Algorithm-based artifact reduction in patients with arms-down positioning in computed tomography. (January 2020)

• Record Type: Journal Article
• Authors: Kawashima, Hiroki; Ichikawa, Katsuhiro; Takata, Tadanori; Mitsui, Wataru
• Is Part Of: Physica medica. Volume 69(2020)
• Page Start: 61
• Page End: 69
• Publication Date: 2020-01
• Subjects: Computed tomography -- Artifact -- Polytrauma -- Image quality -- Reconstruction algorithm; Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2019.11.019
• Languages: English
• ISSNs: 1120-1797

Abstract

Highlights: Arm-artifacts are observed in computed tomography with arms-down positioning. An algorithm of selective arm-induced noise filtering is proposed in this study. The artifact reduction performance was compared with hybrid IR and MBIR. The proposed algorithm significantly reduced the arm-artifacts similarly to MBIR. The spatial resolution preservation performance was superior to MBIR. Abstract: Purpose: Arm-artifact, a type of streak artifact frequently observed in computed tomography (CT) images obtained at arms-down positioning in polytrauma patients, is known to degrade image quality. This study aimed to develop a novel arm-artifact reduction algorithm (AAR) applied to projection data. Methods: A phantom resembling an adult abdomen with two arms was scanned using a 16-row CT scanner. The projection data were processed by AAR, and CT images were reconstructed. The artifact reduction for the same phantom was compared with that achieved by two latest iterative reconstruction (IR) techniques (IR1 and IR2) using a normalized artifact index (nAI) at two locations (ventral and dorsal side). Image blurring as a processing side effect was compared with IR2 of the model-based IR using a plastic needle phantom. Additionally, the projection data of two clinical cases were processed using AAR, and the image noise was evaluated. Results: AAR and IR2 significantly reduced nAI by 87.5% and 74.0%, respectively at the ventral side and 84.2% and 69.6%, respectively, at the dorsal side compared with each filtered back projection ( P < 0.01), whereas IR1 did not. The proposed algorithm mostly maintained the original spatial resolution, compared with IR2, which yielded apparent image blurring. The image noise in the clinical cases was also reduced significantly ( P < 0.01). Conclusions: AAR was more effective and superior than the latest IR techniques and is expected to improve the image quality of polytrauma CT imaging with arms-down positioning.