0684

Preclinical Hyperpolarized ¹²⁹Xe Ventilation Imaging Using 3D Spiral (FLORET) Encoding

Brice J Albert¹, Peter J Niedbalski¹, and Zackary I Cleveland^1,2,3,4
¹Center for Pulmonary Imaging Research, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United States, ²Imaging Research Center, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United States, ³Department of Pediatrics, University of Cincinnati Medical Center, Cincinnati, OH, United States, ⁴Department of Biomedical Engineering, University of Cincinnati, Cincinnati, OH, United States

Synopsis

Center-out trajectories are often used in preclinical HP gas MRI to reduce the impact of physiological motion and magnetization decay on image quality. Recently, implementation of 3D spiral (FLORET) imaging for human ¹²⁹Xe ventilation imaging demonstrated higher accuracy in detecting ventilation abnormalities than traditional sequences. Here we show FLORET sequences provide superior SNR, consume less xenon, and reduce scan time by more than five times when used to image ventilation in mice.

Introduction

In humans, Gradient Recalled Echo (GE) sequences are used commonly for hyperpolarized (HP) ¹²⁹Xe ventilation imaging, because they can be implemented in a single breath and are ubiquitous across scanner platforms(1). In contrast, preclinical images are typically acquired over 10s to 100s of breaths, increasing the deleterious impact of respiratory and cardiac motion on image quality. Of necessity, only a small number of k-space lines are encoded per breath, and signal is intrinsically small. Thus, large flip-angle RF pulses (20-90º) are needed, which cause substantial in-breath signal decay. Further, the high field strengths used in preclinical imaging leads to more rapid T₂* relaxation (~5 ms at 7T), further degrading image quality(2). Because of these challenges, radial sequences have displayed utility in preclinical HP gas imaging, because they oversample low-frequency imaging data, making them robust to motion, data undersampling, view-to-view magnetization variations, and T₂* decay(3-5). Slice-selective 2D strategies have been employed with success, but they generate coarse resolution (~1 mm) in the slice dimension relative to the minute scale of mouse lungs. As such, true 3D sequences are preferable, because they provide an anatomically equivalent to images obtained in human subjects. To date, 3D radial ¹²⁹Xe ventilation imaging had been implemented in both rats and mice to achieve high quality images with high isotropic resolution. These 3D sequences allow magnetization dynamics to be tracked throughout acquisition, allowing signal decay to be mapped and corrected, without addition data collection, to improve quantitative image accuracy(6). However, 3D Radial is inefficient at encoding k-space and requires a large gas volume or a high degree of undersampling. In contrast, 3D spiral sequences, such as Fermat Looped Orthogonally Encoded Trajectories (FLORET), can encode k-space with an efficiency that exceeds conventional Cartesian sequences while retaining the advantages of center-out encoding(7). The utility of FLORET was recently demonstrated in human HP ¹²⁹Xe imaging, and here we extend it to mouse imaging and demonstrate benefits of reduced scan time, decreased ¹²⁹Xe consumption, and increased SNR relative to 3D radial.

Methods

3-D radial and FLORET images (Figure 1) were acquired using a 7 T Bruker BioSpin (Billerica, MA). Acquisition parameters consistent in all scans were: matrix=104³, FOV=26 mm³, bandwidth=64 kHz, dummy pulses=20, α=30°, TE/TR=0.77/30ms. Slab-selection was used to prevent wraparound artifacts. Three separate fully encoded FLORET images, named for readout (RO) duration relative to expected T₂* (i.e., FLORET1: 5ms; FLORET2: 10ms; FLORET3: 15ms), were used. The 3D radial readout duration was 1ms. Points, views, and number of breaths, etc. are summarized in Table 1. ¹²⁹Xe was hyperpolarized (HP) to 35-45 % (Model 9820, Polarean PLC., Durham, NC). Shorter scan durations enabled all three FLORET scans to be acquired in a single 350mL batch of HP ¹²⁹Xe, while a full batch was needed for the 3D radial scan. The order of the FLORET scans was randomized within each mouse to reduce T₁ dependent systematic bias. Five C57BL/6J mice (2 male, 3 female, mass 27.5 ±4.5g, age 36 weeks ±1 week, Jackson Laboratory, Bar Habor, ME) were anesthetized using intraperitoneal injections of 90/9/3 mg/kg ketamine/xylazine/acepromazine. Mice were intubated with 22-ga all plastic catheters, placed on a custom-built, HP-gas compatible, small animal ventilator, and ventilated as described previously, 79/21% mixture of HP ¹²⁹Xe gas/O₂ at expiration or inspiration. Image reconstruction and statistical analysis were performed in MATLAB 2019b (MathWorks, Natick, MA). Signal thresholds (5× standard deviation of the background) followed by erosion-dilation with a spherical structuring element (radius 2 voxels) were used to delineate ventilated volume. Noise was calculated from the standard deviation of artifact-free background, and signal-to-noise ratios (SNR) were adjusted for Rician noise distribution. A slight correction for tidal volume and number of breaths was made to the SNR normalization technique previously reported. Mean polarization, SNR and normalized SNR were compared using ANOVA with post hoc Tukey test.

