RSNA 2011 

Abstract Archives of the RSNA, 2011


LL-GIS-TU2B

Comparison of Model-based Iterative Reconstruction (MBIR), Adaptive Statistical Iterative Reconstruction (ASiR), and Filtered Back Projection (FBP): Evaluation of Radiation Dose and Image Quality at Abdominal CT with an Anthropomorphic Phantom

Scientific Informal (Poster) Presentations

Presented on November 29, 2011
Presented as part of LL-GIS-TU: Gastrointestinal

Participants

Atsushi Nakamoto MD, Presenter: Nothing to Disclose
Tonsok Kim MD, Abstract Co-Author: Nothing to Disclose
Masatoshi Hori MD, Abstract Co-Author: Nothing to Disclose
Hiromitsu Onishi MD, Abstract Co-Author: Nothing to Disclose
Masayuki Kudo PhD,RT, Abstract Co-Author: Employee, General Electric Company
Noriyuki Tomiyama MD, PhD, Abstract Co-Author: Nothing to Disclose
Takahiro Tsuboyama MD, Abstract Co-Author: Nothing to Disclose
Makoto Sakane, Abstract Co-Author: Nothing to Disclose
Mitsuaki Tatsumi MD, PhD, Abstract Co-Author: Nothing to Disclose

PURPOSE

To compare the image quality of abdominal CT images reconstructed with a model-based iterative reconstruction (MBIR), an adaptive statistical iterative reconstruction (ASiR) and a Filtered Back Projection (FBP) at different radiation dose by using an anthropomorphic phantom.

METHOD AND MATERIALS

An anthropomorphic upper abdomen CT phantom (PH-5; KYOTO KAGAKU, Kyoto, Japan) was scanned using a 64-slice CT scanner (GE CT750 HD, GE Healthcare, Milwaukee, WI). Five different tube-currents at two different tube-voltages were used: 400 mA, 200 mA, 100 mA, 50 mA and 25 mA at 120 kVp, and 700 mA, 400 mA, 200 mA, 100 mA and 50 mA at 80 kVp. For each kVp and mA setting, axial images were reconstructed with FBP, 40% ASiR and MBIR. The standard deviation of CT number of material simulating soft tissue was employed as the index of image noise. Contrast-to-noise (CNR) ratio for the liver parenchyma relative to the portal vein was also calculated. As a qualitative analysis, two radiologists assessed the image quality using a five-point rating scale.

RESULTS

At all kVp and mA settings, MBIR yielded 29%-75% noise reduction and 49%-335% CNR increase compared with FBP, and 10%-66% noise reduction and 4%-158% CNR increase compared with ASiR. At 120 kVp setting, noise and CNR for 50mA with MBIR (5.3 and 4.9, respectively) were comparable with those for 400 mA with FBP (5.3 and 4.7, respectively). At 80 kVp setting, noise and CNR at 100 mA with MBIR (5.9 and 5.1, respectively) were comparable with those for 700 mA with FBP (6.5 and 5.3, respectively). The mean visual score of image quality with MBIR was higher than that with FBP or ASiR at lower tube-current (25 mA at 120kVp), although no difference was found at higher tube-currents.

CONCLUSION

MBIR yielded radiation dose reduction of 85.7%-87.5% compared with FBP maintaining CNR and acceptable image noise.

CLINICAL RELEVANCE/APPLICATION

Model-based iterative reconstruction would enable significant reduction in radiation dose maintaining image quality.

Cite This Abstract

Nakamoto, A, Kim, T, Hori, M, Onishi, H, Kudo, M, Tomiyama, N, Tsuboyama, T, Sakane, M, Tatsumi, M, Comparison of Model-based Iterative Reconstruction (MBIR), Adaptive Statistical Iterative Reconstruction (ASiR), and Filtered Back Projection (FBP): Evaluation of Radiation Dose and Image Quality at Abdominal CT with an Anthropomorphic Phantom.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11015061.html