In Computed Tomography (CT), image quality is often quantified by the measuring the amount of noise (SD) in the image in relation to the spatial resolution (lp/cm). The simplest  way to influence the amount of image noise is with the selection of a specific convolution filter in the reconstruction process. Selecting a “smooth” convolution filter will reduce image noise; however, it will also smooth out the details and therefore deteriorate  spatial resolution.

This paper briefly describes Toshiba’s dose reduction technologies and focuses on the latest iterative dose reduction technique, i.e. Adaptive Iterative Dose Reduction 3D (AIDR 3D) which not only reduces image noise but also improves spatial resolution.

Both phantom and clinical studies are presented and discussed. Quantitative analysis of Noise Power Spectra (NPS), Low Contrast Detectability (LCD), and spatial resolution improvements are presented to evaluate the performance of the AIDR 3D algorithm.

The reported results demonstrate that a dose reduction of 75% or more can be achieved while improving the spatial resolution. Results also confirm that the images look natural and clinical details are preserved despite an effective dose of 0.09 mSv for a thorax scan.

In terms of reconstruction speed, there is no significant difference between FBP and AIDR 3D processing. This is particularly  important  for trauma and emergency departments.

In conclusion, it has been demonstrated that the ability to substantially reduce noise in CT images using AIDR 3D  produces “natural looking” image quality with almost no penalty in reconstruction speed.

Download WhitePaper