Jiro Sato MD, Presenter: Nothing to Disclose

Masaki Katsura MD, Abstract Co-Author: Nothing to Disclose

Masaaki Akahane MD, Abstract Co-Author: Nothing to Disclose

Izuru Matsuda MD, Abstract Co-Author: Nothing to Disclose

Sachiko Inano, Abstract Co-Author: Nothing to Disclose

Koji Segawa RT, Abstract Co-Author: Employee, General Electric Company

Kosuke Sasaki MS, Abstract Co-Author: Employee, General Electric Company

Akira Kunimatsu MD, Abstract Co-Author: Nothing to Disclose

Kuni Ohtomo MD, Abstract Co-Author: Nothing to Disclose

Model-based iterative reconstruction (MBIR) is a new reconstruction technique expected to provide a higher quality CT image and allow significant dose reduction, compared with standard filtered back projection (FBP) and adaptive statistical iterative reconstruction (ASIR). The purpose of this study is to compare image quality for abdominal CT images in clinical cases reconstructed with MBIR, FBP and ASIR.

44 patients (26 men, 18 women; mean age 65.5 years) underwent unenhanced body CT (GE Discovery CT 750 HD 64-slice scanner; 120 kVp, 85-200mAs) with automatic tube current modulation technique using a predetermined noise index value of 14 for 5-mm thickness. 2.5 mm-thick images in the abdominal region were reconstructed using standard FBP, 50% ASIR, and MBIR. Mean and standard deviation of CT numbers within regions of interest placed on 2 locations at the level of the right portal vein (portal vein and liver parenchyma) was recorded, and contrast-to-noise ratio (CNR) was calculated. Two experienced radiologists independently graded image noise, visual sharpness of liver edge, conspicuity of the hepatic veins within liver, visualization of the extrahepatic duct (above pancreas) and diagnostic acceptability, using a 1-5 grading score. Statistical analysis was performed using Bonferroni adjusted t-tests for pair-wise comparisons for objective image quality assessment, and Wilcoxon signed rank test for subjective image quality assessment.

MBIR showed significantly less noise than ASIR or FBP (10.6 +/- 1.80, 26.7 +/- 4.00, 39.3 +/- 5.89, respectively (HU); p<0.001) and higher CNR (2.51 +/- 0.85, 0.93 +/- 0.42, 0.63 +/- 0.30; respectively; p<0.001). Subjective image quality scores with MBIR compared to those with ASIR and FBP were significantly higher (p<0.001) for all evaluation areas.

MBIR reduces image noise and improves CNR and subjective image quality as compared with ASIR and FBP.

Application of MBIR to abdominal CT produces higher quality images, with a potential for dose reduction in routine practice.

Sato, J, Katsura, M, Akahane, M, Matsuda, I, Inano, S, Segawa, K, Sasaki, K, Kunimatsu, A, Ohtomo, K, Model-based Iterative Reconstruction Improves Image Quality of Abdominal CT Compared with Filtered Back Projection and Adaptive Statistical Iterative Reconstruction.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11009020.html