Results and Discussion

FLORET greatly reduces preclinical scan time while enabling full k-space encoding and high image quality. This can be seen qualitatively, in Figure 2, which displays maximum intensity projections in the coronal plane for 4-fold undersampled radial and all three FLORET scans. Of note, the trachea and several generations of the large airways are apparent in all images, and uniform signal is observed from the lung parenchyma. Representative coronal slices in Figure 3 reveal only slight blurring with longer RO durations in FLORET2 and 3. Figure 4a shows initial ¹²⁹Xe polarization estimated from T₁ and polarizer calibrations(8). The highest mean polarization, observed in the 3D radial scan, did not result in superior SNR. No significance difference in mean SNR was observed for these four scans. It should be noted that FLORET1 compared to quarter-sampled radial imaging reduces scant time by seven minutes and xenon consumption by 100mL with no apparent change in image quality. By normalizing the SNR, as noted in the methods, significantly higher SNR is observed for FLORET3 (p=0.0008). This highlights the >5x reduction in xenon consumption with long RO durations.

Conclusion

By implementing 3D FLORET, high-resolution images with superior SNR to 3D radial scans, are acquired with comparable image quality while reducing both scan time and xenon consumption by greater than five times.

Acknowledgements

The authors thank the following sources for research funding and support: NIH R01 HL143011 and NIH T32 HL007752.

References

1. He M, Robertson SH, Kaushik SS, Freeman MS, Virgincar RS, Davies J, Stiles J, Foster WM, McAdams HP, Driehuys B. Dose and pulse sequence considerations for hyperpolarized ¹²⁹Xe ventilation MRI. Magn Reson Imaging 2015;33(7):877-885.

2. Niedbalski PJ, Cochran AS, Akinyi TG, Thomen RP, Fugate EM, Lindquist DM, Pratt RG, Cleveland ZI. Preclinical hyperpolarized ¹²⁹Xe MRI: ventilation and T₂* mapping in mouse lungs at 7 T using multi-echo flyback UTE. NMR Biomed 2020;33(7):e4302.

3. Mistry NN, Thomas A, Kaushik SS, Johnson GA, Driehuys B. Quantitative analysis of hyperpolarized ³He ventilation changes in mice challenged with methacholine. Magn Reson Med 2010;63(3):658-666.

4. Cleveland ZI, Virgincar RS, Qi Y, Robertson SH, Degan S, Driehuys B. 3D MRI of impaired hyperpolarized ¹²⁹Xe uptake in a rat model of pulmonary fibrosis. NMR Biomed 2014;27(12):1502-1514. 5. Santyr GE, Lam WW, Ouriadov A. Rapid and efficient mapping of regional ventilation in the rat lung using hyperpolarized ³He with Flip Angle Variation for Offset of RF and Relaxation (FAVOR). Magn Reson Med 2008;59(6):1304-1310.

6. Niedbalski PJ, Willmering MM, Robertson SH, Freeman MS, Loew W, Giaquinto RO, Ireland C, Pratt RG, Dumoulin CL, Woods JC, Cleveland ZI. Mapping and correcting hyperpolarized magnetization decay with radial keyhole imaging. Magn Reson Med 2019;82(1):367-376.

7. Willmering MM, Niedbalski PJ, Wang H, Walkup LL, Robison RK, Pipe JG, Cleveland ZI, Woods JC. Improved pulmonary ¹²⁹Xe ventilation imaging via 3D-spiral UTE MRI. Magn Reson Med 2020;84(1):312-320.

8. Plummer JW, Emami K, Dummer A, Woods JC, Walkup LL, Cleveland ZI. A semi-empirical model to optimize continuous-flow hyperpolarized ¹²⁹Xe production under practical cryogenic-accumulation conditions. J Magn Reson 2020;320:106845.

Figures

Figure 1. High resolution ventilation MRI in mice with 3D encoding. a) Radial encoding oversamples k-space and inefficiently covers k-space. b) Fermat Looped Orthogonally Encoded Trajectories (FLORET) efficiently encodes k-space to allow fully encoded data sets.

Figure 2. Representative max intensity projections in the coronal plane from one mouse. a) 3D radial, 25% percent sampled. b-d) 3D spiral (FLORET), with each spiral arm durations equal to b) 1 x T₂* readout, c) 2 x T₂* readout, and d) 3 x T₂* readout.

Figure 3. Representative coronal slice from each 3D scan. a) 3D radial, 25% percent sampled. b-d) 3D spiral (FLORET), with each spiral arm readout duration equal to b) 1 x T₂* readout, c) 2 x T₂* readout, and d) 3 x T₂* readout. SNR is reported at the bottom of each image. Arrows indicate areas of increased blurring with longer spiral RO duration.

Figure 4. Box plots summarize the values of a) estimated initial polarization, b) Signal-to-noise ratio, and c) the corrected signal-to-noise ratio for each of the four scans. Descriptions for the calculation of SNR and the corrected SNR are detailed in the methods section. * ≡ significant difference (p = 0.0008).

Table 1. Scan parameters for 3D volumetric sequences. RO = readout.

Proc. Intl. Soc. Mag. Reson. Med. 29 (2021)

0684