RF instrumentation for same-breath triple nuclear lung MR imaging of (1)H and hyperpolarized (3)He and (129)Xe at 1.5T.

PURPOSE: The hyperpolarized gases (3)He and (129)Xe have distinct properties and provide unique and complementary functional information from the lungs. A triple-nuclear, same-breath imaging examination of the lungs with (1)H, (3)He, and (129)Xe can therefore provide exclusive functional information from the gas images. In addition, the (1)H images provide complementary co-registered structural information in the same physiological time frame. The goal of this study was to design an RF system for triple nuclear lung MRI at 1.5T, consisting of a dual-tuned transceiver coil for (3)He and (129)Xe, RF switches and a nested (1)H receiver array.
METHODS: A dual-tuned transmit-receive dual-Helmholtz RF coil for (3)He and (129)Xe was designed and constructed to work in unison with a nested (1)H receiver array.
RESULTS: Triple-nuclear imaging (structural and ventilation) and apparent diffusion coefficient mapping of the human lungs was performed in the same breath-hold using the integrated RF system. B1 maps and volumetric ventilation imaging using a three-dimensional, balanced steady-state free precession pulse sequence performed with both hyperpolarized (3)He and (129)Xe indicate good stand-alone performance of the coil for the respective nucleus.
CONCLUSION: Triple-nuclear same-breath lung imaging with a dual-tuned coil ((3)He and (129)Xe) and a nested (1)H array has been demonstrated with a custom RF system.

INTRODUCTION

Imaging the lungs with inhaled hyperpolarized gases 3He and 129Xe has been shown to provide functional information that cannot be accessed with proton (1H) MRI or other imaging modalities 1, 2, 3, 4, 5, 6. The two gases have distinct physical properties, which provide different but complementary functional information 7, 8, 9. The ability to image both nuclei in the same breath alongside the 1H anatomical images adds further structural and functional sensitivity to the acquisition. 3He is highly diffusive when compared with 129Xe 10, 11, 12, and the visualization and quantification of lung ventilation and diffusion with these two gases at the same time can help address important physiological questions such as the position of the diffusion–convection front in the lungs. The capability to measure the diffusivity of both 3He and 129Xe gases in the same lung inflation level also provides added information for measuring and modeling lung microstructure based on their measured apparent diffusion coefficients (ADC) 13. 129Xe is also denser and more viscous than 3He and as such has different fluid dynamic properties that define airflow in the airways, which can be measured with phase contrast MRI 14. 129Xe has the added feature that it is soluble in blood and has a wide range of chemical shift, which enables quantification of perfusion and gas exchange in the lungs 8, 9, 15, 16, 17. In recent years, MRI of perfluorinated 19F gases has also gained interest 18, 19, 20 as another MR‐sensitive gaseous tracer of regional lung function.

Therefore, same‐breath, multinuclear lung imaging with 3He‐129Xe mixtures and 1H MRI provides a unique combination of functional and structural information that is spatio‐temporally coregistered in the same physiological time frame 21, 22. Preliminary studies have used separate and spatially nested transmit‐receive (T‐R) coils for each nucleus 22. The reliance on the 1H MR system's birdcage body coil for signal reception constrains the signal‐to‐noise ratio (SNR) in 1H images of the lung, which is already limited by the low proton density of lung parenchyma. In a recent study 23, we showed that the 1H lung SNR in same‐breath imaging can be improved with a nested 1H receive array, which is compatible with operation with either a 3He or a 129Xe T‐R coil.

The motivation of this study was the design and construction of an integrated radiofrequency (RF) coil and T‐R switching system for triple nuclear lung imaging in the same breath. To achieve this, we developed a new dual‐tuned flexible T‐R RF coil to operate in quadrature for both 3He and 129Xe at 1.5T. For 1H imaging, we incorporated the 1H array developed in our previous study designed to nest within either 3He or 129Xe T‐R RF coils 23. With the developed RF instrumentation, we demonstrated triple nuclear same‐breath lung imaging with hyperpolarized 3He and 129Xe ventilation images and 1H anatomical images. With the same system ADC measurement of mixtures of 3He and 129Xe were performed in the same breath at a particular lung inflation state.

