Related Experiment Video
Updated: Aug 15, 2025
Improved Registration of 3D CT Angiography with X-ray Fluoroscopy for Image Fusion During Transcatheter Aortic Valve Implantation
Published on: June 3, 2018
A token-mixer architecture for CAD-RADS classification of coronary stenosis on multiplanar reconstruction CT images
Marco Penso1, Sara Moccia2, Enrico G Caiani3
1Cardiovascular Imaging Department, Centro Cardiologico Monzino IRCCS, Milan, Italy; Department of Electronics, Information and Biomedical Engineering, Politecnico di Milano, Milan, Italy.
Insights
A deep learning algorithm using ConvMixer accurately classifies coronary artery stenosis from CCTA images. This tool aids in precise stenosis evaluation, potentially reducing variability in clinical decision-making for coronary artery disease.
Area of Science:
- Cardiovascular Imaging
- Artificial Intelligence in Medicine
- Medical Image Analysis
Background:
- Accurate assessment of coronary artery stenosis is crucial for managing suspected Coronary Artery Disease (CAD).
- Coronary Computed Tomography Angiography (CCTA) is a key imaging modality for evaluating coronary stenosis.
- Standardized reporting systems like CAD-RADS improve consistency in stenosis classification.
Purpose of the Study:
- To develop and evaluate an automatic deep learning algorithm for classifying coronary stenosis severity.
- To apply the algorithm to multiplanar reconstruction images from CCTA scans.
- To classify lesions according to the Coronary Artery Disease Reporting and Data System (CAD-RADS).
Main Methods:
- A retrospective study included 288 patients with suspected CAD who underwent CCTA.
- A token-mixer architecture (ConvMixer) was employed to model long-range semantic information in coronary artery images.
- ConvMixer utilizes a patch embedding layer and convolutional blocks to learn pixel dependencies; performance was assessed using Grad-CAM.
Main Results:
- The ConvMixer algorithm achieved 87% accuracy and 90% sensitivity in classifying significant coronary artery stenosis (≥50% luminal narrowing).
- For CAD-RADS 0 vs. 1-2 vs. 3-4 vs. 5 classification, ConvMixer demonstrated 72% accuracy and 75% sensitivity.
- ConvMixer showed a superior performance-complexity trade-off compared to pyramid-shaped convolutional neural networks.
Conclusions:
- The developed deep learning algorithm shows promise for automated coronary artery stenosis evaluation.
- This tool may serve as decision support for clinicians, potentially reducing interobserver variability.
- The findings suggest a role for AI in improving the consistency and precision of CAD assessment.
Background And Objective:
In patients with suspected Coronary Artery Disease (CAD), the severity of stenosis needs to be assessed for precise clinical management. An automatic deep learning-based algorithm to classify coronary stenosis lesions according to the Coronary Artery Disease Reporting and Data System (CAD-RADS) in multiplanar reconstruction images acquired with Coronary Computed Tomography Angiography (CCTA) is proposed.
Methods:
In this retrospective study, 288 patients with suspected CAD who underwent CCTA scans were included. To model long-range semantic information, which is needed to identify and classify stenosis with challenging appearance, we adopted a token-mixer architecture (ConvMixer), which can learn structural relationship over the whole coronary artery. ConvMixer consists of a patch embedding layer followed by repeated convolutional blocks to enable the algorithm to learn long-range dependences between pixels. To visually assess ConvMixer performance, Gradient-Weighted Class Activation Mapping (Grad-CAM) analysis was used.
Results:
Experimental results using 5-fold cross-validation showed that our ConvMixer can classify significant coronary artery stenosis (i.e., stenosis with luminal narrowing ≥50%) with accuracy and sensitivity of 87% and 90%, respectively. For CAD-RADS 0 vs. 1-2 vs. 3-4 vs. 5 classification, ConvMixer achieved accuracy and sensitivity of 72% and 75%, respectively. Additional experiments showed that ConvMixer achieved a better trade-off between performance and complexity compared to pyramid-shaped convolutional neural networks.
Conclusions:
Our algorithm might provide clinicians with decision support, potentially reducing the interobserver variability for coronary artery stenosis evaluation.
More Related Videos
Related Concept Videos
01:28Imaging Studies for Cardiovascular System V: CT
01:25Imaging Studies for Cardiovascular System VI: Calcium -Scoring CT