1. Introduction
In the last years the application of laser-polarized (LP) noble gases received growing interest in clinical imaging, particularly for the study of the otherwise 'invisible' air spaces in the lung. An organ which could not be detected easily before, because the large number of air-tissue interfaces causes large susceptibility differences resulting in rapid decaying proton signals. In addition the low concentration of protons in the lung tissue renders a very weak NMR signal. This problem was overcome by the use of LP-3He or LP-129Xe, where the low density was compensated by an artificial increase of the polarization by 5 orders of magnitude. It was demonstrated by several authors that both nuclei are suitable for lung imaging (1,2), however they possess unique characteristics which makes them very different in their physical and physiological behavior. 3He has a high gyromagnetic ratio and polarization levels of over 70% can be achieved, thus providing high sensitivity and in addition it has no physiological effects. On the other hand, 129Xe is lipophilic; it passes the air/blood barrier and dissolves in the blood (3-5), even acts as an anesthetic and has a large chemical shift range (6). From the clinical point of view 3He is ideal for detailed observation of the anatomy of the lung’s air spaces, detection of pathological changes of the microscopic morphology via restricted self diffusion maps (7-9) and O2 concentration via paramagnetic relaxation (10,11). In contrast, 129Xe is suited for functionality and perfusion studies (5), the gas dissolves in blood, a process which can be accurately monitored through its chemical shift. Recently xenon was used for the study of gas exchange between the xenon dissolved in the lung tissue and the gas in the alveoli (12,13).
The very different gyromagnetic ratios of both gases easily allow double-resonance experiments and hence direct excitation and detection of each individual isotope in consecutive experiments; in this way the unique information provided by each nuclei can be obtained upon the delivery of a single bolus of the admixed gases. In this work we report first attempts to image both isotopes in a gas mixture. We show an in vitro experiment where a mixture of 3He and xenon with molar fraction 0.5 is introduced in a sample with a small amount of toluene, which substitutes the blood as a xenon absorbing site. A 2D image of each nucleus is recorded to show the individual behavior of each noble gas and its effect on the NMR image.
The second experiment is performed on a dried mouse lung prepared as described in (14). Three sets of 3D images are acquired, two with pure helium and xenon as well as a third with a gas mixture. An analysis of the gain in sensitivity and contrast enhancement for the large airways is presented.
2. Materials and methods
2.1. In vitro experiments
Toluene is frozen in the bottom of a sample tube used for polarizing xenon. This tube consists of two concentrical tubes, the inner one is for inflow and the outer is for outflow of a gas mixture carrying LP-129Xe. The principal design is similar to that of a gas bubbler (cf. Fig. 1a). The total length of the tube is 16 cm while the imaged section is about 3.7 cm. One side of the container has a notch to prevent toluene from leaving the imaged section when it melts. 10 ml of toluene is filled into this tube, which is then connected to the xenon polarizer and immersed in a liquid nitrogen bath. Directly after the liquid is frozen a flux of gases is run through the tube, causing the LP-xenon to freeze on the walls and thus being stored in the hyperpolarized state.
LP-xenon was frozen for a time of 5 min at a flow rate of 300 ml/min. After removing the buffer gases, 600 mbar of LP-3He are introduced in the sample tube, which is immediately placed in the bore of the imaging magnet. As soon as the toluene and xenon are both molten 2D images were acquired, first of the 129Xe and immediately afterwards of 3He. In order to quantify the signals, an FID was acquired for each isotope prior to the image acquisition. The positioning of the sample tube and the final distribution of toluene are sketched in Fig. 1b. The final pressure of the sample was of 2.2 bar.
Fig. 1. a) Initial position of the sample tube during collection of for LP-129Xe. It consists of two concentrically arranged tubes of diameters Æ1 = 10 mm and Æ2 = 20 mm. The inner tube is for inflow of the gas and the external one for outflow. The total length of the tube is 160 mm and the imaged section is 37 mm. A notch was made in the tube to prevent toluene from leaving the imaged section once it melts. b) Position of the sample tube in the horizontal bore magnet. The read gradient for the 2D images is transversal (y) to the external magnetic field, B0, while the phase gradient is applied in the direction of the magnetic field (z). g denotes the gravity acceleration.
