Erschienen in:
02.03.2017 | Diagnostic Neuroradiology
Reduction of metal artifacts due to dental hardware in computed tomography angiography: assessment of the utility of model-based iterative reconstruction
verfasst von:
Keita Kuya, Yuki Shinohara, Ayumi Kato, Makoto Sakamoto, Masamichi Kurosaki, Toshihide Ogawa
Erschienen in:
Neuroradiology
|
Ausgabe 3/2017
Einloggen, um Zugang zu erhalten
Abstract
Purpose
The aim of this study is to assess the value of adaptive statistical iterative reconstruction (ASIR) and model-based iterative reconstruction (MBIR) for reduction of metal artifacts due to dental hardware in carotid CT angiography (CTA).
Methods
Thirty-seven patients with dental hardware who underwent carotid CTA were included. CTA was performed with a GE Discovery CT750 HD scanner and reconstructed with filtered back projection (FBP), ASIR, and MBIR. We measured the standard deviation at the cervical segment of the internal carotid artery that was affected most by dental metal artifacts (SD1) and the standard deviation at the common carotid artery that was not affected by the artifact (SD2). We calculated the artifact index (AI) as follows: AI = [(SD1)2 − (SD2)2]1/2 and compared each AI for FBP, ASIR, and MBIR. Visual assessment of the internal carotid artery was also performed by two neuroradiologists using a five-point scale for each axial and reconstructed sagittal image. The inter-observer agreement was analyzed using weighted kappa analysis.
Results
MBIR significantly improved AI compared with FBP and ASIR (p < 0.001, each). We found no significant difference in AI between FBP and ASIR (p = 0.502). The visual score of MBIR was significantly better than those of FBP and ASIR (p < 0.001, each), whereas the scores of ASIR were the same as those of FBP. Kappa values indicated good inter-observer agreements in all reconstructed images (0.747–0.778).
Conclusions
MBIR resulted in a significant reduction in artifact from dental hardware in carotid CTA.