Automated deep learning–radiomics pipeline for non-calcified coronary plaque detection using non-contrast calcium score CT
-
By
-
Wen Chen
-
Qing Tao
-
Can Chen
-
Su Hu
-
Feirong Yao
-
Jie Chen
-
Jinggang Zhang
-
Kejun Gu
-
Li Su
-
Wei Xing
-
Chunhong Hu
-
June 1, 2026
-
Clinical Scorecard: Automated Deep Learning and Radiomics Approach for Identifying Non-Calcified Coronary Plaques via Non-Contrast Calcium Score CT
At a Glance
| Category | Detail |
|---|
| Condition | Coronary Artery Disease (CAD) |
| Key Mechanisms | Detection of non-calcified plaques using deep learning and radiomics on non-contrast CACS images. |
| Target Population | Patients undergoing coronary CT angiography for suspected CAD. |
| Care Setting | Clinical practice, specifically in cardiology and radiology departments. |
Key Highlights
- CACS quantifies calcification but lacks sensitivity for non-calcified plaques.
- Automated pipeline combines deep learning and radiomics for plaque detection.
- SegResNet models achieved optimal performance in coronary segmentation.
- Radiomics models showed moderate to good diagnostic performance with AUCs from 0.700 to 0.855.
- Combined-region models outperformed or matched PCAT models in most settings.
Guideline-Based Recommendations
Diagnosis
- Use non-contrast CACS for evaluating coronary artery calcification.
- Consider automated deep learning and radiomics for detecting non-calcified plaques.
Management
- Implement automated detection methods to enhance CAD screening efficiency.
Monitoring & Follow-up
- Regular assessment of coronary artery and PCAT metrics for CAD progression.
Risks
- Non-calcified plaques are associated with increased risks of adverse cardiovascular events.
Patient & Prescribing Data
Individuals at risk for CAD, particularly those with suspected coronary artery disease.
Non-contrast CT is a safer alternative for CAD screening compared to contrast-enhanced methods.
Clinical Best Practices
- Utilize automated segmentation techniques to reduce manual workload in image analysis.
- Incorporate radiomics features to improve diagnostic accuracy for CAD.
Related Resources & Content
