Jian He MD, PhD, Presenter: Nothing to Disclose

Kefeng Zhou, Abstract Co-Author: Nothing to Disclose

Shangwen Yang MD, Abstract Co-Author: Nothing to Disclose

Bin Zhu, Abstract Co-Author: Nothing to Disclose

To compare the qualities of plain and contrast enhanced images at the shoulder reconstructed by three different algorithms: Filtered Back Projection (FBP), Adaptive Statistical Iterative Reconstruction (ASIR) and Model-Based Iterative Reconstruction (MBIR).

Ten patients underwent humeral CT scan with a low radiation dose (120kVp, 200mA, CTDIvol 14.57mGy, DLP 324.53 mGy•cm, Dose Eff. 90.53%) before and after the injection of contrast media. Images were reconstructed by three different algorithms: FBP, ASIR and MBIR. CT values (HU) of thyroid gland, axillary fat subclavian artery and muscles were measured and recorded as mean ± standard deviation (SD). Contrast-noise ratio (CNR) and noise index (NI=SD) were calculated and compared among three sets of images with analysis of variance. Subjective image quality was assessed in terms of subjective image noise, artifacts, the visibility of small structures, image contrast and subjective visual lesion conspicuity and diagnostic confidence by two radiologists independently.

On plain and congrast enhanced images of the shoulder, significantly increased CNR and decreased NI were documented at MBIR and ASIR compared with FBP images (p < 0.05). And in some areas, CNR increased and NI decreased significantly at MBIR than at ASIR images (p < 0.05). Subjective image noise and artifacts were significantly reduced by MBIR than ASIR and FBP (p<0.001). Image contrast of FBP is better than MBIR (p=0.041), and the latter is better than ASIR (p=0.040). There were no significant differences on the visibility of small structures, subjective visual lesion conspicuity and diagnostic confidence among those three groups.

Significantly reduced objective and subjective image noise can be achieved by MBIR and ASIR at the shoulder at a relative low radiation dose. However, the visibility of small structures, subjective visual lesion conspicuity and diagnostic confidence remained due to compromised image contrast by MBIR and ASIR and the change of image appearance.

Only at an adequately reduced radiation dose, benefits of iterative reconstruction can be achieved by proper weight of ASIR and appropriate utilization of MBIR.

He, J, Zhou, K, Yang, S, Zhu, B, Is ASIR or MBIR Always Better than FBP Algorithm at Shoulder CT Images?.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12024179.html