Johannes M. Voigt, Presenter: Nothing to Disclose

Christian Guldner, Abstract Co-Author: Nothing to Disclose

Stefan Schaefer, Abstract Co-Author: Nothing to Disclose

Martin Fiebich, Abstract Co-Author: Nothing to Disclose

This study shows that consequent optimization following the ALARA principle is able to lower patient dose dramatically (vendor standard settings CTDIvol = 9.0 mGy vs. optimized settings CTDIvol = 0.93 mGy) but are also time consuming and deliver individual results. But we could also show that there is saturation in reachable image quality (CTDIvol ~ 3.5 mGy). Facing this fact a dose reduction of about factor 3 compared to the standard settings is possible without causing difficulties in diagnostics.

Dose optimization in CT following ALARA-principle is an iterative and time-consuming procedure. This study points out the patient dose reduction caused by optimizing image acquisition and reconstruction parameters.

Paranasal sinus CT scans (Siemens Definition Dual Source) of three cadaveric heads had been performed under variation of kVp (80kVp, 100kVp and 120kVp), mAs (starting from 7mAs up to a value where a CTDIvol less or equal 9.0mGy, which is the vendors standard setting, was reached) and reconstruction kernel (H47, H50, H60 and H70). All series were anonymized, randomized and evaluated by three different observers. Diagnostic Quality of selected anatomic structures was graded from 1: excellent; 2: adequate; 3: difficult; 4: not visible). For each reconstruction kernel the mean value of all scores was displayed as function of indicated CTDIvol. The minimum CTDIvol at which diagnostic image quality is reached was defined to be grated less or equal 2.0.

The subjective image quality differs significantly between the convolution kernels. For an acceptable image quality the lowest CTDIvol = 0.93 mGy was delivered by H47 kernel, 100 kVp and 10 mAs. The next lowest CTDIvol = 1.52 mGy was achieved with the H50 kernel, 120 kVp and 10 mAs. H60 and H70 kernels could be excluded because of too bad image quality (minimum CTDIvol = 4.0 mGy). The application of the H47 and H50 kernels showed saturation in image quality (mean score ~ 1.4). Hence, the best image quality (1.0) could not be reached. The saturation for the H47 and H50 kernels lies at a CTDIvol of about 3.5 mGy for all skulls and all observers.

Voigt, J, Guldner, C, Schaefer, S, Fiebich, M, Dose Optimization in CT of the Paranasal Sinus.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14019193.html