Hyperpolarized 3He MRI provides high spatial and temporal resolution images of the lung air spaces in subjects with COPD. These images allow us to directly visualize the lung regions accessed by the hyperpolarized gas during a breath-hold, as well as the regions of the lung not accessed which appear as signal voids and are referred to as “ventilation defects.” The lung micro-structure can also be probed using diffusion- weighted imaging which takes advantage of the rapid 3He and 129Xe atom Brownian motion to generate surrogate measurements of alveolar size. Since the establishment of inhaled gas MRI for pulmonary imaging in 1994, there have been significant advancements in the type and scope of images that can be acquired, however, in order for broader translation of this imaging technology to occur, and for the potential translation of hyperpolarized gas MRI for clinical use, it must be demonstrated that these measurements relate to clinically meaningful outcomes. The overarching objective of this thesis was to generate hyperpolarized gas MRI measurements and measurement tools for the regional quantitative evaluation of hyperpolarized gas MRI with sufficient precision to evaluate COPD disease progression and treatment response. Another important step in the broader translation of this imaging technology is transitioning to
129
Xe MRI, and comparison of hyperpolarized 129Xe with 3He MRI is important in order to realize the potential of 129Xe MRI for respiratory research. The specific objectives and hypotheses tested in each chapter of this thesis are described below.
To better understand the potential for hyperpolarized 3He MRI to provide quantitative longitudinal COPD endpoints, the objective of Chapter 2 was to quantitatively evaluate a small group of COPD ex-smokers and healthy volunteers over 2 years using hyperpolarized 3He MRI. As described above, much of our current understanding of the natural history of COPD arises from the landmark study of Fletcher and colleagues56 who showed that lung function measured using FEV1 declines as we age, and that this decline
is accelerated in smokers with COPD; in ex-smokers, however, they showed that this decline in FEV1 returned to normal rates. Therefore, based on the Fletcher curve
the COPD ex-smokers would be similar to those observed in elderly healthy never- smokers.
This previous work evaluating longitudinal changes in 3He MRI measurements was performed using manual segmentation approaches, which clearly increases segmentation time and, importantly, introduces the potential for inter- and intra-observer variability. For serial evaluation of 3He MRI it is critical that the change measured over time represents physiological change that has occurred and is not due to measurement error. Therefore, the objective of Chapter 3 was to generate a semi-automated segmentation method for 3He MRI to enable its use in longitudinal and serial studies, and to compare the reproducibility and spatial agreement of the developed semi-automated segmentation algorithm to manual segmentation. We hypothesized that semi-automated measurements would not be statistically significantly different from manual measurements and have significantly reduced inter- and intra-observer variability.
Using the segmentation algorithm developed in Chapter 3, the objective of Chapter 4 was to evaluate a group of COPD subjects using 3He MRI prior to and immediately following bronchodilator therapy. A significant bronchodilator response is currently defined as an increase in post-bronchodilator spirometry, however, based on the findings in Chapter 2 we hypothesized that 3He MRI would provide the necessary sensitivity as well as precision to detect any potential regional functional lung changes after bronchodilator therapy in COPD subjects regardless of spirometry based responder classification.
In the same group of COPD ex-smokers evaluated using 3He MRI pre- and post- bronchodilator therapy in Chapter 4, the objective of Chapter 5 was to further evaluate the regional effects of bronchodilator administration in COPD using 3He MRI ADC measurements. Regional evaluation of tissue micro-structure using 3He MRI ADC would provide important insight into the lung alterations that accompany the improvements in regional 3He gas distribution after bronchodilator administration that were previously observed. In order to do this we first developed image registration/segmentation methods for quantifying ADC in the lung regions newly ventilated post-bronchodilator. We hypothesized that the regions of the lung that were newly ventilated following
bronchodilator therapy were not more emphysematous than the remaining lung tissue, and that the 3He ADC could measure significant reductions in regional gas trapping following bronchodilator therapy.
Despite the unique potential of 3He MRI and the previous work demonstrating its potential for use in evaluating COPD outcomes, the broader adoption of hyperpolarized gas MRI requires a transition from 3He gas, which has limited and unpredictable global quantities and high cost, to 129Xe gas, which is substantially more abundant in nature existing in measurable quantities in the atmosphere. The objective of Chapter 6 was to quantitatively compare hyperpolarized 3He and 129Xe MRI acquired within a few minutes in healthy volunteers and subjects with COPD, and to evaluate the correlations between
3
He and 129Xe MRI measurements with standard measurements of pulmonary function. We hypothesized that the different properties of 129Xe gas would result in significant differences in 129Xe compared to 3He gas distribution measurements in COPD but not in healthy volunteers.
To better understand the morphological determinants for the ventilation differences observed between hyperpolarized 3He and 129Xe MRI in COPD in Chapter 6, the objective of Chapter 7 was to evaluate the same group of COPD subjects using the image registration/segmentation methods for quantifying ADC on a regional basis developed in Chapter 5. We hypothesized that emphysematous regions in the lung would more readily fill with 3He as compared to 129Xe gas, leading to decreased 129Xe MRI ventilation.
While it is important to demonstrate that 3He MRI measurements relate to important clinical outcomes, such as disease progression and treatment response, another goal of imaging is to identify early disease changes. The objective of Chapter 8 was to evaluate and compare well-established clinical, physiological and emerging imaging measurements in ex-smokers with normal spirometry and abnormal DLCO with a group of
ex-smokers with normal spirometry and DLCO and ex-smokers with GOLD stage I
COPD. We hypothesized that ex-smokers with normal spirometry but abnormal DLCO
would have significantly worse symptoms, exercise capacity and 3He MRI ADC than ex- smokers with normal DLCO.
The objective of Appendix A was to establish imaging measurement reproducibility in adult cystic fibrosis subjects over a short period of time (7 ± 2 days) in the absence of therapeutic intervention using hyperpolarized 3He MRI. We hypothesized that 3He MRI would provide the necessary and sufficient spatial and temporal sensitivity to detect day- to-day changes in lung function.
A case report of a COPD ex-smoker that was evaluated serially over 4 years using hyperpolarized 3He MRI, twice prior to and twice following an acute exacerbation requiring hospitalization is provided in Appendix B. Based on our previous studies demonstrating the high sensitivity of 3He MRI for measuring improvements in the absence of FEV1 improvements, we hypothesized that following hospitalization and
treatment, visibly obvious improvements in 3He gas distribution would be observed prior to improvements in FEV1.
Finally, in the last chapter of this thesis I will provide an overview and summary of the important findings and conclusions of Chapters 2-8. I will also address the study specific limitations as well as general limitations of the hyperpolarized gas MRI studies presented, and provide some potential solutions. Finally, based on the findings and limitations discussed, I will outline a roadmap for future hyperpolarized gas MRI studies.