METHODS

3He and 129Xe Dual‐Tuned Coil Design

A dual‐tuned (3He‐129Xe) flexible quadrature T‐R coil was constructed in‐house. The conducting elements were made from self‐adhesive copper tape (FE‐5100–5276‐7; 3M, Bracknell, UK) of 66‐μm thickness and 6‐mm width, which was fixed on a substrate of 0.5‐mm‐thick polytetrafluoroethylene (Direct Plastics, Sheffield, UK) as shown in Figure 1b. The capacitors used on the resonant circuit were of 10C package (Dalian Dalicap Technology Co., Ltd, Dalian, China). The thickness of the array with the foam was 6 mm (3 mm each side). The dual‐tuned flexible T‐R coil was a dual Helmholtz‐like pair of quadrature design in which the Helmholtz for the in‐phase resonance of the quadrature spans the anterior right lung to posterior left lung, connected over left trapezius. Similarly, the Helmholtz for the quadrature‐phase resonance spans the anterior left lung to posterior right lung, via the right trapezius. The cross‐over of the copper strip for each of the Helmholtz pairs (which forms a “figure eight” topology) was positioned such that it was within the other resonant element (anterior) and was balanced on either side to minimize coupling as shown in Figure 1a and 1b. The schematic of the dual‐tuned T‐R coil circuit is shown in Figure 1a, and a photograph is shown in Figure 1b. The assembled topology of the flexible coil constitutes a bib design wrapped around the subject longitudinally, as shown in Figure 1c. Both the elements of the dual Helmholtz were fitted with two traps; one trap at the 1H frequency to enable 1H imaging with this coil in situ and the other trap to dual‐tune the coil to the 129Xe and 3He Larmor frequencies. The trap design was based on the formalism established in our earlier study for multituned resonators 23, the frequency of the trap for dual‐tuning was 47.81 MHz. A high‐pass matching circuit was used to match the coil at both resonant frequencies of 3He (48.62 MHz) and 129Xe (17.65 MHz) at 1.5T. The 1H trap was tuned with a 47‐pF capacitor and a seven‐turn wire wound inductor with a diameter of 6 mm. The trap for dual‐tuning the coil was tuned with a 56‐pF capacitor and a nine‐turn wire wound inductor with a diameter of 6 mm. Wire wound inductors were constructed from 21 AWG insulated copper wire. RF measurements were performed with an Agilent 5061B Network Analyzer (Keysight Technologies, Santa Rosa, California, USA). For the RF measurement, the dual‐tuned coil was wrapped longitudinally around the thorax of the subject, as shown in Figure 1c. This coil was designed to work with full functionality when the four‐channel 1H chest receiver array from our earlier study 23 was nested inside for in situ high SNR 1H lung imaging.

Figure 1

a: Schematic of the dual‐tuned flexible T‐R coil for 129Xe and 3He. The in‐phase and quadrature‐phase ports are marked with 0° and 90°, respectively. b: Picture of the dual‐tuned flexible T‐R coil for 129Xe and 3He. The in‐phase and quadrature‐phase ports are marked with 0° and 90°, respectively. The plastic housing at the port (0°, 90°) consists of high pass matching circuits (90 pF, 90 pF, and 300 nH) marked in the schematic and the two loop capacitors (47 pF and 220 pF). c: Illustration of application of dual‐tuned T‐R coil on the subject for same‐breath ADC measurement. d: Illustration of 1H array and dual‐tuned coil nested for triple nuclear imaging. e: Picture of the setup on the scanner. The picture indicates the mouthpiece and the two Tedlar bags affixed to it.

MR Imaging Methods for 3He, 129Xe and 1H

All in vivo imaging with 3He and 129Xe was performed with approval from the National Research Ethics Committee. The imaging was performed on a healthy male volunteer (age, 31 years; height, 185 cm; weight, 89 kg). Lung MRI was performed on a GE whole body 1.5T Signa HDx system with 3He and 129Xe gas polarized with spin exchange optical pumping 24. The gas dosage and the imaging and pulse sequence parameters used for all three nuclei are shown in Table 1. The hyperpolarized 3He and 129Xe gas was delivered in separate Tedlar bags and was mixed at the mouth piece at the time of inhalation, as illustrated in Figure 1e. 3He had polarization of 25% (≈100% of He is 3He). 129Xe had a polarization of 40%–50% (87% of Xe is 129Xe).

