Introduction

Pulmonary septal wall thickness and gas transit time measurements using hyperpolarized 129Xe (HXe) MRI have enabled quantitative assessment of lung function. Previous studies using animal models of radiation-induced lung injury (RILI), for example, have shown such measurements to be sensitive to structural and functional changes during disease onset and progression¹. However, such studies are limited by the fact that the intubation/tracheotomy procedures necessary for gas delivery are usually terminal. In this study, we demonstrate an imaging technique for non-invasively measuring global and regional gas uptake in free-breathing mice.

Methods & Materials

Two C57BL/6 mice were anesthetized using 1-2% isoflurane and imaged in a 9.4T vertical-bore micro-imaging MRI system (Bruker Inc.) while free-breathing a normoxic gas mixture consisting of O₂, isoflurane, and HXe gas administered via homemade gas delivery system. Gas and dissolved-phase images were acquired simultaneously using a respiratory-gated CSI sequence combined with a user-defined phase-encode/flip-angle table that utilized a rectangular RF pulse centered on the dissolved-phase such that the dissolved/gas peaks were subject to a 90º/5º flip angles, respectively. HXe uptake at different points in the gas exchange process was measured by acquiring images with different repetition times (TR values of 13 ms, 20 ms, 25 ms, 30 ms, 40 ms, 50 ms, 60 ms, 75 ms, and 100 ms). Individual k-space points were acquired in a series at end-exhale (3-9 points per breath depending on the TR), preceded by a dummy pulse that was used to eliminate magnetization accumulated during the non-imaged, inhale period. Gas- and dissolved-phase images were both normalized to the average gas signal within the lungs to account for the T₁ decay of HXe during the course of imaging. Dissolved-to-gas (D:G) ratio maps were then fitted to a theoretical gas exchange and uptake model similar to that described in² in order to derive regional septal wall thickness measurements.

Results

Figure 1 shows a representative plot of the average D:G ratio signal for images acquired with TRs increasing from 13 ms to 100 ms. The average D:G ratio increases with TR as expected, representing the gradual accumulation of signal as magnetization flows from the gas to dissolved compartments. This trend is corroborated by Figure 2, which displays representative projection axial D:G ratio maps for the same TRs as in Figure 1: similarly, as TR is increased, the global ratio for each of the maps increases. Figure 3A shows a septal wall thickness map derived from the images in Figure 2; as is apparent, there is a fairly homogenous distribution of values, where deviations are largely seen around the edges of the lung, as well as a few regions within the lung. Figure 3B shows a probability density function (PDF) displayed in a histogram plot for the septal wall thickness values shown in Figure 3A: while a few outliers can be seen, the majority of the values lie within 10 and 18 μm.

Discussion

Septal wall measurements using HXe MRI have been carried out in many previous studies for the assessment of lung function in rodent models of RILI^{2, 3}. However, no work has demonstrated such measurements in free-breathing mice without the use of mechanical ventilation. Acquisition of the functional measurements shown here are particularly useful since it reduces the number of animals used per study, and mitigates inter-cohort variability. Furthermore, the results shown here demonstrate the feasibility of acquiring such measurements under free-breathing conditions. Given the potential for differences, particularly in intrapulmonary blood volume during mechanical ventilation⁵, these conditions may be more representative of normal structure and physiology. Given a fixed flip-angle and increasing TR, the dissolved signal is expected to increase as demonstrated by both Figures 1 and 2. The mode of our septal wall thickness derived from the map in Figure 3A was 6 μm, and while it is in disagreement with a stereology-based assessment of the mouse septal wall thickness of approximately 4.91 ± 0.97 μm⁴, it is similar to previous HXe studies which calculated an average septal wall thickness of 12 ± 1.1 μm³. The discrepancy in our measurements are partially due to our model failing to account for pulmonary vasculature and non-symmetric geometric configurations of septal-wall-capillary boundaries, as well as intrinsic variations in animal breathing conditions during the course of imaging.

Conclusion

In this work, we demonstrated a non-invasive imaging technique for measuring gas uptake over a range of TRs in free-breathing mice. Measurements of regional septal wall thickness were also obtained by fitting acquired images to a theoretical gas uptake model. Further work will apply this technique to rodent models of RILI and fibrosis to assess and characterize the onset and progression of lung injury longitudinally.

Acknowledgements

No acknowledgement found.