract
: not found
Article
: not found
Assessment of Convolutional Neural Networks for Automated Classification of Chest Radiographs
Author(s):
Jared A. Dunnmon
1
,
Darvin Yi
1
,
Curtis P. Langlotz
1
,
Christopher Ré
1
,
Daniel L. Rubin
1
,
Matthew P. Lungren
1
Publication date
Created:
February 2019
Publication date
(Print):
February 2019
Journal:
Radiology
Publisher:
Radiological Society of North America (RSNA)
Read this article at
ScienceOpen
Publisher
Further versions
oa repository (via OAI-PMH doi match)
oa repository (via pmcid lookup)
Powered by
Review
Review article
Invite someone to review
Cite as...
There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.
Abstract
Purpose To assess the ability of convolutional neural networks (CNNs) to enable high-performance automated binary classification of chest radiographs. Materials and Methods In a retrospective study, 216 431 frontal chest radiographs obtained between 1998 and 2012 were procured, along with associated text reports and a prospective label from the attending radiologist. This data set was used to train CNNs to classify chest radiographs as normal or abnormal before evaluation on a held-out set of 533 images hand-labeled by expert radiologists. The effects of development set size, training set size, initialization strategy, and network architecture on end performance were assessed by using standard binary classification metrics; detailed error analysis, including visualization of CNN activations, was also performed. Results Average area under the receiver operating characteristic curve (AUC) was 0.96 for a CNN trained with 200 000 images. This AUC value was greater than that observed when the same model was trained with 2000 images (AUC = 0.84, P < .005) but was not significantly different from that observed when the model was trained with 20 000 images (AUC = 0.95, P > .05). Averaging the CNN output score with the binary prospective label yielded the best-performing classifier, with an AUC of 0.98 (P < .005). Analysis of specific radiographs revealed that the model was heavily influenced by clinically relevant spatial regions but did not reliably generalize beyond thoracic disease. Conclusion CNNs trained with a modestly sized collection of prospectively labeled chest radiographs achieved high diagnostic performance in the classification of chest radiographs as normal or abnormal; this function may be useful for automated prioritization of abnormal chest radiographs. © RSNA, 2018 Online supplemental material is available for this article. See also the editorial by van Ginneken in this issue.
Related collections
Radiology Science
Author and article information
Journal
Title:
Radiology
Abbreviated Title:
Radiology
Publisher:
Radiological Society of North America (RSNA)
ISSN (Print):
0033-8419
ISSN (Electronic):
1527-1315
Publication date Created:
February 2019
Publication date (Print):
February 2019
Volume
: 290
Issue
: 2
Pages
: 537-544
Affiliations
[
1
]
From the Departments of Computer Science (J.A.D., C.R.), Biomedical Data Science (D.Y., D.L.R.), and Radiology (C.P.L., D.L.R., M.P.L.), Stanford University, 300 Pasteur Dr, Stanford, CA 94305.
Article
DOI:
10.1148/radiol.2018181422
SO-VID:
6711a0f9-bae5-474a-a4a3-69534214f3a5
Copyright ©
© 2019
History
Data availability:
Comments
Comment on this article
Sign in to comment
scite_
Similar content
2,071
[Sudden death caused by a less lethal weapon chest-wall injury (Commotio cordis)].
Authors:
C Contargyris
,
E Peytel
ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases
Authors:
Xiaosong Wang
,
Yifan Peng
,
Le Lu
…
Pediatric cervical spine in emergency: radiographic features of normal anatomy, variants and pitfalls.
Authors:
Omar Adib
,
Emeline Berthier
,
Didier Loisel
…
See all similar
Cited by
62
Preparing Medical Imaging Data for Machine Learning
Authors:
Martin Willemink
,
Wojciech A. Koszek
,
Cailin Hardell
…
Diagnostic accuracy of deep learning in medical imaging: a systematic review and meta-analysis
Authors:
Ravi Aggarwal
,
Viknesh Sounderajah
,
Guy Martin
…
Assessing Radiology Research on Artificial Intelligence: A Brief Guide for Authors, Reviewers, and Readers—From the Radiology Editorial Board
Authors:
David A. Bluemke
,
Linda Moy
,
Miriam A Bredella
…
See all cited by
Version 1
- Current
Version 1