Abstract
OBJECTIVE The aim of this study was to compare the degree of artifact reduction in magnetic resonance imaging achieved with slice encoding for metal artifact correction (SEMAC) in combination with view angle tilting (VAT) and multiacquisition variable resonance image combination (MAVRIC) for standard contrast weightings and different metallic materials. METHODS Four identically shaped rods made of the most commonly used prosthetic materials (stainless steel, SS; titanium, Ti; cobalt-chromium-molybdenum, CoCr; and oxidized zirconium, oxZi) were scanned at 3 T. In addition to conventional fast spin-echo sequences, metal artifact reduction sequences (SEMAC-VAT and MAVRIC) with varying degrees of artifact suppression were applied at different contrast weightings (T1w, T2w, PDw). Two independent readers measured in-plane and through-plane artifacts in a standardized manner. In addition, theoretical frequency-offset and frequency-offset-gradient maps were calculated. Interobserver agreement was assessed using intraclass correlation coefficient. RESULTS Interobserver agreement was almost perfect (intraclass correlation coefficient, 0.86-0.99). Stainless steel caused the greatest artifacts, followed by CoCr, Ti, and oxZi regardless of the imaging sequence. While for Ti and oxZi rods scanning with weak SEMAC-VAT showed some advantage, for SS and CoCr, higher modes of SEMAC-VAT or MAVRIC were necessary to achieve artifact reduction. MAVRIC achieved better artifact reduction than SEMAC-VAT at the cost of longer acquisition times. Simulations matched well with the apparent geometry of the frequency-offset maps. CONCLUSIONS For Ti and oxZi implants, weak SEMAC-VAT may be preferred as it is faster and produces less artifact than conventional fast spin-echo. Medium or strong SEMAC-VAT or MAVRIC modes are necessary for significant artifact reduction for SS and CoCr implants. KEY POINTS
Filli, Lukas