J Magn Reson Imaging. 2026 May 23. doi: 10.1002/jmri.70350. Online ahead of print.
ABSTRACT
BACKGROUND: Motion can degrade image quality during Ultrashort Time-to-Echo (UTE) pulmonary MRI and is particularly prevalent in patients with lung disease. Comprehensive assessment of the impact of motion compensation techniques on image quality and clinical interpretation is needed.
PURPOSE/HYPOTHESIS: To compare the impact of retrospective motion compensation schemes on image quality and clinical interpretation of pulmonary UTE MRI in idiopathic pulmonary fibrosis (IPF).
STUDY TYPE: Prospective.
POPULATION: 21 (male = 18; mean age, 69.9 ± 8.1 years) participants with suspected IPF.
FIELD STRENGTH/SEQUENCE: 1.5 T/3 T, 3D center-out radial (gradient-echo) UTE sequence with 2× radial oversampling, while free-breathing.
ASSESSMENT: Images were reconstructed to 1.25 mm isotropic resolution using five retrospective schemes: no gating, hard-gating, soft-gating, motion-resolved (XD-GRASP), and an iterative approach (iMoCo). Signal-to-noise ratios (SNR) were estimated within the lung parenchyma, airways, aorta, muscles, and liver. Contrast-to-noise ratios (CNR) were estimated using the mean airway signal as reference. Image sharpness was estimated using the maximum derivative of a line profile across the diaphragm and a wavelet-based autofocus measure. Three radiologists evaluated image quality, motion artifacts on a 5-point Likert scale, and diagnostic classification of usual interstitial pneumonia (UIP).
STATISTICAL TESTS: The Kruskal-Wallis non-parametric test was used for qualitative reader scores and one-way ANOVA for the quantitative metrics, with p < 0.05 as the threshold for significance.
RESULTS: CNR was highest using the iMoCo reconstructions (lung parenchyma: 1.64 ± 1.41 vs. 0.88 ± 0.81 via XD-GRASP). Image sharpness was significantly improved using compressed sensing (CS)-based techniques (XD-GRASP and iMoCo), compared to the other methods, using both diaphragm profile (CS: 6.28 ± 3.70 vs. non-CS: 3.73 ± 2.06) and wavelet metrics (CS: 2.33 ± 0.42 vs. non-CS: 2.05 ± 0.35). CS methods also demonstrated greatest image quality based on reader scores.
CONCLUSION: Motion compensation using compressed sensing methods can improve image quality and clinical utility of UTE-MRI in the identification and diagnostic classification of typical parenchymal fibrotic patterns.
EVIDENCE LEVEL: Level 2-Prospective study, with a reference standard determined during the course of the study (CT imaging).
TECHNICAL EFFICACY: Stage 1.
PMID:42175722 | DOI:10.1002/jmri.70350
