Quantification of coronary artery calcium using virtual non-contrast images derived from dual-layer spectral CT

To evaluate the feasibility of deriving coronary artery calcium (CAC) scores from Virtual-Non-Contrast (VNC) reconstructions obtained with dual-layer computed tomography (DLCT).

A retrospective study was conducted on 100 patients who underwent coronary computed tomography angiography (CCTA) with a DLCT scanner. Conventional true non-contrast (TNC) images were reconstructed for CAC quantification (CAC_TNC). Post-contrast spectral datasets were processed to generate VNC reconstructions, and a 130 HU threshold was applied to calculate the volume of coronary calcified plaques (VOL_VNC). CAC values from VNC images (CAC_VNC) were derived through linear regression analysis. Correlation and agreement between CAC measurement methods were assessed using the intraclass correlation coefficient (ICC) and Bland-Altman analysis. Patients were assigned to risk categories according to Coronary Artery Calcium Data and Reporting System (CAC-DRS) to evaluate potential misclassifications.

The median CAC_TNC was 95 Agatston Unit (AU) [IQR:1-309], while the median CAC_VNC was 76 AU [IQR:0-340], with no statistically significant difference between the two methods (p = 0.653). An excellent correlation was found between CAC_TNC and CAC_VNC (ICC = 0.977), with good agreement demonstrated by Bland-Altman analysis, showing a mean difference of - 1.2 AU. CAC_VNC led to misclassification in 16% of patients, primarily underestimating the CAC-DRS risk category, yet maintaining strong agreement with CAC_TNC derived risk stratification.

VNC reconstructions from DLCT show excellent correlation and good agreement with conventional TNC images for CAC quantification. This approach has the potential to eliminate the need for a dedicated unenhanced scan, thereby reducing radiation exposure, and acquisition time while preserving diagnostic accuracy.