Introduction

Standard spirometry only provides a global scalar measurement of the pulmonary ventilation. It cannot access the regional function of the organ and is prone to fail for proper diagnostic when the pulmonary alteration is local and blurred in the flow-volume measurement performed at the mouth. Besides, ventilation is essentially a three-dimensional mechanism that can be affected by simply the body position in the Earth gravity field. The above weight promotes more efficient alveolar gas evacuation and increased alveolar density through lung tissue deformation. However, this gravity dependence is usually averaged out over the whole lung^1,2.
Regional ventilation was recently explored with 3D MR spirometry³. Gravity dependence was used here to challenge the sensitivity of the new technique. It is an established phenomenon, easy to set up, control, and differentiate while the subject is her or his own reference by switching from supine to prone positions. Hence, we carried a preliminary prospective clinical study with 3D MR spirometry onto seven healthy volunteers in supine and prone positions.

Methods

Ten-minute dynamic lung MRI was performed for seven healthy asymptomatic volunteers (3 females, 4 males; 51.5 ± 14.1 y/o; 23.5 ± 3.1 kg/m² of BMI). The volunteers were asked to spontaneously breath while they underwent 4 successive acquisitions: two repeated in supine position and two repeated in prone position. The study was performed at 3T (GE Signa PET/MR scanner) using a UTE sequence combined with an AZTEK trajectory⁴, and a 30-channel thoracic coil for signal reception. Main parameters were TE = 14 µs, TR = 2ms, flip angle = 3°, bandwidth = 100 kHz, voxel size = 1.5 mm isotropic, matrix = (212×212×142). During the acquisition, the respiratory function was monitored using the outcome of an abdominal belt.
A self-navigator was extracted from the center of the acquired k-space to retrospectively rephase the MR data into respiratory phases, resulting into 20 dynamic images of an integrated respiratory cycle over the 10-minute acquisition. The strain tensor was estimated by elastic registration between the images of each temporal phase. Parametric maps characterizing the respiratory function were then derived for each temporal phase: local ventilatory volumes were inferred from the Jacobian of the strain tensor and local flow rates were temporally derived from these local volumes. Ultimately, 3D MR spirometry makes it possible to obtain flow-volume loops in every reconstruction kernel throughout the lung. It yields flow-volume 3D maps and full characterization of the local ventilation.
Analyses were performed into two anterior-posterior identical volumes. The areas of the flow-volume loops of the anterior-posterior regions were computed. Slopes of the mean volume ratios along the gravity direction were computed using a linear fit. These values were compared between acquisitions repeated in the same position and between supine and prone positions. Trends of the gravity dependence and volumes of flow-volume loop area were also compared after rescaling between 0 and 1 to cope for any ventilation variations between scans. Repeatability was evaluated using Bland Altman, and correlation analyses and Intraclass correlation coefficient (ICC). Sensitivity was established upon paired Wilcoxon tests on the supine and prone flow-volume loop area and slope measurements.

Results

The values of the respiratory monitoring vary by 4% in amplitude and 11% in frequency between successive acquisitions with a greater variability in supine position. Volume ratio slopes and flow-volume loop areas are similarly repeatable but flow-volume loop areas are not repeatable in supine. However, rescaled measurements are repeatable (Figure 4).
Maximal values of inhaled volume ratios and exhaled flows are found in posterior and anterior regions for supine and prone positions respectively (Figure 1 a,b). The inhomogeneity of the flow-volume maps underlines the gravity dependence of the lung. The dependence is consistently observed on the area of flow-volume loops ($$$p < 10^{-3}$$$) (Figure 1 c and Figure 4). The anterior-posterior slope also differs ($$$p < 10^{-3}$$$) (Figure 4). Finally, rescaled parameters show strong statistical difference between dependent and nondependent regions (Figure 4).

Discussion

Three-dimensional MR spirometry is moderately repeatable before parameters are rescaled possibly because monitored respiratory breathing conditions and patterns evolve from one acquisition to another. Simultaneous standard spirometry measurements could account for physiological variations in spontaneous breathing and a more elaborated approach than rescaling, implemented here, could be used to produce quantitative repeatable measurements. The ventilation gradient is stronger in supine than in prone essentially because the mediastinum applies a greater force on the parenchyma^5,6. This gradient is reversed in the most dependent regions at the periphery of the lung where alveoli are expected to remain deflated⁷.

Conclusion

Three-dimensional MR spirometry is fairly repeatable when the measurement outcomes are properly rescaled. The technique is sensitive to gravity effects either in supine or prone positions whereas the subject is spontaneously breathing. Ventilation is less governed by gravity and it more homogeneously distributed in the lung in prone position.

Acknowledgements

Peder Larson for UTE sequence. The PET/MR platform is affiliated to the France Life Imaging network (grant ANR-11-INBS-0006).

References

1. Bryan, A. C., Milic-Emili, J., & Pengelly, D. (1966). Effect of gravity on the distribution of pulmonary ventilation. Journal of Applied Physiology, 21(3), 778-784.Bryan Emili
2. Hopkins, S. R., Henderson, A. C., Levin, D. L., Yamada, K., Arai, T., Buxton, R. B., & Prisk, G. K. (2007). Vertical gradients in regional lung density and perfusion in the supine human lung: the Slinky effect. Journal of applied physiology, 103(1), 240-248.
3. Boucneau, T., Fernandez, B., Larson, P., Darrasse, L., & Maître, X. (2020). 3D magnetic resonance spirometry. Scientific Reports, 10(1), 1-12.
4. Boucneau, T., Fernandez, B., Besson, F. L., Menini, A., Wiesinger, F., Durand, E., ... & Maître, X. (2021). AZTEK: Adaptive zero TE k‐space trajectories. Magnetic Resonance in Medicine, 85(2), 926-935.
5. Henderson, A. C., Sá, R. C., Theilmann, R. J., Buxton, R. B., Prisk, G. K., & Hopkins, S. R. (2013). The gravitational distribution of ventilation-perfusion ratio is more uniform in prone than supine posture in the normal human lung. Journal of applied physiology, 115(3), 313-324.
6. Klimeš, F., Voskrebenzev, A., Gutberlet, M., Kern, A., Behrendt, L., Kaireit, T. F., ... & Vogel‐Claussen, J. (2019). Free‐breathing quantification of regional ventilation derived by phase‐resolved functional lung (PREFUL) MRI. NMR in Biomedicine, 32(6), e4088.
7. Tobin, A., & Kelly, W. (1999). Prone ventilation—it's time. Anaesthesia and intensive care, 27(2), 194-201.