Effect of different reconstruction algorithms on coronary artery calcium scores using the reduced radiation dose protocol: a clinical and phantom study

Background: This study aimed to evaluate the effects of different iterative reconstruction (IR) algorithms on coronary artery calcium (CAC) score quantification using the reduced radiation dose (RRD) protocol in an anthropomorphic phantom and in patients. Methods: A thorax phantom, containing 9 calcification inserts with varying hydroxyapatite (HA) densities, was scanned with the reference protocol [120 kv, 80 mAs, filtered back projection (FBP)] and RRD protocol (120 kV, 20–80 mAs, 5 mAs interval) using a 256-slice computed tomography (CT) scanner. Raw data were reconstructed with different reconstruction algorithms [iDose4 levels 1–7 and iterative model reconstruction (IMR) levels 1–3]. Signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and Agatston score (AS) were calculated for each image series. The correction factor was derived from linear regression analysis between the reference image series and other image series with different parameters. Additionally, 40 patients were scanned with the RRD protocol (50 mAs) and reconstructed with FBP, iDose4 level 4, and IMR level 2. AS was calculated for the 3-group image series, and was corrected by applying a correction factor for the IMR group. The agreement of risk stratification with different reconstruction algorithms was also analyzed. Results: For the phantom study, the iDose4 and IMR groups had significantly higher SNR and CNR than the FBP group (all P<0.05). There were no significant differences in the total AS after comparing image series reconstructed with iDose4 (level 1–7) and FBP (all P>0.05), while AS from the IMR (level 1–3) image series were lower than the FBP group (all P<0.05). The tube current of 50 mAs was determined for the clinical study, and the correction factor was 1.14. For the clinical study, the median AS from the iDose4 and IMR groups were both significantly lower compared to the FBP image series [(112.89 (63.01, 314.09), 113.22 (64.78, 364.95) vs. 118.59 (65.05, 374.48), both P<0.05]. After applying the correction factor, the adjusted AS from the IMR group was not significantly different from that of the FBP group [126.48 (69.62, 355.85) vs. 118.59 (65.05, 374.48), P=0.145]. Moreover, the agreement in risk stratification between FBP and IMR improved from 0.81 to 0.85. Conclusions: The RRD CAC scoring scan using the IMR reconstruction algorithm is clinically feasible, and a correction factor can help reduce the AS underestimation effect.