"Structure-Function Imaging of Lung Disease Using Ultrashort Echo Time MRI"

Abstract

Rationale and objectives: The purpose of this review is to acquaint the reader with recent advances in ultrashort echo time (UTE) magnetic resonance imaging (MRI) of the lung and its implications for pulmonary MRI when used in conjunction with functional MRI technique.

Materials and methods: We provide an overview of recent technical advances of UTE and explore the advantages of combined structure-function pulmonary imaging in the context of restrictive and obstructive pulmonary diseases such as idiopathic pulmonary fibrosis (IPF) and cystic fibrosis (CF).

Results: UTE MRI clearly shows the lung parenchymal changes due to IPF and CF. The use of UTE MRI, in conjunction with established functional lung MRI in chronic lung diseases, will serve to mitigate the need for computed tomography in children.

Conclusion: Current limitations of UTE MRI include long scan times, poor delineation of thin-walled structures (e.g. cysts and reticulation) due to limited spatial resolution, low signal to noise ratio, and imperfect motion compensation. Despite these limitations, UTE MRI can now be considered as an alternative to multidetector computed tomography for the longitudinal follow-up of the morphological changes from lung diseases in neonates, children, and young adults, particularly as a complement to the unique functional capabilities of MRI.

Keywords: Adult; Child; Cystic fibrosis; Hyperpolarized gas; Idiopathic pulmonary fibrosis; Infant; Longitudinal studies; Lung; Magnetic resonance imaging; Perfusion; Xenon; Young adult.

Figure 1:

Ultra-short echo time (UTE) MRI in scleroderma. The pulse sequence timing showing first and second echo times (TE’s) at 0.08 ms (center-out) and 2.1 ms (radial-in), respectively (left). Corresponding 3D radial images acquired at the UTE of 0.08 ms (a) and conventional TE of 2.1 ms (b) demonstrated much improved visualization of fibrosis on UTE. In the 3D isotropic resolution UTE images, full chest coverage affords visualization of lung structures such as fissures (c) normally not seen on lung MRI. Not only the fibrosis, but also the dilated, fluid and debris-filled esophagus is visualized. Adapted from (16).

Figure 2:

UTE MRI in Cystic Fibrosis (CF). High-resolution computed tomography (HRCT) images (a) acquired 14 months prior to UTE MRI (b) in a 17 year-old with CF. Arrow indicates a region of bronchiectasis and mucus plugging well visualized on both HRCT and UTE MRI. Images were acquired using 3D radial UTE as in (16).

Figure 3:

Parenchymal density measures in obstructive lung disease comparing HRCT (a) and UTE (b) scans of a 63-year-old woman with severe asthma. Histogram analysis is commonly used in HRCT for detection of air trapping (−856 HU threshold) at expiratory lung volume. A similar histogram analysis of signal values normalized to a reference tissue (e.g. left ventricular blood pool) may be useful with UTE MRI for detecting regions of air trapping (arrows). Images were acquired using 3D radial UTE as in (16).

Figure 4:

The global dissolved phase hyperpolarized ¹²⁹Xe gas spectrum in the lungs after gas inhalation and breath-hold. The chemical shift at approximately 215 ppm for the red blood cells (RBC) and 197 ppm for the barrier (tissue + blood plasma) relative to ¹²⁹Xe gas phase resonance frequency. Adapted from (25).

Figure 5:

Structure-function associations in a 64-year-old healthy volunteer. Note the dependent portion of the lungs exhibits increased perfusion as expected. Larger vessels can be seen as hyperintense regions on UTE images and as regions with increased blood flow on perfusion images (red arrows). UTE images were acquired with an optimized 3D radial UTE sequence (16). Perfusion images were acquired using an interleaved variable density SPGR sequence (38).

Figure 6:

Structure-function associations in IPF lung disease. A 63 year old female with IPF with regions of apparent hyperintense signal on UTE consistent with fibrotic injury (upper row) and bilateral basal perfusion defects at the same locations (white arrows, middle row). Corresponding CT scans reveal honeycombing and reticulation in the same regions (black arrows, bottom row). The white arrows point to hyperintense regions in the UTE images that are spatially associated with perfusion deficits while the red arrows point to hyperintense regions that do not show perfusion deficits, demonstrating the potential of structure-function imaging to distinguish disease characteristics that may not be gleaned from anatomic images alone. UTE images were acquired with an optimized 3D radial UTE sequence (16). Perfusion images were acquired using an interleaved variable density SPGR sequence (38).

Figure 7:

Oxygen-enhanced (OE) UTE showing percent signal enhancement (PSE) for the 0.08 ms (A) and 2.1 ms (B) echo times, respectively. Signal contrast for OE UTE is primarily T1 weighted (C) such that the paramagnetic effects of hyperoxic (100% O2) vs. normoxic (21% O2) breathing makes it possible to visualize oxygen dissolved in the tissues and blood of the lungs. Adapted from (48).

Figure 8:

Structure-function association between hyperpolarized ¹²⁹Xe MR spectroscopic imaging of the dissolved phases and UTE MRI in IPF lung disease. Structural UTE images (a) showing typical hyperintense fibrotic regions (arrows) in a 64 year-old male subject diagnosed with IPF that correspond to regional thickening of the barrier (b) and reduced RBC signal (c) relative to the polarized gas compartment. Note the heterogeneity of the parametric maps compared to a healthy normal subject in Figure 9. Images were acquired using the strategy detailed in Kaushik et al. (31).

Figure 9:

Structure-function association between hyperpolarized ¹²⁹Xe MR spectroscopic imaging of the dissolved phases and UTE MRI. The chemical shift at approximately 215 ppm for the RBC and 197 ppm for the Barrier (tissue + blood plasma) relative to Xe gas phase resonance frequency (a) is leveraged using spectroscopic imaging to produce separate barrier:gas (b) and RBC:gas (c) ratio images, which can be fused to structural UTE images (b). Note the homogeneous distribution throughout the lungs for both barrier:gas and RBC:gas ratios in a healthy normal 45 year-old female subject. Images were acquired using the strategy detailed in Kaushik et al. (31).

Figure 10:

Structure-function association between oxygen-enhanced (OE) and UTE MRI in a healthy normal subject using UTE. Image volumes are acquired during normoxic and hyperoxic free-breathing and subtracted to result in the percent signal enhancement (PSE and color bar) based on T1-shortening in the presence of higher concentration of paramagnetic O₂. Enhancement in the lungs, left ventricle, left atrium, and aorta can are readily visualized with 3D isotropic resolution. Relevant pulse sequence parameters are described in detail in Zha et al. (62, 69).

Figure 11:

The OE MRI derived using the same UTE images is spatially registered. Spatial associations between structural CF abnormalities on UTE MRI for a 37-year-old woman with cystic fibrosis (a & c) juxtaposed with (b & d) PSE defects on OE MRI simultaneously depict morphological (air trapping, pleural finding, bronchiectasis, and pleural finding - arrows) and physiological (ventilation defects – green outline) abnormalities. Color bar indicates percent signal enhancement (PSE). Relevant pulse sequence parameters are described in detail in Zha et al. (62, 69).