High-resolution intracranial and cervical MRA at 3.0T: technical considerations and initial experience.

Initial experience with intracranial and cervical MRA at 3.0T is reported. Phantom measurements (corrected for relaxation effects) show S/N (3.0T) = 2.14 +/- 0.08 x S/N (1.5T) in identical-geometry head coils. A 3.0T 3DTOF intracranial imaging protocol with higher-order autoshimming was developed and compared to 1.5T 3DTOF in 12 patients with aneurysms. A comparison by two radiologists showed the 3.0T to be significantly better (P < 0.001) for visualization of the aneurysms. The feasibility of cervical and intracranial contrast enhanced MR angiography (CEMRA) at 3.0T is also examined. The relaxivity of the gadolinium contrast agent decreases by only about 4-7% when the field strength is increased from 1.5 to 3.0T. Cervical 3.0T CEMRA was obtained in eight patients, two of whom had 1.5T studies available for direct comparison. Image comparison suggests 3.0T to be a favorable field strength for cervical CEMRA. Voxel volumes of 0.62-0.73 mm(3) (not including zero-filling) were readily achieved at 3.0T with the use of a single-channel transmit-receive head or cervical coil, a 25 mL bolus of gadoteridol, and a 3D pulse sequence with a 66% sampling efficiency. This spatial resolution allowed visualization of intracranial aneurysms, carotid dissections, and atherosclerotic disease including ulcerations. Potential drawbacks of 3.0T MRA are increased SAR and T(*)(2) dephasing compared to 1.5T. Image comparison suggests signal loss due to T(*)(2) dephasing will not be substantially more problematic than at 1.5T. The dependence of RF power deposition on TR for CEMRA is calculated and discussed. Copyright 2001 Wiley-Liss, Inc.