Recently, ultra-fast balanced steady-state free precession (ufSSFP¹) imaging has been proposed for morphological chest imaging at 1.5T providing high signal-to-noise ratio in the lung parenchyma, as well as an improved visualization of lung vasculature and nodules².
For certain groups of patients with very low proton densities in the lung parenchyma, e.g. with diffuse emphysematous disease of the lung, however, the residual signal may be on the detection limit. The purpose of this work is thus to investigate, whether the $$$T_2$$$ and $$$T_1$$$ shortening effects of contrast agents may have a beneficial effect on the ufSSFP signal intensity and image quality in the lung.

Simulations
Numerical simulations of the transverse steady state magnetization³ $$$M_{xy}(\alpha,T_1,T_2,\mathrm{TR})$$$ were performed in the limit $$$\mathrm{TR}<<T_2<T_1$$$, given by:

$$M_{xy}(\alpha,T_1,T_2,\mathrm{TR})=M_0 \frac{(1-E_1) \sin \alpha}{1-(E_1-E_2) \cos \alpha-E_1E_2} \mathrm{,} \qquad\text{[1]}$$

where $$$E_{1,2}= \mathrm{e}^{\frac{\mathrm{TR}}{-T_{1,2}}}$$$.

Generally, the effect of shortening the tissue relaxation times after intravenous contrast agent administration enters Eq. [1] through:

$$\frac{1}{T_{2,\mathrm{post}}}=\frac{1}{T_{2,\mathrm{pre}}}+r_2\cdot C \qquad \textrm{and} \qquad \frac{1}{T_{1,\mathrm{post}}}=\frac{1}{T_{1,\mathrm{pre}}}+r_1\cdot C \mathrm{,}$$

where $$$r_1$$$ and $$$r_2$$$ are the relaxivities of the contrast agent and $$$C$$$ is its concentration in tissues. From this, a relative signal enhancement ($$$\mathrm{SE}$$$) between pre- and post-contrast agent injection can be defined $$ \mathrm{SE}:=\frac{M_{xy,\mathrm{post}}-M_{xy,\mathrm{pre}}}{\underset{\alpha}{\mathrm{max}}\left\{ \right.{M_{xy,\mathrm{pre}}}\left. \right \}}=\frac{M_{xy}(\alpha,T_{1,\mathrm{post}},T_{2,\mathrm{post}},\mathrm{TR})-M_{xy}(\alpha,T_{1,\mathrm{pre}},T_{1,\mathrm{pre}},\mathrm{TR})}{\underset{\alpha}{\mathrm{max}}\left \{ \right.{M_{xy}(\alpha,T_{1,\mathrm{pre}},T_{1,\mathrm{pre}},\mathrm{TR})}\left. \right \}}\mathrm{.}$$

Simulation parameters were chosen as follows: repetition time $$$\mathrm{TR}$$$=1.2 ms, relaxation times of the lung tissue $$$T_1$$$=1000-1300 ms and $$$T_2$$$=40-90 ms at 1.5T (Refs. 4-6). Contrast agent relaxivities⁷ $$$r_1$$$ =3.6 L/(s*mmol) and $$$r_2 $$$=4.3 L/(s*mmol), concentration $$$C \cong $$$ 0.5 mmol/L (administration of 0.2 ml/kg Gd-DOTA (0.5 mmol/ml) dissolved in the extracellular water⁸ 187ml/kg).

Measurement protocol and image post-processing
Measurements were performed on a whole-body 1.5 T MR-scanner (MAGNETOM Avanto, Siemens Healthcare, Germany) using a 12-channels thorax and a 24-channels spine coil. Two healthy volunteers were scanned using 3D ufSSFP with the following parameters: TE/TR = 0.47/1.19 ms, flip angle α= 23°, RF pulse length = 80 μs, 1563 Hz/pixel bandwidth, field-of-view = 400x400x250 mm³, two averages, isotropic resolution = 3.1 mm³, reconstruction matrix = 128²x80, parallel imaging GRAPPA factor 2, total acquisition time = 16 s. The scans were performed in the end-expiratory breath-hold. Two acquisitions were performed before contrast agent (native images) and two after the single-dose intravenous injection of Gd-DOTA (0.2 ml per kg body-weight).
Calculations of exemplary relative signal enhanced maps were performed after median filtering⁹ (filter size 5x5x5 voxels) in order to reduce signal noise and
 vasculature overlaying the pulmonary tissue. Signal analysis was then performed in segmented three-dimensional regions of interest (ROIs).

A simulation of the bSSFP signal and the expected possible range of relative signal enhancement as a function of the flip angle are presented in Fig. 1. The maximal signal for the lung tissue occurs around $$$\alpha \approx$$$17-35° whereas the maximal signal enhancement between $$$\alpha \approx$$$ 45-65°.
Exemplar coronal thorax images before and after the contrast agent injection are shown in Fig. 2a and 2b; the corresponding signal enhancement map in Fig. 2c. Mean $$$SE$$$ in regions-of-interest were: 92% lungs, 56% blood (aorta), 35% liver, 29% muscle and 2% fat. The contrast-to-noise ratio between blood vessels, lung parenchyma and airspaces was 23:8:1 on native images, and 32:14:1 after the gadolinium injection.
Detailed coronal views of the lungs are shown in Fig. 3 with a side-by-side comparison of pre- and post-contrast administration: the signal intensity increase of gadolinium is perceivable in both the parenchyma and the vasculature. In this work, we have shown that the signal of balanced SSFP sequences can be considerably increased following the intravenous contrast agent administration. In agreement with the simulations, we observe an almost doubling of the parenchymal signal intensity and an improved depiction of vasculature structures.
In conclusion, the use of contrast agents in combination with bSSFP imaging leads to a considerable increase in parenchymal signal. Since the overall achievable signal-to-noise for lung imaging is typically limited by breath holding, the use of contrast-agents to enhance the residual pulmonary bSSFP signal appears appealing especially in the combination with pulmonary functional imaging of pathologies with very low tissue density, such as for emphysematous lung destruction or air-trapping.