Table 1

Gas Mixture Dosage, Imaging Parameters, and Pulse Sequence Parameters Used in the Study

Measurement	Lung structure and ventilation	ADC	Whole lung ventilation	Flip angle map
Physiological details	Triple‐nuclear, same breath	Dual‐nuclear, same breath	Single‐nuclear, separate breath	Single‐nuclear, separate breath
RF coil	1H array, dual‐tuned coil	Dual‐tuned coil	Dual‐tuned coil	Dual‐tuned coil
Nuclei	1H, 3He, 129Xe	3He, 129Xe	3He, 129Xe	3He, 129Xe
Dosage (mL)
1H	—	—	—	—
3He	350	300	200	50
129Xe	500	500	500	100
Flip angle
1H	50°	—	—	—
3He	8°	9°	10°	—
129Xe	9°	10°	10°	—
TE (ms)
1H	0.9	—	—	—
3He	1.1	4.8	0.6	1.1
129Xe	3.6	12.5	2.1	3.6
TR (ms)
1H	2.9	—	—	—
3He	3.6	10	1.9	3.6
129Xe	18.9	27	6.4	18.9
Matrix
1H
Phase	192	—	—	—
Frequency	256	—	—	—
3He
Phase	104	48	82	52
Frequency	80	64	80	44
129Xe
Phase	78	48	82	52
Frequency	64	64	80	44
Slice thickness (mm)
1H	15	—	—	—
3He	15	15	4	200
129Xe	15	15	10	200
Number of slices
1H	3
3He	3	2	46	1
129Xe	3	2	24	1
Field of view (cm)
1H	40
3He	40	44	40	40
129Xe	40	44	40	40
Axis	2D, coronal	2D, coronal	Coronal	2D, coronal
Pulse sequence	bSSFP, FSGRE	FSGRE	3D bSSFP	FSGRE
Imaging time (s)
1H	1	—	—	—
3He	2	6	7	0.9
129Xe	4	8	13	2.4
Multiphase
1H	—	—	—	—
3He	—	—	—	6
129Xe	—	—	—	6
b Value (s · cm−2)
1H	—	—	—	—
3He	—	1.6	—	—
129Xe	—	8	—	—
Corresponding figure
1H	3a, 3d, 3e	—	—	—
3He	3b, 3d	2e	4a	2c
129Xe	3c, 3e	2f	4b	2d

Abbreviations: ADC, apparent diffusion coefficient; bSSFP, balanced steady‐state free precession; FSGRE, fast spoiled gradient echo.

RF Signal Routing and Calibration

To route the transmit RF signal (3He‐129Xe) from the appropriate T‐R switch on the scanner to the dual‐tuned coil and to route the received RF signal (3He‐129Xe) from the dual‐tuned coil back to the appropriate T‐R switch on the scanner, a 2‐kW rated coaxial antenna RF switch (CX‐SW2PL; Watson, Essex, UK) was used. The RF power required for the desired flip angle for the 3He and 129Xe sequences was calculated based on a standard calibration procedure, whereby the rate of depletion of polarization was calculated from the decay of signal resulting from a set of hard RF pulse‐acquires of equal amplitude. The period to prescribe calibration values on the spectrometer between the end of imaging a particular nucleus and initiation of the sequence for imaging the next nucleus was less than 4 s. The time required to operate the RF switch manually between acquisitions was 3 s.

Same‐Breath ADC (3He and 129Xe) and Triple Nuclear (3He, 129Xe, and 1H) Structure and Ventilation Lung Imaging Methods

For same‐breath ADC measurement, the dual‐tuned 3He‐129Xe coil was wrapped longitudinally as shown in Figure 1c, without the 1H array nested inside. To demonstrate same‐breath ADC maps, two sets of ADC measurements were acquired back‐to‐back in a single breath, with 3He ADC measurement followed by 129Xe measurement. The imaging parameters are shown in Table 1.

For triple‐nuclear lung imaging, the dual‐tuned 3He ‐129Xe coil and the 1H array from our earlier study 23 were nested as shown in Figure 1d. To demonstrate imaging of all three nuclei in the same breath, three sets of images were acquired back‐to‐back in a single breath in the order, with 3He imaging followed by 129Xe imaging, in turn followed by 1H imaging. The imaging parameters are shown in Table 1.

