Susan Notohamiprodjo MD, Presenter: Nothing to Disclose

Zsuzsanna Deak MD, Abstract Co-Author: Nothing to Disclose

Fabian Mueck, Abstract Co-Author: Nothing to Disclose

Felix Meurer, Abstract Co-Author: Nothing to Disclose

Maximilian F. Reiser MD, Abstract Co-Author: Nothing to Disclose

Stefan Wirth MD, Abstract Co-Author: Speaker, General Electric Company

CCT is a frequently needed examination in emergency medicine. Compared to the current clinical standard of image reconstruction Adaptive Statistical Iterative Reconstruction (ASIR), MBIR is a more advanced algorithm promising improved spatial resolution and reduced image noise. The aim of the study was to compare the effects of MBIR in CCT imaging to ASIR on identical dose levels.

Raw data sets of anonymized 100 trauma patients receiving CCT according to the institutional standard protocol (120 kV, 260 mAs, 20 mm detector collimation; 0.984 pitch) were reconstructed with ASIR and MBIR, multiplanar reformations of 2.5 mm axial, coronar and sagittal slices were calculated. Two radiologists blinded to the reconstruction independently rated IQ by the depiction of different parenchymal structures and the effect streak artifacts of photon starvation using a semi-quantitative scale (0: non-diagnostic, 1: impaired, 2: sufficient, 3: good, 4: excellent). Mean attenuation value (MAV; [HU]) and standard deviation (SD;[HU]) were measured for liquor space (LS) and white matter (WM) supratentorial (ST) and in the posterior fossa (PF). Data were analyzed using ICC, Mann-Whitney-U and ANOVA testing.

MBIR significantly decreased streak artifacts in PF (p<0.01). IQ of MBIR images were rated significantly superior comparing to ASIR images (p<0.01), with a median of IQ scores of 3 (MBIR) and 2 (ASIR). ICC ranged from 0.8 to 0.9. SD of WM in PF in sagittal and coronar images of MBIR (3.6±0.8 and 3.5±0.7) was significantly lower compared to ASIR (4.1±0.9 and 4.7±0.9); (p< 0.05). MAVs of WM were ST comparable for both algorithms (MBIR: 23.5±1.9 and ASIR: 24.6±1.8; p=0.075) but differed significantly in PF (MBIR: 28.6±4.6 and ASIR: 34.2±3.6; p<0.05), whereas the difference was significantly higher with ASIR (10HU) than with MBIR (5HU).

Our results suggest significant improvement of IQ with MBIR in comparison to ASIR in CCT of trauma patients.

MBIR significantly improves IQ and could represent an effective method to decrease radiation dose of CCT imaging, which is one of the most important causes for increase of public radiation exposure.

Notohamiprodjo, S, Deak, Z, Mueck, F, Meurer, F, Reiser, M, Wirth, S, Improvement of Image Quality (IQ) with Model Based Iterative Reconstruction (MBIR) Algorithm in Cranial CT (CCT) in Trauma Patients.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13024453.html