2.2. Mouse lung experiments
The procedures for rodents lung fixation are described in detail in (14). The lungs/heart complex was dissected out from a mouse killed by ether overdose and the specimen was rinsed with isotonic saline and fixation solution under low vacuum conditions. After fixation, it was dried using alcohol and stored in a non-collapsed state (i.e. state of inhalation). A mouse lung prepared in such a way was sealed in a small epoxy cylinder of 25 mm that fits into the double resonant rf coil. The dimensions of the lung are 45 mm length, including the trachea and a maximum width of 24 mm. In order to deliver the gases to the lung, a set of pneumatic valves (Festo AG and Co. KG, Esslingen Germany) were modified to replace all magnetic components. These are controlled by magnetic air valves (Festo AG) which are in turn driven from the spectrometer’s console. In this way, vacuum in the lung and in different sections of the line can be controlled as well as the supply of gas itself. The gas delivery procedure consists in placing a bottle with hyperpolarized gas in the bore of the imaging magnet and after evacuating the lung and the connection lines to 10-6 bar; the gas is allowed to fill the lung.
Three different experiments were performed, the two first consist of images of 129Xe and 3He in a pure state, 2 bar of each gas were used respectively. For the third experiment a gas mixture consisting of 2 bar of LP-129Xe and 2 bar of LP-3He was prepared, in this way the signal to noise ratio of the images with pure gases can be directly compared with the one of the mixture. In each case the lung was previously evacuated to 10-6 bar and then filled with the corresponding gas mixture.
2.3. NMR procedures
All images were acquired in a 4.7 T horizontal, 20 cm bore magnet. Actively shielded gradients capable of 300 mT/m from Bruker (Bruker Biospin GmbH, Germany), driven by amplifiers from Copley (Copley Controls Corp., USA) were used. A double resonant coil with an inner diameter of 26.5 mm from Bruker was used for rf excitation and detection, with frequencies of 55.59 MHz and 153.096 MHz for 129Xe and 3He, respectively. The gradients and the rf were controlled by a Maran DRX console (Resonance Instruments Ltd, Witney, UK) which runs under a Matlab12 (MathWorks Inc., USA) home made software environment.
2D-FLASH Cartesian trajectories of the k-space in a 64 ´ 64 matrix were acquired using a 5 ms hard rf pulse which corresponds to a tip angle of 3° for both 3He and 129Xe. The intensities in the rf excitation pulses were changed in order to maintain the length of the pulses constant. The acquired data were then filled with zeros to result in a 128 ´ 128 matrix and Fourier transformed. The read gradient was applied in the y direction, perpendicular to the external magnetic field and the phase gradient was applied parallel to the external magnetic field, z, as shown in Fig. 1b. A spectral width of 50 kHz was used, and the field of view (FOV) was chosen to be 4 ´ 4 cm2. The delay between experiments was 10 ms, 4 scans were accumulated and the total imaging time was of 3 s for each image.
3D-FLASH Cartesian trajectories were used for lung images, a 64 ´ 4 ´ 64 matrix was acquired and zero filled to a 128 ´ 8 ´ 128 matrix. A 6° rf hard pulse of 5 ms duration was used for both nuclei. In Fig. 2 a sketch of the positioning of the lung in the imaging system is shown, which corresponds to a prone position. Sagittal and transversal images with 129Xe were acquired to confirm proper location in the coil and magnet. The read gradient was set in y-, the first phase gradient was in x- and the second phase gradient was in the z-direction. The used FOV was 3 ´ 4 ´ 3 cm3. The central slices of 7.5 mm thickness with a pixel size of 470 ´ 620 mm2 are shown.
|
Fig. 2. Schematic drawing of the positioning of the lung in the magnet. 3D-FLASH Cartesian trajectories in a 64 ´ 4 ´ 64 matrix were acquired. The read gradient was set in y- and the phase gradients in x- and in z-direction.
2.4. Noble gas polarization and delivery
Optical polarization of xenon was achieved by means of a home made apparatus similar to that described in (15), located in the Max Planck-Institute for Polymer Research, Mainz. A mixture consisting on 1% of Xe (natural isotope distribution), 5% N2 and 94% 4He was used. LP-129Xenon is frozen by means of a liquid nitrogen bath located in the bore of a magnet consisting of an arrangement of 16 permanent bar-magnets in a Halbach configuration producing a magnetic field of 0.3 T (16). The polarization levels were measured to be higher than 30%.