The T1 of hyperpolarized gases when inhaled into the lungs is sensitive to the oxygen partial pressure in the lung during the breath‐hold 25. 3He is more sensitive to this effect because the gyromagnetic ratio of 3He is approximately three times larger than that of 129Xe, as such the dipolar coupling to the electrons in the paramagnetic oxygen molecule is stronger. This rationale for the order of acquisition is 3He followed by 129Xe, in turn followed by 1H.

Flip Angle Mapping and High‐Resolution Imaging Performance of the Coil as a Stand‐Alone T‐R Coil for 3He and 129Xe

Flip angle maps of the dual‐tuned coil at the 3He and 129Xe frequencies were calculated by measuring the depletion of polarization of the hyperpolarized gas 3He and 129Xe at each voxel in the lungs by repeated imaging at breath‐hold with a two‐dimensional spoiled gradient echo sequence. The imaging parameters for this measurement are shown in Table 1, and T1 relaxation was neglected when calculating the flip angle. In addition, to demonstrate the coil's performance as a stand‐alone 3He or 129Xe T‐R coil (without the 1H array in situ), high‐resolution, three‐dimensional (3D) imaging data sets were acquired with a 3D balanced steady state sequence 26 with imaging parameters as shown in Table 1.

RESULTS

Dual‐Tuned Coil RF Performance

The two traps on the coil at 47.81 MHz and 63.86 MHz (1H trap) generated three resonant modes at 17.65 MHz (129Xe Larmor frequency), 48.62 MHz (3He Larmor frequency), and 79.2 MHz. The isolation between the two ports of the Helmholtz was less than −15 dB. The quality (Q) factor of the dual‐tuned coil at the 129Xe Larmor frequency (17.65 MHz) was 61 in the unloaded condition and 17 in the loaded condition. The Q factor of the dual‐tuned coil at the 3He Larmor frequency (48.62 MHz) was 32 in the unloaded condition and 7 in the loaded condition. Thus, the ratio of Q factor unloaded to loaded condition was 3.5 at the 129Xe Larmor frequency (17.65 MHz) and 4.5 at the 3He Larmor frequency (48.62 MHz). Under the loaded condition, the dual‐tuned coil was matched to less than −20 dB at both ports at both the 129Xe (17.65 MHz) and 3He (48.62 MHz) Larmor frequencies, as shown in Figure 2a. The isolation between the dual‐tuned coil and the 1H array was less than −15 dB, as shown in Figure 2b. Flip angle maps from the dual‐tuned coil for the transmit RF power prescribed for the nominal flip angles used for triple nuclear same‐breath imaging and ADC measurement (Table 1) are shown in Figure 2c for 3He and Figure 2d for 129Xe. The standard deviation of the flip angle map was calculated to be 0.7° (mean = 8°) for 3He and 0.3° (mean = 9°) for 129Xe.

Figure 2

a: Matching of dual‐tuned coil at 17.65 MHz and 48.62 MHz. b: Isolation between dual‐tuned coil and 1H array over frequency span of 10–80 MHz. c: Flip angle map of dual‐tuned coil for 3He. d: Flip angle map of dual‐tuned coil for 129Xe. The color bars indicate the flip angle in degrees. e, f: ADC measurement was taken in the same breath for 3He (e) and 129Xe (f). The color bars indicate the ADC in cm2 · s−1.

Multinuclear Lung Imaging

Same‐breath ADC measurement of 3He and 129Xe performed in the same lung‐inflation state is shown in Figure 2e and 2f. The 3He ADC map shown in Figure 2e and the 129Xe ADC map shown in Figure 2f were acquired in the same breath.

Same‐breath triple nuclear lung (structure and ventilation) images are shown in Figure 3. The 1H images shown in Figure 3a, 3He images shown in Figure 3b, and 129Xe images shown in Figure 3c, all of which were acquired in the same breath, are coregistered as shown in superimposed images in Figure 3d and 3e.

Figure 3

a: 1H images from lungs. b: Same‐breath 3He images from lungs. c: Same‐breath 129Xe images from lungs. d: 3He images superimposed over 1H images. e: 129Xe images superimposed over 1H images.

Hyperpolarized gas images of the lungs from 3D balanced steady‐state free precession sequence with dual‐tuned coil in a stand‐alone configuration (without 1H array nested). a: Hyperpolarized 3He gas. b: Hyperpolarized 129Xe gas.

