Towards Comprehensive Mapping of Lung Function Using Hyperpolarized 129Xe MRI: Developments and Applications

Abstract

Hyperpolarized 129Xe magnetic resonance imaging (MRI) and spectroscopy (MRS) have emerged as a set of effective tools for evaluating pulmonary function, offering unique insights into ventilation, gas exchange, and cardiopulmonary dynamics without exposing patients to ionizing radiation. This thesis aims to advance the field of hyperpolarized 129Xe MRI by developing and optimizing novel techniques that enable comprehensive mapping of lung function, with a particular focus on quantitative assessment of ventilation, gas transfer, and microvascular blood flow.One of the primary challenges in hyperpolarized 129Xe ventilation MRI is the presence of bias field inhomogeneity, which can lead to inaccurate quantification and interpretation of the imaging data. Chapter 3 addresses this issue by introducing two innovative approaches for bias field correction. The first method employs a direct mapping technique that utilizes the RF-depolarization of the 129Xe signal to estimate and correct for the bias field. The second approach involves the construction of a template that models the RF coil sensitivity distribution, allowing for the correction of bias field inhomogeneity across different imaging protocols. These novel methods are compared to the widely used N4ITK algorithm, and the results demonstrate their superior ability to remove RF inhomogeneity artifacts while preserving physiologically relevant ventilation gradients, thereby improving the accuracy and reliability of quantitative ventilation analysis. Chapter 4 focuses on the quantitative mapping of cardiopulmonary oscillations using hyperpolarized 129Xe gas exchange MRI. These oscillations, which arise from the variation in pulmonary capillary blood volume during the cardiac cycle, have been previously observed in whole-lung spectroscopic studies. However, by leveraging the oscillations embedded in the gas exchange imaging data, it is possible to spatially map these fluctuations, providing regional insights into pulmonary hemodynamics. To optimize such an imaging technique, we employ digital phantom simulations to fine-tune keyhole reconstruction methods and establish robust healthy reference values. The refined methods are then applied to patients with chronic thromboembolic pulmonary hypertension (CTEPH) before and after pulmonary thromboendarterectomy (PTE), a surgical intervention aimed at restoring pulmonary blood flow. The results showcase the potential of this approach to assess regional changes in microvascular flow and provide valuable insights into the complex pathophysiology of pulmonary hypertension. Chapter 5 further advances cardiopulmonary oscillation amplitude mapping by investigating compressed sensing reconstruction techniques. Compressed sensing is an approach that enables the recovery of high-quality images from undersampled data by exploiting the inherent sparsity of the image. This novel reconstruction technique is evaluated using digital phantoms and in vivo data from healthy subjects and patients with pulmonary arterial hypertension. The results demonstrate the improved ability of compressed sensing to mitigate the effects of undersampling and enhance the diagnostic value of oscillation mapping, paving the way for more robust and accurate assessment of pulmonary hemodynamics. In addition to these core advancements, Chapter 6 presents a set of supplemental innovations designed to facilitate the standardization and dissemination of 129Xe MRI methods across research centers. These include the development of open-source software tools for streamlined 129Xe MRI analysis, the optimization of RF-pulse shapes to minimize off-resonance artifacts, and the application of deep learning techniques for image enhancement. By providing a comprehensive toolset and promoting collaboration, these initiatives aim to accelerate the translation of 129Xe MRI into clinical practice. In conclusion, this thesis makes significant contributions to the field of hyperpolarized 129Xe MRI, enabling comprehensive and robust mapping of lung function and gas exchange dynamics. The developed core techniques, spanning bias field correction, quantitative oscillation mapping, and compressed sensing reconstruction represent a powerful set of tools for non-invasive assessment of pulmonary function. Given these developments, there now exists the ability to accurately quantify ventilation, gas transfer, and microvascular blood flow in a single imaging session, without the need for ionizing radiation or invasive procedures. By addressing key technical challenges and optimizing imaging protocols, this work positions hyperpolarized 129Xe MRI as a comprehensive tool for understanding the complex causes of dyspnea whether of obstructive, interstitial, or vascular origins.