Publications by authors named "Luc Nicodeme"
Automatic coronary artery calcium scoring from unenhanced-ECG-gated CT using deep learning.
Diagn Interv Imaging
November 2021
Article Synopsis
- The study aimed to create an algorithm using convolutional neural networks (CNN) that can automatically estimate coronary artery calcium (CAC) from unenhanced ECG-gated CT scans.
- Researchers trained a CNN with a 3D U-Net architecture on 783 CT scans to detect and segment calcifications, calculating the Agatston score and comparing it to radiologist assessments.
- The final model achieved a high accuracy (C-index of 0.951), although it struggled with small or low-density calcifications near the mitral valve, potentially enhancing workflow by automating the CAC scoring process.