MAGNETIC RESONANCE IMAGING OF THE LUNG AT 3 AND 1.5 TESLA

Abstract

Routine lung imaging is most often done using nuclear medicine techniques, and more recently using hyperpolarized gas MR. The former suffers from radiation and poor spatial resolution, while the latter requires expensive hardware and costly 3He. With increasing numbers of clinical MRI magnets >1.5T this thesis presents data investigating whether the advantages of higher magnetic field could be applied in lung imaging. Two different approaches to lung perfusion were examined: a non contrast free breathing technique with respiratory and cardiac gating was compared to breath held Gd-enhanced lung perfusion MR imaging. The two techniques were evaluated at two field strengths 3 Tesla and 1.5 Tesla. Healthy volunteers were scanned using both a 3 Tesla and a 1.5 Tesla MRI system each with 8 parallel receivers, using a cardiac gated Fast Spin Echo pulse sequence. Acquisition was cardiac triggered, with different time delays incremented to cover the entire cardiac cycle. To reduce motion artifacts acquired k-space data was recon structed using minimal variance algorithm according to physiological data recorded from respiratory bellows and ECG leads. Contrast injected (Gd-DTPA-BMA) perfusion measurements were performed us ing both SPGR and EC-TRICKS pulse sequences. Images were acquired in one breath hold of 30 seconds. Non-contrast ECG gated FSE perfusion images were assessed by measuring percent signal change between images acquired in the systolic and dias tolic phases of the cardiac cycle. Gd-based perfusion was done through measurement of time to peak of the bolus arrival and signal enhancement integral. Comparable absolute signal magnitude changes were observed through the entire lung from both methods, although the methods differ temporally. Despite worsening susceptibility at higher field, a 3T MR scanner can be used for evaluation of lung perfusion. We suggest increased SNR at higher field allows non-contrast based MR perfusion imaging comparable to Gd-based bolus methods. Thus it is possible to perform perfusion imaging in clinical populations, where the use of breath holds is often intolerable. The nature of the FSE-based signal change is likely due to difference in blood flow between systolic and diastolic phases. The contrast based methods offer a significant increase in signal but are compromised by cardiac motion produced artifacts. Although the longitudinal relaxivity of Gd decreases with increasing field (from 4.06±0.3 at 1.5 Tesla to 3.88±0.16 at 3 Tesla) the higher polarization of spins at higher fields allows the use of half dose to produce MRA images that are of comparable quality to full dose at 1.5T.