LAUSR

The amount of arterial blood flow to a tissue, defined as perfusion, is closely related to the supply of oxygen and nutrients to an organ and consequently to its function. Perfusion can be altered in the case of pathology. For the lung in particular, perfusion is compromised in diseases like pulmonary hypertension, chronic obstructive pulmonary disease (COPD), or cystic fibrosis (CF). Therefore, assessing perfusion as a functional parameter is important in the diagnosis of pulmonary diseases. Currently, scintigraphy and single-photon emission computed tomography (SPECT) in conjunction with the administration of Technetium-labeled albumin are the standard for assessing lung perfusion in patients. Since these nuclear imaging techniques are limited by their poor spatial resolution and the use of radioactivity, magnetic resonance imaging (MRI) approaches have also been developed. Threedimensional perfusion assessment by dynamic contrastenhanced (DCE) MRI, first proposed in the early 2000s, is nowadays available on most clinical scanners. DCE-MRI is based on detecting signal changes in the parenchyma during the first passage of a gadolinium-containing contrast material bolus through the lung capillaries. Although gadolinium contrast agents are usually well tolerated by most individuals, the risk of causing acute allergic reactions or severe complications such as nephrogenic systemic fibrosis in patients with impaired renal function as well as the potential for gadolinium accumulation in various tissues may restrict the generalized use of DCE-MRI. Alternatives not relying on the administration of contrast agents are being explored as well. Arterial spin labeling (ASL) aims at labeling water protons in blood with dedicated inversion pulses and using them as an endogenous tracer. A control and a tag image are acquired and then subtracted in order to suppress the static background. The heterogeneity of lung perfusion has been quantified by ASL. However, its use in clinics is challenging, because of the generally low signal and the necessity to subtract two images, making the technique very sensitive to respiratory and cardiac motion. Fourier decomposition (FD), introduced by Bauman et al, is a 2D approach allowing the regional assessment of perfusion and ventilation during free breathing without the need for inhalation of fluorinated or hyperpolarized He or Xe gases. Variations of the parenchymal MRI signal during the respiratory and cardiac cycles are mapped by ultrafast imaging. At inspiration, the lung volume increases and the parenchymal signal decreases, while the opposite happens during expiration. The signal intensity can thus be scaled to the tidal volume breathing region. On the other hand, the parenchymal signal is also modulated during the cardiac cycle, as unsaturated blood entering the image slice increases the signal. As both cycles occur at different frequencies, they can be spectrally separated by, eg, Fourier decomposition of the time-resolved data. However, prior to this regional spectral analysis the acquired datasets need to be subjected to a nonrigid image transformation to compensate for the respiratory motion. Ventilation, perfusion, and blood inflow maps are then generated from the amplitudes and phases of the respiratory and cardiac signal modulations. Extensive validation of FD in animal studies as well as comparisons to SPECT, DCE-MRI and He MRI in healthy volunteers opened the avenue for the first successful applications of the technique to patients. Two approaches were initially adopted to achieve the necessary temporal resolution in order to resolve the cardiac phases to depict lung perfusion: ultrafast accelerated parallel imaging and self-gating. While ultrafast imaging is limited by the signal-to-noise ratio and spatial resolution, self-gating requires long acquisition times. Moreover, both approaches need advanced sequence design. To address these challenges, Voskrebenzev et al recently introduced a postprocessing method for the reconstruction of the complete respiratory and cardiac cycles, enabling phase-resolved functional lung (PREFUL) imaging from a conventional spoiled gradientecho sequence with a temporal resolution comparable to the