Ke Li PhD, Presenter: Nothing to Disclose

Adam Budde MS, Abstract Co-Author: Employee, General Electric Company

Daniel Gomez-Cardona, Abstract Co-Author: Nothing to Disclose

Jiang Hsieh PhD, Abstract Co-Author: Employee, General Electric Company

Guang-Hong Chen PhD, Abstract Co-Author: Research funded, General Electric Company Research funded, Siemens AG Research funded, Varian Medical Systems, Inc Research funded, Hologic, Inc

Inappropriate patient centering in CT imaging can lead to increase in image artifacts, noise level, and an elevation of radiation dose. The purpose of this work is to investigate whether the use of the model based statistical image reconstruction method equipped with a multi-slice CT scanner would alleviate the impact of patient mis-centering on CT image quality and radiation dose.

A 64-slice clinical CT equipped with a MBSIR recon engine (Veo, GE Healthcare) was used to scan an anthropomorphic LUNGMAN phantom and a uniform polyethylene oval phantom using clinical scanning protocols. Custom-written software was used to perform online phantom centering from localizer radiographs. After centering and scanning, the mis-centered scans were produced by lowering the patient couch by 3 cm and 6 cm from the centered position and scanning with the same protocol. Both FBP and Veo reconstruction were performed on each scan. Quantitative image analyses were then performed using noise standard deviation (STD) and noise nonuniformity index (NNI) as metrics. A Monte-Carlo method was used to estimate dose delivered across each phantom.

For the chest phantom, there was no significant difference in the overall STD between the centered and the 3 cm mis-centered FBP images, but the 6 cm FBP image led to 14% increase in STD (p<2e-6) versus the centered FBP images. No significant difference in overall STD was found between the centered and off-centered Veo images (<0.3 HU). NNI increased from 13% (0 cm) to 43% (3 cm) or 68% (6 cm) in FBP images but remained constant in Veo images. For the oval phantom, the overall STD values were 7% and 25% higher (p<0.002) in the 3 and 6 cm off-centered FBP images, respectively, but no significant difference was found between the centered and off-centered Veo images (<0.4 HU). Mis-centering increased the overall dose by 1.5% (3 cm) or 4.5% (6 cm) in the chest phantom, and by 2.0% (3 cm) or 6.2% (6 cm) in the oval phantom.

Unlike the conventional CT images reconstructed by FBP, the noise magnitude and spatial uniformity of MBSIR images is insensitive to patient mis-positioning within 6 cm. However, significant mis-positioning could potentially limit the capability to reduce radiation dose to the patient.

This work addresses a current concern about the importance of patient centering while using the increasingly popular MBSIR during routine CT exams.

Li, K, Budde, A, Gomez-Cardona, D, Hsieh, J, Chen, G, Impact of Patient Posititioning on CT Image Quality: With and without Model Based Statistical Image Reconstruction Method .  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14014029.html