Abstract
SummaryBackgroundAcute 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.MethodsCNNs were pretrained on 595 506 radiographs from two centres to identify common chest findings (eg, opacity and effusion), and then trained on 8072 radiographs annotated for ARDS by multiple physicians using various transfer learning approaches. The best performing CNN was tested on chest radiographs in an internal and external cohort, including a subset reviewed by six physicians, including a chest radiologist and physicians trained in intensive care medicine. Chest radiograph data were acquired from four US hospitals.FindingsIn an internal test set of 1560 chest radiographs from 455 patients with acute hypoxaemic respiratory failure, a CNN could detect ARDS with an area under the receiver operator characteristics curve (AUROC) of 0·92 (95% CI 0·89–0·94). In the subgroup of 413 images reviewed by at least six physicians, its AUROC was 0·93 (95% CI 0·88–0·96), sensitivity 83·0% (95% CI 74·0–91·1), and specificity 88·3% (95% CI 83·1–92·8). Among images with zero of six ARDS annotations (n=155), the median CNN probability was 11%, with six (4%) assigned a probability above 50%. Among images with six of six ARDS annotations (n=27), the median CNN probability was 91%, with two (7%) assigned a probability below 50%. In an external cohort of 958 chest radiographs from 431 patients with sepsis, the AUROC was 0·88 (95% CI 0·85–0·91). When radiographs annotated as equivocal were excluded, the AUROC was 0·93 (0·92–0·95).InterpretationA CNN can be trained to achieve expert physician-level performance in ARDS detection on chest radiographs. Further research is needed to evaluate the use of these algorithms to support real-time identification of ARDS patients to ensure fidelity with evidence-based care or to support ongoing ARDS research.FundingNational Institutes of Health, Department of Defense, and Department of Veterans Affairs.
Highlights
Acute respiratory distress syndrome (ARDS) is a common critical illness syndrome characterised by the acute onset of severe hypoxaemia and lung oedema of non-cardiac cause in patients with conditions such as sepsis, pneumonia, or trauma
In an internal test set of 1560 chest radiographs from 455 patients with acute hypoxaemic respiratory failure, a convolutional neural network (CNN) could detect ARDS with an area under the receiver operator characteristics curve (AUROC) of 0·92
The University of Michigan training set included 8072 chest radiographs from 1778 patients, with 2665 (33%) consistent with ARDS based on physician review
Summary
Acute respiratory distress syndrome (ARDS) is a common critical illness syndrome characterised by the acute onset of severe hypoxaemia and lung oedema of non-cardiac cause in patients with conditions such as sepsis, pneumonia, or trauma. Despite research investm ent, current treatment for ARDS remains largely supportive and mortality remains at 35%.1. Patients who develop ARDS often go unrecognised and do not receive evidence-based care.[2,3] Bilateral airspace disease on chest radiograph is a key criterion in the definition of ARDS, and a major driver of the definition’s lower reliability.[4] Intensivists in clinical practice have poor agreement when identifying ARDS findings on chest radiographs (κ 0·13),[5] and ARDS clinical research study coordinators do little better (κ 0·27).[6] New approaches for identifying ARDS findings on chest radiographs are needed to support ARDS care
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
