Three-dimensional isotropic T2-weighted cervical MRI at 3T: comparison with two-dimensional T2-weighted sequences.

AIM To compare three-dimensional (3D) isotropic T2-weighted magnetic resonance imaging (MRI) sequences and reformation with two-dimensional (2D) T2-weighted sequences regarding image quality of the cervical spine at 3T. MATERIALS AND METHODS A phantom study was performed using a water-filled cylinder. The signal-to-noise and image homogeneity were evaluated. Fourteen (n=14) volunteers were examined at 3T using 3D isotropic T2-weighted sagittal and conventional 2D T2-weighted sagittal, axial, and oblique sagittal MRI. Multiplanar reformation (MPR) of the 3D T2-weighted sagittal dataset was performed simultaneously with image evaluation. In addition to artefact assessment, the visibility of anatomical structures in the 3D and 2D sequences was qualitatively assessed by two radiologists independently. Cohen's kappa and Wilcoxon signed rank test were used for the statistical analysis. RESULT The 3D isotropic T2-weighted sequence resulted in the highest signal-to-noise ratio (SNR) and lowest non-uniformity (NU) among the sequences in the phantom study. Quantitative evaluation revealed lower NU values of the cerebrospinal fluid (CSF) and muscles in 2D T2-weighted sagittal sequences compared to the 3D volume isotropic turbo spin-echo acquisition (VISTA) sequence. The other NU values revealed no statistically significant difference between the 2D turbo spin-echo (TSE) and 3D VISTA sequences (0.059<p<0.959). 3D VISTA images showed significantly fewer CSF flow artefacts (p<0.001) and better delineated intradural nerve rootlets (p=0.001) and neural foramina (p=0.016) compared to 2D sequences. CONCLUSION A 3D T2 weighted sequence is superior to conventional 2D sequences for the delineation of intradural nerve rootlets and neural foramina and is less affected by CSF flow artefacts.