0521

Reproducibility of tailored and universal non-selective excitation pulses at 7T for human cardiac body imaging: A 3-year and an inter-day study

Manuel Fernando Sánchez Alarcón^1,2, Sebastian Dietrich¹, Jean Pierre Bassenge^1,2, Jeanette Schulz-Menger², Sebastian Schmitter^1,3,4, and Christoph Stefan Aigner¹
¹Physikalisch-Technische Bundesanstalt (PTB), Berlin, Germany, ²Charité – Universitätsmedizin Berlin, Berlin, Germany, ³Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, ⁴Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States

Synopsis

Keywords: RF Pulse Design & Fields, High-Field MRI, Reproducibility, Universal pulse, 7T, pTx

Motivation: Addressing the issue of reduced spatial variability in flip angle (FA) patterns in ultra-high field 3D imaging of the human heart for inter-year and inter-day studies, ensuring a high level of reproducibility.

Goal(s): Find the correct RF parallel transmission (pTx) excitation scheme that reduces FA spatial variability in 7T 3D imaging of the human heart for long-term and short-term studies.

Approach: Default, tailored pulses (TP), and pre-computed universal pulses (UP) were evaluated to optimize FA homogeneity in B₁⁺datasets across three years.

Results: The study highlights UPs' robustness in managing FA variations across subjects and coil placements in 3D body imaging at 7T.

Impact: This study confirms pre-computed UPs' suitability for 7T cardiac flip angle homogenization. Different MRI operators maintained consistent RF performance across three years and inter-day tests, with no significative differences in scans at various coil positions for both tests.

Introduction

Ultra-high field (UHF) body MRI faces challenges due to inherent spatial variations in flip angle (FA) patterns¹. Subject-specific parallel transmission (pTx) methods, relying on personalized $$$B_1^+$$$-maps, ensure high FA uniformity in the cardiac region^2,3,4. In contrast, using cardiac universal pulses (UPs) offers time-saving benefits and robustness for various RF coil placements, while also achieving FA uniformity in surrounding tissues, including the aorta^5,6. This study assesses the reproducibility and variability of three excitation methods in six subjects: default excitation, 4kT-points subject-tailored pTx (TP), and calibration-free UP. In addition to a previous study⁶, three different MRI operators conducted re-scans at one- and two-year intervals, including inter-day scans by two different operators.

Methods

MRI was conducted using a Siemens Magnetom 7T with an MRI.TOOLS 32-element body array used in an 8Tx/32Rx mode. A total of 46 channel-wise 3D $$$B_1^+$$$-datasets were acquired during free-breathing⁴ and subsequently categorized into five groups: i) UP-library (Library: 15M/7F, 21-66y, 19.8-28.3kg/m²), ii) unseen test-cases (Scan 1: 3M/3F, 25-33y, 19.5-35.3kg/m²), iii) re-scans after one year (Scan 2), iv) re-scans after two years (Scan 3), and v) intra-day re-scans of Scan 3 group (Scan 4).

A 4kT-point UP was computed based on Library group⁷ and applied to all subjects and all four scans. TPs were calculated for each subject³ in each Scan and resulting TPs were grouped in four TP configurations, labeled Config. 1 for TPs in Scan 1 to Config. 4 for TPs in Scan 4. Pulses from each Config were then applied for each Scan yielding 16 permutations MRI operators consistently positioned the posterior half of the coil approximately 2 cm from the volunteer's chin, over the volunteer's heart.

The postprocessing encompassed $$$B_1^+$$$-dataset reconstruction⁴, manual 3D cardiac region selection, RF pulse design, and pulse file generation. Pulse performance was evaluated through FA predictions and coefficient-of-variation (CV) within cardiac volumes. CV cluster points for each scan group were examined using parameters including inertia ($$$I_{CV}$$$), distance from the CV space origin to the cluster centroid ($$$|\bar{x}|$$$), and perpendicular distance from the cluster centroid to the 4D identity line ($$$d_\perp$$$).

Results and Discussion

Fig. 1 illustrates FA predictions for three pTx settings in one representative subject. TPs achieved CV values below 5% for Config matching Scan number, but up to 19.6% otherwise, indicating limited reproducibility. In contrast, UP consistently produced similar CV values for all scan groups, demonstrating high reproducibility.

Fig. 2 presents predicted CV and FA values for all subjects and excitation methods over the entire cardiac volume. Notably, Config. 3 notably reduces median CV values for Scan 4, contrasting with Scan 1 and Scan 2 groups. This suggests TP's applicability in inter-day studies done by two different operators, but less for inter-year ones. A similar trend is observed for Config. 4. The UP setting maintained median CV values of approximately 10% in all cases, demonstrating reproducibility across all scan groups.

Fig. 3 shows 4D data CV predictions across the cardiac volume for different scan groups. In the inter-year study (see Figure 3g), the UP consistently yields lower $$$I_{CV}$$$, $$$|\bar{x}|$$$, and $$$d_\perp$$$, signifying reduced CVs and strong long-term reproducibility. In the inter-day study (see Figure 3h), UP does not achieve the lowest $$$I_{CV}^{34}$$$, $$$|\bar{x}^{34}|$$$ and $$$d_\perp^{34}$$$ individually, as seen in the inter-year study. Nonetheless, it maintains a $$$d_\perp^{34}$$$ value close to zero while keeping $$$I_{CV}^{34}$$$ and $$$|\bar{x}^{34}|$$$ reasonably low in comparison to other pTx settings, ensuring high reproducibility with low CV values for the inter-day study.

