Abstract
Magnetic resonance imaging (MRI) of the lung is challenging due to the low proton density of lung parenchymal tissue relative to that of other soft tissue structures within the body, the high concentration of air-tissue interfaces and associated short transverse relaxation time following radiofrequency (RF) excitation within lung parenchyma, and the need to avoid or mitigate motion-related effects associated with breathing. This thesis aims to produce advancements to a number of MRI-based approaches for assessing lung structure and function.
The first work shown in this thesis demonstrates a method for performing 3D multi-phase MRI of grid-tagged hyperpolarized 3He gas in the lungs during exhalation. This technique is promising for direct measurement of volume change during the breathing cycle on both a global and regional basis, for quantification of lung biomechanical quantities related to pulmonary compliance, such as regional strain values, and for visualization and assessment of global and regional biomechanical abnormalities. The approach described herein takes advantage of the predictable distribution of k-space energy imposed by using an RF pulse train to apply a tag pattern to inhaled hyperpolarized 3He. Dynamic images of tagged 3He were collected during exhalation using this technique, and multiple-time-point displacement and strain maps and lobar strain profiles were calculated from tagged 3He images collected as described.
The second work shown in this thesis seeks to improve dissolved-phase 129Xe MRI of the lung, a technique for visualization and quantification of pulmonary gas exchange efficacy, by characterizing dependence of quantitative gas-exchange metrics derived from dissolved-phase 129Xe MRI on lung volume during imaging in healthy individuals and in individuals with chronic obstructive pulmonary disease (COPD). Linear relative-difference relationships between gas-exchange metrics and lung volume were observed and characterized in healthy and diseased subject samples. Significant differences in some of these metrics between healthy individuals and individuals with COPD were observed, but were largely eliminated upon correcting for lung-volume contribution by projecting signal ratios to expected values at reference volumes specific to each target inflation level. This result demonstrates the need for careful consideration of volume-related effects when comparing results in individuals with COPD, who frequently present with chronic lung hyperinflation, with those in healthy individuals.
The third work shown in this thesis seeks to demonstrate an optimized proton lung MRI approach, in which an ultrashort echo time, balanced steady-state free precession pulse sequence designed to maximize sampled signal in lung tissue is combined with temporally-constrained compressed sensing reconstruction, permitting generation of motion-resolved, high-resolution, high-quality 3D image frames at a range of positions in the breathing cycle. This approach produces high-signal, high-resolution lung images at end-of-exhalation collected during free breathing, with work ongoing to improve image quality in non-end-of-exhalation frames by reducing blurring of moving features at these respiratory phases.
