4090

Dynamic pTx pulse design with accelerated B₁⁺-mapping for cardiac imaging at 7T

Nico Egger¹, Sophia Nagelstraßer¹, Judith Schirmer¹, Jürgen Herrler², Laurent Ruck¹, Saskia Wildenberg^1,3, Titus Lanz⁴, Ralph Kimmlingen², Michael Uder¹, Sebastian Schmitter^5,6,7, Andreas K. Bitz³, and Armin Michael Nagel^1,6
¹Institute of Radiology, University Hospital Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Erlangen, Germany, ²Siemens Healthcare GmbH, Erlangen, Germany, ³Electrical Engineering and Information Technology, FH Aachen - University of Applied Sciences, Aachen, Germany, ⁴Rapid Biomedical GmbH, Rimpar, Germany, ⁵Physikalisch-Technische Bundesanstalt (PTB), Berlin, Germany, ⁶Division of Medical Physics in Radiology, German Cancer Research Centre (DKFZ), Heidelberg, Germany, ⁷Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States

Synopsis

Keywords: Parallel Transmit & Multiband, Parallel Transmit & Multiband

Motivation: Subject-specific pTx methods can counteract excitation inhomogeneities in UHF body imaging, but require lengthy calibration times. Calibrationless universal approaches exist but yield reduced homogeneity for similar SAR levels.

Goal(s): Our aim was to evaluate the possibility of shortening the calibration time by accelerating the B₁⁺-mapping process.

Approach: For 35 heart subjects, tailored pTx pulses were designed on B₁⁺-maps with different undersampling and their performance was compared to universal approaches regarding flip angle homogeneity and SAR.

Results: Tailored pulses designed on accelerated, undersampled B₁⁺-maps (TA=30s) result in improved homogeneity compared to universal pulses or in a roughly 2.5-fold reduction of SAR at comparable homogeneity.

Impact: Subject-specific pTx pulses optimized on accelerated, undersampled B₁⁺-maps allow for improved homogeneity or reduced SAR compared to universal pulses. Reduced SAR could benefit body imaging at 7T, since SAR limitations often prevent the optimal choice of acquisition parameters at UHF.

Introduction

Body imaging at 7T suffers from flip angle (FA) inhomogeneities caused by the shorter, field strength dependent excitation wavelength. Parallel transmission (pTx) methods¹ can counteract these inhomogeneities, but subject-specific solutions require long calibration times for mapping the channel-wise transmit (B₁⁺) fields. In this work, we investigate the possibility of accelerating the B₁⁺-mapping via undersampling and compare our approach to calibrationless universal pTx approaches^2,3.

Methods

For the measurement of 3D channel-wise relative B₁⁺-maps, a method based on the small FA regime⁴ was used. The sequence was implemented with a density-adapted 3D-radial readout⁵ (DA-3D-Rad) and measured with following parameters: TE/TR = 2.02/4.5ms, FOV = (320mm)³, nominal resolution = (4mm)³, nominal FA = 8°, projections = 10000, TA = 6min. The projections were measured in the golden angle scheme⁶ to enable retrospective undersampling. All measurements were carried out at 7T (MAGNETOM Terra, Siemens Healthcare GmbH, Erlangen, Germany) with an 8Tx/16Rx transceiver array (Rapid Biomedical GmbH, Rimpar, Germany). B₁⁺-maps of the heart were acquired for 35 healthy subjects (12 female, 23 male, age = 28.1 ± 8.2 years, BMI = 23.4 ± 3.2 kg/m²). The raw data of the relative B₁⁺-maps were reconstructed with different USFs (1, 2-20, steps of 2) and the sum of magnitudes of the fully sampled maps were used to define a 3D ROI covering the heart. All B₁⁺-maps were normalized by the ROI-mean signal intensity of their sum of magnitudes and multiplied with a scaling factor calculated from an AFI⁷ measurement of one subject, to obtain absolute B₁⁺-estimations⁸. Dynamic pTx pulses with a kT-point trajectory⁹ were optimized, consisting of four rectangular subpulses with a duration of 0.15 ms each, three intervening gradient blips of 0.09 ms³, target FA of 5° and initial specific energy dose (SED, proportional to SAR) limit of 0.1256 J/kg. At first, a universal pulse² (UP) was designed with the complete library of 35 heart subjects (UP35-4kT). The UP was then used as a starting value to optimize tailored-4kT pulses for each subject and USF. All pulses were evaluated by simulating them on the corresponding fully sampled B₁⁺-maps and calculating the CoVs of the resulting FA distributions. The pulses were also compared to a vendor provided default phase shim, intended for cardiac imaging. Additionally, for USF1 and USF12, tailored-4kT pulses were designed for which the SED limit was gradually lowered until the resulting median CoV equaled that of the UPs. Validation measurements of this approach were performed for one subject with a DA-3D-Rad sequence (TE/TR = 2.02/7 ms, FOV = (320 mm)³, nominal resolution = (1.4 mm)³, nominal FA = 5°, TA = 7 min).

