Related Experiment Video
Updated: Sep 20, 2025
Author Spotlight: Advancing Cardiovascular Imaging - Introducing the Spatially Weighted Calcium Score for Early Disease Detection
Published on: September 22, 2023
Generative models for reproducible coronary calcium scoring
Sanne G M van Velzen1,2,3,4, Bob D de Vos1,2,3, Julia M H Noothout1,2,3
1Amsterdam UMC location University of Amsterdam, Department of Biomedical Engineering and Physics, Amsterdam, The Netherlands.
Insights
A new generative adversarial network (GAN) method improves coronary artery calcium (CAC) quantification by eliminating the need for intensity thresholds. This enhances reproducibility for more reliable coronary heart disease risk assessment.
Area of Science:
- Radiology
- Artificial Intelligence
- Cardiovascular Imaging
Background:
- Coronary artery calcium (CAC) scoring is a key predictor of coronary heart disease (CHD) events.
- Current CAC scoring methods have limited interscan reproducibility due to fixed intensity thresholds, especially in non-ECG-synchronized CT.
- Cardiac motion and partial volume effects exacerbate reproducibility issues.
Purpose of the Study:
- To develop and evaluate a novel CAC quantification method that does not require a segmentation threshold.
- To improve the interscan reproducibility of CAC scoring.
- To enhance the reliability of CHD risk categorization and prediction accuracy.
Main Methods:
- Utilized a cycle-consistent generative adversarial network (GAN) to decompose CT images into CAC and non-CAC components.
- Trained the GAN model using a dataset of 626 low-dose chest CTs and 514 radiotherapy treatment planning (RTP) CTs.
- Compared interscan reproducibility of the proposed method against clinical calcium scoring in 1662 patients with two scans each.
Main Results:
- The proposed GAN method achieved a significantly lower relative interscan difference in CAC mass (47%) compared to manual clinical scoring (89%).
- The intraclass correlation coefficient for Agatston scores was 0.96 with the proposed method, versus 0.91 for automatic clinical scoring.
- Demonstrated superior interscan reproducibility for CAC quantification.
Conclusions:
- The developed GAN-based CAC quantification method offers improved interscan reproducibility.
- Enhanced reproducibility can lead to more reliable CHD risk stratification.
- This method holds potential for improving the accuracy of CHD event prediction.
Abstract:
Purpose: Coronary artery calcium (CAC) score, i.e., the amount of CAC quantified in CT, is a strong and independent predictor of coronary heart disease (CHD) events. However, CAC scoring suffers from limited interscan reproducibility, which is mainly due to the clinical definition requiring application of a fixed intensity level threshold for segmentation of calcifications. This limitation is especially pronounced in non-electrocardiogram-synchronized computed tomography (CT) where lesions are more impacted by cardiac motion and partial volume effects. Therefore, we propose a CAC quantification method that does not require a threshold for segmentation of CAC. Approach: Our method utilizes a generative adversarial network (GAN) where a CT with CAC is decomposed into an image without CAC and an image showing only CAC. The method, using a cycle-consistent GAN, was trained using 626 low-dose chest CTs and 514 radiotherapy treatment planning (RTP) CTs. Interscan reproducibility was compared to clinical calcium scoring in RTP CTs of 1662 patients, each having two scans. Results: A lower relative interscan difference in CAC mass was achieved by the proposed method: 47% compared to 89% manual clinical calcium scoring. The intraclass correlation coefficient of Agatston scores was 0.96 for the proposed method compared to 0.91 for automatic clinical calcium scoring. Conclusions: The increased interscan reproducibility achieved by our method may lead to increased reliability of CHD risk categorization and improved accuracy of CHD event prediction.
Related Concept Videos
01:25Imaging Studies for Cardiovascular System VI: Calcium -Scoring CT
01:28Imaging Studies for Cardiovascular System V: CT