Iterative beam-hardening correction with advanced modeled iterative reconstruction in low voltage CT coronary calcium scoring with tin filtration: Impact on coronary artery calcium quantification and image quality

Objective: We investigated the impact of iterative beam-hardening correction (IBHC) with advanced modeled iterative reconstruction (ADMIRE) of ultra-low radiation-dose tin filtered (Sn100 kVp) CACS acquisitions on image quality, calcium quantification, and risk classification. Methods: CT data of 60 patients (55% male, age 62.3 +/- 9.8years) who underwent a 3rd generation dual-source CACS scan using a prospectively ECG-triggered 100 kVp sequential acquisition protocol with tin filtration (Sn100 kVp) were reconstructed using IBHC with filtered back projection (FBP) and ADMIRE with strength levels of three and five. Image noise was calculated and Agatston scores were derived from all reconstructions. Image noise, Agatston score categories, and percentile-based cardiac risk categorization of the respective reconstruction techniques were compared. Results: The mean estimated radiation dose equivalent of CACS acquisitions in the study population was 0.20 +/- 0.07 mSv. Mean image noise significantly decreased with ADMIRE compared to FBP (both p < 0.0001). Agatston scores derived from the respective reconstructions were significantly different in a paired comparison (median [25th and 75th percentile]): FBP 34.7 [1.9,153.6], ADMIRE 3 28.6 [1.1,134.5], ADMIRE 5 22.7 [0.3, 116.8]; both p < 0.0001). Agatston score categories and cardiac risk categorization showed excellent agreement of ADMIRE 3 and ADMIRE 5 with FBP (kappa = 0.92 [0.86-0.98] and kappa = 0.86 [0.79-0.94]; kappa = 0.94 [0.87-1.00] and kappa = 0.91 [0.83-0.99]; however, modest cardiac risk reclassifications of 3% and 7% for ADMIRE 3 and ADMIRE 5, respectively, were observed. Conclusion: Iterative reconstruction using IBHC ADMIRE in low voltage, ultra-low dose CACS with tin filtration significantly decreased image noise. However, it also reduced Agatston scores compared to FBP, which may have an impact on subsequent cardiac risk classification, although risk reclassification occurred only in a modest number of subjects. (C) 2017 Society of Cardiovascular Computed Tomography. Published by Elsevier Inc. All rights reserved.