The purpose of this thesis was to detect fetal cardiovascular magnetic resonance imaging motion through both 2D and 3D CINE dataset reconstructions. An evaluation of reconstruction methods that compensate for fetal motion was also performed. Motion artifacts are noted to be one of the major problems in imaging the fetus. They occur mainly due to fetal heart contractions, involuntary fetal movement as well as maternal breathing and motion. In the present study, radial k-space acquisition was used. The radial sampling technique along with breath-hold acquisition led to the development of 2D datasets with minimal artifacts. In addition, 3D-datasets that were obtained under free breathing conditions were analyzed. Further, image processing techniques such as non-uniform fast Fourier Transform, Compressed Sensing, Coil and Density Compensation were developed. These techniques were developed to improve image quality. An additional component of the thesis was focused on the extraction of a self-gating cardiac signals from the raw k-space data using Principal Component Analysis and filtering techniques. Objectives included assessing self-gating's practicality as a compliment to Doppler ultrasound. Self-gated signals were comparable to the ultrasound derived peaks. In summary, this research shows the practicality and constraints concerning enhanced radial acquisition methods and self-gating fetal cardiac imaging techniques.
Ειρήνη Γ. Ξανθοπούλου (Wed,) studied this question.