Matthias Klaus Werner MD, Presenter: Nothing to Disclose

Cristina Lois PhD, Abstract Co-Author: Nothing to Disclose

Jürgen Kupferschläger, Abstract Co-Author: Nothing to Disclose

Carina Fredriksson, Abstract Co-Author: Employee, ContextVision AB

Henriette Heners, Abstract Co-Author: Nothing to Disclose

Christina Pfannenberg MD, Abstract Co-Author: Nothing to Disclose

Thomas Beyer, Abstract Co-Author: CEO, cmi-experts GmbH

In PET/CT imaging CT is used for diagnosis, attenuation correction (CTAC) and anatomical correlation of the PET emission data. In select cases minimum CT tube currents are used; thus, leading to noise propagation in combined data. We evaluate the effects of low-dose CT on noise in CT and PET following CTAC before and after post-reconstruction filtering of CT images.

Phantom scans were performed on a clinical PET/CT (Biograph HiRez16, Siemens). An abdominal phantom (NEMA-2001, 35x40x25cm3) containing 6 concentric spherical lesions of 11mm to 38mm diameter was used. Two 10cm diameter plastic cylinders simulating the arms were positioned at the lateral sides of the phantom. The phantom and arm cylinders were filled with water+[18F]-FDG (5kBq/mL). Lesion-to-background ratio was 1.5-to-1. The phantom was centred and CT and PET data were acquired following a whole-body PET/CT protocol: 120kVp CT, 5mm slices, 10min 3D-emission scan. Six CT image volumes were acquired with tube currents (mAs) of 17, 30, 80, 130, 180 and 230 and post-processed using a non-linear adaptive filtering method (ContextVision, Sweden). PET emission images were reconstructed iteratively (OSEM2D, 4i,8s) following CTAC using the original CT (orgCT) and filtered CT (filCT). CT and AC-PET images were analysed by means of regions-of-interests placed in the arms (A), abdomen (AB), central background (CB) and the largest 3 spheres.  

CT image noise varied with the tube current whereas AC-PET image noise did not. Noise (HU) in orgCT/filCT was 23/14 (A), 26/16 (AB) and 28/17 (CB) for 17mAs, and 7/6 (A), 8/7 (AB) and 9/7 (CB) for 230mAs. Corresponding noise values in AC-PET were the same for orgCTAC and filCTAC: 10% (A), 8% (AB) and 8% (CB) for 17mAs and 10% (A), 8% (AB) and 8% (CB) for 230mAs. Using filCT for CTAC had a small effect on the noise level of the AC-PET of the lesions. Noise in the 3 largest regions was reduced from 11% to 10%, 11% to 10% and 9% to 7% following orgCTAC and filCTAC (17mAs). Similar noise level reduction was observed for 230mAs.  

CT image noise can be reduced using post-reconstruction filtering. Subsequent benefits on noise propagation through CT-AC are not clinically relevant.

Noise propagation from low-dose CT in PET/CT is clinically insignificant but post-recon filtering of the CT data may help to enhance the quality of PET/CT data in ultra-low-dose CT applications.

Werner, M, Lois, C, Kupferschläger, J, Fredriksson, C, Heners, H, Pfannenberg, C, Beyer, T, Noise Propagation in PET/CT through CT-based Attenuation Correction: Effects of Low-Dose CT Imaging.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034633.html