Key result
A deep neural network-based algorithm for automated pulmonary embolism detection on CT angiograms achieved 84.6% sensitivity, 95.1% specificity, and 93.8% overall accuracy.
Why the study?
The study was conducted to assess the performance of a deep neural network-based prototype algorithm for automated pulmonary embolism detection on CTPA scans.
Does a DNN-based algorithm accurately detect pulmonary embolism on CTPA scans in patients with suspected PE?
Population
1,000 patients (903 analyzed) undergoing CTPA for suspected PE
Comparison
DNN-based prototype algorithm vs clinical reports as ground truth
Design
Retrospective diagnostic accuracy study
Authors
Loading...
DNN shows high retrospective accuracy on CTPA; leaves open prospective validation before any clinical use.
Observational (n=903)
Does a DNN-based algorithm accurately detect pulmonary embolism on CTPA scans in patients with suspected PE?
A deep neural network algorithm demonstrated high accuracy and specificity for detecting pulmonary embolism on CTPA, suggesting potential utility for assisting radiologists and prioritizing exams.
Zsarnóczay et al. (2025) conducted an observational in Suspected pulmonary embolism (n=903). Deep Neural Network (DNN)-based prototype algorithm vs. Clinical reports (ground truth) was evaluated on Diagnostic performance (sensitivity, specificity, PPV, NPV, and accuracy). A deep neural network-based algorithm for automated pulmonary embolism detection on CT angiograms achieved 84.6% sensitivity, 95.1% specificity, and 93.8% overall accuracy.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: