Key result
Deep learning automated detection of pulmonary embolism (sensitivity 79.6%, specificity 95.0%) combined with an electronic notification system did not significantly reduce patient turnaround times.
Why the study?
It was unknown whether deep learning automated detection of PE on CTPA combined with worklist prioritization and an electronic notification system improves communication times and ED patient turnaround.
Does a deep learning-based automated detection algorithm for pulmonary embolism on CTPA improve communication and patient turnaround times in the Emergency Department?
Observational (n=1,808)
Does a deep learning-based automated detection algorithm for pulmonary embolism on CTPA improve communication and patient turnaround times in the Emergency Department?
The implementation of a deep learning algorithm for PE detection on CTPA and an electronic notification system did not significantly improve clinical workflow times in the emergency department.
No takes yet. Share an insight, caveat, or question.
Caution against expecting faster PE turnaround from DL detection plus notifications; challenges assumptions of AI-driven workflow gains in radiology.
Schmuelling et al. (2021) conducted an observational in Pulmonary embolism (n=1,808). Deep learning automated detection of PE and electronic notification system vs. Standard workflow without deep learning or electronic notification system was evaluated on Radiology report reading times, communication time, time to anticoagulation, and patient turnaround times. Deep learning automated detection of pulmonary embolism (sensitivity 79.6%, specificity 95.0%) combined with an electronic notification system did not significantly reduce patient turnaround times.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: