Image Quality and Artifact Reduction of a Cochlear Implant With Rotatable Magnets.

Abstract

To determine if metal reduction magnetic resonance imaging sequences and changes in implant placement minimize artifact from cochlear implants and improve visualization of intracranial structures. Cadaveric study. Tertiary referral center. Five cadaveric heads. Specimens were implanted with Advanced Bionics HiRes Ultra3D devices at nasion-external auditory canal angles of 90, 120, and 160 degrees, and distances from the external auditory canal of 9 or 12 cm. Standard brain/internal auditory canal (IAC) sequences with metal artifact reducing technique were acquired in a 1.5T scanner. The primary outcome was visibility of 14 intracranial structures graded on a 4-point scale (1, structures <50% visible; 2, >50% visible with some areas nonvisible from artifact; 3, artifact present but adequate for diagnosis; and 4, high quality). Scores were determined by experienced head and neck radiologists and compared with one-way analysis of variance. Imaging sequences included axial 5-mm whole-brain turbo spin echo (TSE) T2 with right to left and anterior to posterior encoding, fluid-attenuation inversion recovery high bandwidth, axial 5-mm whole-brain slice-encoding metal artifact correction (SEMAC), axial IAC constructive interference in steady state, and axial 3-mm T1 IAC with and without fat saturation. T1 IACs in axial and coronal planes were best for ipsilateral structures overall (mean [standard deviation {SD}], 3.8 [0.6] and 3.8 [0.5]). SEMAC (mean [SD], 3.5 [0.8]) was superior to TSE with anterior to posterior encoding (mean [SD], 3.5 [0.9) for ipsilateral cortex, cerebellopontine angle, and brainstem/cerebellum, and equivalent for the inner ear. Constructive interference in steady state and T1 with fat saturation were poor for all ipsilateral structures (mean, 2.8 [ p < 0.01]; mean, 3.1 [ p < 0.01]). The 120 degrees/12 cm position was overall best, although the 120 degrees/9 cm position still afforded visualization of ipsilateral structures; other angles and distances conferred slight advantages for specific structures of interest. SEMAC and T2 TSE with anterior to posterior encoding sequences provide artifact suppression while retaining excellent image quality. Different placement angles did not confer improvement in visualization, although placement distances provided slight advantages for some structures.