UCSF

UTE proton MRI is getting more attention in structural and functional lung imaging. It improves signal-to-noise ratio (SNR) from low T2* tissues such as the lung parenchyma, and is motion robust compared to traditional Cartesian acquisitions. For radial UTE scans, the repeated acquisition of the center of k-space can serve as a self-navigator respiratory motion during free-breathing, enabling the motion-resolved 3D reconstruction at multiple respiratory states. The volume change can then be assessed by the lung tissue deformation and intensity variation across the different respiratory phases, which is a straightforward biomarker for ventilation quantification. Compared to existing ventilation quantification methods such as pulmonary function testing (PFT), computed tomography (CT), and hyperpolarized noble gas MRI, 1H MRI offers local functionality via ventilation analysis, is non-ionizing, and does not require special equipment.

This dissertation aims to exploit the simultaneous structural and functional aspects of UTE 1H MRI. The first project explored ventilation quantification with free-breathing motion-resolved 3D UTE 1H lung MRI through the tissue deformation-based Jacobian determinant method. In the second project, we advanced on the motion-compensated low-rank constrained reconstruction, which jointly improves the motion field estimation and the reconstruction to get high-quality ventilation maps and structural images. Third, we compared the 3D UTE 1H ventilation calculated from motion-compensated low-rank constrained reconstruction (MoCoLoR) with the HP 129Xe ventilation for validation.

In addition, we investigated the feasibility of using a convolutional neural network for motion-compensated proton lung MRI reconstruction. It substantially accelerates the reconstruction for 3D radial UTE data, shortening the required reconstruction time from hours to minutes. It showed the potential to shorten the scan time, thus facilitating the clinical application of proton pulmonary UTE MRI. Moreover, we evaluated the imaging quality of UTE lung MRI in the pediatric population through a reader study. Lastly, we pushed the structural-functional proton 3D UTE lung MRI to clinical applications and reported results for pediatric patients with pectus deformity.