The 3He polarization was carried out using a home built large scale polarizer located at the department of Physics in the University of Mainz, which can produce up to 70% of polarization at 3.3 bar×l/h and 80% at 1.2 bar×l/h (17). The polarization is achieved by metastable spin exchange method (1,18). Transport cells of iron-free glass (Supremax glass, Schott, Mainz, Germany), were filled with 2.1 bar of 60% LP-3He. Subsequently, the cells were placed in a shielded container enclosing a permanent low magnetic field (0.8 mT) (17) and transported to the MRI laboratory in the Max Planck Institute. Helium is then stored in very homogeneous magnetic field of 2.5 mT, generated by five coaxial coils of 45 cm diameter and of a total length of 70 cm.
3. Results and Discussion
3.1. In vitro experiments
Fig. 3a shows the 129Xe spectrum obtained from the FID acquired previous to the image in Fig. 3b. A high intensity peak corresponding to free xenon gas appears at -12 ppm while xenon absorbed in liquid toluene has a chemical shift of 188 ppm [19]. Fig. 3b shows the 2D image of 129Xe, acquired with the same spectral width as Fig. 3a. The read direction, where the spectroscopic information is contained, is displayed on the vertical axis while the horizontal axis corresponds to the phase direction. Free xenon appears in the center of the spectrum while xenon absorbed in toluene appears at a lower position in the image, with the same chemical shift as shown in Fig. 3a. Figure 3c shows the spectrum of 3He recorded prior to the image in Fig. 3d. As expected, a single line is observed which is neither affected by the increased pressure nor the presence of xenon and toluene. Fig. 3d shows the resulting image for helium. It is clearly observed that helium occupies only the residual volume left by the toluene, in which it does not dissolve.
Fig. 3. a) spectrum of 129Xe in the presence of liquid toluene acquired prior to the image The difference in the chemical shift between xenon absorbed in toluene and free xenon gas is 199 ppm at the used pressure and temperature conditions. The used spectral width was 50 kHz. b) 2D image of 129Xe in the presence of toluene, free gas appears near the center of the vertical axis (read direction) while 129Xe absorbed in toluene has a chemical shift corresponding to 188 ppm. c) spectrum of 3He in the same conditions as a). d) 2D image of 3He in the presence of toluene, helium only occupies the residual volume above the liquid.
3.2. Mouse lung experiments
The results obtained by imaging the prepared mouse lungs are summarized in Fig. 4. Figures 4a and 4b show the images of a mouse lung filled with pure 3He and 129Xe at 2 bar respectively, while Fig. 4c and 4d show the images acquired for the gas mixture of equal partial pressures (2 bar) of 3He and 129Xe.
 |
Fig. 4. Central slices from 3D images of the prepared mouse lung a) pure 3He at 2 bar, b) pure 129Xe at 2 bar c) 3He detected in a mixture of 2 bar 3He with 2 bar xenon d) 129Xe detected in the same mixture. The grayscale in Figs. 4a and 4c is set to the maximum intensity of Fig. 4c while the maximum intensity of Fig. 4b is used for 4b and 4d.
In order to avoid influence of depolarization on the data, an FID was recorded previous to the acquisition of each image and the data were rescaled to the FID’s amplitude before performing the Fourier Transform. For pure 3He the alveolar region can be clearly observed while the trachea has a very low intensity. This is due to the attenuation of the NMR signal in the bigger cavities due to rapid (unrestricted) diffusion. In the case of the gas mixture it can be observed that the intensity in the alveolar region has increased, and that the trachea is presented in a comparable intensity. For pure xenon (Fig. 4b) the features of the lung are no longer resolved, the alveolar spaces are clearly observed while the trachea is within the noise level. The image of 129Xe acquired in the gas mixture (Fig. 4d) presents similar characteristics.
A pixel by pixel comparison was performed for each set of images corresponding to the same detected nucleus. Figure 5a shows the ratio of the image of 3He corresponding for the mixture of gases (Fig. 4c) with respect to that of pure helium (Fig. 4a). A threshold value for those pixels with signal to ratio (SNR) smaller than 2 was used to mask the noise; i.e. all values below this value were set to zero. It is clearly observed that the gain in signal intensity in the trachea is much more pronounced than in the alveoli. Figure 5b shows a histogram of the distribution of this signal gain for the unmasked part of Fig. 5a.
