Rationale and objectives: The objective of this study was to evaluate a 2-dimensional spoiled gradient echo (GRE) imaging approach using a very short in-phase TE for routine T1-weighted imaging of the brain at 3 T.
Materials and methods: Patient examinations were compared from a 3 T magnetic resonance (MR) unit located immediately adjacent to a similarly equipped 1.5 T unit. Pre- and postcontrast T1-weighted images were evaluated and compared at 1.5 versus 3 T with a 2-dimensional (2-D) spin echo sequence used at 1.5 T and a 2-D GRE sequence at 3 T. The 2 MR systems used are from the same vendor, use similar 8-channel coils, and use identical gradients. The T1-weighted GRE sequence, used at 3 T, relies on a short TE (2.4 ms) to limit flow-related and susceptibility artifacts. Region-of-interest analysis was performed on 16 different sagittal patient examinations at both field strengths (32 total) and similarly on 10 different pre- and postcontrast axial examinations (40 total). Four blinded neuroradiologists also evaluated these studies.
Results: Using an off-midline sagittal slice depicting the caudate nucleus (signal-to-noise ratio [SNR] 163 +/- 28 vs. 70 +/- 7, 3 T vs. 1.5 T) and corona radiata (SNR 214 +/- 35 vs. 82 +/- 10), 3 T markedly outperformed 1.5 T in both SNR and contrast-to-noise ratio (CNR) (51 +/- 14 vs. 12 +/- 5). On axial imaging, despite a reduction in slice thickness (5 to 3 mm) and scan time (5 to 1 minute), there was no significant difference pre- or postcontrast in SNR and CNR comparing 3 and 1.5 T. On blinded film review, 3 T performed slightly better on sagittal scans than 1.5 T in regard to motion artifacts (reduced), gray-white matter differentiation, and overall image quality. On axial scans, 3 T performed markedly better in all 3 categories both pre- and postcontrast. In regard to overall image quality, 3 T was preferred 9:2 precontrast and 4:1 postcontrast.
Conclusions: High-quality, thin-section (3-mm) T1-weighted imaging can be readily performed at 3 T using a short TE 2-D GRE technique. This approach offers superior SNR and CNR with reduced motion artifacts and scan time as compared with imaging at 1.5 T and is advocated for routine brain imaging at 3 T. It is robust (used in over 1500 patients to date) and does not experience significant specific absorption ratio limitations, poor tissue contrast, or accentuated motion artifacts like encountered with spin echo T1-weighted imaging at 3 T.