(Wednesday, 8/1/2018) 7:30 AM - 9:30 AM
Room: Room 202
Purpose: The goal of this study is to modify T2SE multi-slice acquisition and reconstruction to permit sub-mm through-plane spatial resolution. Current slice thicknesses in multislice MRI are typically no smaller than 1.7 mm.
Methods: The basis of the new method is to perform the image reconstruction in kz-space. Acquisition is done using slices with modest slice-to-slice overlap. The acquired data are next transformed along the slice direction (Z) into kz-space. In kz-space we then precisely account for the slice profile created by the RF excitation and for the kz sampling effects caused by the slice-to-slice spacing. We have tested the new method on 2D multi-slice images of a resolution phantom, an SNR phantom, and in vivo prostate MRI. Axial images were acquired, and performance was assessed in reconstructed sagittal reformats. Typical acquisition parameters were 3.2 mm slice thickness and 0.8 mm (75% overlap) and 1.6 mm (50%) slice spacing. In the reconstruction process, to compensate for the slice profile, inversion problems due to zero crossing were addressed by Tikhonov regularization. 4x zero padding was typically used.
Results: Reconstructions for 0.8 and 1.6 mm slice spacings show improved resolution vs. the case of abutting slices (spacing = slice thickness = 3.2 mm). Similar results were observed in the prostate phantom and in vivo prostate images. Adjusting the regularization parameter provides improved modulation with some cost of increased noise. Use of 0.8 mm spacing provided clear visualization of the 0.5 lp/mm resolution pattern, confirming the desired performance. Sagittal and coronal reformats are competitive with direct acquisition in those orientations.
Conclusion: We have demonstrated that the newly proposed kz-space-based technique provides finer through-plane resolution than the slice thickness and is superior to simple zero padding. The method is particularly applicable to multislice scans in which averaging is used.
Keywords
Taxonomy
IM- MRI : Image Reconstruction
Contact Email
