Why the study?
Coronary artery stenosis is a pivotal indicator in CAD diagnosis, and the study investigated the feasibility of an anchor-free deep learning method (FCOS) for its automatic detection.
Does the anchor-free deep learning method FCOS improve the automatic detection of coronary stenosis in ICA images compared to anchor-based DL models?
Population
2786 invasive coronary angiography images from 130 patients
Comparison
FCOS vs SSD vs Faster R-CNN vs YOLOv3
Design
Model evaluation with 10-fold cross-validation
Authors
Loading...
Anchor-free DL enables feasible automated stenosis detection on angiography; leaves open clinical utility pending prospective validation.
Does the anchor-free deep learning method FCOS improve the automatic detection of coronary stenosis in ICA images compared to anchor-based DL models?
The anchor-free deep learning model FCOS is highly feasible and outperforms standard anchor-based models in the automatic detection of coronary stenosis from invasive coronary angiography images.
Yue et al. (2024) studied this question.
Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context: