|
|
|
“Fast Flexible Dynamic Three-Dimensional Magnetic Resonance Imaging”
by Yinghua Zhu
May 2016
Magnetic resonance imaging (MRI) is a non-invasive 3D imaging modality capable providing unique contrast for anatomical and functional assessment. MRI is slow (compared to other modalities, such as ultrasound and computed tomography) because of physical and physiological limitations. Therefore, critical trade-offs have to be made between the spatial coverage, spatial resolution and temporal resolution. In conventional 2D time-resolved (dynamic) examinations, one parameter is usually sacrificed for maintaining the other two. Such trade-off would be even more critical in dynamic 3D imaging. For a long time, MRI speed prevented the application of MRI to dynamic 3D imaging where the motion of tissues (muscles, organs) or the movement of some contrast media is diagnostically important. The desired MRI signals reside in a 4D space (3D object + time), and the contemporary scanners are able to capture a very small fraction of the data, resulting in temporal blurs if spatial coverage and resolution are prioritized, or image artifacts if insufficient data in one time frame. Speeding up the MRI acquisition therefore has long been an important area of research, as it enabled many new applications and more flexible use of the MRI scanner. This dissertation work aims to dynamic 3D MRI specifically for speech imaging and dynamic contrast-enhanced imaging. At first, an engineering method is proposed to combine dynamic 2D data to form dynamic 3D speech visualization, using audio-based alignment. Later on, the dissertation work proposes more advanced techniques that provide 3D dynamics in a â€fast†manner such that the data acquisition sequence can be generated rapidly, and useful 3D information can be such that scanning parameters can be freely chosen, and the temporal resolution can be adjusted in retrospect. The acceleration is achieved by using efficient data acquisition and a combined parallel imaging and constrained reconstruction method; the flexibility is obtained by applying a golden angle data acquisition scheme, and variable-density sampling in the data space. The imaging speed and flexibility have been shown to be useful for investigations of human speech production and investigations of contrast agent kinetics in brain tumors.