Deep learning to detect acute respiratory distress syndrome on chest radiographs: a retrospective study with external validation.
Journal:
The Lancet. Digital health
PMID:
33893070
Abstract
BACKGROUND: Acute respiratory distress syndrome (ARDS) is a common, but under-recognised, critical illness syndrome associated with high mortality. An important factor in its under-recognition is the variability in chest radiograph interpretation for ARDS. We sought to train a deep convolutional neural network (CNN) to detect ARDS findings on chest radiographs.
Authors
Keywords
Aged
Algorithms
Area Under Curve
Datasets as Topic
Deep Learning
Female
Hospitals
Humans
Lung
Male
Middle Aged
Neural Networks, Computer
Pleural Cavity
Pleural Diseases
Radiographic Image Interpretation, Computer-Assisted
Radiography
Radiography, Thoracic
Respiratory Distress Syndrome
Retrospective Studies
United States