Joerg Blobel PhD, Presenter: Employee, Toshiba Corporation

Jurgen Mews DIPLENG, Abstract Co-Author: Employee, Toshiba Corporation

Joanne Schuijf, Abstract Co-Author: Employee, Toshiba Corporation

Willem Overlaet, Abstract Co-Author: Employee, Toshiba Corporation

We performed a phantom study to calibrate the Volume calcium score using Adaptive Iterative Dose Reduction (AIDR 3D) at 80, 100 and 120kVp instead of Filtered Back-Projection (FBP) at 120kVp (reference protocol).

ECG-gated volume scans of an anthropomorphic thoracic phantom containing 200, 400 and 800mg HA/cm3  calcium spheres of 1, 3 and 5mm diameter were performed on two 320-row CT`s (Aquilion ONETM (CT1)/Aquilion ONETM ViSION ED. (CT2)). Using 32 dose steps with 120kVp (CT1) for FBP, with 120/100kVp (CT1) and 80kVp (CT2) for AIDR 3D reconstructions, the lowest volume CTDI values that provide accurate Volume scores were determined for these 4 groups. The score averages and standard deviation (σ) were calculated over all included dose steps. The CT number thresholds for calcium identification were calibrated as follows. First, the scores were determined using thresholds between 120-160HU for 100kVp and 130-175HU for 80kVp. Secondly, a linear regression of the score averages was performed to determine at which CT number the reference score is crossed. Using these calibrated CT number thresholds, the score averages of the 4 remaining groups were statistically tested for 3σ-outliers, for normal distribution by Kolmogorov-Smirnov test and for statistical difference between the 4 groups by Kruskal-Wallis test. The 4 groups were finally averaged and with the 2σ-standard deviation graphed by Bland-Altman plot.

The CT number thresholds were calibrated to be 130HU (120kVp), 132HU (100kVp) and 140HU (80kVp) for the AIDR 3D groups and ensure an equivalent score average with 585.8±9.9mm3 (p=0.417) of the 4 groups compared to 585.5±9.7mm3 of the reference protocol. By using AIDR 3D instead of FBP reconstruction, radiation exposure can be potentially reduced with 69% (120kVp), 73% (100kVp) and 83% (80kVp). The related pixel noise threshold were determined for FBP with 26HU (120kVp) and for AIDR 3D with 23HU (120kVp), 23HU (100kVp) and 31HU (80kVp).

Volume scoring can be performed at reduced kVp levels in combination with AIDR 3D using kVp specific calibrated CT number thresholds.

By applying iterative reconstruction and lower kVp, considerable radiation dose reduction can be achieved for Volume calcium score with good accuracy compared to reference protocol. This approach should be clinically validated.

Blobel, J, Mews, J, Schuijf, J, Overlaet, W, Calibration of Coronary Volume Calcium Score Using Iterative Image Reconstruction (AIDR 3D) at 120, 100 and 80kVp instead of the Reference Protocol (FBP at 120kVp)..  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14010960.html