Andrei S. Purysko MD, Presenter: Nothing to Disclose

Andrew Primak PhD, Abstract Co-Author: Employee, Siemens AG

Mark E. Baker MD, Abstract Co-Author: Research Consultant, Bracco Group Researcher, Siemens AG Research support, Siemens AG

Jennifer Bullen MSc, Abstract Co-Author: Nothing to Disclose

Nancy A. Obuchowski PhD, Abstract Co-Author: Research Consultant, Siemens AG Research Consultant, Hologic, Inc Research Consultant, CVUS Research Consultant, Elucid Bioimaging Inc

Erick Marc Remer MD, Abstract Co-Author: Nothing to Disclose

Christopher Peter Coppa MD, Abstract Co-Author: Nothing to Disclose

Namita Sharma Gandhi MD, Abstract Co-Author: Nothing to Disclose

Myra Kay Feldman MD, Abstract Co-Author: Nothing to Disclose

Devaraju Kanmaniraja MD, Abstract Co-Author: Nothing to Disclose

Shetal N. Shah MD, Abstract Co-Author: Nothing to Disclose

Larry J. Lee MD, Abstract Co-Author: Nothing to Disclose

Dianna Yu Ning Yang DO, Abstract Co-Author: Nothing to Disclose

Dipika Patel MD, Abstract Co-Author: Nothing to Disclose

Brian Robert Herts MD, Abstract Co-Author: Research Grant, Siemens AG

To compare image quality and CT numbers between true-noncontrast (TNC) and virtual noncontrast (VNC) images of the abdomen from second-generation dual-source dual-energy CT (DSCT) scanner

IRB-approved, HIPPA-compliant retrospective study of 42 subjects (mean age 59.7 yrs. [SD=9.1]) who underwent 3-phase CT of the liver on a DSCT scanner (Flash, Siemens Healthcare) using single energy (SE) (120kVp) and dual energy (DE) protocol (100/Sn140kVp), on different occasions. Four sets of images (TNC-SE, TNC-DE and VNC reconstructed from the arterial [VNCa] and portal-venous [VNCp] phases from DE exams) of the abdomen from each patient were randomly reviewed by 5 readers, who graded the image quality (1. Very good; 2. Good; 3. Fair; 4. Poor; 5. Very poor) and the presence of artifacts (1. None; 2. Do not affect interpretation; 3. Affect interpretation but diagnosis still possible; 4. Inadequate for use). Acceptable image quality was defined as a score ≤3, and acceptable artifact as a score ≤3. Regression analysis was used to assess differences in scores for image quality and artifacts, and differences in CT numbers obtained in the liver, kidneys, spleen, inferior vena cava, aorta and retroperitoneal (RP) fat.

Compared to TNC-DE, the mean proportion of acceptable image quality was 0.114 lower for VNCa (95% CI: 0.077, 0.152) and 0.119 lower for VNCp (95% CI: 0.073, 0.165). The difference between the VNCa and VNCp was small and not statistically significant (0.005, 95% CI: -0.031, 0.041). VNC sets had high frequencies of acceptable artifacts (VNCa: 0.995, VNCp: 0.986), and the differences among them were not statistically significant. For the liver, the mean difference in CT number between TNC-DE and VNCa was -0.7 (95% CI: -1.7, 0.3), and between TNC-DE and VNCp was -0.7 (95% CI: -1.7, -0.3). The largest differences in the mean CT numbers between VNCa and TNC-DE was in the aorta (50.2 vs. 40.8, 95% CI: 6.8, 11.9), and the largest difference between VNCp and TNC-DE was in the RP fat (-87.2 vs. -92.4, 95% CI: 4.0, 6.6). In the liver, spleen and RP fat both VNC sets showed a better agreement with TNC-DE than with TNC-SE.

VNC images had high frequency of acceptable image quality and artifacts, with small differences in CT numbers compared to TNC-DE.

VNC is an alternative to TNC images of the abdomen, which helps decrease patients’ exposure to radiation.

Purysko, A, Primak, A, Baker, M, Bullen, J, Obuchowski, N, Remer, E, Coppa, C, Gandhi, N, Feldman, M, Kanmaniraja, D, Shah, S, Lee, L, Yang, D, Patel, D, Herts, B, Comparison of Image Quality and CT Numbers between True-noncontrast and Virtual Noncontrast Images of the Abdomen from Second-generation Dual-source CT Scanner.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14001923.html