Fig. 5 a) Pixel by pixel comparison of the 3He image of the gas mixture respect to the pure helium case. b) Histogram of the distribution in signal gain. An enhancement by a factor of 1.5 is obtained for the alveolar region and values up to 6 for the trachea. c) Pixel by pixel comparison of the 129Xe image of the gas mixture with respect to the case of pure xenon. d) Histogram of the signal gain distribution. The broad distribution of SNR values shown in the histogram is due to the high noise level in the image. The maximum number of pixels are close to a unity value, showing that using a gas mixture doesn’t effect the image contrast considerably in the case of xenon.
The mayor contribution is due to the alveolar region, were a gain of a factor of 1.5 is obtained. For the trachea a gain of a factor up to 6 is achieved, but the number of pixels is to low to show up distinctively in the histogram. Figure 5c shows the masked ratio image for xenon (image of Fig 4d divided by Fig 4b). The region corresponding to the trachea is masked out as the SNR is below to the selected threshold value. Figure 5d shows the corresponding histogram of the signal intensity gain for xenon. A broad distribution of values is obtained due to the high noise level of the images. The maximum intensity corresponding to the signal from the alveoli appears at 0.85, indicating that the used gas mixture results in a reduced signal level for the alveoli.
4. Conclusions
In this work we show the possibility to image a mixture of LP-3He and LP-129Xe. The results presented in the in vitro sample (Fig. 3) demonstrate how spatial dependent chemical interactions of 129Xe can be monitored by MRI. The presence of LP-3He permits the acquisition of an image with higher sensibility for the free spaces. No appreciable change in the relaxation times of each gas in the presence of the other was observed. In a second experiment we analyze how the mixture of gases of very different diffusion coefficient can influence the contrast on lung MRI (Figs. 4 and 5). For the setup presented the diffusion coefficient of 3He is reduced by the combined effect of the mixture and the pressure increase. On the other hand, for xenon an increase in the diffusion coefficient is expected due to the mixture with helium, which is partially compensated by the increase of the pressure. From this it can be expected, that when experiments are performed at the same pressure conditions the gain for the 3He contrast would decrease while in the xenon images the attenuation of the NMR signal would be even greater. Since the used lung phantoms were dried and sealed with latex, no difference in the chemical shift of 129Xe was observed. However, for a real lung this would be a valuable experiment, which could also run as a chemical shift resolved image on a coarser spatial resolution, because the positions could be registered by a highly resolved 3He image.
A detailed study on the influence on the mixture of gases of very different diffusion coefficients on the contrast in an image will be presented in a future communication.
Acknowledgments
We want to thank Manfred Hehn and Hanspeter Raich for their help in the design and construction of the gas handling system at the MPI-P. Financial support by DFG (Forschergruppe “Bildgestützte zeitliche und regionale Analyse der Ventilations - Perfusionsverhältnisse in der Lunge”) and a special grant of the Max Planck society made this work possible.
References
1. Boyd M. Goodson, Nuclear Magnetic Resonance of Laser-Polarized Noble Gases in Molecules, Materials, and Organisms, J. Magn. Reson., 155, 157-216 (2002).
2. Harald E. Möller, X. Josette Chen, Brian Saam, Klaus D. Hagspiel, G. Allan Johnson, Talissa A. Altes, Eduard E. de Lange, and Hans-Ulrich Kauczor, MRI of the Lungs Using Hyperpolarized Noble Gases, Mag. Reson. Med., 47, 1029-1051 (2002).
3. C.H. Tseng, S. Peled, L. Nascimben, E. Oteiza, R.L. Walsworth and F.A. Jolesz, NMR of Laser-Polarized 129Xe in Blood Foam, J. Magn. Reson., 126, 79-86 (1997).
4. Jan Wolber, Andrea Cherubini, Martin O. Leach and Angelo Bifone, Hyperpolarized 129Xe NMR as a Probe for Blood Oxygenation, Mag. Reson. Med., 43, 491-496 (2000).
5. Sven Mansson, Jan Wobler, Bastiaan Driehuys, Per Wollmer and Klaes Golman, Characterization of Diffusing Capacity of 3He Gas in Healthy and Diseaed Lungs, Mag. Reson. Med., 50, 1170-1179 (2003).
