Browsing by Author "Bier, Elianna Ada"

129Xe MRI has emerged as a promising means of imaging ventilation distribution. Recently, interest has turned from only imaging ventilation to exploring solubility and chemical shifts to characterize gas exchange, including images of gas exchange patterns that are sensitive to a wide array of disease. The next challenge is to differentiate the underlying causes of gas exchange impairment to begin enabling not just detection, but diagnosis. This prompted us to investigate the relatively unexplored dynamics of 129Xe spectroscopy.

Hyperpolarized 129Xe is a powerful noninvasive tool for studying the physiological environment of the lung due to its distinct chemical shifts as xenon diffuses from the airspaces, through the alveolar membrane, and into red blood cells (RBCs). Previous work has shown that the spectral characteristics (amplitude, chemical shift, line width, and phase) of each resonance are sensitive to disease. In principle, the dynamic variations in four spectral parameters of each 129Xe resonance detected over inhalation, breath hold, and cardiac cycle can contain vital information on the origins of reduced lung function. In this work, our objective was to a) develop a framework for evaluating dynamics, b) assess temporal differences in spectroscopic parameters between patients with idiopathic pulmonary fibrosis (IPF) and healthy volunteers, c) understand the interplay between lineshape and dynamics, and d) create metrics to test the sensitivity of dynamic spectroscopy to underlying lineshapes and signal to noise.

The developed analysis methods were used to characterize the spectral dynamics of 14 healthy volunteers and 10 subjects with IPF. In all subjects, the breathing maneuver (inhale, breath hold, exhale) was reflected in the chemical shift and line width of the 129Xe RBC, barrier, and gas resonances. When comparing a representative IPF and healthy subject, we found that the subject with IPF had an RBC chemical shift that was initially 2 ppm lower and decreased by 0.5 ppm throughout a 5 s breath hold. Over the breath hold, we observed oscillations at the cardiac pulsation frequency in all of the spectral parameters of the RBC resonance. A mean peak-to-peak change of 7.9% ± 2.8% in amplitude of the RBC resonance was consistent across both cohorts. However, the oscillations in the chemical shift and phase of subjects with IPF were three times higher than in healthy volunteers. In light of the known relationship between RBC chemical shift and blood oxygenation levels, these findings suggest that we are sensitive to pathologically driven changes in blood oxygenation on the timescale of the cardiac cycle.

To analyze the effects of our spectral analysis in comparison to a known spectral model, we developed a digital dynamic spectroscopy phantom that simulated clinical free induction decays (FIDs) similar to what we collect during a dynamic spectroscopy acquisition based on specified spectral parameters. This phantom was used to validate our analysis techniques and explore the dependence of RBC signal oscillations on spectral structure. Additionally, motivated by a recent report that dissolved-phase xenon spectra is best fit to three dissolved-phase peaks, we performed a lineshape analysis on stationary spectra from a data set with higher spectral resolution. It was concluded that fitting to non-Lorentzian line shapes reduced the residual error of the fit and that the barrier resonance can potentially be represented by a linear combination of two Lorentzians where the spectral characteristics of the second peak are completely dependent on the first.

129Xe dynamic spectroscopy can be used to evaluate the temporal variations in gas exchange, although additional work must be done to properly characterize the spectral structure. The derived dynamic metrics have the potential to be useful biomarkers for disease progression and help discriminate between different pathologies that affect gas exchange.

Chronic respiratory diseases are one of the leading causes of death in the US and a driving factor in their mortality rate is the presence of comorbid cardiovascular diseases such as pulmonary hypertension (PH). As an increasing number of patients exhibit concomitant cardiac and pulmonary disease it becomes progressively more difficult to determine disease etiology and thus the optimal treatment course. The current standard diagnostic methods are insensitive to the underlying cause of gas exchange impairment, are unable to differentiate between phenotypes, and have limited utility in assessing disease progression or therapy response. The primary diagnostic tools for assessing pulmonary function are collectively referred to as pulmonary function tests (PFTs). While these tests are simple and non-invasive, they are also a global measurement that is effort-dependent and has poor reproducibility. Furthermore, PFTs cannot separate the contribution of concomitant disease on their measurements. The diagnosis of PH and subsequent determination of World Health Organization (WHO) classification requires invasive right heart catheterization (RHC) to meet strict hemodynamic cutoffs. However, the RHC interpretation can be challenging in patients with complex disease because the effect of comorbidities on RHC measurements is unknown. Therefore, new non-invasive diagnostic tools must be developed that can assess gas exchange impairment and pulmonary hemodynamics in tandem for patients to receive optimal treatment.

Hyperpolarized 129Xe MR imaging (MRI) and spectroscopy (MRS) have emerged as a powerful tool for assessing the pulmonary environment due xenon exhibiting distinct chemical shifts as it diffuses from the airspaces, through the alveolar membrane, and interacts with red blood cells (RBCs). This unique property allows the 129Xe signal to be decomposed in order to separately measure or image the three gas exchange compartments (gas, barrier, and RBC). 129Xe gas exchange imaging is beginning to show exquisite sensitivity to a range of obstructive and restrictive diseases. Still, despite this sensitivity to disease burden, 129Xe imaging techniques are unable to probe pulmonary hemodynamics. Thus, it does not provide sensitivity to PH, one of the possible causes of dyspnea. Previous work has demonstrated that in 129Xe MRS the characteristics of the spectral peaks can detect diffusion impairments present in interstitial lung disease (ILD). Yet current 129Xe MRS techniques only investigate static measurements of an inherently dynamic process. It is possible to extend 129Xe MRS and collect spectra as a time-series in dynamic spectroscopy to assess the cardiogenic changes in spectral parameters that may be associated with the hemodynamic changes in PH.

