Preview

PDF, English Download (19MB) | Terms of use

Abstract

Fourier Decomposition (FD) magnetic resonance imaging is a non-invasive method for assessing ventilation and perfusion in the lungs. However, the technique struggles with a low signal to noise ratio (SNR). This work analyzed and optimized the standard balanced steady-state free precession (bSSFP) sequence used in the FD framework. The biggest drawbacks of the standard sequence were the long echo time compared to the short T2* of the lung parenchyma and the restriction of the utilized flip angles, due to specific absorption rate (SAR) limitations. To achieve the necessary improvements, advanced techniques were used, specifically, (utra-)fast bSSFP, two speed variablerate selective excitation (VERSE) pulses and variable flip angle (VFA) patterns. All steps of the optimization process were tested in both simulations and phantom measurements. The finished sequence was tested in two human studies, one featuring healthy volunteers and the other featuring patients with different types of lung cancer. In both studies the average SNR of the morphological images was increased by 47%, while the SNR of the functional images was increased by 53%. Furthermore, due to the higher SNR of the morphological images, both the effective resolution and the robustness of the functional images were increased. None of the employed techniques introduced any transient artifacts and possible blurring was minimized. Due to these improvements, this work reinforces both the position of FD MRI as a research tool and brings the technique closer towards a clinical application.