Microstructural function in the lung : static and dynamic studies of ventilation and diffusion using hyperpolarized helium-3 MRI at two field strengths

Lung diseases such as chronic obstructive pulmonary disease and asthma affect mil- lions of people worldwide, and the World Health Organization predicts this trend will continue in upcoming decades. Many of these lung diseases are regionally heteroge- neous and may not be easily detected at early stages by pulmonary function tests. Lung imaging techniques are better suited for regional assessment of the extent and severity of disease. Most lung imaging modalities suffer from poor SNR, limited spatial resolution, and significant exposure to ionizing radiation. Hyperpolarized helium-3 MRI, how- ever, is a new technique which can detect a range of functional measures including regional ventilation and pulmonary microstructural changes. The technique is safe, is noninvasive and reproducible, and does not require ionizing radiation. This thesis used hyperpolarized helium-3 MRI to compare, for the first time, the sensitivities of the helium-3 apparent diffusion coefficient and T*2 in rat lungs at 0.5 T and 1.5 T. At 0.5 T, ADC measurements detected changes induced in the lung microstructure with p < 0.05 while T*2 was statistically unchanged. At 1.5 T, the reverse occurred: T*2 showed great differences with p < 0.05 due to induced microstructural changes, and differences in ADC were not statistically significant. Since most clinical MRI scanners operate at or above 1.5 T, T*2 appears to be the most appropriate parameter to characterize the lung microstructure in a clinical environment. T*2 maps could eventually be a viable tool for investigating lung diseases where there are changes in lung tissue composition, such as emphysema or idiopathic pulmonary fibrosis. This thesis also performed exploratory studies in dynamic imaging. The first ever dynamic T*2 measurement across the respiratory cycle was performed using spiral MRI, and the calculated T*2 was correlated with the lung inflation state. In addition, one of the first demonstrations of 3D dynamic imaging of the lungs was acquired using k-t BLAST undersampling, and resulting images spatially and temporally matched a fully-sampled reference scan. Since the pathophysiology of many lung diseases results in altered gas flow patterns, the ability to dynamically visualize pulmonary ventilation could be useful for understanding and clinically evaluating disease. Given that emphysema is characterized by abnormally large values of ADC and T*2, it was of paramount importance to avoid systematic error which could produce incorrect clinical conclusions. As a result, the pulse sequence parameters and imag- ing methodology in most of this thesis were based on the results of Monte Carlo simulations. To maximize the accuracy of ADC and T*2 estimation, these simulations showed that three or more echoes should always be used, and the measured signal should be non-linearly fit to the mean value of a signal-decaying Rician distribution.EThOS - Electronic Theses Online ServiceGBUnited Kingdo