Key result
An automated Convolutional Neural Network detected atrial epicardial fat from cardiac CT images with 89.22% accuracy, 90.18% sensitivity, and 88.52% specificity.
Why the study?
Atrial fat can lead to atrial fibrillation, motivating the development of an automated method using a convolutional neural network to detect epicardial fat in the atrium.
Does a Convolutional Neural Network (CNN) accurately detect atrial epicardial fat on cardiac CT images?
Does a Convolutional Neural Network (CNN) accurately detect atrial epicardial fat on cardiac CT images?
Answers are grounded in this paper and cite their sources.
A Convolutional Neural Network can automatically and accurately detect atrial epicardial fat from cardiac CT images, providing a robust tool for assessing this atrial fibrillation risk factor.
Atrial Fibrillation (A-fib) is a common cardiac rhythm problem in the population these days in which irregular heartbeat leads to blood clots, heart failure, stroke, and other significant clinical complications. Researchers have found that the atrial fat can lead to AF in most patients. To develop an automated method for detecting the epicardial fat present in the atrium using a Convolutional Neural Network. Cardiac Computed Tomography (CT) images of ten patients were pre-processed to remove the unwanted structure around the heart. An automated pixel value masking was done to locate the epicardial fat in the atrium and a 3D view of the heart was constructed for correct visualization of the location of the fat. A fast and fully automated Convolutional Neural Network (CNN) was applied to detect the atrial epicardial fat through feature selection from the CT images. We achieved 89.22% accuracy, 90.18% sensitivity, and 88.52% specificity in the detection of atrial epicardial fat using our CNN architecture. Our results showed that this CNN-based method can be helpful in atrial epicardial fat detection. Since Deep learning techniques add robustness, rapidness, and reliability, this study provides an unutilized way to detect the atrial fat tissue.
No takes yet. Share an insight, caveat, or question.
Deepa et al. (2021) studied Atrial Fibrillation (n=10). Convolutional Neural Network (CNN) was evaluated on Detection of atrial epicardial fat. An automated Convolutional Neural Network detected atrial epicardial fat from cardiac CT images with 89.22% accuracy, 90.18% sensitivity, and 88.52% specificity.
Synapse has enriched 10 related papers. Consider them for comparative context: