Takashi Takahata RT, Presenter: Nothing to Disclose

Tomokatsu Tsukamoto, Abstract Co-Author: Nothing to Disclose

Hiroki Mori MD, Abstract Co-Author: Nothing to Disclose

Kazunari Mesaki MD, Abstract Co-Author: Nothing to Disclose

Keisuke Nishihara MD, Abstract Co-Author: Nothing to Disclose

Katsuhide Ito MD, Abstract Co-Author: Nothing to Disclose

Xiao Zhu Lin MD, Abstract Co-Author: Nothing to Disclose

To assess the CT dose reduction potential and image quality improvement with model-based iterative reconstruction algorithm (Veo) using autopsy imaging by comparing image noise and DOSE (DLP mGy-cm) with the adaptive statistical iterative reconstruction (ASiR) and the filtered back projection (FBP) reconstructions.

With institutional review board approval, 8 autopsy imaging (AI) underwent HDCT (Discovery CT750 HD, GE) with different noise index (For Brain scan: NI: 2.8, 3.2, 4.5, 6.0, 8.5, and For Body scan: NI: 8.5, 10.5, 14.5, 20.5, 30.5) was included. For comparison, the 3 sets of 0.625mm slice thickness CT images were reconstructed with FBP, 50% ASiR and Veo. The image noise (SD) was measured with the same size of regions of interest at the same slice in 3 locations for head, lung, liver and pelvis. The image noise reduction ratio was defined by SD (at large NI)/SD (at small NI )(NI8.5/NI2.8 for brain, NI30.5/NI8.5 for Body). Using a 5-point score (1: poor; 3: diagnosis, 5 excellent), 3 radiologists independently and graded overall noise and delineation of the brain and body image.

For the comparison with same image slice thickness, the average image noise reduction at different NI with Veo compared with current algorithm (FBP and 50%ASiR) for the brain, lung, liver and pelvis were (33.2±14.6% and 17.0±15.4%), (50.1±14.9% and 55.6±12.4%), (62.0±13.0% and 48.5±15.5%) and (64.7±12.9% and 50.2±18.4%), respectively. The reduction ratio of image noise with different algorithms (Veo, 50%ASiR and FBP) for the brain, lung, liver and pelvis were (1.7, 2.7 and 2.9), (0.8, 1.8 and 1.8), (1.5, 3.2 and 3.9), and (1.5, 3.9 and 3.9), respectively. All differences were statistically significant between Veo and FBP (p<0.05). The average scorings with different algorithms (Veo, 50%ASiR and FBP) for the brain, lung, liver and pelvis were (2.9±1.1, 2.5±1.2 and 2.1±0.9), (3.5±0.7, 3.3±1.2 and 3.1±1.1), (3.4±1.0, 2.6±1.2 and 2.3±1.0), and (3.5±0.9, 2.7±1.2 and 2.4±1.0), respectively. Veo greatly improve image quality from scoring result compared with ASiR and FBP.

Veo advanced reconstruction algorithms greatly reduced image noise to compare FBP and ASiR .

Veo reconstruction technique has the ability to reduce radiation dose through their improvement in image quality compared with the current algorithms such as FBP and ASiR.

Takahata, T, Tsukamoto, T, Mori, H, Mesaki, K, Nishihara, K, Ito, K, Lin, X, CT Dose Reduction Potential and Image Quality Improvement with Model-based Iterative Reconstruction Using Autopsy Imaging: Evaluation of Image Noise and DOSE Estimation with Different Noise Index.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12043916.html