The aim of this study was to determine the dose reduction potential of iterative reconstruction (IR) algorithms in computed tomography angiography (CTA) of the circle of Willis using a novel method of evaluating the quality of radiation dose–reduced images.
Materials and Methods
This study relied on ReconCT, a proprietary reconstruction software that allows simulating CT scans acquired with reduced radiation dose based on the raw data of true scans. To evaluate the performance of ReconCT in this regard, a phantom study was performed to compare the image noise of true and simulated scans within simulated vessels of a head phantom. That followed, 10 patients scheduled for CTA of the circle of Willis were scanned according to our institute's standard protocol (100 kV, 145 reference mAs). Subsequently, CTA images of these patients were reconstructed as either a full-dose weighted filtered back projection or with radiation dose reductions down to 10% of the full-dose level and Sinogram-Affirmed Iterative Reconstruction (SAFIRE) with either strength 3 or 5. Images were marked with arrows pointing on vessels of different sizes, and image pairs were presented to observers. Five readers assessed image quality with 2-alternative forced choice comparisons.
In the phantom study, no significant differences were observed between the noise levels of simulated and true scans in filtered back projection, SAFIRE 3, and SAFIRE 5 reconstructions.
The dose reduction potential for patient scans showed a strong dependence on IR strength as well as on the size of the vessel of interest. Thus, the potential radiation dose reductions ranged from 84.4% for the evaluation of great vessels reconstructed with SAFIRE 5 to 40.9% for the evaluation of small vessels reconstructed with SAFIRE 3.
This study provides a novel image quality evaluation method based on 2-alternative forced choice comparisons. In CTA of the circle of Willis, higher IR strengths and greater vessel sizes allowed higher degrees of radiation dose reduction.