A Pipeline for Zoomed Fetal MRI

Abstract

Fetal Magnetic Resonance Imaging (MRI) is heavily constrained by unpredictable and substantial fetal motion that causes image artifacts and limits the set of viable diagnostic image contrasts. Fast, single-shot MRI of the fetal brain, such as HASTE, is a commonly applied acquisition method to mitigate artifacts from fetal motion by imaging over a sub-second timeframe per slice. For this application, although the fetal brain of interest is only a small fraction of the total field of view across the gravid abdomen, conventional slice-selective excitation followed by Cartesian sampling carries significant overhead in encoding time, which could be reduced with restricted slice RF excitation.

In this thesis, we propose a pipeline to exploit novel zoomed or restricted slice RF excitation for imaging for fetal MRI with benefits that include the mitigation of motion artifacts, reduced RF power deposition, and shorter overall scan time. The three dominant contributions of this thesis are 1) a model for fetal pose estimation which benefits the tracking of ROI; 2) selective excitation design which aims at exciting only the ROI for zoomed imaging and 3) a model for post-processing to improve the SNR; which taken together open new possibilities for improved MRI in pregnancy.