Reproducibility of functional lung parameters derived from free-breathing non-contrast-enhanced 2D ultrashort echo-time

BACKGROUND: Imaging the lung parenchyma with magnetic resonance imaging (MRI) is challenging due to cardiac and respiratory motion, the low proton density and short T(2)* relaxation time, and therefore not well established in the clinical routine. As a further step in facilitating lung MRI for longitudinal monitoring, this study aimed to assess the reproducibility of 2D ultrashort echo time (UTE)-derived lung function parameters in healthy subjects. METHODS: In this study, a 2D UTE technique was combined with tiny golden angle (tyGA) ordering. Data were acquired either during breath-holds (BH) or continuously during free-breathing (FB) at a field strength of 3T. Retrospective self-gating (image- and k-space-based) was used to reconstruct respiratory and cardiac multistage images from the FB acquisitions. The reproducibility of functional lung parameters derived from BH and FB acquisitions was assessed for three independent examinations (M1–3). M1 and M2 were acquired within 2 h, whereas M3 was acquired at least 14 d after M1/2. Different respiratory and cardiac phases were reconstructed for three coronal slices. Quantitative analysis including proton fraction (f(P)), apparent signal-to-noise ratio (apparent SNR), fractional ventilation (FV), and perfusion (f) was performed by two independent observers, and inter-measurement and inter-observer repeatability were assessed. RESULTS: All scans could be performed successfully in all volunteers. Intraclass correlation coefficients (ICC) of inter-measurement and inter-observer variability, and Bland-Altman analysis showed good to very good reproducibility. Larger breathing amplitudes were observed in the BH acquisitions, which also showed lower reproducibility when compared with the FB acquisitions. For the FB approach, the ICC ranged between 0.70 and 0.98 for all measurements, and ranged between 0.86 and 0.97 for the two observers. No bias or significant differences were observed between the three measurements or the two observers in healthy volunteers. CONCLUSIONS: The study proves the feasibility of FB 2D tyGA UTE for lung imaging. Functional parameters derived from FB acquisitions are reproducible in healthy volunteers, allowing for further investigation of this technique in patients with various underlying diseases.