Fig. 4 depicts CV and FA predictions for the three distinct pTx settings explored in this study, obtained from four separate scans conducted on one of the six subjects on the same day, each scan involving different coil positions. TPs were customized for the first scan (Coil. pos 1). In Figure 4b, UP predictions demonstrate exceptional resilience to body coil position variations, as evident from their minimal CV spread. This robustness surpasses other pTx settings for this inter-day study. Figure 4c further accentuates UP's strength, consistently achieving FA values close to the 10° target, outperforming alternative pTx settings in FA accuracy.

Fig. 5 shows 3D GRE images with UP aligning with $$$B_1^+$$$-predictions, validating coil placement robustness and calibration-free pTx effectiveness in the human heart. Results from 2022 were obtained with a Cartesian sequence, while 2020 and 2023 results used a non-Cartesian sequence⁵.

Conclusion

This study demonstrates the robustness of pre-computed UPs, which effectively address FA variations among subjects and variations in coil placements in both inter-year and inter-day studies. This property distinguishes UPs from the default pTx setting and TPs, making UPs ideal for 3D body applications at 7T.

Acknowledgements

We gratefully acknowledge funding from the German Research Foundation SCHM 2677/2-1, SCHM 2677/4-1 and GRK2260, BIOQIC.

References

[1] Ladd, E, Bachert, P, Meyerspeer, M, et al. Pros and cons of ultra-high-field MRI/MRS for human application, Prog. Nuc. Magn. Reson. Spec. 2018; 109:1-50. doi: 10.1016/j.pnmrs.2018.06.001

[2] Padormo, F, Beqiri, A, Hajnal, JV, and Malik, SJ. Parallel transmission for ultrahigh‐field imaging. NMR Biomed. 2016; 29: 1145– 1161. doi: 10.1002/nbm.3313

[3] Aigner, CS, Dietrich, S, Schmitter, S. Three-dimensional static and dynamic parallel transmission of the human heart at 7 T. NMR in Biomedicine. 2021; 34:e4450. doi: 10.1002/nbm.4450

[4] 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. doi: 10.1002/mrm.28602

[5] Gras, V., Vignaud, A., Amadon, A., Le Bihan, D. and Boulant, N. (2017), Universal pulses: A new concept for calibration‐free parallel transmission. Magn. Reson. Med., 77: 635-643. doi:10.1002/mrm.26148

[6] Aigner, CS, Bassenge, JP, Dietrich, S, Lutz, M, Krüger, F, and Schmitter, S. Reproducibility and variability of tailored and universal nonselective excitation pulses at 7T for human body imaging: Re-scans after one year. ISMRM 2023, 4590.

[7] Aigner, CS, Dietrich, S, Schaeffter, T, Schmitter, S. Calibration-free pTx of the human heart at 7T via 3D universal pulses. Magn Reson Med. 2021; 87: 70–84. https://doi.org/10.1002/mrm.28952

Figures

Figure 1: a) 3D flip angle predictions for the human heart using Default excitations, TP (Config. 1) and UP. The data corresponds to all four scans made to one of the six subjects considered in this study. Notably, the results highlight the efficiency of the UP, saving 10-15 minutes of critical scan time when compared to TPs, achieving CV values below 14% for all scan groups.

Figure 2: CV and FA predictions for the entire cardiac volume, employing Default excitation, TP for all four configurations UP. Notably, the use of TP leads to an increased spread in CV predictions (c) for groups where the TP was not originally designed, highlighting a lack of reproducibility. Conversely, when UP was applied, the CV value distribution for all scan groups remained relatively consistent, indicating good reproducibility. Similar trends can be observed for FA predictions obtained with TP (e) and UP (f).

Figure 3: 2D scatter plots (figures (a) to (f)) obtained from the 4D data generated by the CV predictions for the entire cardiac volume for each subject in each one of the scan groups. Colored X markers represent the centroid of each cluster of points. To quantify the behavior of CV predictions, three parameters were measured for each cluster of points: Inertia, distance to the centroid, and perpendicular distance from the centroid to the 4D identity line. 2D Analog variables were defined for CV predictions obtained from Scan 3 and Scan 4 groups in (h).

Figure 4: In (a), coil position parameters were deliberately adjusted to assess measurement robustness to changes in the relative body coil position within a single subject for an inter-day measurement. In (b), the minimal CV spread in UP results underscores UP predictions' resilience to body coil position variations compared to other pTx settings. This robustness is further evident in (c), where UP predictions exhibit median FA values closely approximating the target FA of 10°, outperforming other pTx settings in this regard.

Figure 5: FA predictions for a single sagittal slice within a 3D cardiac volume, utilizing both default excitation and the UP, for one subject in all scan groups. Notably, the 3D images exhibit a remarkable absence of breathing artifacts, underscoring the feasibility of achieving uniform calibration-free FA homogeneity throughout the entire heart, despite some minor discrepancies near the coil elements. Any remaining signal variations observed in the anterior-posterior (AP) direction are attributed to receive (B₁^-) variations.

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

0521

DOI: https://doi.org/10.58530/2024/0521