Abstract

Quantitative errors (due to magnetic susceptibility artifacts) in the measurement of the cervical spinal neural foramina with fast gradient-echo (GRE) magnetic resonance imaging were assessed. Cylindric phantoms of different materials were used to demonstrate the nature of magnetic susceptibility artifacts, emphasizing the dependence of the artifact on tissue geometry. Neural foramina diameters measured on thin, sagittal GRE and spin-echo (SE) images through the neural foramina of a fresh human cervical spine specimen were then compared with direct measurements with calipers. The GRE images showed more apparent narrowing than did the SE images. The absolute distortion of seven neural foramina was rather constant (less than two pixels) on the GRE images; therefore, the relative distortion was inversely proportional to the size of the neural foramen, ranging up to 10% in the upper cervical region at a short TE. The absolute and relative distortion increased as TE increased. At a constant TE, the structural distortion did not change with different TRs or flip angles. The shortest possible TE is recommended in evaluation of the cervical spine.