6. Kuniyoshi Sakai, Anastacia M. Bilek, Eduardo Oteiza, Ronald L. Walsworth, Dilip Balamore, Ferenc A. Jolez, and Mitchell S. Albert, Temporal Dynamics of Hyperpolarized 129Xe Resonances in Living Rats, J. Mag. Reson. B, 111, 300–304 (1996)
7. Harald E. Möller, X. Josette Chen, Mark S. Chawla, Gary P. Cofer, Bastiaan Driehuys, Laurence W. Hedlund, Steve A. Suddarth and G. Allan Johnson, Sensitivity and Resolution in 3D NMR Microscopy of the Lung With Hyperpolarized Noble Gases, Mag. Reson. Med., 41, 800-808 (1999).
8. Brian T. Saam, Dmitriy A. Yablonsky, Vikram D. Kodibagkar, Jason C. Leawoods, David S. Gierada, Joel D. Cooper, Stephen S. Lefrak and Mark S. Conradi, MR Imaging of Diffusion on 3He Gas in Healty and Diseased Lungs, Mag. Reson. Med., 44, 174-179 (2000).
9. Michael Salerno, Talissa A. Altes, James R. Brookeman, Eduard E. De Lange, and John P. Mugler III, Rapid Hyperpolarized 3He Diffusion MRI of Healthy and Emphysematous Human Lungs Using an Optimized Interleaved-Spiral Pulse Sequence, J. Mag. Reson. Imag., 17, 581-588 (2003).
10. A.J. Deninger, B. Eberle, M. Ebert, R. Grossmann, W. Heil, H.-U, Kauczor, L. Lauer, K. Markstaller, E. Otten, J. Schmiedeskamp, W. Schreiber, R. Surkau, M. Thelen, and N. Weiler, Quantification of Regional Intrapulmonary Oxygen Partial Pressure Evolution during Apnea by 3He MRI, J. Magn. Reson., 141, 207-216 (1999).
11. Anselm J. Deninger, Balthasar Eberle, Michael Ebert, Tino Grossmann, Gorden Hanisch, Werner Heil, Hans-Ulrich Kauczor, Klaus Markstaller, Ernst Otten, Wolfgang Schreiber, Reinhard Surkau and Norbert Weiler, 3He-MRI-based measurements of intrapulmonary pO2 and its time course during apnea in healthy volunteers: first results, reproducibility, and technical limitations, NMR in biomedicine, 13, 194–201 (2000).
12. Kai Ruppert, James R. Brookeman, Klaus D. Hagspiel, and John P. Mugler III, Probing lung physiology with xenon polarization transfer contrast (XTC), Magn. Reson. Med., 44, 349-357 (2000).
13. Kai Ruppert, Jaime F. Mata, James R. Brookeman, Klaus D. Hagspiel, and John P. Mugler III, Exploring lung function with hyperpolarized 129Xe nuclear magnetic resonance, Magn. Reson. Med., 51, 676-687 (2004).
14. Peter Blümler, Rodolfo H. Acosta, Angelika Thomas-Semm, and Stefan Reuss, Lung fixation for the preservation of air spaces, Exp. Lung Res., 30, 1-10 (2004)
15. Nadim J. Shah, Timur Ünlü, Hans-Peter Wegener, Horst Halling, Karl Zilles and Stephan Appelt, Measurement of rubidium and xenon absolute polarization at high temperatures as a means of improved production of hyperpolarized 129Xe, NMR in biomedicine, 13, 214–219 (2000).
16. H. Raich and P. Blümler, Design and Construction of a Dipolar Halbach Array with an Homogeneous Field from N ´ 8 Identical Bar-Magnets ― NMR-Mandhalas―, Conc. Magn. Reson. B, submitted
17. E. J. R. van Beek, J. Schmiedeskamp, J. M. Wild, M. N. J. Paley, F. Filbir, S. Fichele, F. Knitz, G. H. Mills, N. Woodhouse, A. Swift, W. Heil, M. Wolf and E. Otten, Hyperpolarized 3-helium MR imaging of the lungs: testing the concept of a central production facility, Eur Radiol 13, 25832586 (2003).
18. J. Becker, W. Heil, B. Krug, M. Leduc, M. Meyerhoff, P. J. Nacher, E. W. Otten, Th. Prokscha, L. D. Schearer, and R. Surkau, Study of mechanical compression of spin-polarized 3He gas, Nucl. Instrum. Methods A 346, 45–51 (1994).
19. Mathias Haake, Boyd M. Goodson, David D. Laws, Eike Brunner, Michelle C. Cyrier, Robert H. Havlin, Alexander Pines, NMR of supercritical laser-polarized xenon, Chem. Phys. Lett. 292, 686-690 (1998).