Achieva 1.5 tesla scanner

Fetal DTI was conducted without sedation on a Philips 1.5-Tesla Achieva scanner using a 32-channel phased array cardiac coil. Single-shot echo-planar DTI sequence parameters were: b value 0 (b₀) and 500 s/mm², 15 non-collinear directions, TE 121 ms, TR 8500 ms, FoV 290 × 290 × 128 mm³, voxel size 2.3 × 2.3 × 3.5 mm³, slice gap − 1.75 mm, and number of slices 62–66 (dependent of gestational age), and acquisition order was set to: odd–even slice, ascending order. Overlapping slices allow for oversampling the fetal brain, which increases the likelihood of sampling enough data for reconstruction, even in cases of significant motion [9 (link)] (acquisition time: 5 min 6 s). To assist in subsequent registration operations, three additional stacks of b₀ spin-echo EPI images, two axial and one coronal with respect to fetal brain anatomy were acquired using matched parameters (acquisition time: 1 min 42 s). Static magnetic field (B₀) maps (TE1 4.6 ms, TE2 9.2 ms; TR 10 ms, Flip Angle 10 degree, voxel size 2.27 × 2.27 × 10 mm³, FoV 400 × 400 × 150 mm³) covering the region of the fetal head were collected just prior to the start and at end of each full DTI acquisition (acquisition time: 30 s each). Total scan acquisition time was approximately 12 min.

MRI data were obtained using a SENSE acquisition on a Philips 1.5-Tesla Achieva scanner (Philips Healthcare, Best, Netherlands). Anatomical data included 3DSSFP sequences acquired using a respiratory navigator following intravenous injection of contrast agent. Patients were given 0.1 mmol/kg body weight of either gadopentetate dimeglumine (Magnevist, 41 Berlex Laboratories, Wayne, NJ, USA) or gadoterate meglumine (Dotarem, Guerbet, Villepinte, France). An acceleration factor of 2 was employed with a flip angle of 40° and a breath-hold time 20–30 s. Images had 1.2–1.7-mm isotropic voxel size.
Full-field aortic blood flow was acquired from a free-breathing, prospectively ECG-triggered PC-MRI sequence with velocity encoding of 120 cm/s. A spatial resolution of 2.0-mm isotropic voxels and a temporal resolution below 35 ms (corresponding to 24–30 phases) were employed (mean field of view 300 × 70 × 150 mm; TR = 3.8 ms; TE = 2.4 ms; flip angle 5°; acceleration kt + , 8; and bandwidth = 500 Hz). Respiratory gating for motion correction was applied and data were reconstructed using an in-house implemented kt-principal component analysis method [9 (link), 13 (link)]. Automatic eddy current correction was applied to all data.