Results

Figure 1 depicts the coefficients of variation (CoVs) of the different pulses for all 35 heart subjects. Compared to the default phase shim, UP35-4kT achieves a reduction of the median CoVs to 10.6% as well as smaller intersubject-variations. Tailored-4kT pulses with USF1 result in a further reduction to a median of 5.7%. The median CoVs and intersubject-variations of the tailored-4kT pulses increase with rising USF. From USF2 onward one strong outlier can be observed. A B₁⁺-mapping time of 30s (≙ USF12) results in a median CoV of 8.4%. Figure 2 shows the SED values of tailored-4kT pulses with USF1 and USF12 optimized to yield an equal median CoV as the UP35-4kT. Relative to the UP, these pulses result in a mean SED of 29.84% ± 0.02% and 39.75% ± 0.04% respectively. A FA simulation of all four pulses is shown in Figure 3 for the subject of the validation measurements and the corresponding DA-3D-Rad images are depicted in Figure 4. In general, simulations and measurements show a good agreement. For the default shim, signal dropouts in the heart are visible, whereas the dynamic pulses result in a homogeneous excitation, although at different SED levels.

Discussion and Conclusion

Compared to the vendor provided shim, the homogeneity for cardiac MRI could be substantially improved with the universal 4kT-pulse. Tailored pulses achieved a further improvement, but come at the cost of additional calibration time. As the results illustrate, calibration time could be shortened by accelerating the B₁⁺-mapping via undersampling. Tailored pulses with a B₁⁺-mapping time of 30s still showed improved homogeneity compared to the UP. However, a more important benefit of tailored pulses might be the reduction of SED. Compared to the UP35-4kT, tailored-4kT pulses based on B₁⁺-maps acquired in 30s resulted in a roughly 2.5-fold SED reduction while maintaining comparable homogeneity. The strong outlier visible in Figure 1 may have been caused by patient movement but should be further investigated.

Acknowledgements

No acknowledgement found.

References

1. Webb AG, Collins CM. Parallel Transmit and Receive Technology in High-Field Magnetic Resonance Neuroimaging. Int J Imag Syst Tech 2010;20:2-13.

2. Gras V, Vignaud A, Amadon A, Le Bihan D, Boulant N. Universal pulses: A new concept for calibration-free parallel transmission. Magn Reson Med 2017;77:635-643.

3. Aigner CS, Dietrich S, Schaeffter T, Schmitter S. Calibration-free pTx of the human heart at 7T via 3D universal pulses. Magn Reson Med 2022;87:70-84.

4. Van de Moortele P-F, Ugurbil K. Very fast multi channel B1 calibration at high field in the small flip angle regime. In Proceedings of the 17th Annual Meeting of ISMRM, Honolulu, HI, 2009. p. 367.

5. Nagel AM, Laun FB, Weber MA, et al. Sodium MRI using a density-adapted 3D radial acquisition technique. Magn Reson Med 2009;62:1565-1573.

6. Winkelmann S, Schaeffter T, Koehler T, Eggers H, Doessel O. An optimal radial profile order based on the Golden Ratio for time-resolved MRI. IEEE Trans Med Imaging 2006;26:68-76.

7. Yarnykh VL. Actual flip-angle imaging in the pulsed steady state: a method for rapid three-dimensional mapping of the transmitted radiofrequency field. Magn Reson Med 2007;57:192-200.

8. Dietrich S, Aigner CS, Kolbitsch C, et al. 3D Free-breathing multichannel absolute B1+ Mapping in the human body at 7T. Magn Reson Med 2021;85:2552-2567.

9. Cloos M, Boulant N, Luong M, et al. kT‐points: short three‐dimensional tailored RF pulses for flip‐angle homogenization over an extended volume. Magn Reson Med 2012;67:72-80.

Figures

Distribution of the CoVs in the heart for all 35 subjects when applying the default phase shim (Def.), UP35-4kT and tailored-4kT pulses optimized on B₁⁺-maps with increasing USFs. UP35-4kT results in a considerable improvement in homogeneity compared to the default shim. The tailored-4kT pulses with USF1 result in a further improvement. For rising USFs, the CoV ranges and medians of the tailored-4kT pulses increase, but the median for USF20 is still lower than for UP35-4kT. From USF2 onward, one strong outlier can be seen.

a) CoVs of the 35 heart subjects with UP35-4kT and the tailored-4kT pulses with USF1 and 12 that were optimized with more restrictive SED limits, to achieve an equal median CoV of 10.6% for all three pulse types. b) Mean SED of these pulses relative to the SED of UP35-4kT. For an equal median CoV to UP35-4kT, the tailored-4kT pulses with USF1 (B₁⁺-mapping time = 6min) result in a mean relative SED of 29.8% and the tailored-4kT pulses with USF12 (B₁⁺-mapping time = 30s) in a mean relative SED of 39.8%.

FA simulations of the default phase shim, UP35-4kT and tailored-4kT pulses calculated with USF1 and USF12 for the subject of the validation measurement. For the optimization of the tailored pulses, the more restrictive SED limits were used, which lead to an equal median CoV over all 35 subjects as UP35-4kT. The CoVs of the FAs within the ROI are given at the top and the ROI itself is outlined in black. The FAs of the default phase shim were scaled to reach a mean target FA of 5°. The dynamic pulses all result in a smooth FA distribution and improved homogeneity compared to the default shim.

DA-3D-Rad images of one heart subject, acquired with the default phase shim, UP35-4kT and tailored-4kT pulse calculated with USF1 and USF12. For the optimization of the tailored pulses, the more restrictive SED limits were used, which lead to an equal median CoV over all 35 subjects as UP35-4kT. The SED values relative to UP35-4kT are specified at the top. The dynamic pulses all show similar image quality and improved homogeneity in the heart compared to the default phase shim, but lead to different amounts of SED. The images show a good agreement to the FA estimations in Figure 3.

Proc. Intl. Soc. Mag. Reson. Med. 32 (2024)

4090

DOI: https://doi.org/10.58530/2024/4090