The objective of this work is to establish methods using 129Xe MRI/MRS to differentiate between diverse cardiopulmonary diseases. To this end, we develop the technique of 129Xe dynamic spectroscopy and assesses its utility in differentiating pre-capillary and post-capillary PH. We also investigate quality assurance metrics and tools including the repeatability of spectroscopic measurements and a thermally polarized xenon phantom to help facilitate the transition of 129Xe MRI/MRS into a clinical tool.

The foundation of 129Xe dynamic spectroscopy is the decomposition of each static spectrum into its three separate components. This is achieved by fitting each spectrum in the time-series to a mathematical model that describes the shape of each peak. In Chapter 3, to characterize the spectral parameters more accurately, we analyzed 6 different mathematical models for 129Xe dissolved-phase MR spectroscopy. We demonstrate that the optimized spectroscopic fitting model is a barrier Voigt model where the RBC peak has a Lorentzian lineshape and the barrier peak is a Voigt profile. This model was used in dynamic spectroscopy to extract the area, chemical shift, linewidth, and phase of each peak.

In principle, the dynamic variations in the spectral parameters of each 129Xe resonance detected during the cardiac cycle can contain vital information on pulmonary hemodynamics. Thus, in Chapter 4 we developed techniques to quantify and assess the temporal changes in the spectroscopic parameters during inhale, breath-hold, and exhalation. We observed a distinct cardiogenic oscillation in the amplitude and chemical shift of the RBC peak. This oscillation was quantified by its peak-to-peak height. Furthermore, we identified static and spectral parameters that are statistically different between healthy volunteers and subjects with idiopathic pulmonary fibrosis (IPF). This study demonstrated that that 129Xe dynamic spectroscopy is sensitive to disease.

The initial characterization of a diverse array of diseases is essential to understand the relationship between 129Xe spectroscopy and the cardiopulmonary environment. Thus, Chapter 5 characterizes 129Xe MRI/MRS in healthy volunteers and subjects with chronic obstructive pulmonary disease (COPD), IPF, left heart failure (LHF), and pulmonary arterial hypertension (PAH). The chosen cohorts provide two forms of chronic lung disease (IPF, COPD) and two forms of PH (LHF, PAH) that have different impedance locations with respect to the pulmonary capillary bed. LHF is a form of post-capillary PH because the impedance to blood flow is downstream of the pulmonary capillary bed as left ventricular dysfunction leads to a sustained increase in left atrial pressure. On the other hand, PAH is a form of pre-capillary PH caused by occlusions upstream of the capillary bed. We found that while gas exchange imaging is essential in the discrimination of obstructive and interstitial disease, only the height of oscillations in the RBC amplitude was able to differentiate between the different types of PH.

To test the utility of 129Xe MRI/MRS in differentiating PH status, we designed a diagnostic algorithm in Chapter 6 to distinguish between pre-capillary PH, post-capillary PH, no PH, and interstitial lung disease (ILD). Algorithm performance was tested in a single-blind reader study in which three expert readers used 129Xe MRI/MRS to determine the PH status of 32 test subjects. The algorithm performed well on straightforward cases of PH. For subjects with concomitant disease, the combination of MRI/MRS provided additional insight to the complex pathophysiology that cannot be quantified by hemodynamic measurements alone. This demonstrated that 129Xe dynamic MRS and gas exchange MRI can be used in tandem to uniquely provide non-invasive assessment of both hemodynamics and gas-exchange impairment to aid in the differentiation and detection of PH.

For 129Xe MRI/MRS to be adopted into a clinical setting it is essential to understand the underlying measurement variability. Chapter 7 presents an assessment of the repeatability of the dynamic spectroscopy sequence and quantification methods by acquiring two dynamic spectroscopy acquisitions during a single MR study. We also use these paired scans to develop quantitative criteria to assess the scan quality for inclusion in dynamic analysis. Additionally, as 129Xe MRI/MRS is more broadly implemented it is imperative to have standards for day-to-day validation and for comparing performance at different 129Xe imaging centers. Therefore, Chapter 8 present our development of a thermally polarized xenon phantom assembly and associated imaging protocol to enable rapid quality‐assurance (QA) imaging.

The work in this thesis develops a robust 129Xe dynamic spectroscopy protocol for evaluating the temporal dynamics in the RBC resonance. In particular, the height of RBC amplitude oscillations is found to be sensitive to PH and can be used to differentiate between pre- and post-capillary forms. 129Xe dynamic spectroscopy and 129Xe gas exchange MRI can differentiate between diverse cardiopulmonary diseases and together provide a complete evaluation of pulmonary hemodynamics and gas exchange impairments. This research lays the groundwork for the use of 129Xe MRI/MRS in clinical practice to diagnose and monitor PH and transforms 129Xe MRI/MRS into a more comprehensive tool for investigating the pathogenesis of unexplained dyspnea.