Purpose

Imaging plays a major role in assessment of pulmonary diseases, and a broad spectrum of imaging modalities are available. Of these modalities, the main workhorse is still computed tomography (CT), which allows fast and high-resolution assessment of the lung parenchyma and surrounding structures. However, the soft tissue contrast and gain of direct functional information is limited. Lung MRI is difficult owing to the low proton density of the tissue and, consequently, low intensity on conventional ¹H MRI. At the same time, large differences in the magnetic susceptibility between lung tissue and air, respectively, reduces the signal strength even further. To overcome these limitations, we have investigated the possibility of imaging mouse lungs by use of fluorinated gases¹ such as CF₄ and have compared this novel technique with the gold standard hyperpolarized ¹²⁹Xe MRI2. Biocompatibility, high gyromagnetic ratio of ¹⁹F, and extremely short relaxation times (T₁ and T₂ in order of a few milliseconds) are the key benefits of fluorinated gases. In addition, the low cost of the gases, and no requirement for expensive polarization equipment, make perfluorocarbons ideal for lung MRI. Finally, the absence of fluorine in the body results in the lungs appearing bright with no background signal.

Methods

¹⁹F and hyperpolarized ¹²⁹Xe MR images were acquired using an Ultra Short Echo (UTE) sequence on a 7.0T preclinical spectrometer (Agilent Inc, USA). ¹²⁹Xe images were acquired using a home-built Helmholtz coil, and ¹⁹F images with a quadrature coil (PulseTeq, UK). For both nuclei, 3D radial trajectories were used; number of points along each trajectory np=128, and number of trajectories ntraj=8192. The acquired points were gridded onto a 128³ matrix. Although the imaging sequence in both cases was identical, the imaging strategies were different. Hyperpolarized ¹²⁹Xe MRI used isotopically enriched gas (80% ¹²⁹Xe) polarized to 20% (Polarian Inc, USA). During each breath-hold, 8 radial trajectories were acquired, with excitation angle a=20°, TR=20ms and TE=10μs, making the total experimental time Texp=16min. For ¹⁹F MRI, a very short TR=2ms was used, and 80 trajectories were acquired during each breath-hold, with a typical excitation angle a=60°. The thermally-polarized gases have inherently low signal and, therefore, require significant signal averaging. We limited our total experimental time to Texp=2h, which corresponds to NT=100. Proton images were also acquired using the UTE sequence and the same breath-hold technique, with the following imaging parameters: TR=10ms, TE=10μs, a=20°, NT=1, Texp=16min.

Results and Discussions

In this pilot study ex vivo naïve mice were used, in which the lungs were ventilated using a custom-built MR- and HP-gas compatible small animal ventilator (Duke University, USA) and data were acquired during breath-hold within an in vivo compatible breathing cycle (50ms – controlled inhalation, 250ms – breath-hold, 450ms – passive exhalation). Good quality images of the lungs were acquired using both ¹⁹F (Figure 1) and hyperpolarized ¹²⁹Xe (Figure 2) MRI techniques, and the typical Signal-To-Noise (SNR) in both cases was broadly similar. Interestingly, the two approaches provide complimentary information; ¹²⁹Xe MRI can provide dynamic/functional information on the lungs, whilst influx of fresh ¹⁹F does not influence the MR Image so much and, therefore, more static/physiological information is obtained. This information, combined with proton MR images, offers a unique lung diagnostic tool and the next step will be to move to in vivo studies, which would require overcoming additional challenges (e.g. replacement of non-recovery tracheotomy with recovery intubation for longitudinal studies; maintenance of the injectable anesthetic depth, by refining initial and supplementary dosage; fine synchronization of various experimental stages, etc.)

Conclusion

The ability to monitor lung function is undoubtedly important. ¹⁹F based perfluorocarbon and hyperpolarized ¹²⁹Xe MRI provide bright and detailed lung images, which enable visualization of lung structure and function to a degree that is not possible with proton MRI. The ¹⁹F approach is easier and cheaper to implement than the hyperpolarized gas setup and is immediately applicable to a range of pre-clinical work. However, ¹²⁹Xe MRI provides a unique dynamic tool that may have important applications warranting the additional effort and cost involved.

Acknowledgements

This work was supported by Cancer Research UK (grant number C5255/A12678) and the CRUK & EPSRC Cancer Imaging Centre in Oxford (grant number C5255/A16466). The authors also thank Stavros Melemenidis (Stanford University, USA), Peter Thelwall (Newcastle University, UK) and Konstantinos Papoutsis (MR Solutions, UK) for their time and effort on the initial stages of this long lasting journey.