Volumetric ventilation images from the 3D balanced steady‐state free precession sequence for the coil in operation as a stand‐alone transceiver for 3He and 129Xe are shown in Figure 4a and Figure 4b, respectively.

DISCUSSION

The construction of the flexible dual‐tuned coil is in the form of a bib, which enables a close fit to the subject's thorax irrespective of body type. The design was optimized to the typical subject size mentioned earlier. As the shape/form deviates from the optimal design with other body types, the distributed inductance and T‐R efficiency of the dual‐tuned coil changes accordingly. The B1 field homogeneity of the dual Helmholtz design is inherently inferior to that of a birdcage design 27, and the flexibility of the dual‐tuned coil adds some variability in this respect. Considering the typical anatomy of a torso, the distance between the RF coil and lung air spaces generally increases from superior (upper) to inferior (lower). This means that sensitivity in the lower lung is reduced for two reasons: first, due to proximity of the conducting elements to the lungs, and second, as the parallel condition for a Helmholtz pair is disrupted. Despite these factors, the observed B1 transmit homogeneity (variation in flip angle, 3% for 129Xe and 9% for 3He) is comparable to 28 or better than 29 studies reported previously using single‐tuned flexible T‐R coils for 3He and 129Xe lung imaging.

Because the RF switches are currently manually operated, and the spectrometer has an inherent delay time for precalibration for each nucleus, the method is not currently compatible with repetition time resonant frequency interleaved imaging, as demonstrated in our earlier study with same‐breath 3He‐1H lung imaging 22. It should be noted that this limitation is not due to the RF coil design or configuration; instead, it is due to the MR system, which supports only one spectrometer T‐R switch (single‐nucleus) to be actively connected at any given point in time (in addition to 1H). Both the dual‐tuned RF coil and the nested 1H array from the earlier study 23 are capable of operating simultaneously. If we consider the coil's operation as part of the system for triple nuclear imaging, 50%–60% of the time (ie, 18–20 s of the breath‐hold) is consumed by switching the spectrometer between the nuclei. This can be reduced with the appropriate spectrometer software engineering and using electrically driven RF switches (eg, PIN diodes and Field Effect Transistor).

The free diffusion (in air) of 3He is 0.88 cm2 · s−1 30, 31; in this study, we report 0.85 cm2 · s−1 for 3He in the trachea (slightly lower than 3He free diffusion). The free diffusion of 129Xe is 0.14 cm2 · s−1 31; in this study, we report 0.22 cm2 · s−1 for 129Xe in the trachea. The higher ADC value for 129Xe in the trachea, we presume is due to its mixture with the highly diffusive 3He (as shown in Table 1). In the ventilation images, any observed asymmetry beyond what can be attributed to the measured variation/asymmetry in the flip angle was verified to be caused by the distribution of 3He and 129Xe as a gas mixture in the lung (variation in the local concentration of the gases). These findings are currently being investigated in future work studying the physiology of gas mixing in the lung with the two gases.

In contrast to our previous triple nuclear same‐breath lung imaging experiments demonstrated at 3T on a Philips system using the 1H body T‐R coil, a 3He birdcage T‐R coil, and a nested 129Xe T‐R vest coil 22, the design used in this study at 1.5T has several potential benefits. First, from the coil perspective, the use of the dual‐tuned 3He‐129Xe coil minimizes the number of individually tuned coils, and the nested 1H array 23 improves the 1H SNR by closer proximity to the lung. Applications of this triple nuclear RF system for lung MRI are manifold and allow the different physical and physiological properties of the two gases to be explored in the same time course with added provision of high‐quality and coregistered 1H structural images.

In conclusion, we have demonstrated a system for triple nuclear same‐breath lung imaging of 1H with hyperpolarized gases 3He and 129Xe at 1.5T using a custom integrated RF system. This system incorporates a new design of dual‐tuned RF coil for 3He and 129Xe and RF switches, together with a nested receiver array for 1H imaging. With this system, we have demonstrated high‐quality, same‐breath 1H with 3He and 129Xe ventilation imaging and the capability for ADC mapping of 3He and 129Xe in the same lung‐inflation state. In addition, the image quality on all three nuclei is comparable with those acquired with separate RF coils for the given nucleus.