Performance of single-energy metal artifact reduction in cardiac computed tomography: A clinical and phantom study

Abstract

BackgroundTo investigate the performance of a reconstruction algorithm, single-energy metal artifact reduction (SEMAR), against standard reconstruction in cardiac computed tomography (CT) studies of patients with implanted metal and in a defibrillator lead phantom. MethodsFrom a retrospective, cross-sectional clinical study with institutional review board approval of 118 patients with implanted metal, 122 cardiac CT studies from November 2009 to August 2016 performed on a 320-detector row scanner with standard and SEMAR reconstructions were included. The maximum beam hardening artifact radius, artifact attenuation variation surrounding the implanted metal, and image quality on a 4-point scale (1-no/minimal artifact to 4-severe artifact) were assessed for each reconstruction. A defibrillator lead phantom study was performed at different tube potentials and currents with both reconstruction methods. Maximum beam hardening artifact radius and average artifact attenuation variation were measured. ResultsIn the clinical study, SEMAR markedly reduced the maximum beam hardening artifact radius by 77% (standard: 14.8 mm [IQR 9.7–22.2] vs. SEMAR: 3.4 mm [IQR 2.2–7.1], p < 0.0001) and artifact attenuation variation by 51% (standard: 130.0 HU [IQR 75.9–184.4] vs. SEMAR: 64.3 HU [IQR 48.2–89.2], p < 0.0001). Image quality improved with SEMAR (standard: 3 [IQR 2–3.5] vs. SEMAR: 2 [IQR 1–2.5], p < 0.0001). The defibrillator lead phantom study confirmed these results across varying tube potentials and currents. ConclusionsSEMAR reconstruction achieved superior image quality and markedly reduced maximum beam hardening artifact radius and artifact attenuation variation compared to standard reconstruction in 122 clinical cardiac CT studies of patients with implanted metal and in a defibrillator lead phantom study.