To reduce acquisition time and improve image quality and robustness of ventilation assessment in a single breath-hold using H-guided reconstruction of fluorinated gas ( F) MRI.
Reconstructions constraining total variation in the image domain, L1 norm in the wavelet domain, and directional total variation between F and H images were compared in order to accelerate F ventilation imaging using retrospectively undersampled data from a healthy volunteer. Using the optimal constrained reconstruction in 8 patients with chronic obstructive pulmonary disease (16-seconds breath-hold), ventilation maps of various acceleration factors (2-fold to 13-fold) were compared with maps of the full data set using the Dice coefficient, difference in volume defect percentage and overlap percentage, as well as hyperpolarized Xe gas MRI.
The reconstruction constraining total variation and directional total variation simultaneously performed best in the healthy volunteer (RMS error = 0.07, structural similarity index = 0.77) for a measurement time of 2 seconds. Using the same reconstruction in the patients with chronic obstructive pulmonary disease, the Dice coefficient of defect volumes was 0.86 ± 0.05, the mean difference in volume defect percentage was -1.0 ± 1.7 percentage points, and the overlap percentage was 87% ± 2% for a measurement time of 6 seconds. Between volume defect percentage of F and Xe, a linear correlation (r = 0.75; P = .03) was found, with F volume defect percentage being significantly higher (mean difference = 11%; P = .04).
H-guided reconstruction of pulmonary F gas MRI enables reduction of acquisition time while maintaining image quality and robustness of functional parameters.

© 2020 The Authors. Magnetic Resonance in Medicine published by Wiley Periodicals, Inc. on behalf of International Society for Magnetic Resonance in Medicine.